Rilley's Head Injury Complete

HEAD INJURY
Frontispiece
Copy to follow.
Frontispiece
Copy to follow.
The illustrations on this page are black and white

visuals of illustrations from Chapter 9 which will be
colour on this page Top left is Figure 9.8, top right is
Figure 9.43(a), bottom left is Figure 9.43(b) and bottom
right is Figure 9.58
HEAD INJURY
Pathophysiology and management of
severe closed injury
EDITED BY
Peter Reilly
MD, BMedSc, FRACS
Department of Neurosurgery, Royal Adelaide Hospital, Adelaide, Austalia

and
Ross Bullock
MD, PhD
Division of Neurological Surgery, Medical College of Virginia, Richmond,

Virginia, USA
Published by Chapman & Hall, 26 Boundary Row, London SE1 8HN,

UK
Chapman & Hall, 26 Boundary Row, London SE1 8HN, UK
Chapman & Hall GmbH, Pappelallee 3, 69469 Weinheim, Germany
Chapman & Hall USA, 115 Fifth Avenue, New York, NY 10003, USA
Chapman & Hall Japan, ITP-Japan, Kyowa Building, 3F, 221 Hirakawacho, Chiyoda-ku, Tokyo 102, Japan
Chapman & Hall Australia, 102 Dodds Street, South Melbourne, Victoria
3205, Australia
Chapman & Hall India, R. Seshadri, 32 Second Main Road, CIT East,
Madras 600 035, India
1997 Chapman & Hall
Typeset in 10 on 12 pt Palatino by Genesis Typesetting, Rochester, Kent
Printed in Great Britain by Cambridge University Press, Cambridge
ISBN 0 412 58540 5
Apart from any fair dealing for the purposes of research or private study,
or criticism or review, as permitted under the UK Copyright Designs and
Patents Act, 1988, this publication may not be reproduced, stored, or
transmitted, in any form or by any means, without the prior permission in
writing of the publishers, or in the case of reprographic reproduction only in
accordance with the terms of the licences issued by the Copyright Licensing
Agency in the UK, or in accordance with the terms of licences issued by the
appropriate Reproduction Rights Organization outside the UK. Enquiries
concerning reproduction outside the terms stated here should be sent to the
publishers at the London address printed on this page.
The publisher makes no representation, express or implied, with regard to
the accuracy of the information contained in this book and cannot accept any
legal responsibility or liability for any errors or omissions that may be made.
A catalogue record for this book is available from the British Library
Library of Congress Catalog Card Number: 97-66517
While every effort has been made to trace copyright holders and obtain
permission, this has not been possible in all cases; any omissions brought
to our attention will be remedied in future editions.
CONTENTS
List of contributors
Preface
PART ONE: THE INJURY

1
Pathology
Peter C. Blumbergs
39
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
39
40
40
41
46
47
48
Epidemiology
Michael R. Fearnside and Donald A. Simpson
1.1
1.2
1.3
1.4
3
3
3
3.9
4
6
7
8
10
10
12
14
15
16
21
3.10
3.11
3.12
3.13
3.14
3.15
3.16
3.17
3.18
3.19
3.20
1.5
1.6
1.7
1.8
1.9
1.10
1.11
1.12
1.13
1.14
ix
xi
Introduction
Definitions in epidemiology
Source data
Definitions and classification of head
injury
Deaths from trauma
Severity of trauma
Population-based national studies
Population-based regional studies
Causation
Children
Minor head injury
Counting the cost
Reducing the burden
References
Biomechanics of closed head injury

A. J. McLean and Robert W. G. Anderson
25
2.1
2.2
2.3
2.4
25
27
28
2.5
2.6
2.7
Impact to the head

Response of the head to impact
Methods of investigation
Toward an understanding of brain
injury mechanisms
Tolerance of the head to impact
The state of the art of head injury
biomechanics
References
29
34
36
36
3.21
3.22
3.23
3.24
3.25
3.26
4
Introduction
Assessment of severity of brain injury
The mechanism of brain injury
Axonal injury
Concussive syndromes
Hypoxicischemic damage in humans
Brain swelling
Neurotransmitter agonistreceptor
abnormalities
Hippocampal pathology in traumatic
brain injury
Traumatic vascular injury
Lacerations
Traumatic intracerebral hemorrhage
Extradural (epidural) hemorrhage
Acute subdural hematoma
Chronic subdural hematoma
Other subdural fluid collections
Subarachnoid hemorrhage
Diffuse vascular injury
Brain-stem lesions
Brain damage secondary to raised
intracranial pressure
Long-term effects
Post-traumatic vegetative state
Post-traumatic epilepsy
Head injury and Alzheimers disease
Brain injuries due to boxing
References
50
50
50
55
55
58
59
60
60
60
61
61
63
64
64
65
65
65
66
Primary and secondary brain injury

A. David Mendelow and Peter J. Crawford
71
4.1
4.2
4.3
71
72
73
Introduction
Primary brain damage
Secondary brain damage
vi
CONTENTS
4.4
4.5
4.6
4.7
4.8
4.9
4.10
4.11
4.12
4.13
5
Brain metabolism and cerebral blood flow

Alois Zauner and J. Paul Muizelaar
5.1
5.2
5.3
5.4
The basic principles of brain metabolism

and blood flow
Normal values for CBF and metabolism
CBF and metabolism following head
injury
References
Intracranial pressure and elastance

Ian Piper
6.1
6.2
6.3
7
Extradural hematoma
Intradural hemorrhage
Intracerebral hemorrhage
Herniation
Brain swelling
Infection
Post traumatic vascular damage
Pyrexia following head injury
Conclusion
References
The problem: raised intracranial

pressure after head injury
The principles: physiology and
pathophysiology of intracranial
pressure
References
77
78
79
79
80
83
85
85
86
86
89
94
9.1
9.2
9.3
9.4
9.5
9.6
9.7
9.8
9.9
9.10
168
168
169
171
172
173
173
185
197
95
97
9.11
9.12
101
9.13
9.14
89
10
101
105
117
7.1
7.2
121
7.5
7.6
7.7
7.8
7.9
7.10
121
123
209
209
210
210
213
214
215
216
11.5
11.6
11.7
Clinical examination and grading

Donald A. Simpson
145
8.1
8.2
145
146
12
Introduction
Historical aspects
Transducers
Methods of measuring ICP
Which system?
Interpretation of ICP monitoring
Conclusion
References
Measuring cerebral blood flow and

metabolism
11.4
143
206
206
206
10.1
10.2
10.3
10.4
10.5
10.6
10.7
10.8
138
139
140
140
Introduction
197
200
209
11.1
11.2
11.3
143
Introduction
The initial examination
11
Introduction
The role of plain film radiography
CT in head injury
MRI in head injury
SPECT in head injury
Classification
Intracerebral lesions
Extracerebral collections
Pneumocephalus
Raised intracranial pressure and
herniation
Patterns of ischemia
Radiology in the diagnosis of brain
death
Conclusion
References
Intracranial pressure monitoring

Brian North
126
128
133
PART TWO: MEASURING AND

MONITORING INJURY
156
158
164
168
121
7.3
7.4
The definitive examination

Evaluation of injury severity
References
Imaging the injury

Evelyn Teasdale and Donald M. Hadley
Injury and cell function

Ross Bullock
Introduction
Biomechanical characteristics of
neurotrauma
Biomechanical effects and age
Major vascular damage secondary to
shear injury
Metabolic consequences of TBI
Intracellular mechanisms
Brain swelling and cellular events
after neurotrauma
Conclusion
Acknowledgments
References
8.3
8.4
8.5
Overview of CBF measurements

Xenon CBF measurements
Further direct clinical methods for
obtaining CBF
Indirect methods for obtaining CBF
and metabolism
Direct measurement of cerebral
metabolism
Comprehensive neuromonitoring
References
217
217
218
221
222
223
226
226
Electrical function monitoring

R. J. Moulton
229
12.1
12.2
12.3
12.4
12.5
229
229
230
240
240
Goals
Problems and limitations
Methods and modalities
Conclusions
References
CONTENTS
13
Transcranial Doppler
Peter J. Kirkpatrick and Kwan-Hon Chan
13.1
13.2
13.3
13.4
13.5
13.6
13.7
13.8
13.9
14
Introduction
The theory of TCD sonography
TCD measurements
Signal processing and data collection
Results of analysis using TCD in
head-injured patients
Role of TCD in monitoring therapy
TCD in the diagnosis of brain death
Summary
References
Magnetic resonance spectroscopy

Robert Vink
14.1
14.2
14.3
14.4
14.5
14.6
Introduction
Principles of magnetic resonance
spectroscopy
MRS studies of neurotrauma
MRI studies of brain function
Conclusion
References
249
254
255
255
258
261
262
265
267
267
269
15
From accident site to trauma center

J. E. Gilligan
271
15.1
15.2
15.3
271
271
16
Fluid, electrolyte and metabolic

management
Peter D. Thomas
16.1
16.2
16.3
16.4
16.5
Introduction
Rationale of metabolic support
Basic principles
Fluid resuscitation
Effects of intravenous fluids on the
brain
16.6 Metabolic response to injury
16.10
16.11
17
272
273
276
280
281
293
293
294
302
305
307
20
338
340
343
357
358
Sedation and anesthesia

Guy L. Ludbrook
363
18.1
18.2
18.3
18.4
18.5
18.6
18.7
363
363
363
367
369
371
Management of intracranial pressure and

cerebral perfusion
Peter Reilly
Introduction
Intracranial pressure
Cerebral perfusion pressure
Intracranial volumes
Volumepressure relations
Principles of management
Treatment
Surgical treatment
Plan of management of raised ICP
and reduced CPP
19.10 References
293
326
327
327
333
334
Introduction
Physiology
Effects of head injury upon oxygen
delivery
Management: resuscitation
Management: maintenance of cardiopulmonarycerebral homeostasis
Conclusion
References
19.1
19.2
19.3
19.4
19.5
19.6
19.7
19.8
19.9
281
283
289
290
315
17.1
17.2
17.3
Introduction
Cerebral pharmacology
Intravenous anesthetic agents
Muscle relaxants
Inhaled anesthetic agents
Other agents used in neuroanesthesia
Conduct of anesthesia in severe
head injury
18.8 Emergence and recovery
18.9 References
19
312
333
17.6
17.7
18
Fluid therapy in uncomplicated postoperative and post-traumatic states

Disorders of water and electrolyte
balance
Special fluid and electrolyte
problems in the neurosurgical patient
Conclusions
References
Respiratory and cardiovascular support

John A. Myburgh
17.4
17.5
261
Introduction
15.9
15.10
15.11
16.9
261
269
15.4
15.5
15.6
15.7
15.8
16.8
243
244
245
247
PART THREE: TREATMENT
General aspects of trauma

Management at the accident site
Transportation to and between
hospitals
The Level 1 trauma center
Care of the patient in the hospital
Setting surgical priorities
Radiological examination
The effects of concurrent medical
conditions
Transport modalities
Summary
References
16.7
243
vii
373
379
380
385
385
385
388
388
388
390
393
400
401
405
The role of surgery for intracranial mass

lesions after head injury
Nigel Jones, Ross Bullock and Peter Reilly
409
20.1
20.2
409
409
Introduction
Post-traumatic lesions on CT
viii
CONTENTS
20.3
Indications for evacuation of

intracranial hematomas
20.4 Techniques for craniotomy
20.5 Specific surgical problems
20.6 References
21
Neuroprotection in head injury

Graham M. Teasdale and Paul E. Bannan
423
21.1
21.2
423
Introduction
From preclinical research to clinical
benefit
21.3 Treatments undergoing clinical
evaluation
21.4 Conclusion
21.5 References
22
409
411
416
421
424
425
435
436
Outcome after severe head injury

Bryan Jennett
439
22.1
22.2
22.3
22.4
22.5
439
439
442
443
444

Glasgow Outcome Scale
When to assess outcome
Ethical issues
Brain death
22.6
22.7
22.8
22.9
22.10
22.11
22.12
22.13
22.14
The vegetative state

Neurophysical sequelae in
conscious survivors
Cerebral hemispheres
Cranial nerve deficits
Delayed complications
Mental sequelae
Deficits of intellectual (cognitive)
function
Personality change
References
Appendices
447
449
449
450
451
454
454
457
459
463
A Antibiotics recommended for infections of

the CNS
B Seizure management in acute head injury
C Possible causes of status epilepticus after
head injury
D Cardiovascular drugs used for augmentation
of cerebral perfusion pressure
465
Index
467
463
464
464
CONTRIBUTORS
Robert W. G. Anderson BE
Road Accident Research Unit,
University of Adelaide,
Adelaide,
Australia
Michael R. Fearnside MS, FRACS

Department of Neurosurgery,
Westmead Hospital,
Westmead,
Australia
Paul E. Bannan FRACS

Royal Perth Hospital,
Perth,
Australia
J. E. Gilligan FANZCA, FFICANZCA

Intensive Care Unit,
Royal Adelaide Hospital,
Adelaide,
Australia
Peter C. Blumbergs FRACP, FRCPA

Department of Neuropathology,
Institute of Medical and Veterinary Sciences,
Adelaide,
Australia
Donald M. Hadley PhD, FRCR

Department of Neuroradiology,
Institute of Neurological Sciences,
Southern General Hospital,
Glasgow,
UK
Ross Bullock MD, PhD

Division of Neurological Surgery,
Medical College of Virginia,
Richmond,
Virginia,
USA
Bryan Jennett CBE, MD, FRCS

Glasgow,
UK
Kwan-Hon Chan MS, FRCS

University of Hong Kong,
Queen Mary Hospital,
Hong Kong
Nigel Jones DPhil, FRACS

Adelaide,
Australia
Peter J. Crawford BSc, FRCS

Newcastle General Hospital,
Newcastle upon Tyne,
UK
Peter J. Kirkpatrick BSc, MSc, FRCS(SN)

Academic Department of Neurosurgery,
Addenbrookes Hospital,
Cambridge,
UK
CONTRIBUTORS
Guy L. Ludbrook FANZCA

Department of Anaesthesia,
Adelaide,
Australia
A.J. McLean BE, ME, SMHygSD
Adelaide,
Australia
A. David Mendelow PhD, FRCS(SN)
Newcastle General Hospital,
Newcastle upon Tyne,
UK
R. J. Moulton MD, FRCS(C)
Division of Neurosurgery,
University of Toronto,
Toronto,
Canada
J. Paul Muizelaar MD
Richmond,
Virginia,
USA
John A. Myburgh DA(SA), FANZCA, FFICANZCA
Adelaide,
Australia
Brian North RFD, FRCS, FRACS
Adelaide,
Australia
Ian Piper PhD
Department of Clinical Physics,
Glasgow,
UK
Peter L. Reilly MD, BMedSc, FRACS

Adelaide,
Australia
Donald A. Simpson AM, MS, FRCS, FRACS
Adelaide,
Australia
Evelyn Teasdale MRCP, FRCR
Department of Neuroradiology,
Glasgow,
UK
Graham M. Teasdale FRCP, FRCS (Edinb. and Glasg.)
Glasgow,
UK
Peter D. Thomas FRACP, FANZCA, FFICANZCA
Adelaide,
Australia
Robert Vink PhD
Department of Physiology and Pharmacology,
James Cook University of North Queensland,
Townsville,
Australia
Alois Zauner MD
Richmond,
Virginia,
USA
PREFACE
Brain injury remains one of the most difficult and

challenging problems facing many clinicians, particularly neurosurgeons and intensivists. In major trauma
centers there has been a steady decline in the mortality
rate due to severe head injury of about 10% per decade
since the mid 1970s, but this does not seem to be
reflected in an overall population-based decline, even
in developed countries (Jennett et al., 1977; Klauber et
al., 1989; Marshall et al., 1991; Waxweiler et al., 1996).
This suggests that a more widespread application of
modern principles of neurotrauma care could save
many lives throughout the world. It is the goal of this
book to provide a theoretical and practical foundation
upon which such care can be based.
We wish to thank our colleagues who have contributed their tremendous expertise to these chapters. We
owe a strong debt of gratitude to the Glasgow school,
and in particular to Bryan Jennett and Graham
Teasdale, both of whom have contributed fundamentally to our own views on head injury and indeed to
all neurosurgeons. We also wish to acknowledge the

outstanding contribution to neurotrauma made by our
friend the late Douglas Miller. The originality and
clarity of his thoughts and words have challenged and
enlightened all those seeking to understand more
clearly the complexity of brain injury.
Peter Reilly and Ross Bullock
Adelaide, South Australia and Richmond, Virginia,
December 1996
References
Jennett, B., Teasdale, G., Gailbraith, S. et al. (1977) Severe head injuries in 3
countries. Journal of Neurology, Neurosurgery and Psychiatry, 40, 291298.
Klauber, M. R., Marshall, L. F., Leursen, T. G. et al. (1989) Determinants of head
injury mortality: importance of the low risk patient. Neurosurgery, 24,
3136.
Marshall, L. F., Gautile, T., Klauber, M. R. et al. (1991) The outcome of closed
head injury. Journal of Neurosurgery, 85, 528536.
Waxweiler, R., Thurman, D., Spiezek, J. et al. (1996) Monitoring the impact of
traumatic brain injury: a review and update, in Traumatic Brain Injury:
Bioscience and Mechanics, (eds F. A. Bandak, R. H. Eppinger and A. K.
Ommaya), M. A. Liebert, New York, pp. 18.
INDEX
Numbers in italics refer to tables; numbers in bold refer to illustrations

Abbreviated injury scale (AIS) 78
Abdominal
examination 279
investigations 281
paradox 338
Acceleration of head 278, 30, 31, 323, 34,
41, 45, 59
duration of force 30, 312, 121
Penn 1 and 2 impact model 121, 126
Accident
insurance, Australia 16
site
management at 2712
transfers from, controlling ICP and
CPP 393
ACE inhibitors 317, 323
Acetazolamide 90, 254, 388
Acidosis, cerebral 131
Acute
hyponatremia 317
treatment 319
lung injury 345, 349, 350
physiological and chronic health
evaluation score (APACHE II) 282
respiratory distress syndrome (ARDS)
3501
subdural hematoma see Subdural
hematoma
Adenosine 91
Adenosine triphosphate (ATP) 91, 128
Adrenalin 356, 357, 372, 373
Adrenoreceptors 94
Advance directives 444
Advanced
Pediatric Life Support program (APLS)
72
Trauma Life Support system (ATLS) 71,
72, 271, 289
Aeromedical service 273, 283
Aerosol nebulizers 348
Afterload 337
Age
age-dependent distribution, ICP 102
age-specific mortality rate 4
biomechanical effects and 1236
and prognosis 163
reaction to injury 459
Agitated patients 341
Agonist-gated ion channels 124, 125, 133,

134, 432
Airbags 18, 19
Airliners, passenger 285, 286, 287
Airway management
after initial resuscitation 344
initial resuscitation 277, 3402
Akinetic mutism 159
Albumin infusions 296, 2989, 303, 342, 356
Alcohol
abuse 283
coagulopathy 420, 421
control 17
intoxication, subarachnoid hemorrhage
61
levels, GCS depression 150
Alfentanil 367, 380
Alpha
coma 232
frequency, return of 234
Altitude, effects of 2889
Alveolar ventilation (VA) 3356
Alveoli
distribution of O2 from atmosphere
3346
O2 delivery from to tissues 3367
overdistension 345
Alzheimers disease 65
Ambient pressure, effect of altitude 289
Ambulance officers case report 2745, 276
American Association of Neurological
Surgeons 71, 269
American Brain Injury Consortium 436
American College of Surgeons
Committee on Trauma 150
fluid management recommendations 305
Amiloride 323
Amino acids 307, 308
required for protein synthesis 313
Amino steroids 399, 429
Amnesia 14, 21, 44, 47, 1567, 15961, 454,
456
AMPA-receptor antagonists 83, 432
Amplitude transfer function, systems
analysis 113, 114, 115
Amyloid precursor protein (APP) markers
42, 43, 45, 47
Anaerobic metabolism 91, 128, 129, 137, 301
Anal tone, lax 279

Analgesia 286, 342, 37980
at accident site 272
opioids 309, 343, 345, 367, 373, 374
Anemia 298
Anesthesia
assessment and preparation 373
emergence and recovery 37980
induction 341, 3734
inhaled anesthetic agents 36971
intravenous agents 3637
maintenance 3747
Aneurysms 85, 176
Angiography 173, 201
brain death 206
internal carotid dissection 201, 203
traumatic aneurysm 176
Angle of isonation (TCD) 244, 245
Angular accleration of head 278, 30, 31,
33, 34, 45, 59
boxing 65
role in injury 323
Anosmia 450
smell testing 157
ANP, sodium regulation 312
Anterior cerebral arteries, TCD 245, 246
Antibiotics 355
for CNS infections 463
dural fistula 420
empirical use 350
Anticoagulants, low dose 3512
aspirin therapy 421
Anticonvulsants 3789
discontinuation 452
prophylactic 453, 454
Antidiuretic hormone (ADH), release of
309, 311, 312
Antilock electronic braking systems 18
Anxiety 454
Apallic state 159
Apathy 457, 458
Apnea 338
Apoptosis 135, 137
Apparent diffusion coefficients (ADC) 82
Appendices 4634
Aptiganel HCH 434
Arachidonic acid metabolism, modulators
of 399, 430
468
INDEX
Arachnoid
granulations 388
tears 60
Arm weakness 155
Arterial
catheterization, pressure measurement
343
gas tension 901
see also PaCO2; PaO2
oxygen content (CaO2) 3367
Arteriovenous oxygen difference (AVDO2)
95, 96, 144, 218, 222
fall in, hyperventilation 395
Aspiration, gastric contents 349, 373, 374
Aspirin therapy 421
Assisted mechanical ventilation 346
Asthma 282, 333
Astrocytes 92, 94, 124, 128
anaerobic glycolysis 12930
swelling of 131, 132, 206, 266
Astrocytic end feet, swelling of 126, 138
Ataxia 158, 441
Ataxic respiration 338
Atracurium 368
Attention, tests of 455
Augmentation duraplasty 417
Australia
compulsory insurance scheme 16
mortality 6, 7
children and infants 13
population-based study 10
Auto PEEP 3467
Autoregulation 97, 143, 294, 388
and cerebral blood flow (CBF) 91
cerebrovascular pressure transmission
215
impairment of 74, 8990, 104, 115, 298, 365
lower limit 393
therapeutic window 253
transcranial Doppler assessment 2502,
253
Autoregulatory threshold 252, 253
Avitene 415
AVPU coma scale 150
Axonal injury 416, 80
effect of shear forces 1256
quantitation 40
see also Diffuse: axonal injury (DAI)
Axonal injury sector score (AISS) 40, 44
Axoplasmic stasis 126
Ayalas index 106
Back, examination 279
Barbiturates 97, 133, 139, 355, 3656, 3979
actions of 90, 398
clinical evaluation 4278
complications and contraindications 398
dose and duration 3989
indications 3978
monitoring 398
pupillary responses 398
SSEP responses 239
Barotrauma 345, 346
Basal
cerebral arteries, Doppler signals 244
circulatory arrest 255
see also Middle cerebral arteries
cisterns, imaging 200
ganglia hemorrhages 567, 185
imaging, CT 187, 188, 189, 206
Base of skull fractures 26, 55, 62, 84, 156, 374
cribriform plate 340, 344
diabetes insipidus 320
infection 84
internal carotid artery injury 85
petrous 84, 156, 237, 451

sphenoid 168, 173, 451
surgery 420
Basilar artery 89
ischemia 201, 203
Battles sign 84, 156
Bedside nursing charts 105
Behavioural disorders 454
frontal lobe syndromes 456
Benign intracranial hypertension 386
Benzodiazeines 345, 365, 3667, 379
Beta A4 deposits 65
Beta-blockers 339, 357, 372
Bicoronal flap 413, 415
Bifrontal craniotomies 417, 418
Bimolecular leaflet cell membrane 124
Biomechanics 257, 1213
brain injury mechanisms 2934
effects of age 1236
experimental studies 289
tolerance of head to impact 346
Blindness, cortical 450
Blood
alcohol concentration (BAC) 17
flow velocity (FV) see Flow velocity (FV)
pressure 1556
arterial catheterization 343
during anesthesia 3756
mean arterial (MAP) 303, 337, 343, 375
relationship with ICP 105
see also Hypertension; Hypotension
transfusions 298, 337, 342
coagulopathy 420
hemorrhagic shock models 305
Blood-brain barrier (BBB)
astrocytic membranes 124
breakdown 61, 73, 76, 80, 306
imaging 180
transient opening 61, 138, 266
effect on brain metabolism 92
physiology 297
Bohr equation, modified 335
Boxing injuries 12, 656, 127
Brachial plexus paralysis 155
Bradycardias 339
Bradykinin antagonists 435
Brain
abscess 845
pH 90, 131, 2634, 265
swelling 40, 4850, 803, 1389, 385,
388, 390, 392, 409
children 823, 387
compliance measurements 389
intraoperative 138, 139, 204, 398,
41718, 419
lobectomy 417
resolution 139
surgical decompression 417
volume 388
Brain death 4447
diabetes insipidus 320
diagnosis 4456
effects of barbiturate treatment 398
oculovestibular reflex 154
radiology 206
transcranial Doppler 255
validity of criteria 4467
withdrawal of treatment 443, 444
Brain injury
assessment of severity 40, 15864
biomechanics of 2536
effects of age 1236
definition 5
mechanism of 401
pattern of 392
primary 39, 40, 713, 385

relationship to raised ICP 1025
see also Contusions; Diffuse axonal
injury (DAI); Extradural
hematoma; Intracerebral
hemorrhage; Lacerations;
Subdural hematoma
secondary 39, 40, 63, 64, 1356, 266, 357,
385, 423
extracranial 735
intracranial 757
relationship to raised ICP 1025
see also Brain: swelling; Hypoxicischemic brain injury; Ischemia
Brain Injury Foundation 269
Brain-stem
auditory evoked potentials (BAEPs) 229,
2367, 239
depression, reversible causes 445
failure 445
lesions 614, 340
arterial hypotension 156
clinical signs 154
focal 434
imaging 175, 178
pyrexia 85
syndrome 449
Breathing patterns, head injured patients
338
Bridging veins 32, 122
rupture 34, 41, 65, 126, 127
acute subdural hematoma 59
Burns 45
Burrholes, exploratory 79, 41113
Burst lobes 5960, 128, 190
frontal 59
temporal 55, 59
Burst suppression, EEG 232
barbiturates 397, 398, 427
Cabin pressurization 288
Calcarine cortex, lesions of 450
Calcium
channel antagonists 75, 76, 77, 134,
4302, 435
influx 81, 1334, 137
prevention of 75, 77
levels, spinal injuries 327
Calories 313
Calorimetry, indirect 3378
Camino transducer 212, 213
Canada, mortality 7
Capillary
density 89
hydrostatic pressure 306
Capnography 335, 341, 375
Carbamazepine 454
Carbon dioxide 90, 91, 923
control of 3478
see also Hypercapnia;
Hyperventilation; Hypocapnia
production 335
reactivity 90, 97, 253, 377, 388, 398
hyperventilation 394
propofol 364
TCD assessment 254
see also PaCO2
Carbon monoxide poisoning 283
Cardiac
arrest 48, 111, 444
dysrhythmias 339
output 337
response to acute injury 338
in shock 302
tamponade 279
INDEX
Cardiogenic
edema 354
shock 356
Cardiorespiratory
arrest, brain-stem death 445
instability, MRI 171
parameters 1556
physiology 3348
Cardiovascular
conditions, pre-existing 282
support 33361
Carotid
arteries 89
occlusions 85
bifurcation, Doppler tracing 245, 246
cavernous fistula 85
Case fatality rate 4
population-based studies 9
Catecholamines 94, 356, 357
Catheter-tip transducers, ICP monitoring
212, 213, 215
Causation, head injury 1012
Cell
cell membranes 295, 297
and ion channels 1234
function and injury 12141
Central
pontine myelinolysis 318
venous catheterization 353
complications and contraindications
354
venous pressure (CVP) 303, 343, 352, 376
Centroid, high-frequency 114
Cephalic reference electrodes, EEG 230
Cerebellar coning 63, 80
Cerebral
atrophy 60, 123, 137, 185, 204, 457
compression, stages of 105, 106
edema see Edema: cerebral
hemispheres
injury to dominant 157
sequelae after injury to 44950, 4556
swelling of one 49
perfusion
electrical activity as marker of 103
see also Cerebral perfusion pressure
(CPP)
pharmacology 363
resistance vessels 394
Cerebral blood flow (CBF) 89, 102, 243
effects of hematocrit on 90
flow probes 144
following head injury 957
flow-metabolism uncoupling 96, 131,
133
measurements
clinical methods 21718
indirect methods 2223
laser Doppler flowmetry 221
MRI 2212
overview 21718
relationship to MCA blood flow 245
thermal diffusion technique 221
xenon 21821
neurotransmitters and 94
normal values 94
perfusion imaging 266
regulation of 901
relationship with CPP and ICP 1034,
109
studies 47, 74
mechanism of tissue damage 1367
Cerebral blood volume (CBV) 96, 388, 389
dynamic CT technique 2201
Cerebral function monitor 232
Cerebral metabolism 914

cerebral blood flow regulation 901
see also Cerebral blood flow (CBF)
cerebral hemodynamics 8990
following head injury 957
magnetic resonance spectroscopy (MRS)
2618
measurements 21719
intracerebral microdialysis 2234
positron emission tomography 224,
225, 226
normal values 945
Cerebral perfusion pressure (CPP) 74, 294,
354, 356, 385
children 83
critical threshold 115
definition 388
during anesthesia 376
pressure-passive 357
relationship with ICP 8990, 386, 3901
and CBF 1034, 109
trancranial Doppler 115, 250, 251, 2523
treatment for reduced
goals of treatment 392
optimal CPP 356
protocols 4024
ventilation 354
Cerebromalacia 180, 182, 184
Cerebrospinal fluid (CSF)
contribution to raised ICP 109, 387
displacement of 389
lactate concentration 264
leakage 156, 197, 420
outflow resistance 107, 1089, 116
pressure wave form 214
removal 106
ventricular catheters 394
spinal pressure as measure of ICP 209
volume 388
Cerebrovascular
function, effects of therapy on 3545
pressure transmission analysis 21415
reactivity 90, 97, 253, 254
resistance 8990, 102
Cervical spine
imaging 167, 168, 170, 171, 281
injury
abdominal paradoxical breathing 338
endotracheal intubation 277, 278, 341,
374
fractures 272
Chemical shift, MRS 262
Chest X-ray 168, 281
ARDS 350, 351
Cheyne-Stokes breathing 338
Chiasmal lesions 450
Child abuse 32, 78
retinal hemorrhages 155
shaken baby syndrome 204
Children
brain death, diagnosing 446
brain swelling 49, 823
cause of injury 11, 1214
EEG measurements 231
extradural hemorrhage 59, 79
hyperemia 83, 254
outcome 163
epilepsy risk 453
final examination 164
pediatric coma scales 1501
proton MRS study 263
raised ICP 102
pulse rate 155
retrospective PTA measurements 160
seat belts 14, 19
469
testing cognition 157

xenon CBF measurements 945
see also Infants
Childrens Orientation and Amnesia Test
(COAT) 161
Chlorpromazine 309
Chronic
hyponatremia 317
treatment 317, 318, 319
obstructive pulmonary disease (COPD)
282, 333
subdural hematomas 60, 79
venous extradural hematomas 78
Chylomicrons 313
Circle of Willis 63, 89
Classification, head injury 56
Clinical
evaluation
applications 161, 163
treatments undergoing 42536
examination 14558, 2767
definitive 145, 1568
initial 14556
and outcome 1634
roles and limitations 145
topic reports 4
Closed
barrier edema see Cytotoxic edema
head injury, biomechanics see
Biomechanics: brain injury
mechanisms
Clothing
protective 276
removal of patients 278
Coagulopathy 79, 4201
congenital 283
diffuse intraoperative bleeding 419
hypothermia 400
Codeine phosphate 380
Codman MicroSensor 212, 213
Cognition
deficits of intellectual function 442,
4547
magnesium homeostasis 264
tests of 157
conclusions about 457
Collateral vessels 89
Colloid
osmotic pressure (COP) 295, 306
solutions 278, 298, 3023, 342, 356, 377
effect on normal brain 306
effects after brain injury 3067
synthetic 299, 300
versus crystalloid solutions 3045
Coma 40, 148, 149
duration of 146, 159, 163
EEG measurements 2302
power spectral analysis 2334
hyperglycemic hyperosmolar non-ketotic
327
loss of consciousness 122, 136
Coma scales 146, 14750
advantages and disadvantages 144
pediatric 1501
Community Trauma Center 20
Compartment syndromes 281
Compliant cavities, gas expansion 288, 289,
369
Compound depressed fractures 83, 84
risk of epilepsy 452, 453
Compounds under evaluation 430, 433,
434, 435
Computed tomography (CT) 712, 77, 143,
145, 167, 16971, 206, 343, 409
abdominal injuries 281
470
INDEX
Computed tomography (CT) contd

atrophy, correlation with performance
IQ 457
bifrontal decompressive craniotomy 418
blood-brain barrier (BBB) breakdown
180
brain death 206
brain edema 812
brain engorgement studies 139
cervical spine 168
scout film 170
contusions 174, 177, 180, 1813, 184, 187,
198, 418
diagnostic categories, types of
abnormality 409, 410
diffuse brain injury 40, 57
extradural hematomas 58, 186, 189, 190,
191, 194, 414
indications 169, 171
intracerebral hematomas 173, 174-5, 176,
177
raised ICP 391, 392
scout films 170, 413, 414
subarachnoid hemorrhage 174, 192, 194,
195, 196
subdural hematomas 82, 175, 178, 181,
184, 191, 192, 193, 194, 195, 197
three-dimensional (3D) reconstruction
168, 169
transtentorial herniation 199
vegetative state 447
xenon techniques 945, 96, 97, 217,
21921, 224
Concurrent medical conditions 2813, 290,
333
Concussion 30, 31, 32, 122
ICD-9 CM rubric 6
mitochondrial changes 80
MRI 180, 186
syndromes of 41, 467
Confusion, post-coma disturbance (PCD)
161
Congenital coagulopathies 283
Congestive brain swelling 48, 49, 80, 387
Coning 63, 80
fall in CPP 257
Consciousness
ICD-9 CM rubric 6
level of 146, 147, 148, 158
loss of 122, 136
Consensual light reflex 151, 153
Contact injuries 41
Continuous positive airway pressure
(CPAP) 347
Contralateral hematomas 418
Contrecoup injuries 267, 41
contusions 53, 54
Controlled ventilation 277
Contusion
hemorrhages 51
index 40
necrosis 51, 52
Contusions 72, 122, 126, 197, 410
boxing 65
brain-stem 62
clinicopathological correlations 55
experimental injuries 545
ICD-9 CM rubric 6
imaging 173, 180, 200
CT 174, 177, 1813, 184, 187, 198, 418
MRI 172, 179, 184
pathology 514
removal 416
swelling around 49
Convective oxygen transport 336
Core body temperature 376, 377

effect on ICP 402
Corectopia 153
Corneal reflex 158
Corpus callosum, focal lesions 42, 43, 46
Cortical
blindness 450
blood flow measurements 221
electrical activity 1023
hemorrhages 52
Corticosteroids 138
Costs 1516
ICP monitoring 213
Count, epidemiology 3, 4
Coup contusions 41, 53, 62
Cranial nerve
function 154
testing 158
palsies 79, 152, 153, 441, 4501
Craniectomy 412, 413, 415, 417
Craniomaxillofacial injury 4
Craniospinal volume and pressure
relationships 106, 109, 110, 111
mathematical model 107
Craniotomy 293, 409
acute subdural hematoma 41516
bifrontal 417, 418
exploratory burrholes 41113
extradural hematoma 41315
for ICP reduction 400, 417
preoperative preparation 411
Cribriform plate fractures 340, 344
Cricothyrotomy 277, 278, 341
Crush
injuries 121
syndrome 279
Crystalloid solutions 298, 300, 301, 3023,
377
effects after brain injury 3067
effects on normal brain 306
versus colloid solutions 3045
CSF see Cerebrospinal fluid (CSF)
CT see Computed tomography (CT)
Curare 368
Cushing response 74, 339, 389, 394
Cyclists see Pedal cycle accidents
Cyclooxygenase inhibitors 430
Cytokines 309
systemic inflammatory response
syndrome (SIRS) 338
Cytotoxic edema 49, 73, 80, 81, 82, 131,
138, 139, 180, 200, 266, 390
Data
bias in mortality 3
collection, head injury surveillance 20
Dead space 335
Death
certificates 7
genes 435
Deceleration injury 121
Deep vein thrombosis 351
Definitions
epidemiology 3, 4
head injury 45
Definitive examination 145, 1568
organization 156
timing 156
Dehydration 355
mannitol treatment 396, 397
Delayed
complications, outcome 4514
intracerebral hemorrhage 58, 175, 401
Delta activity, EEG 234
Dementia pugilistica 12, 65

Denitrogenation, encysted air 289
Depolarizing muscle relaxants 3678
Depression 454, 458, 459
Desflurane 363, 371
Dexamethasone 81, 428
Dextrans 299, 303, 306
Dextrose solutions 300, 301
Diabetes 282, 377
insipidus 308, 312, 320
treatment 321
mellitus 323
Diazepam 3667, 379
Diffuse
axonal injury (DAI) 40, 41, 426, 57, 72,
73, 121, 122, 1256, 136
boxing 65
brain-stem 63
clinicopathological features 44
grading 44
imaging 180, 185, 186, 187, 188, 189
immunocytochemical markers 445
in infancy 45
late sequelae 46
macroscopic markers 434
magnetization transfer imaging 266
microscopic diagnosis 423
non-traumatic 46
pathogenesis 456
vegetative state 645, 448
intraoperative bleeding
management 41920
prevention 418, 419
vascular injury 40, 57, 61
Diffusion gradient 336
Diffusion-weighted imaging (MRI) 81, 82,
266
Diffusive oxygen transport 336, 337
Diplopia 450
Disability Rating Scale 441
Disinhibition 456
Disseminated intravascular coagulation
(DIC) 420
Dizocilpine (MK801) 434
Dobutamine 356, 357
Documentation and charting
ambulance officers case report 2745,
276
during patient transportation 286
neurological observation sheet 147
see also Coma scales
Dolls eye test 152, 154
Donnan equilibrium effect 295
Dopamine 356, 357
receptors (D2) 94
Doppler shift 244
Dr ABC acronym 271
Driver behavior 19
Drug overdose, EEG measurements 232
Drugs
affecting response to trauma 282
illicit use 283
Dura, opening 413
Dural
fistula 420
hitching sutures 415
sinus pressure 387
venous sinus thrombosis 202
Duraplasty 400, 416, 417
Duret hematoma 175, 178
Dynamometry 158
Dysarthria, testing for 157
Dysphagia 158
Dysphasia 440, 450
testing for 157
INDEX
Ectopic pupil 153
Edema
cardiogenic 354
cerebral 52, 73, 803, 108, 1389, 355,
390
CT appearance 171
diffusion-weighted imaging (MRI) 266
hypertonic solutions 302
pathophysiology 82
plasma osmotic pressure 306
treatment 81, 83
types of 49
gastrointestinal tract 305
interstitial 49, 812, 301, 337
peripheral 305
pulmonary 296, 305, 345
neurogenic 3267, 340, 351, 354
Eighth cranial nerve damage 451
Elastance 107
brain tissue 111, 38990
craniospinal system 106, 109, 110
Elderly patients 123, 238
water replacement 312
Electrical
function monitoring
goals 229
methods and modalities 23041
problems and limitations 22930
noise 230
Electrocardiography (ECG) 339, 343, 375
Electrocerebral silence 232
Electroencephalography (EEG) 229, 2304
brain death 446
Electrolytes 293, 294
disorders of water and electrolyte
balance 31526
effect of stress on movement 309
ICF and ECF 2945
osmolality and tonicity 2978
requirements 31213
solutions of 3007
composition 300
see also Fluids
Eliprodril (SL820715) 434
Emergence from anesthesia 37980
Emotional lability 458
Emphysema, subcutaneous 279
Enadoline 435
End-tidal CO2 335, 341
Endothelial
cells 92, 299
membrane 61, 62
microvacuolation 126
Endothelium-derived relaxing factor
(ERDF) 91
Endotracheal intubation 145, 149, 159,
2778, 3401, 344, 374
preparation and requirements 278
Enflurane 370
Enteral nutrition 314
Epidemiology 323
definitions 3, 4
Epidural hemorrhage see Extradural
hemorrhage
Epilepsy
post-traumatic 65, 441, 451, 4524
prediction of fits 452, 453
pre-existing 333
status epilepticus 48, 232, 365, 464
see also Seizures
Epinephrine 356, 357, 372, 373
Esmolol 357, 372
Ethacrynic acid 397
Ethanol level, GCS depression 150

Ethmoidectomy 420
Etomidate 366, 374, 397
European Brain Injury Consortium 436
Evoked potentials 229, 231, 234, 23540
advantages and disavantages 144
with ICP measurements, outcome 103
multimodality monitoring 23940
technical principles 234
Excitory amino acids (EAA) 81, 125, 127,
223
Exercise 314
Extracellular fluids (ECF) 294, 295
of brain parenchyma 385, 386
hyponatremia 315, 316, 317
sodium movements, response to stress
309, 31112
tonicity 297
Extradural hematoma 77, 78, 79, 4001
asymptomatic patients 410
conservative therapy 410
effect of operative delay 75
ICD-9 CM rubric 6
imaging 59, 185, 186, 188
CT 189, 190, 191, 194, 414
pathology 589
surgery
craniectomy 413
craniotomy 41315
exploratory burrholes 41112
Extrinsic PEEP 347
Eye
movements 154, 450
opening
eliciting 148, 149
return of after coma 159
Facial
fractures 168
nerve, testing 158
palsy 451
Faciomaxillary trauma 4, 340
Falls
children 12, 13
diffuse axonal injury (DAI) 44
Family see Relatives
Fast Fourier transform alogorithm (FFT) 233
Fat use in starvation 307
Fentanyl 343, 345, 367, 373, 374, 380
Fibreoptic
intubation 277, 278
sensors 144
Fick equation 3378
Fighting the ventilator 3445
Fistuale 85, 420
Fixed dilated pupils 151, 152
Fixed-wing aircraft 283, 284, 285, 2878, 290
Flaccid arm or leg weakness 155
Flail chest 279, 338
Flexion, abnormal 149
Flow velocity (FV) data 115, 243, 244, 245
abnormally high 254
CPP assessment 252
derivates of 247
oscillating flow pattern, brain death 255
raw baseline 24950
relationship between MCA FV and CBF
245
Fluid challenge technique 304, 352
Fluids 293
abnormal losses 312
disorders of water and electrolyte
balance 31526
distribution in body 2947
of administered fluids 298302
471
effects
on brain, i.v. fluids 3057
of stress on movement of 309, 31112
fluid therapy 31215
resuscitation 272, 3025
special problems 3267
see also Electrolytes
Flumazenil 367
Focal brain injuries 39, 40, 1223, 4545
brain stem 434
cerebral contusions 513
see also Contusions
corpus callosum 42, 43
Foraminal impaction see Cerebellar coning
Force of impact 27
Fractional concentration of oxygen (FiO2)
334
Fracture contusions 53
Fractures 77, 411
brain injury severity 31
compound depressed 83, 84
risk of epilepsy 452, 453
infection after 834
open depressed 55
radiological appearance 1678
see also Base of skull fractures
Free
magnesium, MRS studies 264, 265
radicals 81, 134
scavengers of 81, 83, 299, 4289
transcellular fluid 295
Freeways 1718
Frontal
impacts 27, 28, 31, 35, 54
mathematical modelling, pressure
waves 122, 123
motor vehicles 11, 18
optic nerve injury 152
lobe syndromes 456
Fronto-orbital blink reflex 154
Frusemide (furosamide) 317, 355, 371, 388,
396, 397
Functional MRI 2667
Fundal examination 154, 155
Gadd Severity Index 35
Gait, assessing 158
Galveston Orientation and Amnesia Test
(GOAT) 160, 161
Gamma hydroxbutyrate 399
Gamma-aminobutyric acid (GABA) 94
Gas expansion, cavities and spaces
effects of altitude 288
nitrous oxide 369
Gastric feeding 314
Gastrointestional tract, edema 305
Gene expression, regulation and signalling
135
Glasgow Coma Scale (GCS) 8, 14, 40, 147,
148, 149, 150, 155, 163, 279
effect on outcome, intracranial
hematoma 77
Pediatric (PGCS) 151
variant forms 148
Glasgow motor scale 159
Glasgow Outcome Scale 43942
categories, reliability and validity 4401
Glasgow-Liège (GLS) score 154
Glial cells, energy consumption 92
Gliding contusions 53, 54, 122
imaging
CT 187, 198
MRI 180, 186
Global oxygen consumption 3378
472
INDEX
Glucose
administered 307, 308, 313, 377
see also Dextrose solutions
brain utilization 91, 92, 94, 95
after brain injury 93, 97, 128, 129, 130,
131, 225, 226
comatose patient 128
cerebral metabolic rate (CMRG) 95
transporter gene 135
Glutamate
receptors 432, 433
release 128, 223, 224, 293
Glyceral trinitrate 357, 372
Glycolytic pathway 90, 128, 307
Gosling Pulsatility Index 247
Gradient echo T2-weighted sequences,
MRI 171, 172
Grading of injury see Severity, trauma
Guedal airway 277
Guidelines, head injury management 71,
269, 3901
Haber-Weiss reaction 134
Haemaccel 299, 300, 303
Halothane 309, 36970, 374
Hand-held ventilation assemblies 342, 395
Hartmanns solution 278, 300, 301
Head elevation 350, 355
Head Injury Criterion (HIC) 31, 346
Health Interview Survey 14
Hearing
loss 451
testing 157
Heart
disease
myocardial injury 33940
pre-existing 333
rate 337
Heat shock protein HSP70 135
Helicopters 283, 284, 287, 290
Helmets 10, 30
bicycle 20
motorcycle 1920, 32
Hematocrit 375
effect on CBF 90
minimum acceptable 298
Heme oxygenase 135
Hemianopia
bitemporal 450
homonymous 79
Hemiparesis 411, 441, 449, 450
Hemodynamic
management in ICU 3527
indices from pulmonary artery
catheters 353
resuscitation 342
Hemoglobin 92, 298, 337
Hemorrhage
delayed 58, 175, 401
expansion, mannitol treatment 397
maturation, MRI appearance 176, 177, 179
Hemorrhagic
lesion score 40
shock models, fluid resuscitation 3045,
3067
Hemosiderin 176, 180
Herniation 7980, 388, 389, 391
coning 63, 80, 257
contusions 53, 54
imaging 197200
intracranial artery occlusion 85
relationship with ICP 386, 392
through craniotomy 417, 419
see also Tentorial herniation
Hetastarch 303, 306, 307

Hippocampus
ischemic neuronal loss 137
long-term potentiation 125
pathology 50
History 145, 146, 276
HIT II trauma trial 77
Hollow skull bolt 211
Holmes-Adie syndrome 152
Homatropine 152
Homonymous hemianopia 79
Horseback riding 12
Humidifcation, inspired gases 348
Hydralazine 372
Hydrocephalus 61, 78, 84, 199, 200
controlled CSF drainage 116
CT 175, 177
post-traumatic 451
systems analysis 114
Hydrogen ions, role of 134
Hydromas 60
Hydrostatic edema 49
Hyperbaric oxygen 400
Hypercalcemia, spinal injuries 327
Hypercapnia 90, 92, 106, 113, 252, 335
CBF increase 91
use in ARDS 351
Hyperemia 83, 967, 206, 254, 387, 390, 421
Hyperglycemia 282, 307, 377
Hyperglycemic hyperosmolar coma,
non-ketotic 327
Hyperkalemia 3234
Hypernatremia 312, 31922
clinical features 320
diagnosis 3201
hyperosmolar therapy 321, 322
treatment 321
Hyperphosphatemia 324
Hypertension 156, 282, 33940, 355, 375
benign intracranial 386
neurogenic 339
pre-existing 333, 392
pulmonary 336
severe pain 380
treatment 357, 372
Hyperthermia 279, 377
Hypertonic
hyponatremia 315, 316
saline 301, 302, 307, 342, 377, 399
acute hyponatremia 319
hypernatremia 322
Hyperventilation 75, 80, 156, 3478, 357,
3778, 386, 389, 394
children 83
combined with THAM 395, 426
indications for 342
prolonged 901, 97, 3478, 3945
prophylactic 3945
ventilation parameters 395
Hypoalbuminemia 299
Hypoaldosteronism 323
Hypocalcemia 3245
Hypocapnia 75, 80, 90, 92, 342, 347
Hypoglycemia 282, 377
Hypokalemia 3223, 325
mannitol treatment 397
Hypomagnesemia 324, 3256
Hyponatremia 75, 297, 298, 312, 315
clinical features 316, 317
diagnosis 317, 318
treatment 31719
Hypophosphatemia 325
Hypotension 73, 74, 102, 1034, 1556,
302, 342, 375
barbiturate therapy 398, 427
cerebral perfusion 104

mannitol treatment 397
nursing staff/computer identification 105
propofol 364, 365
refractory 342
thiopentone 366
Hypothalamic
control, pituitary secretion 308, 310, 312
lesions 339
Hypothermia 272, 279
effect of barbiturate treatment 398
intentional 85, 133, 138, 3767, 400, 423
clinical status 4256
rationale 425
Hypotonic hyponatremia 316
Hypoxia 73, 74, 106, 243, 342
Hypoxic-ischemic injury 40, 478, 65, 73,
74, 77, 81
see also Ischemia
Ibuprofen 430
ICP waveform pulse amplitude
(ICPUplseu) 11113, 115
Imaging 167207, 281
Immediate early (IE) gene expression 135
Immunological conditions, pre-exisitng 282
Impact to head 25
biochemical consequences 133
contusion patterns 53, 54
experimental studies 289, 313, 45, 124,
126
Penn 1 and 2 primate impact model
121, 126
impact velocity 256
location of impact 267
brain injury severity 27
recording 156
mathematical studies 29
observational studies 29
response of head 278, 122
tolerance of head 346
Incidence 3, 4
Indomethacin 399
Induction of anesthesia 3734
rapid sequence 341
Infants
pediatric coma scales 150, 151
premature neonates 123
prognosis 163
retinal hemorrhages 155
shaking 32, 204
Infarction 137, 155, 204
Infection 835, 333
central venous catheters 354
intraventricular pressure recording 211
systemic inflammatory response
syndrome (SIRS) 338
ventriculostomy-related 111
Inhaled anesthetic agents 36971
Injury severity score (ISS) 8, 282
Innerspace transducer 212, 213
Inotropes 356, 372, 373
Insight, lack of 458, 459
Inspired oxygen tension (PIO2) 3345
effects of altitude 2889
Insulin levels 308
Insurance, accident 16
Intensive care
computerized data collection system
104, 105
role of 3334
INDEX
Intermediate coup contusions 41, 53
Intermittent T-piece weaning 349
Internal
carotid artery
dissection 201, 203
occlusion 85, 155
jugular veins 89
International Classification of Diseases
(ICD) 56
Interstitial
COP 296
edema 49, 812, 301, 337
Intracarotid xenon-133 method 218
Intracellular
fluids 294
mechanisms 1338
Intracerebral
air, entrapped 289
hemorrhages 73, 78, 79, 417
delayed hematomas 58, 175, 401
imaging 17380
pathology 5563
surgery 41011
swelling around 49
Intracranial pressure (ICP) 3857
and CPP, relationship between 1034,
109, 386, 3901
effects of
barbiturates 365
head elevation 355
opioids 367
patient manipulations 355
PEEP 347
propofol 364
suxamethonium 368
volatile agents 369, 370
gradients 386
monitoring 144, 20916, 3912
choosing system 21314
criteria for 392
duration of 403
historical aspects 20910
interpretation 21415
methods 21013
Intracranial pressure (ICP), raised 48, 60,
1015
brain damage secondary to 634, 74, 75
causes of 385, 387, 390, 4012
children 83
clinical signs 155, 391
hippocampal pathology 50
imaging 197200
management 3902
controlling during transportation 393
definitive treatment 393400
plan of raised ICP and reduced CPP
4014
surgical 4001, 417
treatment goals 392
physiology and pathophysiology 10517
pressure wave fluctuations 106, 389
relationship
to outcome 1012
to primary and secondary injury 1025
Intracranial volume 385, 386, 388
measuring reserve 389
and pressure
current concepts 10717
historical concepts 1056
Intradural hemorrhage 789
see also Subdural hematoma
Intraocular air, entrapped 289
Intraoperative brain swelling 138, 139,
41718
barbiturates 398
Intravenous anesthetic agents 3637

Intraventricular
hemorrhage (IVH) 79, 416
imaging 177, 185, 187, 196, 197
pathology 57
pressure 106
recording 211
Intravoxel incoherent motion (IVIM) 222
Inverted horseshoe flap 413, 415
Ion channels, neuronal 1234, 125, 126,
133, 134, 432
Iowa Safety Restraint Assessment 1819
Ipatropium bromide 348
Ischemia 8990, 956, 102, 105, 131, 1356,
180, 243, 423, 424
genesis of 136
glucose infusions 301
hyperventilation causing 377, 395
hypothermia 376
imaging 201, 202, 203, 204, 205, 206
CT 171, 200, 201, 203
MRI 201, 202, 203
splanchnic 337
see also Hypoxic-ischemic injury
Ischemic threshold, perfusion imaging 266
Ischemic-reperfusion 47, 139
Isoflurane 370, 371
Isoprenaline 356, 357
Isotonic hyponatremia 315
Japan Automobile Research Institute
(JARI) 323
human head tolerance curve 36
Jet air ambulance 285
Jugular
bulb pressure 96, 109
venous saturation (SUjvuO2) 95, 96, 115,
218, 222, 243, 337, 357, 376, 390, 392
Julia Farr Centre Post-traumatic Amnesia
Scale 161, 162
Ketamine 366, 374, 433
Ketones 307
Kety and Schmidt equation 217, 219
Krebs cycle 128, 129, 307
Krypton-85 studies 217
Kyphoscoliosis 352
Laboratory data 280
Lacerations 51, 55, 72
boxing 65
brain-stem 62
ICD-9 CM rubric 6
Lacrimation, Schirmers test 158
Lactate
cerebral metabolic rate (CMRL) 95, 96
levels after brain injury 93, 128, 129,
1301, 137, 308, 338
MRS studies 2634
Lap belts 19
Laparotomy, emergency 280
Laryngeal trauma 340
Laser Doppler flowmetry 218, 221
Lateral gaze, forced 154
Lateral impacts 27, 28, 33, 54
motor vehicles 11
Law of Poiseuille 90
Le Fort III fractures 420
Learning capacity 455, 456
Leeds screw 211
Left
ventricular end-diastolic pressure and
volume 352, 353
ventricular preload 3523
Leg elevation 374
473
Legislative penalties 19
Leukocyte adherence, intravascular 126
Leukopenia 83
Lidocaine 399
Liège Reflex Scale 154
Lignocaine 373
Limb
injuries 281
movements and reflexes, testing 145,
155, 158
Lindegaard ratio 254
Linear acceleration of head 27, 30, 31, 33,
34
role in injury 323
Lipid peroxidation, inhibitors of 42830
Lipids, intravenous 313
Lobar intracerebral hemorrhages 55
Lobectomy 416, 417, 41920
Log-rolling technique 272
Lubeluzole 435
Lucid interval 44, 55, 56, 77, 78, 125, 136
Lumbar
CSF pressure 106
puncture 105, 209
Lumped craniospinal compliance 107, 109,
111, 112
Lung
cysts, effect of altitude 289
mechanics 336, 348
volumes, assessment 348
Lymphatic drainage 297
Magnesium
daily requirements 313
free magnesium, MRS studies 264, 265
levels, refractory hypokalemia 322, 323
Magnetic resonance imaging (MRI) 167,
1712, 206
abdominal injuries 281
brain death 206
brain edema 81, 82, 138, 139, 266
brain function studies 265
applications to trauma 266, 267
CBF measurements 218, 2212
contusions 172, 179, 180, 184
diffuse axonal injury (DAI) 185, 186
extradural hematomas 186, 188
intracerebral hematomas 76, 175, 179,
180
hemorrhage maturation, MRI
appearance 176, 177, 179
ischemia 201, 202, 203
mannitol studies 396
subarachnoid hemorrhage 196
subdural hematomas 172, 190, 191, 193,
194
transtentorial herniation 198, 199
white matter lesions, correlation with
neuropsychological tests 457
Magnetic resonance spectroscopy (MRS)
2618
principles of 2612
Magnetization transfer imaging 266
Maintenance of anesthesia 374
anesthetic agents 374
CVS homeostasis 3745
monitoring 3757
Major accident sites 272
Malnutrition 308, 315
Management of head injury 3434,
390404
accident site to trauma center 27191
fluids and electrolytes 293331
hemodynamic 3527
initial resuscitation 3403
474
INDEX
Management of head injury contd

limiting and withdrawing treatment
4434, 449
respiratory 34452
surgery 4001, 40921
treating raised ICP and reduced CPP
392, 393400, 4014
duration of monitoring 403
first-tier therapy 402
refractory intracranial hypertension 402
treatment algorithms 4023
Mannitol 97, 133, 139, 390
complications 397
dose 3967
during transfer between hospitals 393
hypernatremia 321, 322
modes of action 3545, 371, 396
effect on brain compliance 386, 389
effect on flow velocity (FV) 254, 255,
256
Marmarou weight-drop model 126
Mathematical
models
cerebrovascular bed 11617
CPP and autoregulatory status 253
CSF system 1078
frontal impact 122
studies, closed head injury 29
Maturation of hemorrhage 176, 177, 179
Maxillofacial trauma 4
Mean arterial pressure (MAP) 303, 337,
343, 375
Medevac 288
Medial parahippocampal gyri 64
Medical College of Virginia, staircase
protocol 402, 403
Medicolegal considerations
cerebral contusions 54
checklist of complaints 163, 164
external findings, recording 156
retrograde amnesia 160
Medullary coning see Cerebellar coning
Memory
defects 454, 455, 4567
tests 157
see also Alzheimers disease
Meningitis 84, 420, 451
Meningocerebral cicatrix 55
Mental
sequelae 439, 441, 4549
state 157
Meperidine 367
Metabolic
stress 307, 308, 309
Metabolism 914
clinical measurements 21718, 2236
intracerebral microdialysis 2234
magnetic resonance spectroscopy (MRS)
2618
management of
rationale 2934
see also Electrolytes; Fluids
normal values 945
response to injury 957, 12833, 30712
Metaraminol 356, 373
Methemoglobinuria 279
Methohexitone 366
Methylprednisolone 428, 435
Microchip transducers, implanted 212, 213,
215
Microdialysis, cerebral 144, 218, 2234
placement of probe 225
Microvasculature, effect of shear force 126
Midazolam 345, 367, 374, 379

Midbrain lesions, clinical signs 151, 152, 154
Middle cerebral arteries
distortion 57
flow velocity 243, 244, 245, 255, 392
correlation with CPP 115
derivatives of FV 247
high levels 254
relationship with cerebral blood flow
245
role in monitoring therapy 254
ischemia 201, 202
occlusions 85
Midline shift 40, 109, 391
Minor head injury
epidemiology 8, 1415
children 13
sports related 12
extradural hemorrhage 77
Glasgow Coma Scale 163
postconcussive sequelae 47
seizures 378
skull films 167
Missile injuries 123
Mivacurium 368
Modal analysis 334
Monitoring
comprehensive 226
duration of 143
during
anesthesia 3757
initial resuscitation 343
electrical function 22941
fluid management 303, 304
ICP 101, 102, 143, 20916, 3912, 403
multiparameter 92, 130
transcranial Doppler, role of 2545
volume status 3523
Morphine 309, 345, 367, 380
Mortality 67, 1013, 271, 442
age factors 163
bias in data 3, 7
predictors of 104, 150, 238
Motor
evoked potentials 103
responses 148, 149, 155
best motor score 159
Motor Accidents Authority 16
Motor vehicle
accidents 10, 12, 271
extrication from vehicle 272
patterns of injury 11
spinal injury 272
design 18
Motorcycle accidents 11
helmets 1920, 32
Mouth, examination of 340
Movement of head, response to impact 27
MR spectroscopy 82
MRI see Magnetic resonance imaging
(MRI)
Multiparameter monitoring 92, 130
Multiple organ dysfunction syndrome 338,
351
Muscle relaxants 373
depolarizing 3678
non-depolarizing 343, 345, 3689
Muscle tone, assessing 155
Myocardial injury 33940
N-acetyl aspartate (NAA), levels 265
Nailbed pressure 148
Naloxone 380
Nasogastric intubation 277, 340

Nasopharyngeal airways 340
Nasotracheal intubation 277, 340, 344,
374
National Coma Data Bank, US
severe head injury definition 1589
see also Traumatic Coma Data Bank
(TCDB)
National head and spinal cord injury
studies 4
National Acute Spinal Cord Injury Study
(NASCIS II) 428
National Head and Spinal Cord Injury
Study (NHSCIS) 9
children 13, 14
costs of head injury 15
National public health records 3, 4
Near infrared spectroscopy (NIRS) 75,
144, 218, 223, 376
Nebulization, 2-agonists 324, 348
Necrotic brain, removal 416
Needle thoracostomy 278
Nephrogenic diabetes insipidus 320
Neuroendocrine axis 308, 310
Neurofilament proteins 445
Neurogenic
hypertension 339
pulmonary edema 3267, 340, 351,
354
Neurological
observation 278
sheet 147
see also Coma scales
rehabilitation 21
Neuromuscular blockade see Muscle
relaxants
Neuronal
death 75, 135
delayed and selective loss 137
energy consumption 92
gene regulation 135
ischemia 79
membranes, effect of shearing forces
124, 128
Neuroprotection 42338
main targets 424
Neurosurgical diagnosis 343
Neurosyphilis 152
Neurotransmitter
agonist-receptor abnormalities 50, 60
antagonists 81, 83
Neurotransmitters 128, 223
excitatory 125, 293
metabolic effects 94
New Zealand, mortality 6, 7
Nimodipine 75, 76, 77, 134
Nitric oxide (NO) 91, 135
Nitrogen excretion 308
Nitroglycerin 247
Nitrous oxide (N2O) 369
CBF studies 217, 363
NMDA receptor 432
activation 50, 60
antagonists 81, 83, 432, 433
free magnesium 264
Nodes of Ranvier 45, 46, 126
Non-accidental injury see Child abuse
Non-depolarizing muscle relaxants 343,
345, 3689, 373
Non-hemorrhagic contusions 51
Non-impact injury 32
INDEX
Non-ketotic hyperglycemic hyperosmolar
coma (NHHC) 327
Non-steroidal anti-inflammatory drugs
(NSAIDs) 380
Norepinephrine (noradrenalin) 90, 356,
357, 373
Nottingham Health Profile 441
Nursing staff
recognition abnormal values 105
trauma team 273
Nutritional
problems, pre-existing 282
support 31315, 352
Obesity 282, 352
Observational studies, closed head injury
29
Occipital impacts 267, 28, 31, 33, 54
Ocular
deviation, forced 154
rupture at altitude 289
Oculocardiac reflex 154
Oculocephalic reflexes 154
Oculomotor
nerves, disorders of 4501
paralysis 152
Oculovestibular reflex 154, 445
Odontoid peg fracture 170
Omega conotoxins 134
Open barrier edema see Vasogenic edema
Opiates, effects on pupils 152
Opioids 309, 345, 365, 367, 373, 374, 379,
380
Optic nerve injury 152, 153, 154, 450
Oral intubation 344
Orbital
fractures, blow-out 168
swelling 156
Organ donation 445, 4467
Orientation 1567
Oropharyngeal airways 340
Osmolality 297, 355
total plasma 2956
Osmotic
brain edema 49
gradient, mannitol therapy 396
Outcome 645, 43961
brain death 4447
see also Brain death
clinical examination and 1634
delayed complications 4514
Glasgow Outcome Scale 43942
nutrition 315
operative delay 75, 76
predictors of mortality 104, 150, 238
Glasgow Coma Score 77, 150
raised intracranial pressure (ICP) 1012
vegetative state 44, 65, 73, 159, 440,
4479
when to assess 4423
Oval pupil 153
Oxygen
delivery (DO2) 298
and consumption (VO2) 2934
effects of head injury on 33840
from alveoli to tissues 3367
distribution from atmosphere to alveoli
3346
hyperbaric 400
peripheral extraction and utilization
3378
see also PaO2; Supplemental oxygen
Oxygen-hemoglobin dissociation curve
3367
Oxygenation, brain tissue 92
P50 3367
PaCO2 92, 93
adequacy of ventilation 335
brain death 445, 446
low levels 395
relationship with
CBF 90, 388
ICP 394
Painful stimuli 148, 149
Pancuronium 345, 368
Panda eyes 84
PaO2 92, 336
altitude effects on 288
levels after brain injury 93
weaning from ventilator 348
Papilledema 155
Paramedics 272
Parasagittal vein, bleeding 416
Parenchymal
electronic sensors 144
small vessels, coalescence and rupture
128
Parenteral nutrition 305
daily allowances, nutrients and vitamins
314
glucose administration 313
Partial pressures see PaCO2; PaO2
Pathology 29, 3970
Patient-controlled analgesia 380
Patient-ventilator dyssynchrony 3445
Pedal cycle accidents 11, 12, 17
helmets 20
Pedestrians 1112, 17
Pediatric coma scales 1501
Pegboard tests 158
Pelvic
examination 279
X-ray 281
Penetrating injuries 55, 845, 173, 175
entrapped air at altitude 289
missile injuries 123
visual pathway damage 450
Pentobarbitone 366, 399
Percutaneous dilatational tracheostomy
344
Performance IQ tests 455, 457
Perfusion imaging 266
Pericallosal artery ischemia 201, 202
Perineum, examination 279
Periorbital swelling 149, 154
Peripheral edema 305
Peritoneal lavage 280
Perivascular hemorrhage 126
Periventricular lucency 81
Personality changes 441, 442, 443, 454, 4578
five-point scale 458
relationship to previous personality
458
Petechial hemorrhages, MRI 179, 185,
186
Pethidine 367
Petrous bone fractures 84, 156
BAEP absence 237
sixth nerve palsy 451
Phase and coherence data, EEG 234
Phencyclidine (PCP) 137, 433
Phenobarbitone 366, 379
Phenylephrine 356, 373
Phenytoin 379, 453, 454
drug interactions 333
inhibition ADH release 309, 317, 318
Phosphates, MRS studies 2623
Phosphorous MR spectrum 262
Physical disabilities 439, 441
neurophysical sequelae 441, 442, 449
475
Physiological saline 300, 303, 312, 327, 342

hemorrhagic shock models 306, 307
Pituitary gland
hypothalamic control 308, 310
injury to 308, 320
Plain film radiography 1678
Plantar reflexes 155, 158
Plaque jaune 52, 53
Plasma
COP 295, 296
equivalents 2989, 300
osmolality 2956, 297
sodium 297
Plasma-protein derivatives 2989
Platelets 83
disorders 421
Pleural drainage 278
Pneumocephalus
imaging 197
nitrous oxide 369
Pneumonia, nosocomial 34950
Pneumothorax 278, 279
effect of altitude on undrained 289
nitrous oxide 369
Polyamines 134
Polyethylene-glycol-conjugated superoxide
dismutase (PEG SOD) 134
Polygeline 299, 300, 342, 356
Polyneuropathy 345
Pontine lesions, clinical signs 152, 154
Pontomedullary disruption 623, 62
Population-based
national studies 4, 810
regional studies 10
Positioning, patient 350, 374, 401
head elevation 355
leg elevation 374
Positive end-expiratory pressure (PEEP)
3467, 395
Positive pressure ventilation 309, 345
weaning from 3489, 352
Positron emission tomography (PET) 167,
21718, 224, 225, 226
glucose metabolism 97, 128
Post-coma disturbance (PCD) 161
Post-traumatic amnesia (PTA) 157, 15961,
456, 457
concussive syndromes 47
duration of 163, 454
head injury severity 14
rehabilitation 21
Postconcussion syndrome 47
Posterior cerebral arteries
infarction 204
ischemia 200, 201, 202
transcranial Doppler 245, 246
Posterior fossa hematomas 78
surgery 416
Potassium 294
efflux to ECF 124, 125, 132, 312
uptake by astrocytes 132, 139
see also Hyperkalemia; Hypokalemia
Power spectral analysis, EEG 233, 235, 236
in traumatic coma 2334
Pre-existing medical conditions 2813, 290,
333
Precapillary sphincters 90, 91
Prednisolone 428
Preferred hand 157
Pregnancy 283
Preload 337, 3523
476
INDEX
Pressure
controlled ventilation 346
sores, TCD measurements 246
support ventilation (PSV) 346, 349
waves, ICP 214
waveforms 21415
Pressure-volume index (PVI) 96, 107, 109,
110, 111, 394
formulae 108
Prevalence 3, 4
Prevention 1617
Priapism 279
Prigatonos classification of deficits 457
Primary
axotomy 41, 46
brain injury see Brain injury: primary
brain-stem hemorrhages 62
hypoaldosteronism 323
survey 145, 146, 276, 2778
Prognosis 145, 163, 442
EEG 231, 232
evoked potentials 240
brain-stem auditory evoked 236, 237
somatosenory evoked 238
Glasgow Coma Scale 150
pupillary light reflex 153
see also Outcome
Propofol 345, 3635, 373, 374, 379, 397
Prosthetic eye 153
Protein
catabolism 307
molecules 295
requirements 313
synthesis, effect of neurotrauma 135,
308
Proton MRS studies 264, 265
Pseudo-hypoaldosteronism 323
Pseudobulbar syndromes 449
Pseudomemebrane development 79
Psychiatric symptoms 459
Ptosis 152
Public
education 1617
health records, national 3, 4
Pulmonary
capillary wedge pressure 282
edema 296, 305, 345
neurogenic 3267, 340, 351, 354
hypertension 336
thromboembolism 336, 3512
Pulmonary artery
catheterization 353
pressure monitoring 352, 353, 354
wedge pressure (PAWP) 303, 304
Pulsatile blood volume (dV) 112
Pulsatility indices, Doppler 115, 246, 247,
249, 392
standardized 252, 253
Pulse
ampiltude, ICP 214
oximetry 342, 348, 375
rate 155
strength of 343
Punch-drunk syndrome, boxers 656
Pupillary light reflex 151, 153, 154, 445
loss of 152
Pupils 1514
asymmetry 391
barbiturate therapy 398
localizing hematomas 411
Purulent secretions 350
Pyrexia 85
Question mark trauma scalp flap 412, 413
Radiography, plain film 1678

Random breath testing 17
Rate 3, 4
Raven colored matrix test 157
Ravens Progressive Matrices 454
Reabsorption, hemorrhage 52, 53
Red cells
diapedesis 126
transfusions 298, 305, 356
Refeeding syndrome 325
Reflex scale, five-point 154
Regional
anesthetic techniques 286, 380
Resource Trauma Center 20
Rehabilitation 21, 163
Relative motion concept, brain injury
334
Relatives 442
costs to family 15, 16
counseling 459
discussing
organ donation 446
treatment withdrawal 444
Renal
impairment, pre-existing 283, 333
sodium excretion 309, 31112
Renin-angiotensin-aldosterone system 312
Rescue services, deployment of 272
Resonance, definition 2612
Resonium A 324
Respiratory
management
after initial resuscitation 34452
initial resuscitation 277, 3402
quotient (RQ) 335
rate 155
Resuscitation 276, 277
accident site 272
airway management 3402
circulation 342
intravenous therapy 293, 294
monitoring 343
sedation and analgesia 343
ventilation 342
Retinal hemorrhages 155
Retraction balls, axonal 41, 42, 43, 44, 45,
46, 125, 126, 136
Retrograde amnesia 47, 157, 15960, 456
Retromandibular pressure, eye opening
148, 149
Reverse steal 394
Richmond screw 211
Right
atrial pressure 352
catheters, placement 354
ventricular ejection fraction 353, 354
ventricular preload 352
Rigid cavities, effect of altitude 289
Riluzole 435
Ringers lactate 278, 300, 301, 303, 307, 342
hemorrhagic shock models 304, 305, 306
Road construction and traffic safety 17
Rocuronium 3689
Rostral pons
focal lesions 43, 44
secondary hemorrhages 63
Rotary-wing aircraft 283, 284, 287, 290
Rotational acceleration of head see Angular
acceleration of head
Royal Adelaide Hospital treatment
algorithm 402, 404
Royal Flying Doctor Service 284
Rural areas 273
Trauma Hospitals 20
Sagittal sinus
pressure 109
systems analysis 11314
superior, MR venogram 2023, 205
Salbutamol, nebulized 324, 348
Saline
equivalents 298
see also Hypertonic saline; Physiological
saline
Scalp examination 156
Schirmers test of lacrimation 158
Scopolamine 137
Seat belts 11, 1819
bruising 279
children 14, 19
internal carotid artery occlusion 85
Second messenger systems 133
Secondary
axotomy 412, 45
brain injury see Brain injury: secondary
brain stem lesions 63
survey 145, 146, 276, 27880
watershed ischemia 203
Sedation 342, 345, 36383, 402
during transportation 393
Seizures
EEG monitoring 232
hyponatremia 319
management 3789, 464
occult 402
propofol 365
thiopentone 366
see also Epilepsy
Selfotel (GCS 19755) 433, 434
Semicoma 148
Sensation, assessing 158
Sepsis 354
Septic shock 351
Seventh cranial nerve damage 158, 451
Severity, trauma 78
assessment of injury 40, 15864
coma scales 144, 146, 14751
conscious level 146
prospective grading 1589
retrospective grading 15961
Sevoflurane 363, 371
Shaken baby syndrome 204
Shape, striking object 26
Shearing forces 28, 30, 122, 123
effect on
axons 1256
ion channels 124
microvasculature 126
synapses 125
Shivering 377
Shock
fluid management 302, 356
spinal 327
Short Form 36: 441
Short-term memory 456
Shunt 336
Sick cell syndrome 316
Sickness Impact Profile 441
Single photon emission tomography
(SPECT) 167, 1723
brain death 206
CBF and brain metabolism 218
ischemia 201, 205
subdural hematomas 190, 192
vasogenic edema 139
Six-month outcome 442
Size of pupils 153
Skeletal conditions, pre-existing 282
Skew deviation 154
INDEX
Skull
deformation 26, 32
films 77, 1678
fractures see Fractures
Sleep patterns, within EEG 231, 232
Small pupils, bilateral 152
Smell testing 157
Smooth muscle cells, response to
transmural pressure 91
Snellen test 157
Social restraint, lack of 456, 458
Society of British Neurosurgeons 71
Sodium 294, 300, 355
effect of stress on movement of 309,
31112
plasma 297
see also Hypernatremia; Hyponatremia
Sodium nitroprusside 247, 3567, 372
Sodium pentobarbitol 355
Somatosensory evoked potentials 2379, 240
grading 238
median nerve 231
Source data 3, 4
SPECT see Single photon emission
tomography (SPECT)
Speech 157, 454, 457
restoration after coma 159
verbal scores 455
Speed control 1718
Sphenoid fractures 168, 173, 451
Spinal
cord injuries 327
abdominal paradoxical breathing 338
clinical signs 155, 279
endotracheal intubation 277, 278, 341,
374
imaging 167, 168, 170, 171, 281
risk of, accident site management 272
reflexes, brain dead patients 447
Spindle pattern coma 232
Spironolactone 323
Splanchnic ischemia 337
Spontaneous roving eye movements 154
Sporting and recreational injuries 12, 47
Squint, divergent 152
Stab wounds 123
Staphylococcal meningitis 84
Starling forces 296
Starvation 307, 308
Statham pressure transducer 210
Status epilepticus 48, 232, 365
possible causes 464
Steroids 399, 429
see also Corticosteroids
Stiffness, striking object 26, 29
Strain
gauges 210
rate 28
Stress
effect on fluid and electrolyte
movements 309, 31112
genes 135
hyperglycemia 282
metabolic 307, 308, 309
tensile or compressive 28
Striking object, physical properties 26
Stroke volume 337
STYCAR toys 157
Subarachnoid hemorrhage 52, 79, 80, 435
cerebral vasospasm 85
ICD-9 CM rubric 6
imaging
CT 174, 192, 194, 195, 196
MRI 196
nimodipine trials 134, 4312

pathology 5661
secondary brain damage 76, 77
Subarachnoid screws 211
Subconjunctival hemorrhages 84
Subdural bolts 211
Subdural hematoma 789, 126, 1278, 409
asymptomatic patients 410
boxing 65
chronic 61, 79
glutamate elevation 432
ICD-9 CM rubric 6
imaging 188, 190, 193, 203, 204
CT 81, 175, 178, 181, 184, 191, 192, 193,
194, 195, 197
MRI 172, 190, 191, 193, 194
SPECT 192
pathology 59, 602
surgery 413, 41516
effect of operative delay 75, 76
Subfalcine herniation 79, 80
Sucking pneumothorax 278
Sufentanil 367
Sulcal hematoma 51, 52
Superior sagittal sinus, MR venogram
2023, 205
Superoxide
anion 48
dismutase 429
Supplemental oxygen 334-5, 342
Suprorbital pressure, eye opening 148
Surface contusions 51, 52, 54, 55, 122
Surfactant, neurogenic depletion 340
Surgery 40921
airborne neurosurgical services 283, 290
ICP reduction 4001
indications for 40911
intentional hypothermia 376
setting priorities 2801
specific problems 41621
Surgical tracheostomy 344
Surgicell 415
Surveillance, head injury 20
Suxamethonium 3678, 373
Sweating, absence of 155
Swedish Reaction Scale (RLS85) 149
Synaptic function, effect of shear forces on
125
Synchronized intermittent mandatory
ventilation (SIMV) 346
weaning from 349
Syndrome of inappropriate
ADH secretion (SIADH) 312, 316, 317
antidiuresis (SIAD) 316
Synergistic damage models 1378
Systemic
inflammatory response syndrome (SIRS)
3389
vascular resistance (SVR) 337
Systems analysis 11314
T-piece weaning 349
Tachyarrhythmias 339
Tachycardia 155, 342
Tachypnea 156
Taste, testing 157
Teleradiology 273
Temperature 279, 3767
effect on ICP 402
see also Hypothermia
Temporal burrholes 41112
Tendon reflexes 155, 158
Holmes-Adie syndrome 152
Tension pneumothorax 278
477
Tentorial herniation 63, 64, 79, 80, 175, 183,

193, 197200, 386
clinical signs 151, 152, 153
Thalamus, hemorrhage 56, 56
Thermal diffusion techniques 218, 221
Thermoregulation see Temperature
Theta frequency, return of 234
Think First Prevention Programme 16
Thiopentone 365, 366, 373, 374, 379, 3989
Third
nerve palsy 79, 152, 153, 450
ventricle, imaging 199, 200
Thoracic injury 279, 281
Thoracostomy 278, 352
Thrombocytopenia 83
Tidal volumes, alveolar overdistension 345
Tirilazad 134, 399
Tonicity 297, 298
Tonsillar herniation 63, 80
Trace element requirements 31415
Tracheal stenosis 344
Tracheostomy 3412, 344
Transcranial Doppler (TCD) 74, 144, 218,
220, 223, 24359, 376, 392
in diagnosis of brain death 255
long-term measurements on ICU 246,
247
results of analysis 24954
role in monitoring therapy 2545
signal processing and data collection
2479
theory of 2445
Transducers, ICP monitoring 210, 211, 215
catheter-tip 212, 213
implanted microchip 212, 213
Transport-related injuries 1012, 16
Transportation of patients 2723, 290
airborne 2839
deterioration in conscious state 286
preparation of patient 286
selection of aircraft 2869
control of ICP and CPP 393
intrahospital 2801
Trauma
care systems 201
centers, level 1 273, 290
definition 4
general aspects of 271
ICD classification 5
Score 8
teams 273
criteria for assembly 276
Traumatic Coma Data Bank (TCDB) 95,
102, 386, 409
outcome studies 443
see also National Coma Data Bank
Traumatic encephalopathy, boxers 12, 656
Treatment see Management of head injury
Trigeminal nerve, testing 158
Triple lumen bolt 225
TRISS 8
Tromethamine (THAM) 395, 399400
Tube feeding, discontinuing 449
Ubiquitin 45
Ultra-fast echo planar imaging, MRI 172
Unbonded strain gauge 210
United Kingdom
mortality 6, 7
Upper torso accleration 25
Urea 298
478
INDEX
Urethral rupture 279

Urinary catheterization 279
USA
mortality 7, 10
Vascular
component, raised ICP 387
injury 1268
diffuse 40, 57, 61
pathology 505
quantitation 40
Vasoconstriction and vasodilation, effect
on sodium excretion 311, 312
Vasogenic edema 49, 52, 80, 81, 138, 139,
390
diffusion-weighted imaging 266
Vasomotor paralysis 106, 111, 112
Vasopressin 309, 311
Vasopressors 342, 3567
Vasospasm 395
middle cerebral arteries 254
Vault fractures, infection 834
Vecuronium 368, 373
Vegetative state 44, 65, 73, 159, 440, 4479
limiting treatment 443, 444
recovery from 448
Vena caval filters 352
Venous outflow pressure 109, 111
Ventcontrol MTC 213
Ventilation 342, 3778
adequacy of 335
of brain dead patients 447
disconnection
accidental 344
intentional, brain death tests 445

management of 348
modes of 3456
principles of 3445
weaning from 3489
pre-existing physical limitations 352
see also Hyperventilation
Ventilation-perfusion mismatching 336
Ventilator-associated pneumonia 34950
Ventricular
catheters 211, 394
hemorrhage see Intraventricular
hemorrhage (IVH)
puncture 20910
Ventriculostomy 144
related infection 111
Verbal responses 148, 149
Vertebral arteries 89
Vestibular dysfunction 451
Visceral rupture 279
Viscoelastic properties 386, 387, 38990
Vision, assessment 157
Visual
evoked potentials 234, 236
pathway damage 450
Vitamin requirements 31415
Volatile anesthetic agents 36971
Voltage-gated ion channels 1234, 125, 133,
134
Volume
pressure response (VPR) 109, 111, 113,
38890
status monitoring 3523
Volume-controlled ventilation 3456
Volutrauma 345
Wallerian degeneration 45, 126

Water
and electrolyte balance, disorders of
31526
proton nuclear spins, labeling 222
total body 294, 295
Water-bath humidifiers 348
Water-equivalent solutions 298, 300, 301
Wayne State tolerance curve 301, 32, 35
Wechsler Adult Intelligence Scale (WAIS)
454
Westmead PTA scale 161
Whiplash injury 31
Whole
blood 298, 303
hemorrhagic shock models 304
brain death 445
Windscreens 18
Work, return to 439, 443
World Medical Association, brain death
declaration 444
Wound examination 156
Wrights respirometer 348
Xenon CBF measurements 945, 97, 190
stable-xenon-enhanced CT 224
advantages and limitations 220
cerebral blood volume 2201
clinical studies 21920
methodology 219
threshold for ischemia 96
Xenon-133 studies 217, 21819
Zinc, daily requirements 314
APPENDICES
Appendix A: Antibiotics recommended for infections of the CNS (Source: from Reilly,
P. L. and Simpson, D. A. (1995) Craniocerebral injuries, in Craniomaxillofacial Trauma,
(ed. D. J. David and D. A. Simpson), Churchill Livingstone, Edinburgh, pp. 367396.)
Organism
Penicillin/amoxycillin-sensitive
organisms
Haemophilus influenzae
(beta-lactamase-negative)
Neisseria meningitidis
Streptococcus species, including Strep.
pneumoniae
Staph. aureus/Staph. epidermidis
(beta-lactamase negative)
Anaerobic bacteria: Clostridia,
Fusobacteria
Moderately resistant organisms
Haemophilus influenzae
(beta-lactamase-negative)
Strep. pneumoniae (moderately
penicillin-resistant)
Staph. aureus (beta-lactamase positive)
Enterobacteriaceae
(beta-lactamase inducible negative)
Bacteroides species
(beta-lactamase-positive)
Very resistant organisms
Enterobacteriaceae
(beta-lactamase-inducible-positive)
Pseudomonas aeruginosa
Staph. aureus/Staph. epidermidis
(methicillin resistant)
* ceftriaxone or cefotaxime
Treatment
Standard
Alternative
Amoxycillin
Penicillin G
Third-generation cephalosporin*
Chloramphenicol
Chloramphenicol
Chloramphenicol
Chloramphenicol, vancomycin
Penicillin G
Metronidazole
Chloramphenicol
Chloramphenicol
Flucloxacillin
Vancomycin
Metronidazole
Imipenem
Chloramphenicol
Imipenem aminoglycoside
Anti-psuedomonad penicillin and
aminoglycoside
Consult microbiologist
Ceftazidime and aminoglycoside
Imipenem with aminoglycoside
Vancomycin and rifampicin
Consult microbiologist
Penicillin G
Penicillin G
464
APPENDICES
Appendix B: Seizure management in acute head injury

Seizures may arise as a result of the brain injury or from other pre-existing or concurrent factors. These should
be identified and corrected where possible while the seizures are being treated.
Treatment
1. Airway and oxygenation

2. Intravenous glucose
3. Anti-epileptic medication
Initial treatment
Drug
Dose
Rate
Diazepam
510 mg i.v.
(0.050.15 mg/kg)
Maximum 30 mg
1 mg/min
Maximum < 5 mg/min
Phenobarbitone
1020 mg/kg
< 60100 mg/min
1041 g/ml
Phenytoin
1520 mg/kg i.v.
50 mg/min
(over a minimum
of 20 minutes)
1030 g/ml
Maintenance
Phenytoin
begin 68 h after the loading dose

5 mg/kg/24 h (equivalent to 300 mg/d for a 60 kg adult)
Appendix C: Possible causes of status

epilepticus after head injury

Drug withdrawal
Antiepileptic drugs
Alcohol
Benzodiazepines
Short-acting barbiturates
Antibiotic and drug reactions
Penicillin
CSF contrast media
Fluid and electrolyte disturbances
Water intoxication including SIADH
Hypocalcemia and hypomagnesemia
Hypoxia
Hypoglycemia
CNS infections
Meningitis
Encephalitis
Brain abscess
Subdural empyema
Cerebrovascular disease
Venous thrombosis
Thromboembolic arterial occlusion
Aiminoff, M. J. and Simon, R. P. (1980) Status epilepticus. Causes, clinical

features and consequences in 98 patients. American Journal of Medicine, 69,
675666.
Pike, A., Partinene, M. and Kovanen, J. (1984) Status epilepticus and alcohol
abuse: an analysis of 82 status epilepticus admissions. Acta Neurologica
Scandinavica, 70, 443450.
Therapeutic range
APPENDICES
465
Appendix D: Cardiovascular drugs used for augmentation of cerebral perfusion

pressure
General principles: (Refer to Section 17.5.2 (g): pp. 355357
1. Ensure euvolemia: hypotension due to hypovolemia is the most comnmon cause of low cerebral perfusion
pressures and must be assiduously monitored and corrected.
2. The use of inotropes and vasopressors in other than very small doses requires regular hemodynamic
monitoring: intra-arterial line, central venous pressure and where indicated a pulmonary artery catheter.
3. Mean arterial pressure and cardiac output should be interpreted in the context of premorbid cardiac
function.
4. No single inotrope (or mixture or inotropes) has been shown to be superior to another. Selection of an
inotrope or vasopressor is made according to experience and familiarity. If the desired perfusion pressure is
not attained or complications arise change to a different agent.
5. Administration:
(a) All infusions must be administered through a central vein using volumetric infusion pumps.
(b) Start infusion at 35 ml/h and titrate until the desired perfusion pressure is reached.
Agent
1 effects
+ Chronotropy
+ Dromotropy
+ Inotropy
Dose
(solutions in
5% dextrose)
Inotropes
Adrenaline
Noradrenaline
Dopamine
Dobutamine3
Isoprenaline3
6 mg/100 ml1
6 mg/100 ml1
400 mg/100 ml2
500 mg/100 ml2
6 mg/100 ml1
Vasopressors
Metaraminol
Phenylophrine
10 mg/100 ml
10 mg/100 ml
2 effects
+ Inotropy
Vasodilatation
Bronchodilatation
effects predominate at low dose
1 effects
+ Inotropy
Vasoconstriction
2 effects
+ Inotropy
Vasoconstriction
effects predominate at high dose
+
+
+
(+)
(+)
+
+
+
+
+ = Strong effect; (+) = moderate effect; = no effect.

1. Rate in ml/h approximates g/min.
2. Rate in ml/h approximates g/kg/min.
3. These agents are predominantly vasodilators and should be used with caution in head injury.
Part One
THE INJURY
EPIDEMIOLOGY
Michael R. Fearnside and Donald A. Simpson
1.1
Introduction
The last 20 years have seen major advances in the

prevention and treatment of head injury, resulting in a
substantial decrease in mortality. Yet head injury
remains a major health and social problem both for
developed and developing nations. The frequency and
diversity of head injury provides organizational problems for retrieval and early response services, accident
and emergency wards and rehabilitation departments.
The long-term disabilities may be grave and special
difficulties are experienced by the community in
general and by families in particular when the headinjured patient attempts to re-enter and integrate with
society.
In order to meet these challenges, epidemiological
information regarding the frequency of occurrence,
causes, distribution and outcome of head injury is
necessary. Using reliable data, preventive measures
may be undertaken as well as measures to minimize
brain injury when it occurs. Epidemiological studies
are concerned with populations rather than individuals. The methods used, therefore, measure disease rates
and population statistics rather than individual case
records, although these may be the source of data.
Such information for head injury is not always easy
to obtain (Jennett and MacMillan, 1981). Injuries occur
in a widely dispersed geographical pattern, care is
decentralized and non-systematized in many regions,
deaths occur both inside and outside hospitals and the
survivors are cared for in the main by family members
(Fearnside, McDougall and Lewis, 1993). While the
reporting of death and its causation is mandatory in
most countries, it is may not be standardized and is
often incomplete, providing a bias in the collection of
mortality data (Romano and McLoughlin, 1992).
A disease such as head injury results from an

interaction between an individual and an external
agent such as a mechanical force. While clinical
medicine is concerned with the management of the
individual patient, epidemiology deals with the
nature and pattern of the interaction as it applies to a
population in a particular environment (Walton, Beeson and Bodley-Scott, 1986). An epidemiological study
aims to answer questions, therefore, about distribution, causation, age, sex or racial differences that might
be important for measures to prevent or treat head
injury in a more rational manner within a
community.
A number of terms commonly used in epidemiological studies of head injury are shown in Table 1.1.
The count is made more meaningful by the use of a
denominator to relate it to the population in which
the event occurred and the numerator must include
only persons derived from the denominator population. This descriptor of the count related to a
population is termed a rate and all members of the
population must have an opportunity to appear in
the numerator.
The incidence of head injury describes the occurrence of new cases in the population over a period of
time, usually 1 year.
The prevalence of head injury describes all cases in
a population at a particular time and is a measure of
both new and established cases.
1.3
Epidemiological information involving head injuries

may be derived from a number of sources and
levels.
(a)
1.2
Definitions in epidemiology
Epidemiology is a quantitative science that measures

the occurrence of disease in the human population
(Friedman,1987) and is concerned with the patterns of
disease in groups of people rather than in individuals.
Source data
National public health records
Necessarily retrospective, these are, in the main,

mortality studies from death certificates (Ring, Berry
and Dan, 1986; Sosin, Sacks and Smith, 1989). Some
national information may also be available based on
hospital admissions, using the N (diagnostic) or E
Head Injury. Edited by Peter Reilly and Ross Bullock. Published in 1997 by Chapman & Hall, London. ISBN 0 412 58540 5
EPIDEMIOLOGY
Table 1.1
Definitions used in epidemiological studies of head injury
Count
The number of instances an event occurs, e.g. number of head injuries admitted to
a hospital
Frequency of the observed event
Rate
Total number in the at risk population
Incidence rate*
Prevalence rate
Mortality rate
Age-specific mortality rate
Case fatality rate
The number of persons sustaining a head injury

The total population from which they are drawn
The number of persons with a head injury
The number of persons dying from head injury/unit time
The number of persons dying from head injury in a particular age group
The number of persons dying from head injury
The total number in the sample or population with head injury
* Established cases should be excluded from the numerator but not the denominator
May be individualized to other parameters, e.g. gender, socioeconomic class
The CFR refers to the proportion of persons who die with head injury, e.g. following hospital admission. Like the
incidence rate, a time period need not be specified, but it may be.
(external cause) of the ICD system (see below). In

many countries, public authorities maintain records of
cause-related injuries, such as industrial injuries.
(b)
National head injury studies
A number of these have been undertaken in an

attempt to identify broad features such as incidence
and mortality rates on a national basis (Anderson and
McLaurin 1980). Difficulties with data collection may
occur in these larger studies from an inability to
capture all events in a large population and the results
may underestimate the actual rates.
(c)
Population-based studies
These studies investigate a defined, focal population

such as a state or county area (Kraus et al., 1984, Ring,
Berry and Dan, 1986, Hung et al., 1991), an urban
(Chan, Walker and Cass, 1988) or a rural area (Jagger,
Levine and Jane, 1984). The true incidence is more
likely to be approached as the size of the sample
becomes smaller and capture more complete.
(d)
Clinical topic reports
These are based on clinical events or etiological

characteristics. Thus, head injuries may be further
defined by groupings such as mild (Kraus and
Nourjah, 1988) or moderate head injuries (Rimel et al.,
1982). Head injuries may be associated with a particular mode of transport. Thus, Bucklew et al. (1992)
described head injuries following falls and ejections
from pick-up trucks in New Mexico. Studies of head
injuries in sports such as football (Mueller and Cantu,
1988) are further examples of such focused
investigations.
1.4 Definitions and classification of head

injury
1.4.1
DEFINITIONS IN HEAD INJURY
Head injury demands a broad definition. The anatomical term suggests any trauma to the body above the
lower border of the mandible. In general, maxillofacial
trauma is considered separately from head injury,
although the two frequently coexist. David and
Simpson (1995) have used the term craniomaxillofacial injury to group together injuries of the face
and frontal region. The head includes the scalp, skull,
meninges and blood vessels as well as the brain and
its constituent parts. The term craniocerebral injury is
often used to emphasize that the brain should not be
considered in isolation from its integuments. Trauma
refers to an external source of energy, such as a
mechanical force, causing a physical injury to any or
all of the tissues comprising the head. Electrical,
thermal and chemical causes of injury are usually
considered separately, in the category of burns, but it
DEFINITIONS AND CLASSIFICATION OF HEAD INJURY
should be remembered that deep craniocerebral burns

are likely to require neurosurgical management and
may appear in neurosurgical statistics. However, they
are very rare.
There is no agreed definition of head injury for
epidemiological purposes and the definitions in each
of the studies shown in Table 1.2 differed. Such
differences may result in an inability to compare
epidemiological studies. Kraus et al. (1984) emphasized the desirability of separating the broader classification of head injury from brain injury, the latter more
precisely indicating neurological damage. They
stressed the importance of identifying patients having
sustained neuro-trauma of public health consequence where there was a probability of ongoing
neurological impairment, requiring resources for medical or nursing care. This distinction is of obvious
value in the planning of rehabilitation services and
community support schemes.
For epidemiological studies, it is often desirable to
define a broader range of head injury, bearing in
mind that many patients sustain an apparently minor
Table 1.2 Definitions of head injury showing variations

among the series
Field (1976)
Trauma which caused some risk of

damage to the brain. Used ICD-9
rubrics.
Anderson and
McLaurin
(1980)
Trauma to the brain or spinal cord.

Here, trauma refers to physical injury to
living tissue caused by an external force.
Jennett and
MacMillan
(1981)
Patients with a history of a blow to the

head or with altered consciousness after
a relevant injury or with a scalp or
forehead laceration, or who had an X-ray
examinaton.
Klauber et al
(1981)
Patients whose head injury resulted in

skull fracture, unconsciousness, amnesia,
neurological deficit or seizure.
Selecki et al
(1981)
Injuries to the brain or skull.
Kraus et al
(1984)
Physical damage to,or functional

impairment of, the cranial contents from
acute mechanical exchange, excluding
birth trauma.
Jagger, Levine
and Jane
(1984)
Documented head injury with loss of

consciousness, post traumatic amnesia or
skull fracture.
Brookes et al
(1990)
Any injury to the scalp including

swelling, abrasion or contusion as well as
laceration; or a well authenticated history
of a blow to the head; or any patient in
whom a skull X-ray was performed
immediately following trauma, and
patients who had clinical evidence of a
fracture at the base of the skull.
injury on presentation but later exhibit evidence of

brain injury, intracranial hematoma or skull fracture
(Jennett, 1989). Brain injury is best considered as a
subgroup of head injury and attention needs to be
given to these definitions when assessing various
reports in the literature.
1.4.2
CLASSIFICATION OF HEAD INJURY
The World Health Organization (WHO) publishes an

International Classification of Diseases (ICD), where
trauma and head injury are included in the chapter
Injury and poisoning. Prior to 1950, trauma was
classified according to the external cause (E code) and
there are no available methods for extracting information regarding the effect of trauma to the head in a
systematic manner. The fifth revision, ICD-5, introduced a classification based on the nature and
diagnosis of the injury (N code).
However, the fact that all brain injuries were
classified as Intracranial injury without skull fracture
and included hematomas of the scalp meant that a
precise description of the cerebral injury was not
available. No doubt this classification reflected the
preoccupation of surgeons and pathologists of the
time with focal, low-energy impacts such as skull
fracture and extradural hematoma, rather than with
high-energy impacts that damaged the brain rather
than its coverings.
The diagnoses are contained in a series of digital
codes or rubrics. The ninth edition, ICD-9-CM (WHO,
1978, 1979), contains a three-digit rubric for the major
diagnostic groups (e.g. 800 Fracture of skull), to
which is added a further digit which specifies anatomical, pathological or clinical detail (e.g. 800.1 Closed
head injury with cerebral laceration and contusion;
800.2 Closed head injury with subarachnoid, subdural and extradural hemorrhage). Using the Clinical
Modification (CM), a fifth digit may be added where
appropriate, providing information of clinical relevance or further defining a diagnostic statement (Table
1.3).
The tenth ICD revision (ICD-10, WHO, 1992), soon
to be published for general use, contains rubrics that
will more accurately reflect clinical diagnosis when
compared with the rubrics of ICD-9-CM (Table 1.4).
However, the new alphanumeric code will require a
considerable change in coding practice.
The ICD system is used in many countries as the
basis for public health data from which epidemiological statistics are derived. The system is also used
by most hospitals for diagnostic or treatment coding
for the purposes of data recording and retrieval. ICDbased epidemiological studies of head injury are
necessarily population-based and may be either
national (Jennett and MacMillan, 1981; Sosin, Sacks
EPIDEMIOLOGY
Table 1.3 ICD-9 CM rubric for Intracranial injury

excluding those with skull fracture, showing the use of
the 3-, 4- and 5-digit rubrics
850
Concussion
850.0850.6 describe various levels of loss of
consciousness
e.g. 850.4, with prolongd loss of consciousness,
without return to preexisting conscious level
851
Cerebral laceration and contusion

851.0851.9 describe various anatomical sites of
injury with or without open wound
e.g. 851.3, Cortex (cerebral) laceration without open
intracranial wound.
852
Subarachnoid, subdural and extradural hemorrhage

following injury
852.0852.5 describe open or closed injury.
853
Other and unspecified intracranial hemorrhage

following injury
853.0853.1 describe open or closed injury
854
Intracranial injury of other or unspecified nature

854.0854.1 describe open or closed injury
The following fifth digit subclassification is for use with

categories 851 to 854 and adds clinical information:
0
1
2
3
4
Unspecified state of consciousness

With no loss of consciousness
With brief (less than 1 hour) loss of consciousness
With moderate (124 hours) loss of consciousness
With prolonged (more than 24 hours) loss of
consciousness and return to pre-existing conscious level
5 With prolonged (more than 24 hours) loss of
consciousness, without return to pre-existing conscious
level
6 With loss of consciousness, unspecified duration
7 With concussion, unspecified
Table 1.4 Comparison between ICD-9 CM and ICD 10

chapters Intracranial injury, excluding those with skull
fractures
ICD-9 CM
ICD-10
850 Concussion
SO6.0 Concussion
851 Cerebral contusion

and laceration
SO6.1 Traumatic cerebral

edema
SO6.2 Diffuse brain injury
SO6.3 Focal brain injury
852 Subarachnoid,
subdural and extradural
hemorrhage following
injury
SO6.4 Epidural hemorrhage

SO6.5 Traumatic SDH
SO6.6 Traumatic SAH
853 Other and

unspecified hemorrhage
following injury
SO6.7 Intracranial injury

with prolonged coma
854 Intracranial injury of

other unspecified nature
SO6.8 Other intracranial

injuries
SO6.9 Intracranial injury,
unspecified
and Smith, 1989), or locality- or hospital-based mortality or morbidity studies (Kraus et al., 1984; Kraus and
Nourjah, 1988). Like all databases, the ICD system is
no better than the people who code it and experience
has shown that inaccurate or incomplete coding is
more likely to occur when staff are inadequately
trained. Klopfer et al. (1992) found that ICD coding of
external causes of injury had been omitted in the
majority of a large sample of USA hospital discharges
after eye injury.
1.5
Deaths from trauma
Using ICD-9 chapters, deaths from injury and poisoning rank fourth in age-standardized death rates for
males and females in most Western countries, behind
circulatory, neoplastic and respiratory diseases (Table
1.5).
In Australia in 1990, deaths from injury and
poisoning accounted for 8.6% of male and 4.2% of
female deaths (mortality rate 68 per 105 of the
population for males and 25 per 105 of the population
for females). These deaths comprised 49% of all deaths
in the age ranges 144 years (Australian Institute of
Health and Welfare, 1992). Since the early 1970s,
deaths from injury have steadily decreased by annual
decrements of 2%, due largely to a decrease in motorvehicle-related deaths. In 1990, motor vehicle accidents remained the most common cause of death in
this group, accounting for 31% of both male and
female deaths. However, the figure of 2489 motorvehicle-related deaths was the lowest for several
decades and 12% less than in 1989. Suicide accounted
for 38% of male and 18% of female deaths from injury
or poisoning (Table 1.6).
In the UK, a study of coroners records of trauma
deaths in the south-west Thames region revealed that
road traffic deaths were the most common cause of
traumatic death (Daly and Thomas, 1992). The majority of deaths occurred before arrival at a hospital and
were due to chest and multiple injuries, whereas the
majority of those who survived to reach hospital and
subsequently died did so as the result of a head
injury.
In New Zealand (Table 1.5), injuries ranked fourth as
a cause of death and accounted for 32% of potential
years of life lost between the ages of 1 and 70 years. The
leading causes of injury death were from motor vehicle
crashes (37%) and self-inflicted injuries (21%; Langley
and McLoughlin, 1989). However, the two leading
causes of hospital admissions for trauma were falls
(25%) and motor vehicle crashes (19%), where head
injuries contributed 35% to the injury morbidity. For
children, a similar pattern of falls and motor vehicle
accidents emerged as the most common presentations
in Accident and Emergency departments following
SEVERITY OF TRAUMA
Table 1.5 Age standardized death rates (per 100 000 population) by ICD-9 chapters for males and females for selected
countries. Resp = respiratory; Inj = injury and poisoning. (Source: World Health Organization 199192; Australian
Institute of Health and Welfare)
Female
Male
Australia 1990
Canada 1989
New Zealand 1987
UK 1990
USA 1988
Circulatory
Neoplasm
Resp
Inj
Circulatory
Neoplasm
Resp
Inj
401
382
517
469
456
247
260
259
276
246
81
88
130
118
95
67
75
85
48
90
258
221
322
281
283
150
162
186
186
161
36
42
72
61
51
26
30
36
19
32
Table 1.6 Deaths from injury or poisoning (ICD-9), for males and females in Australia, 1990 (Rates per 100 000 of the
population; crude rates for individual causes and standardized rates for All external causes) (Source: Australian
Institute of Health and Welfare, 1992)
Males
Cause of death
Females
Number
Rate
Number
Rate
Motor vehicle traffic accidents

Suicide
Accidents, falls
Homicide
Accidental drowning
Other causes
1751
1735
472
239
227
1164
21
20
6
3
3
14
738
426
558
146
73
406
9
5
7
2
1
5
All external causes
5558
68
2347
25
trauma and the head was the most frequently injured

part of the body, the incidence increasing with age
(Gofin et al., 1989). A comprehensive study by Shackford et al. (1993) of all trauma in San Diego County,
California, USA for 1 year revealed an incidence rate of
27.3 per 105 of the population and among those, motor
vehicle accidents were the most common cause of
injury. Head injuries were the leading cause of death
(48.5%), with a mortality rate of 13.2 deaths per 105of
the population.
Care must be exercised when considering deathcertificate-derived data, as failure to specify the entire
injury inventory is probably common and would
introduce a bias to understate the true incidence.
Romano and McLoughlin (1992) found that, during a
12-month period, the death certificates of 41% of
fatally injured Californian motorcyclists recorded no
specific injuries. When autopsy reports were abstracted, 68% of these motorcyclists had sustained a
significant head injury.
1.6
Severity of trauma
With the development of trauma management systems and trauma registries in many hospitals during
the 1970s and 1980s, there arose a need to develop a
standardized classification of injuries and their
severity.
1.6.1
THE ABBREVIATED INJURY SCALE (AIS)
The first AIS was published in 1971 as an initiative of

the American Medical Association, the Association for
the Advancement of Automotive Medicine and the
Society of Automotive Engineers. The scale, originally
developed for impact injury assessment, has undergone a number of revisions, the most recent in 1990
(Association for the Advancement of Automotive
Medicine, 1990).
The scale uses a numerical method of ranking
injuries by severity and is based on the anatomical
injury alone, thus allowing no descriptor of the
consequences of the injury, such as disabilities. There
is only one AIS score for each injury and the scale does
not provide a single assessment of multiple injuries.
The anatomically based system classifies the single
injury by body region on a six-point ordinal scale from
1 (minor) to 6 (maximum). The 1985 revision, AIS 85,
introduced a unique six-digit code for each injury
based on a body region (digit 1), the type of anatomic
structure (digit 2), the specific anatomic structure or
the specific nature of the injury if external (digits 3 and
4) and the level of the injury within an anatomic
region (digits 5 and 6). The digit to the right of the
decimal point identified the AIS score. The most recent
revision, AIS 90, expanded the descriptors for brain
injury, following data analysis which suggested that
EPIDEMIOLOGY
Table 1.7 Examples of derived Abbreviated Injury Scale

(AIS) scores (Source: Association for the Advancement of
Automotive Medicine, 1990)
1. A patient with a diagnosis of brainstem contusion has
an AIS score of 140204.5, which is derived as:
1 = Head
4 = Organ
02 = Brainstem
04 = Contusion
.5 = Critical
2. A patient with large, multiple bilateral cerebral
contusions has an AIS score of 140624.4, which is
derived as:
1 = Head
4 = Organ
06 = Cerebrum
24 = Large (total volume 3050 ml)
.4 = Severe
serious brain injuries were under-recorded (Gennarelli

et al., 1989). Two examples of the derivation of AIS
scores in head injury are shown in Table 1.7.
The AIS also contains a means of classifying head
injury based on the level of consciousness (Glasgow
Coma Scale) and the duration of coma. These parameters are used only when the clinical features reflect
a more serious injury than the anatomical lesion
suggests, or when no anatomical lesion is identified
using imaging or at autopsy. Ross et al. (1992) tried to
use the AIS for the head region to predict outcome, but
found it inadequate.
1.6.2
THE INJURY SEVERITY SCORE
The AIS provides no means to assess the effects of

multiple injury and, clearly, such a measure is needed.
This requirement was fulfilled by the Injury Severity
Score (ISS) (Baker et al., 1974), which is derived by
calculating the sum of the squares of the highest AIS
score in three different body regions. The six body
regions used in the ISS are:
1. Head or neck
2. Face
Table 1.8
3.
4.
5.
6.
Chest
Abdomen or pelvic contents
Extremities or pelvic girdle
External.
The ISS provides a range from 1 (mild) to 75 (most

severe) and provides a much better correlation
between the injury, the severity and the probability
of survival than does the AIS (Bull, 1975). A derived
example of the ISS for a patient who sustained
multiple injuries including a severe head injury with
diffuse axonal injury and a cerebral contusion, a
penetrating wound of the face and an open chest
injury is shown in Table 1.8. A Trauma Score has
been devised to assess the early effects of injury
(Boyd et al., 1987, Champion et al., 1990). This score
allocates points for the conscious level (05), respiratory rate (05), respiratory expansion (01), systolic
blood pressure (04) and capillary refill (02). The
perfect score is 16. This can be combined with the
ISS, the type of injury (whether penetrating or blunt)
and the patients age to give a severity index called
the TRISS. The system is chiefly of value in assessing
the overall severity of multiple injuries and in auditing the quality of management.
1.7
Population-based national studies
(Table 1.9)
A survey of an entire population with capture of all
head-injury patients at each severity level would be
necessary to satisfy epidemiological requirements to
determine the real incidence and prevalence rates of
head injury for that population. However, the majority of such injuries are towards the minor end of the
severity spectrum and are probably under-reported,
as they may only present as a casualty attendance for
observation (Jennett, 1975; Jennett and MacMillan,
1981). No such comprehensive survey has yet been
reported. However, minor head injury is now recognized as having a potential morbidity (Gronwall and
Wrightson, 1974; Marshall and Ruff, 1989) and
should be regarded as of major social and economic
significance.
Derivation of the Injury Severity Score (ISS) in a patient with head, facial and chest injuries
Injury
AIS score
Diffuse axonal injury

Small intracerebral hematoma
Penetrating injury of the face with blood loss
Open wound of the chest
140628.5
140640.4
216006.3
415000.4
Injury Severity Score = 25 + 9 + 16 = 50
Highest
AIS score
AIS2
ISS group
25
3
4
9
16
Head/neck
Head/neck
Face
Thorax
POPULATION-BASED NATIONAL STUDIES
Table 1.9 Incidence, mortality and case fatality rates in population based studies of head injuries from various
countries
Country or area
Incidence
(/105 pop.)
Britain (Field, 1976)
430
Britain (Jennett and MacMillan,

1981)
270
Rural USA (Jagger, Levine and

Jane, 1984)
208
San Diego (Kraus et al., 1984)
180
Peoples Republic of China

(Wang et al., 1986)
Deaths
(/105 pop.)
Case fatality
rate (CFR) (%)
1.62.5*
Source data
Live hospital adms
GP consulations
Deaths, hosp adms and casualty
attendances
30
6.5
Live hospital adms
17
Severe brain injury including gunshot
56
All head injuries in 6 defined urban

areas
Sweden (Hook, 1988)
300
All registered cases
NSW, Australia (Selecki, 1988)
392
Deaths and hospital admissions
Aquitaine, France (Tiret et al.,

1990)
281
22
Hualien Province Taiwan

(Hung et al., 1990)
333
89
Hospital admissions and deaths
Taipei, Taiwan (Lee et al., 1992)
180
23
Deaths and hospital admissions
91
19.7
Cantabria, Spain (VasquezBarquero et al., 1992)
4.4
Hospital deaths and hospital

admissions
Cross-sectional sample
* CFR was 1.6% for deaths on arrival at hospital and 2.5% when deaths in hospital were included
1.7.1
UNITED STATES OF AMERICA
An enumerative population study (National Head and

Spinal Cord Injury Study, NHSCIS) aiming to determine the frequency, prevalence and economic cost of
head-injured patients requiring hospital admission for
the year of 1974 was undertaken for the National
Institute of Neurological and Communicative Disorders and Stroke (NINCDS; Anderson and McLaurin,
1980), where the sample was the entire US population.
Prevalence was estimated by including those who
required inpatient care between 1970 and 1974
(Anderson, Kalsbeek and Hartwell, 1980). Eligibility
required live hospital admission. Accident and Emergency attenders or those dead on arrival were excluded. The incidence rate for head injury using these
criteria was 200 per 105 of the population, the age
range with highest incidence being 1524 years and
the male incidence being more than twice that for
females. The case fatality rate was 3% and 97% were
alive at discharge.
A review of mortality data in the USA from
19791986 using information from the National Center
for Health Statistics and based on ICD-9 codes was
reported by Sosin, Sacks and Smith (1989). The deaths
associated with head injury represented 2% of all
deaths and 26% of injury deaths, with an annnualized
death rate of 16.9 per 105 residents. The age distribu-
tion was bimodal, with peaks at 1524 years and over

75 years. The younger group was particularly affected
by motor vehicle accidents (77%) and the older group
by falls (73%).
1.7.2
UNITED KINGDOM
In an extensive study of head injury in England and

Wales for the year 1972, Field (1976) relied on hospital
diagnosis defined by ICD-8 rubrics and general
practitioner consultations, the latter estimated from a
sample of general practices with due regard to
regional and urban/rural variations. There were
142 016 admissions for head injury and an estimated
68 000 consultations for head injury, giving an estimated head injury incidence of 430 per 105 of the
population. This was probably an overestimate, as a
significant number were considered likely to have
been admitted to hospital following a consultation
with a general practitioner or to have seen a general
practitioner following discharge from hospital. Males
outnumbered females by more than two to one and
over 50% of admitted patients were younger than 20
years. For those admitted to a hospital, the case
fatality rate was 1.6% but if all who died at the
hospital, whether admitted or not, were included the
case fatality rate rose to 2.5%.
10
EPIDEMIOLOGY
Further information from the UK was provided by

Jennett and MacMillan (1981), who discussed yearly
death rates, hospital admissions in England, Scotland
and Wales and Accident and Emergency attendances
for head injury (Scotland only) for the year 1974. The
ICD-8 codes were used for death records by the
Registrar General and for hospital admission and
diagnosis, whereas attendances at casualties were
identified retrospectively using particular criteria for
head injury (Table 1.2). Head injury accounted for 9
deaths per 105 of the population each year, less than
1% of all deaths. As expected, there was a much higher
incidence in the young; in the age group 1524 years
head injury caused 15% of all deaths. The annual
admission rate to hospitals following a head injury
was 270 per 105 of the population in England and
Wales that year and in Scotland 313 per 105 of the
population. The authors estimated that there was a
Accident and Emergency attendance rate of 1780 per
105 of the population, which accounted for around
10% of all Accident and Emergency attendances.
1.7.3
AUSTRALIA
The Trauma Subcommittee of the Neurosurgical Society of Australasia conducted a retrospective analysis
of patients with neurotrauma (injuries to the head,
spinal cord and nerves), identified by ICD-8 rubrics,
who either died or were discharged from hospitals in
New South Wales (NSW) or South Australia (SA) in
1977. The hospitals were either major teaching hospitals, metropolitan or country-based hospitals. The
mortality rate was 28 per 105 of the population in
NSW and 25 per 105 of the population in SA. The rate
of hospital admission for all neurotrauma was 443 per
105of the population and for head injury 392 per 105 of
the population. The study analyzed data to provide
information regarding the nature, extent, distribution
and cost of neurotrauma (Selecki et al., 1981, 1982a, b;
Simpson et al., 1981).
1.8
Population-based regional studies
Comprehensive data are available when smaller samples are studied, but even then, difficulties in total
Figure 1.1
Causation of head injury in various studies.
capture of events probably result in underestimation

of the true incidence. The observed values vary greatly
among different countries (Table 1.9), with the highest
recorded in Hualien Province, Taiwan (Hung, 1991) of
333 per 105 of the population and a mortality rate of
89 per 105 of the population. This extraordinarily high
incidence is considered to be due to the high usage of
pedal and motor bicycles without mandatory head
protection by the use of helmets. The various studies
also reveal considerable variation in both incidence
and mortality rates, probably because of variability in
definitions of head injury, variations in target populations (e.g. predominantly young or elderly in a
region), and the accuracy of data sources and recording systems.
1.9
Causation
1.9.1
TRANSPORT-RELATED INJURIES
For both mortality studies (Sosin, Sacks and Smith,

1989) and studies with a broader definition of head
injury (Kalsbeek et al., 1980; Kraus et al., 1984) the most
frequent cause of head injuries is motor vehicle
accidents. In the USA study by Sosin, Sacks and Smith
(1989), deaths from head injury were related to motor
vehicle accidents in 57%, firearms in 14% and unintentional falls in 12%. Age-specific incidence rates
were bimodal and related to the causes, with motor
vehicle deaths most frequent in the 1524-year-old age
group (26.7 per 105 of the population) and falls most
frequent in those over 75 years of age (34.1 per 105 of
the population). The annualized death rate was 17 per
105 of the population.
For severe primary head injuries producing coma,
motor vehicles were again the most frequent cause
(Miller et al., 1978; Kalsbeek et al., 1980). Kalsbeek et
al. (1980) found that head injury was most likely to
occur when the injured person was out of doors and
traveling by motor car during the warmer months
of the year at a weekend. In several geographically
localized population studies in developed countries,
motor vehicles were the most frequent cause of
head injury: 48% in San Diego (Kraus et al., 1984)
and 53% in NSW, Australia (Ring, Berry and Dan,
1986; Figure 1.1).
CAUSATION
Of transport-related injuries (Kraus et al., 1984), 92%

were due to on road crashes and of these, 62% were to
occupants of vehicles. Injuries to motorcyclists
accounted for 20%, pedestrians 12% and bicyclists 6%.
Children, on the other hand, were more likely to be
injured as pedestrians (68%) or cyclists (24%) than as
passengers in a motor vehicle (Craft, Shaw and
Cartlidge, 1972).
Patterns may also be influenced by the country, and
in the Norwegian province of Trondelag (Edna and
Cappelen, 1985), bicycle accidents were the most
common cause of road-traffic-accident-related head
injury, reflecting the prevalence of that form of
transport. In this study, the annual incidence of road
traffic related head injury was 89 per 105 of the
population.
Causation is affected by a variety of factors, including age, gender, locality and method of presentation.
Jennett and MacMillan (1981) found assault to be
twice as common as motor vehicle accidents as a cause
among Scottish men aged 1524 years presenting at
Accident and Emergency departments. This study also
found that road accidents were responsible for only a
minority of Accident and Emergency attenders who
were not admitted, but they accounted for over half of
the severe injuries and deaths.
(a)
Patterns of injury from motor vehicles
Analysis of the patterns of damage to motor vehicles

is of value in determining the severity and type of
injury sustained by the passengers. After examining
the effects of 500 accidents Fox et al. (1991) concluded
that the severity of the injury correlated principally
with the incidence of head injury, pelvic and femoral
fractures and abdominal injury. These severe injuries
related to such motor-vehicle damage as dashboard
intrusion, steering wheel deformity, windshield
violation and the vehicle being non-reparable.
Frontal and lateral impacts were associated with
significant intrusion into the passenger capsule
(Mackay et al., 1992) and where the occupant was
restrained with a seat belt maximum loading occurred to the head and chest. Under-run crashes with
trucks constituted about 30% of such frontal impacts.
Where a restrained front seat passenger was situated
on the side opposite to a lateral impact, the injuries
were generally less severe and head injury was
caused by ejection from the shoulder strap of the seat
belt (Mackay et al., 1992). More general factors are
also relevant when considering the effects of road
traffic crashes. Van Beeck et al. (1991) identified such
diverse factors as traffic density and the availability
of advanced trauma care (particularly CT scanning
and neurosurgical facilities) and found an inverse
relationship to death from road traffic accidents in
11
each when geographical regions in the Netherlands

were compared.
(b)
Motorcycles
Motorcyclists contribute substantially to motor-vehicle-related deaths, at around 12% (Sosin, Sacks and
Holmgreen, 1990) and around 20% of all transportrelated brain injury (Kraus et al., 1984). Of the deaths,
head injury was the cause in over half; collision with
another vehicle (52%) and loss of control of the
motorcycle (40%) were the common antecedents. The
compulsory use of helmets for head protection does
reduce the death rate. The evidence for this statement
is discussed below.
(c)
Pedal cycles
Pedal cyclists are more exposed to injury in a collision

than the occupants of motor vehicles and the need for
protection is greater. Collision with a motor vehicle is
the most common cause of all bicycle injuries on the
roads. Off-road bicycle accidents are also common and
usually result from falls or loss of directional control.
In both, the most common region injured is the head
(Sacks, Holmgreen and Smith, 1991; Silverberg, Meer
and Silvinger, 1992). Spence et al. (1993) examined fatal
bicycle accidents in Ontario, Canada. The annualized
death rate for children in this province was 1.44 per
105 of the population. Of 540 deaths, 81 were due to
bicycle accidents and the great majority of these were
from head injuries. None of the victims was wearing a
helmet. Police investigations suggested that, in 70% of
these deaths, errors made by the children were
responsible. This raises the question of the childs
awareness of road safety. There is some evidence to
suggest that, in younger children, immaturity of
judgment prevents full awareness of danger (Sandels,
1977). Sacks, Holmgreen and Smith (1991) reviewed
death certificate and emergency room data for the
entire USA over a 4-year period from 19841988 and
identified 62% of all bicycle deaths as being due to
head injury. Of injuries following a bicycle accident,
41% of the head-injury-related deaths and 76% of all
head injuries occurred among children aged 014
years.
(d)
Pedestrians
Pedestrians fare poorly when exposed to motor

vehicles. Of 115 pedestrians so injured, 22% died
(Brainard, Slauterbeck and Benjamin, 1989) and the
majority who died did so during the initial resuscitation phase as a result of a combination of head, chest
and abdominal injuries.
12
EPIDEMIOLOGY
The average ISS among those who died was 46. The
most frequent injuries were musculoskeletal (77%),
head (34%), abdomen (21%) and chest (15%). Pelvic
and lower limb fractures were the most frequent
musculoskeletal injuries. A similar pattern of injury
was reported by Hill, West and Abraham (1993)
although these investigators found head injury to be
more frequent (66%) in a series of pedestrian injuries
from inner Sydney, Australia. A hospital mortality of
30%, mainly from head injuries or blood loss emphasized the need for coordinated pre-hospital resuscitation and evacuation services, together with an integrated in-hospital trauma team approach to
management.
For pediatric pedestrians injured by motor vehicles,
a similar pattern was reported from New Zealand by
Roberts et al. (1991), where life-threatening injuries
were most commonly to the head and less severe
injuries to the limbs. The mortality among the
children was 14%, all due to head injury.
1.9.2
SPORTING AND RECREATION
Sport is generally considered an uncommon cause of

more than trivial head injury but the incidence may be
expected to increase in the future as technological
advances allow more freedom and the leisure-based
industries expand. Whereas sport accounted for 35%
of head injuries in studies in the 1950s (Rowbotham et
al., 1954) and 1960s (Klonoff and Thomson, 1969) it
had increased to 12% in the 1980s (Kraus et al., 1984),
when sport accounted for 6% and recreation for 6%.
Mild head injuries associated with sport were more
common in males than in females, with a male peak at
1014 years of age but females showing an earlier
peak by 5 years.
(a)
Horseback riding
This has become an increasingly popular recreation

during recent decades. Ho ok
(1988) has estimated
that, in Sweden, there were some 1200 riding accidents, causing a head injury incidence of 0.48% of 10
million riding occasions in a population of 250 000
riders, yet there were only two fatalities. The incidence
was somewhat higher in Alberta, Canada, where, in a
6-year retrospective study, head injuries occurred in
92% of 156 riding injuries and accounted for all 11
deaths. These 11 deaths accounted for 79% of all
deaths associated with horseback riding (Hamilton
and Tranmer, 1993).
(b)
Boxing
This ancient sport has attracted much controversy. The

desirability of tolerating a competition in which
victory may be won by rendering an opponent braininjured is often questioned. Whatever the view taken,
boxing has provided the neurosurgeon with much
information as to the effects of acute and chronic brain
damage from repeated blows to the head (Pincemaille,
1989). Two of the considerations are of the acute effects
of a single blow to the head and the cumulative effects
of recurrent blows during sparring practice and actual
bouts. Increased regulation of the sport has led to
fewer fatalities (Adelson et al., 1991). Between 1918
and 1983 there were 645 deaths from boxing reported
worldwide, of which 190 were amateurs. From
19181945 there were 10.1 deaths per year, from
19701981 there were 4.5 deaths per year and there
were 4.6 deaths per year from 19791985. A knockout
is a head injury producing coma exceeding 10 seconds
and occurs in 14% of boxing matches (McGowan,
1959a, b). Such head injuries can be considered as
frequent, when 8.719% of pugilists are knocked out
in a bout (Larsson et al., 1954; Estwanik, Boitano and
Ari, 1984). In the longer term, traumatic encephalopathy can be identified in around 17% of former
professional boxers (Roberts, 1969). Amateur boxing
appears to be a safer recreation, although occasional
disasters have been reported. Several investigations
have failed to show convincing evidence of intellectual impairment in amateur boxers studied under
control conditions (Butler et al., 1993).
1.10
Children
In all types of pediatric trauma in which head injury is

a component, the patterns differ between children
(015 years) and adults. For children, falls are the most
common cause of injury (Chan, Walker and Cass,
1989).These investigators reported that the majority of
injuries requiring hospital admission were minor (87%
of total), but for more serious injuries, with an ISS of
more than 16, (13% of total), pedestrian injuries from
motor vehicles were the most common cause (31%),
followed by falls (22%), as pedal cyclists (19%) and as
occupants of motor vehicles (19%). The head was the
most commonly injured region in this group of more
severe injuries. Studies of such target populations
(015 years of age) characteristically (Gallagher and
Finison, 1984), but not invariably (Chan, Walker and
Cass, 1989) show a bimodal distribution for frequency
with peaks at 01 year of age from falls and a second
peak in adolescence.
For more severe injuries, where the ISS is more than
16, head injury is the major cause. Walker and Cass
(1987) reported a 91% incidence of head injury in such
a subgroup of 78 (13%) of a sample of 598 children.
Abdominal injury was associated frequently with
head injury and pedestrian accidents were the most
frequent cause of injury. The overall mortality for the
CHILDREN
group was 1.5% and all deaths were due to head

injury, with a case fatality rate of 11% in the more
severe subgroup.
When causation was compared between a sample of
3124 adults and 2118 children who attended Scottish
Accident and Emergency departments in 1985,
Brookes et al. (1990) reported that, for children, falls
were the most frequent cause for the presentation for
all grades of severity of head injury (57%) and were
also the leading cause of adult presentation (33%). The
definition of head injury included scalp injuries, blows
and those who had a skull radiograph following
presentation. For all causes, the frequency rates for
children (4011 per 105 of the population per year) were
more than twice that for adults (1473 per 105 of the
population per year). For falls a fourfold difference
was found between adults and children, but little
difference in frequency was found for road traffic
accidents. Using a particular definition of brain injury
(Table 1.2) and excluding less severe injuries not
requiring hospital admission, but including deaths
and hospital admissions, Kraus et al. (1984) identified
an annual frequency rate for San Diego, CA, of around
190 per 105 of the pediatric population.
More severe head injuries in children are most
frequently related to the use of motor vehicles. Of fatal
pediatric head injuries between 1979 and 1986 in
Newcastle upon Tyne, UK, Sharples et al. (1990)
reported that 76% were caused by motor vehicles,
most often occurring when the children were pedestrians and at play. These investigators found that the
fatalities occurred most often within 2 km of the
childs home (67%) and in the afternoon between 3 pm
and 9 pm (63%). The frequency was noted to be high in
areas of the city considered to be socially deprived.
Simpson et al. (1992) studied a consecutive series of
12 infants (birth to 24 months) and 103 children (214
years) dying after road accidents in South Australia
and found brain injuries in the majority. In this series
of road fatalities (Table 1.10), the victims were most
often car passengers (51.3%), followed by pedestrians
(30.4%) and pedal cyclists (18.3%). Not surprisingly,
the infants were almost all passengers, but the
childhood deaths also showed a preponderance of car
13
occupants. The high proportion of car passengers

reflects the degree of motor vehicle use in Australia,
and perhaps the attention given in areas of high
population density to the design of road systems to
minimize risk to child pedestrians. In both pediatric
groups there was a male preponderance.
The majority of pediatric head injuries are of minor
severity and the incidence may be underestimated as
the victims do not present at Accident and Emergency
departments or require hospital admission. Using the
Glasgow Coma Scale (GCS) as a criterion of severity of
head injury, Henry, Hauber and Rice (1992) reported
that 56.5% of a sample of 138 children were classified
as having sustained a mild injury, 17.4% as moderate
and 26.1% as severe. They found an 8% mortality in
the series.
The Scottish study of Brookes et al. (1990), identified
only 1% of children (and 5% of adults) in the sample as
having evidence of altered consciousness at the time
of presentation to Accident and Emergency departments. In a wide and comprehensive study of deaths
from head injury in Scotland and England, hospital
admissions for head injuries in Scotland, England and
Wales and attendances at casualties in Scotland for
head injury for the year 1974, Jennett and MacMillan
(1981) found that, for children, the overall rate of
attendance at Accident and Emergency departments
following a head injury was 3017 per 105 of the
population and that this figure accounted for 40% of
the attenders. They concluded that the attendance rate
at Accident and Emergency departments following
head injury is a reliable guide to the incidence of head
injury in the community, although access to a hospital
(e.g. in rural areas) might bias the findings.
The NHSCIS study (Kalsbeek et al., 1980) used a
more restricted definition of head injury, by excluding
attenders who were not admitted and including all
deaths from head injury and hospital admissions. This
study identified an annual rate of 230 per 105 of the
population and an annual prevalence rate of 524 per
105 of the population for children (015 years of age).
For all ages, the prevalence rate was 439 per 105 of the
population. Severity was classified as concussion
only (78.9%), while the more severe groups were
Table 1.10 Deaths in infancy and childhood from road crashes in South Australia, 19831988 (Source: from Simpson
et al., 1992)
Age 024 months
Age 214 years
Type of road user
Male
Female
Male
Female
Total
(< 15 years)
Car passengers
Pedestrians
Pedal cyclists
8
1
0
3
0
0
22
19
17
26
15
4
59
35
21
Total
58
45
115
14
EPIDEMIOLOGY
characterized as contusion/laceration (5.7%) and

hematoma (1.2%).
Benefits of seat belt use for children were demonstrated by Osberg and Di Scala (1992) for 413 children
with a head injury, where the mortality for restrained
children was 2.4% and for unrestrained children 4.5%.
The children who were unrestrained exhibited more
body areas injured, injuries of greater severity, longer
hospital stays and were 15% more likely to be
discharged with an impairment. Agran, Winn and
Dunkle (1989) focused on a subset of 49-year-olds,
often too large for child-type safety seats and too small
for adult-type lap-sash belts, where, of the children
injured, 70% sustained head and facial injuries.
1.11
Minor head injury
Epidemiological studies of minor head injury cause

greater methodological difficulties in attempting to
obtain accurate data than those of severe head injury,
where death and hospital statistics provide objective
records. When there are more serious injuries, together
with a minor head injury, hospital codings often omit
reference to the associated minor head injury. Particularly in rural areas or when hospital access is
difficult, it is probable that many minor head injuries
are unreported because the patients do not attend an
Accident and Emergency department. In hospital,
responsibility for care is often shared among neurosurgeons, general or orthopedic surgeons, neurologists, pediatricians, geriatricians and primary care
physicians.
The Health Interview Survey (National Center for
Health Statistics, 1977) of households in the USA was
based on a national probability sample of 116 000
persons in 1975. It yielded an estimated annual rate of
head injury of 6 per 1000 of the population. This was
probably an overestimate as facial injuries were
included in the definition of head injury.
Table 1.11
Definitional issues as to what constitutes a minor

head injury further compound the problem. Three
major population studies (Table 1.11) used differing
definitions for minor head injury and differences at
this level render comparisons between studies difficult. While the Glasgow Coma Scale has provided
useful descriptors for more severe head injury, it was
never intended to classify the different types of minor
head injury (Jennett 1989). Neither is the level of
consciousness on admission or presentation at Accident and Emergency departments necessarily a guide
as to the severity of a head injury, and serious lowenergy injuries such as an extradural hematoma may
be present in patients with no history of loss of
consciousness or only a transient alteration.
Jennett and Miller (1972) found that, of patients who
were admitted to the neurosurgical unit in Glasgow
with a compound depressed skull fracture, 44% had
never lost consciousness. Kraus and Nourjah (1988)
commented that in 5% of patients categorized as
having a minor head injury with a GCS between 13
and 15, a hospital physician made a diagnosis consistent with the code of cerebral contusion, laceration
or hemorrhage and the authors questioned the
adequacy of the GCS as an accurate descriptor for the
less severe end of the head injury spectrum when
describing intracranial pathology. Those with a contusion, laceration or hemorrhage had a longer median
hospital admission with necessarily increased costs.
Russell and Nathan (1946) used the duration of posttraumatic amnesia (PTA) to rank head injuries in
severity and Gronwall and Wrightson (1974) defined a
study group of minor head injuries by a PTA less than
24 hours. This latter definition or an even shorter
period is probably a better criterion, but a comprehensive and uniform definition of minor head injury
has as yet eluded the neurosurgical community and
there is a clear need to establish adequate descriptors
that are generally acceptable and applicable in order
Epidemiological data for minor head injury (LOC = loss of consciousness; PTA = post-traumatic amnesia)
Year of
study
% of all
head injury
Rate/105
population
Annegers, Grabow and

Kurland, 1980
Olmstead County, MN, USA
193574
60%
149.0
LOC or PTA < 30 min

Fractures excluded
Kraus and Nourjah, 1988

San Diego, CA, USA
1981
72% all cases, including

deaths.
82% of hospitalized cases
130.8
Glasgow Coma Scale

1315
Whitman, CoonleyHoganson and Desai, 1984,

Chicago, IL, USA
197980
80%
74163*
LOC < 30 min and

trivial
Author
Definition
A range for three area populations studied; inner city, predominately African American people (163 per 105 population);
Evanston Caucasian (74 per 105 population) and Evanston African American people (227 per 105 population). See text.
COUNTING THE COST
to standardize the terminology so that valid comparison can be made between studies.
The three population studies shown in Table 1.11
reveal that, of all head injuries, those classified as
minor account for between 60% and 80% and these
studies suggest that the annual rate lies between 130.8
and 163 per 105 of the population. No doubt, however,
there are great variations according to the community
under study.
The Chicago study (Whitman, Coonley-Hoganson
and Desai, 1984) compared three areas of Chicago, IL.
Among inner-city African Americans the incidence of
minor head injury was 163 per 105 of the population.
In Evanston, a suburb of Chicago, the population
consisted of 75% Caucasian and 21% African Americans at the time of the study. The incidence of minor
head injury for the former was 74 per 105 of the
population and for the latter 227 per 105 of the
population.
Annegers, Grabow and Kurland (1980) reported
that for minor head injury, the peak incidence was
1524 years for both males and females but females
had only a slightly lower incidence in the 514-year
age group. The study period stretched from 19351970
and during this time there was a steady increase in the
incidence of minor head injury.
Kraus and Nourjah (1988) studied 2435 patients
with minor head injury in San Diego, CA. They found,
like Annegers, Grabow and Kurland (1980), that males
showed a peak incidence for the age group 1524
years, with an annualized rate of 174.7 per 105 of the
population. A differing age pattern was identified for
females, for whom the incidence showed a bimodal
distribution with peaks at 05 years and at over 75
years of age. Overall, males were twice as likely to
sustain a minor head injury, except those under 5
years of age or over 45 years of age where, in the latter
group, males remained in excess, but the male to
female ratio decreased. Motor-vehicle-related accidents were the most frequent cause of minor head
injury in this study (42%), with falls (23%) next most
frequent, followed by assault (14%), sport and recreation (12%) and other causes at 9% (Figure 1.2). Costs
were related largely to the length of hospital admission. In San Diego, 64% remained in hospital less than
3 days, 87% less than 1 week and 5% more than 2
weeks. This last group was considered to be most
probably due to associated injuries. Costs of care were
US$6.3 million dollars (1981 dollar value) or an
average cost of US$2774 per admission, increasing
with the length of stay.
Minor head injury is a serious and underestimated
public health problem, deserving of more attention by
investigators and health planners. While a physical
disability following a minor head injury is unusual
and recovery is the rule, there is good evidence
15
Figure 1.2 Causes of minor head injury according to Kraus

and Nourjah, 1988.
(Gronwall and Wrightson, 1974, Wrightson and Gronwall, 1981) that cognitive and neurobehavioral impairments may cause substantial disruption to individuals
and families, with the loss of many hours of productive work.
1.12
Counting the cost
Monetary costs in head injury are directed towards

provision of medical and hospital care in the acute
phase, rehabilitation and specialized retraining in the
months after the injury and then to many areas such as
community and family support services for the headinjured as they attempt to reintegrate with society.
Costs incurred also include lost earning capacity and
the effect on family units.
The NHSCIS study (Kalsbeek et al., 1980) divided
costs into direct and indirect. Direct costs were
associated with the monetary values of real goods and
services that were provided for health care and
indirect costs were the monetary loss incurred by
society because of interruption of productivity by the
injured person. In 1974 US dollar values, the total cost
for all head injuries studied was $2384 million, of
which $696 million was related to the direct costs of
care and $1688 million to indirect costs. On a per
patient basis, the average cost was $2534. The largest
annual cost was $889 million in the 2544-year age
group where the losses incurred due to productivity
were maximal. Costs associated with motor vehicle
crashes as a cause of head injury were the highest at
$1639 million, reflecting both the high frequency and
severity of these injuries, followed by falls at $316
million and all other causes at $429 million.
Families provide the major support and respite for
head-injured patients after they leave hospital (Jacobs,
16
EPIDEMIOLOGY
1988; Fearnside, McDougall and Lewis, 1993). In the

Los Angeles study (Jacobs, 1988) it was found that
most families experienced substantial financial stresses, ranging from mild to severe and particularly
relating to medical and rehabilitation costs where
insurance was inadequate. In over one-third of families, a member was required to act as a permanent
supervisor, often at the cost of employment, compounding already existing losses and estimated as an
annual cost of around $28 000.
The economics of the care of head injury is most
accurately measured when the costs against the
injured person are provided by insurance. In NSW,
Australia, a compulsory insurance scheme, subsidized
by all drivers, is provided to cover all persons injured
through no fault of their own in motor vehicle
accidents. The scheme is administered by the Motor
Accident Authority, a statutory body independent of
the government.
Between 1989 and 1993 in NSW large claims
(defined as in excess of Aust$0.5 million) for injuries
sustained as a result of motor vehicle accidents
amounted to $29.7 million, of which 54% were for
brain injury, 22% for spinal cord injury, 3% for both
head and spinal cord injury and 11% for other claims
(Motor Accidents Authority, 1993). Although claims in
excess of $0.5 million numbered only 0.7% of all the
motor vehicle claims over that period, they exceeded
30% of the total estimated dollar costs. For claims in
excess of $2 million, 60% were for brain injury. Males
accounted for over two-thirds (69%) and females for
less than one-third (31%) of these large claims.
Costs were incurred in various categories, including
medical and hospital (3.7%), rehabilitation services
Table 1.12 Projected costs of care by severity of disability
statutory discount rate (prepared by Walsh, J. and Cuff,
C. for the Motor Accidents Authority of NSW, 1992)
Category of care
Projected annual cost

(Australian $
1992 value)
Persistent vegetative state.

No level of meaningful
responsiveness
132 235
Profound physical and cognitive

disability, possible ongoing medical
problems
117 385
Severe cognitive disability.

No behavioral disability or ongoing
medical problems
56 074
Little or no physical disability: severe

cognitive disability, no behavioral
disability but no family support
42 686
Moderate disability (includes a range

of outcomes)
27 136
(4%), long-term care and home care (40.7%), past and

future economic loss (28.7%), general damages
(20.8%), legal costs (5.1%), aids and appliances (2.3%),
home and vehicle modifications (0.5%) and other
miscellaneous costs (20.2%). Payments for indirect
costs such as damages and economic losses were by
far the greatest quantum and legal charges were only
slightly less than the cost for the entire medical,
hospital and rehabilitation services provided.
Projected annual costs varied with the level of
neurological disability and functional independence.
Table 1.12 shows the classification devised to categorize these various levels and the projected annual costs
for each level. These costs covered such supportive
areas as day activity, accommodation, attendant care
and respite care and did not include costs of any
ongoing medical problems requiring hospital admission. For categories A and B, those most severely
disabled, permanent accommodation was planned
and for categories D and E the emphasis was on
outpatient day activities.
1.13
Reducing the burden
1.13.1
PREVENTION BY PUBLIC EDUCATION
Prevention is better than cure and much less expensive. Measures aimed at prevention of head injury are
therefore important public policy initiatives, both for
developed and developing countries (Trinca et al.,
1988). Risk reduction as a public health measure
demands a national response. Community awareness
of the problem of head injury generates public and
political education to establish programs to prevent
such injuries.
A program to promote public education and awareness was undertaken by the American Association of
Neurological Surgeons (AANS) and the Congress of
Neurological Surgeons (CNS) in 19851986, the Think
First Prevention Program (Eyester and Watts, 1992).
The aim of this program was to persuade individuals
to alter their risk-taking behavior using a State-based
syllabus of youth-oriented programs, a reinforcement
and public education program and a program to
influence government policy. The program provided
lectures by neurosurgeons or lay members, videos,
brochures and a film entitled Harms Way.
Transport-related accidents are the most frequent
cause of severe head injury and much effort is aimed
at modification of the environment in which the
motorist, motorcyclist, pedal cyclist or pedestrian
interacts with the environment. Public awareness
campaigns are crucial in providing information to the
community such that individuals can use the information to protect themselves or demand better protection. One example of this is the realization by some
REDUCING THE BURDEN
motor vehicle manufacturers that cars can be sold

successfully using safety devices such as airbags and
reinforced passenger capsules as features, rather than
the engine capacity and acceleration rate.
1.13.2
ROAD TRAFFIC SAFETY
Evans (1991) in his thoughtful book Traffic Safety and

the Driver made a judgmental assessment of the factors
that led to the decrease in car-related fatalities recorded in many developed countries after 1975. He
concluded that changes in human behavior had been
much more influential than changes in transport
technology. Within the realm of transport technology,
he saw better roads as more influential than improved
traffic control or safer vehicles. Few road crashes,
however, are monocausal and crashes commonly
result from a mix of behavioral and environmental
factors, often cumulative and imponderable.
1.13.3
ALCOHOL CONTROL
Ethanol (ethyl alcohol) is a well-documented cause of

road crashes. There is also reason to believe that a high
blood alcohol concentration (BAC) reduces the individuals capacity to survive a crash (Waller et al., 1986),
especially if associated with head impact. Alcohol
control is therefore a central issue in head injury
prevention programs.
Borkestein et al. (1964) performed random BAC
estimations on some 1300 drivers in the state of
Michigan, USA. From this large database, a casecontrol study showed that the risk of crash involvement rose rapidly when the BAC exceeded 100 mg/dl.
Later studies have confirmed this finding. Evans
(1991) reviewed recent USA data on BAC findings in
fatally injured drivers. Data from 26 states showed
that concentrations in excess of 100 mg/dl were found
in 24.9% of drivers killed in two-vehicle crashes, but in
54.7% of drivers killed in single-vehicle crashes. Evans
pointed out that the first figure understates the role of
alcohol in two-vehicle crashes, since sober drivers are
often killed by drunk drivers.
The reason for this very clear relationship is, of
course, the effect of alcohol on neuronal function, and
especially on the neurological and neuropsychological
functions relevant in driving. Moskowitz and Robinson (1987) reviewed the English literature on the
relations between BAC and testable skills relevant in
driving. They concluded that some important skills
are impaired well below the level of 100 mg/dl, the
legal BAC limit in many states in the USA. However,
there is much variation in individual tolerance of
alcohol.
Awareness of the importance of alcohol in road
crashes, and indeed in other accidents likely to cause
17
head injury, has prompted much legislation and

educational endeavor. The Scandinavian countries led
the way by making it a crime to drive with a BAC in
excess of a specified level. Norway (Glad, 1987) first
set this level at 50 mg/dl, and many countries have
accepted this figure. In the UK, a level of 80 mg/dl
was fixed and after the enactment of what was then a
very controversial measure, a dramatic fall in drunk
driving was reported (Ross, 1982). The deterrent effect
of such legislation depends largely on the publics
perception of the risk of detection. Frequent random
breath testing has been shown to increase general
awareness of the risks incurred by drinking before
driving. Public support for alcohol control is also
important. In many countries, government-funded
educational campaigns have been conducted. In the
USA, independent bodies of activists such as MADD
(Mothers Against Drink Driving) have been prominent, and Evans (1991) credits these bodies, together
with the media, for much of the change in societal
attitudes to alcohol which have been seen in North
America, mirrored in declining per capita consumption
of alcoholic drinks. Nevertheless, drunk driving
remains a major cause of injury, even in countries
where it is severely penalized (Glad, 1987) or under a
religious prohibition (Mekky, 1984).
1.13.4
ROAD CONSTRUCTION AND SPEED CONTROL
Road engineering has become an exact science, and

crash prevention is one of its objectives. Speed control
is a key part of crash-prevention strategy. In this
strategy, driver education and law-enforcement are
essential elements, but a well designed road system is
also of great importance.
Road design can prevent road crashes in several
ways. Appropriate banking and well-maintained surfaces make loss of directional control less likely. Good
visibility and warning signs make collisions less likely
at intersections. Unidirectional traffic flow in rural
areas makes head-on collisions less likely. In cities,
separation of traffic streams by a median strip helps
pedestrians to cross in safety.
Control of access is an important strategy for all
categories of road. Pedestrians and pedal cyclists are
segregated by footpaths and cyclepaths. Traffic lights
facilitate pedestrian crossings at predictable sites.
Lights also control speed, especially if monitored by
infrared cameras. Urban street speed control can be
achieved by obstacles. Brindle (1992) found this
method of speed control effective but unpopular.
Fast driving is nevertheless an economic imperative
in modern society, especially in thinly populated
countries. Well-designed freeways allow fast driving
under optimal conditions by removing pedestrians,
segregating slow traffic and avoiding intersections.
18
EPIDEMIOLOGY
When crashes do occur on freeways, they are likely to

be more serious. However, there is general agreement
that limited-access freeways can reduce rural road
deaths and severe injuries (Evans, 1991; Kraus et al.,
1993). This achievable only when drivers appreciate
the purpose of the road system: in Nigeria, Asogwa
(1992) found that better roads resulted in excessive
speed and reckless driving, with an actual increase in
the death toll.
1.13.5
MOTOR VEHICLE DESIGN
Improvements in motor vehicle design, combined

with quality control in manufacture, can reduce the
incidence of head injuries in two ways, by making the
vehicle less likely to crash and by making crashes
more survivable. Motor vehicle designs intended to
promote these benefits are (or should be) tested
experimentally and in the field. If successful, the
design is standardized and becomes the basis of a
safety standard which can be tested and enforced by
governmental agencies.
Much effort has gone into the design of improved
engine performance, lights, steering systems and other
aspects of roadworthiness. It is often difficult to prove
that improved car design will prevent road crashes.
Defective or inappropriate car design certainly can
cause crashes, as was argued by Nader (1966) in his
memorable polemic Unsafe At Any Speed. He
denounced the Chevrolet Corvair with fervor, and
linked reported accidents with this car to the design of
the rear suspension. Naders book has been compared
with Uncle Toms Cabin, the novel that helped to
abolish slavery in the USA. It certainly publicized the
dangers of bad car design. It has, however, been
difficult to show that a seemingly good design will
lower the road toll. Thus antilock electronic braking
systems increased vehicle stability in tests demanding
rapid braking on surfaces of various adhesive properties (Rompe, Schneider and Wailrich, 1988). However,
field trials by taxi drivers in Munich showed no
advantage in crash rates, perhaps because these
brakes encouraged drivers to take risks (Evans, 1991).
One of the few certain facts about the relation
between vehicle design and road safety is that the
occupants of large vehicles are safer than the occupants of small vehicles.
Cars and other vehicles can be designed to minimize the risk of injury, and especially head injury, by
absorbing some or all of the impact energy that might
otherwise cause injury to the occupants of the vehicle.
To absorb the energy of a frontal crash, the forward
components of the vehicle can be designed to crumple.
It is harder to design protection against the effects of a
side impact. Car occupants often suffer head injuries
by impact on some structure in the interior of the car
and these structures can be designed to modify the

nature of the impact on the head. Projections capable
of inflicting a penetrating injury are eliminated in
modern cars. As a consequence, compound skull
fractures are now rare in car occupants, except after
very severe crashes with massive intrusion into the car
interior.
1.13.6
WINDSCREENS
These are a common site of head impact and can be

made of laminated glass, which absorbs energy by
yielding and does not often shatter into knife-like
fragments capable of penetrating the brain or the eyes.
This safety measure appears to have performed well
in real-life crashes (Hass and Chapman-Smith, 1976).
Attention is now being given to the steering wheel and
to the hard components of the roof and its supporting
pillars. In an in-depth study of 26 car occupants
sustaining severe or fatal head injury, Paix et al. (1991)
identified the impacting agents in 22. The roof or its
supporting pillars appeared responsible in nine and
the steering wheel in two of the accidents.
1.13.7
AIRBAGS
In frontal crashes, the impact energy transmitted to a

car occupant can be further reduced by inserting an
airbag between the occupant and the striking force.
Zador and Ciccone (1991) assessed the value of
airbags in preventing fatalities by comparing frontal
crashes in cars fitted with airbags with non-frontal
crashes and with crashes in cars fitted with manually
secured seat belts only. They concluded that airbags
alone reduced the risk of driver death by 21% when
compared with the risk to an unsecured driver, and by
9% when compared with the risk to a belted driver.
The advantage was greater with large cars. Airbags
can also be used to mitigate the effect of side impacts.
It has been convincingly shown that the occupants of
vehicles who are ejected are at greater risk than those
remaining in the vehicle. Cars can be fitted with
devices to prevent ejection, such as effective door
locks and seat belts.
1.13.8
SEAT BELTS
The evidence that seat belt use in motor vehicles does

decrease injury is conclusive. The Iowa Safety
Restraint Assessment 198788 (ISRA, 1989) was a
population based study of 16 hospitals in Iowa over a
period of 5 months from November 1987 until March
1988. Of the 1454 persons injured, 48% were wearing a
restraint and 52% were unbelted. Unbelted persons
were 8.4 times more likely to sustain a head injury
and 2.7 times more likely to sustain a fracture. The
REDUCING THE BURDEN
average length of stay and the hospital costs were

both greater for the unbelted. The investigators urged
increased public awareness and education, emphasizing the ability of safety restraints to reduce crash
injuries, the instruction of children about the importance of wearing seat restraints and the education of
the 1624-year-old age group, who are most at risk.
Other studies support the assertion that the use of
restraints in motor vehicles decreases the frequency
and severity of injury (Orsay, Dunne and Turnbull,
1990; Bradbury and Robertson, 1993).
It is now agreed that seat belts should embody
thoracic restraint as well as a pelvic band. The three
point lap-sash system is a convenient and popular
means of achieving this goal, but seat belts must be
properly fitted and sited. There are many reports of
lumbar spinal and abdominal visceral injury from the
lap belt, especially in children (Newman et al., 1990).
Some of these have resulted from failure to site the belt
to lie below the anterior superior iliac spines. Devastating injury to the cervical carotid artery may occur
from a shoulder sash, possibly caused by an inappropriately high siting of the anchorage.
1.13.9
DRIVER BEHAVIOR
There is also ample evidence that individuals do alter

their risk-taking behavior when faced with legislative
penalties. States, Annecharico and Good (1990) compared the frequency and severity of all injuries in New
York State over a period of 18 months, before and after
the introduction of a law making the use of seat belts
in motor vehicles mandatory. The use of seat belts
increased from 11.2% before the law was enacted to
53% afterward. Hospital admission for motor vehicle
trauma fell by 11.9% after, but had increased by 2.6%
in the study period prior to the law.
Salmi et al. (1989) used a regional study before and
after seat belt legislation was introduced in France in
1979 to show a significant decrease in motor-vehiclerelated head injury after the law was enacted.
Thomas (1990) also used a regional study in south
London to confirm a significant decrease in head
injury following the introduction of seat belt legislation in England.
Rutherford et al. (1985) compared hospital attendances before and after the introduction of the mandatory use of seat belts in the UK. There was a large
fall in the number of brain injuries of all types, but a
slight rise in the number of serious (AIS 3) brain
injuries. This appeared to be because seat belts gave
good protection to front seat passengers, but did not
reduce the risk of impact of the drivers head on the
steering wheel. This interpretation provides powerful
support for the use of a drivers-side airbag. All
occupants benefit from seat belt use and airbags
19
should be regarded as complementary rather than as

an alternative to a seat belt.
Child occupants of cars provide special problems
and Simpson et al. (1992) reviewed these in the context
of a series of 115 fatalities. The infants thin, flexible
skull gives little protection against impacts of any
kind. Attention, therefore, has been given to the
design of safety capsules in which the infant is secured
against a sudden dislocation and head impact against
some part of the interior of the car. Children may be
too small to be safely fitted with an adult-size seat belt,
and may suffer a spinal injury from violent deceleration while belted inappropriately. At present, most
authorities advise a special harness for small children
with a four or five point anchorage. Older children
may use an adult belt system if a foam chair-like
support is inserted to give a better fit. In Sweden,
backward-facing child seats are favored and these
may give superior protection, at least in frontal
crashes (Carlsson, Norm and Ysander, 1989).
1.13.10
MOTORCYCLE HELMETS
Studies of the effects of legislative compulsion in the

wearing of helmets by motorcyclists have revealed
findings similar to seat belt studies in that the
incidence and severity of head injury among motorcyclists has decreased. Many studies have confirmed this
hypothesis and the most convincing derive from the
astonishing decision of 27 states in the USA to repeal
laws enforcing mandatory use of helmets. Sosin,
Sacks and Holmgreen (1990), in a mortality study of
National Center for Health Statistics data from
19791986, found that 53% of the 28 749 motorcyclist
deaths were due to head injury. The annual death rate
was 5.5 per 106 of the population for states with full
helmet laws, 10.2 per 106 of the population in states
with partial laws and 10.4 per 106 of the population in
states with no helmet laws. Death rates increased in
three states which changed from full to partial laws
during the study and were lowest in states with
comprehensive helmet laws.
McSwaine and Belles (1990) found that the incidence of injury rose where motorcycle helmet laws
were repealed (Kentucky and Louisiana) and fell
again after their reintroduction (Louisiana). Based on
1989 US$ values, an annual $121 million of additional
medical care and rehabilitation costs were ascribed
directly to the non-use of helmets.
One 12-month study in Maryland, where no helmet
laws were in place (Shankar, Ramzy and Soderstrom,
1992) found that helmet usage was only 35%, and was
30% for those who were fatally injured. While head
injuries are the most common cause of death among
motorcyclists in most studies, injuries to the integument and fractures to the extremities and pelvis are
20
EPIDEMIOLOGY
more frequent non-fatal injuries, as shown in Singapore, where helmet laws were in place (Wong, Phoon
and Lee, 1989; Shankar, Ramzy and Soderstrom,
1992).
1.13.11
BICYCLE HELMETS
Of all road users, the cyclist and pedestrian are the

least well protected from injury on the roads and
attention is now focused on the value of helmets for
pedal cyclists. Of fatal bicycle accidents in the
015-year group in Ontario, Canada, Spence et al.
(1993) found that 91% were considered unsurvivable
and 89% were due to head injury. No victim was
wearing a helmet and the authors concluded that
emphasis should be placed on bicycle safety education
for children and the promotion of helmet use. Helmets
do prevent death from head injury, as McDermott et al.
(1993) showed in Victoria, Australia: there, in a series
of 1710 consecutive injuries in pedal cyclists, the head
injury rate was significantly less for the helmeted
riders (34.8%). The AIS scores were also significantly
less for helmeted than unhelmeted riders. After
legislation was introduced in Victoria making the
wearing of bicycle helmets compulsory, usage
increased from 31% to 75% (Lane and McDermott,
1993). Uniform legislation throughout Australia
requiring compulsory bicycle helmet usage was introduced sequentially in 1991 and 1992 and the effects on
head injury are yet to be evaluated. Data from the
Victorian Traffic Accident Corporation has shown
that the number of cyclists admitted to Victorian
Public Hospitals fell by 37% after the law was
enacted.
1.13.12
HEAD INJURY SURVEILLANCE
In order to plan prevention and treatment programs

for the head-injured and to evaluate such programs,
data collection is essential. The increasing use of head
injury surveillance or registry systems will provide
information regarding the changing patterns in a
geographical area, which should allow for equitable
resource distribution. Head injury surveillance uses
passive methods of data collection such as death
records or hospital diagnosis at separation, reported to
a central data collection point. Surveillance systems
may have low capture rates, whereas head injury
registries identify new cases using a researcher, often a
nurse investigator specially trained in methods of data
collection and recording. Head injury registries are
generally more expensive to maintain than surveillance, but data capture is more complete.
While such data collection is successfully established in many developed countries (Woodward,
Dorsch and Simpson, 1984; Parkinson, Stephenson
and Phillips, 1985), particular problems emerge in

developing countries where injury now outranks
infectious disease as a cause of death (Mock et al.,
1993). These authors identified the need for trauma
registries in developing countries when they compared trauma profiles at a rural African hospital in
Berekum, Ghana and a Level 1 center in Seattle, WA,
USA. They found that vital registry data was incomplete in Ghana as only a minority of deaths, namely
those occurring in hospital, were recorded, and that
there was a need for accurate registrations. The study
identified a lack of pre-hospital care and delays in
transport to hospital and considered these less costly
to improve and more likely to improve mortality rates
than expensive improvements in hospital services. A
simple method to obtain incidence data based on a
capture-recapturetechnique has been suggested by
Chiu et al. (1993) which might be useful for developing
countries with limited resources.
1.13.13
TRAUMA CARE SYSTEMS
Recognition of the importance of the coordination of

resuscitation, triage, evacuation and primary and
secondary care for the patient with severe or multiple
injuries has led to the development of trauma care
systems that aim to provide systematized management from the time of injury through initial resuscitation and definitive treatment to rehabilitation. The
system is regionalized to ensure that the severity of
the injury is appropriate to the resources available in a
particular geographical area (Eastman et al., 1987).
Such systems are complex and include pre-hospital
personnel such as ambulance officers and police,
systems management personnel and staff at acute care
and rehabilitation facilities (American College of
Emergency Physicians, 1987). A process for the stepwise establishment for such a system of trauma care
has been described (West, Williams and Trunkey, 1988;
US Department of Transportation, 1989).
In the USA and elsewhere, involvement of hospitals
in such programs has led to their categorization at
various levels, depending on the resources available:
Level 1 Regional Resource Trauma Center. These

are capable of providing total care at all levels as
well as education and research programs.
Level 2 Community Trauma Center. These
centers provide total care but not the education and
research programs.
Level 3 Rural Trauma Hospital. These centers
serve local communities and provide care dependent upon resources available, often stabilization and
transfer only.
Head injuries are responsible for the majority of

trauma deaths (Chan, Walker and Cass, 1989; Sosin,
REFERENCES
Sacks and Smith, 1989; Daly and Thomas, 1992) and

the involvement of neurosurgeons is pivotal in the
design of these systems. The Joint Section of the AANS
and the CNS has recommended minimal standards for
an institution to qualify as a designated center for the
treatment of severe neurological trauma (Pitts, Ojemann and Quest, 1987):

a specifically named neurosurgeon on call at all

times;
an emergency room staffed 24 hours per day by a
physician certified in Advanced Trauma Life Support (ATLS);
an operating department capable of rapid acceptance of a patient for a craniotomy or spinal surgery
at all times;
availability of a CT scanner and a technician at all
times;
an appropriately equipped and staffed intensive
care unit;
a clearly defined bypass plan in the event of an
unavoidable lack of availability of a neurosurgeon.
The establishment of trauma systems does result in

improvements in mortality (Shacksford et al., 1986),
although problems of underfunding, difficulties in
recruitment of residents to careers in trauma care and
problems of the establishment of such systems in rural
areas have impeded development in many areas of the
USA (Eastman et al., 1991; Esposito et al., 1991).
1.13.14
REHABILITATION
A comprehensive trauma care system for head injury

includes rehabilitation services that aim to minimize
any disability and maximize the potential for functional and meaningful recovery. Neurological rehabilitation has a longer time profile and consists of several
stages, characterized initially by involvement of the
rehabilitationist during the acute phase of management of the head-injured patient and the family. After
the acute phase, subacute programs are generally
designated for slower stream patients who remain in
coma or post-traumatic amnesia (PTA) and consist of
coma management and neurobehavioral therapy.
Later, after the PTA resolves, an individual program is
prepared to provide for transitional living, day care
and activities, and supported or sheltered employment. Thus the trauma service forms an integrated
therapy system with the rehabilitation unit, providing
facilities for research, education and a standardized
data base for the assessment of performance.
There is evidence that early intervention with
rehabilitation therapy does decrease the length of
admission and treatment in a rehabilitation facility
(Cope and Hall, 1982). The establishment of designated and purpose built brain injury units such as
21
established by the Motor Accident Authority in NSW,

Australia, will focus attention on the particular needs
of the brain injured and hopefully result in improved
outcomes.
The quality and cost-effectiveness of rehabilitation
programs has now become an important issue in an
era of cost containment. If patients can become
independent and even return to some form of gainful
employment as a result of early intensive rehabilitation, the cost savings over nursing or group home care
would be substantial (Aronow, 1987). However, concerns might arise as to the selection of patients for
such programs if they were perceived to be potentially
slow stream. Some who were denied the equality of
access to such a program might not achieve their full
potential for recovery.
There is now good evidence of improved survival
following head injury compared with outcomes 1020
years ago (Bowers and Marshall, 1980; Fearnside et al.,
1993). It will be of considerable interest and importance to see whether the many improvements in head
injury care and neurological rehabilitation programs
can be translated into objective improvements in the
quality of life and its value to the survivors of head
injury and their families.
1.14
References
Association for the Advancement of Automotive Medicine (1990) Abbreviated

Injury Scale, 1990 revision, Association for the Advancement of Automotive
Medicine, Des Plaines, IL, USA.
Adelson, P., Thomas, S., Hovda, D. et al. (1991) Boxing fatalities and brain
injury. Perspectives in Neurological Surgery, 2, 167188.
Agran, P., Winn, D. and Dunkle D. (1989) Injuries among 4 to 9-year-old
restrained motor vehicle occupants by seat location and crash impact site.
American Journal of Diseases of Childhood, 143, 13171321.
American College of Emergency Physicians (1987) Guidelines for trauma care
systems. Annals of Emergency Medicine, 16, 459463.
Anderson, D., Kalsbeek, W. and Hartwell, T. (1980) Report on the National
Head and Spinal Cord Injury Survey, Journal of Neurosurgery (Supplement),
53, S11S18.
Anderson, D. and McLaurin, R. (1980) Report on the National Head and
Spinal Cord Injury Survey. Journal of Neurosurgery (Supplement), 53,
S1S43.
Annegers, J., Grabow, J and Kurland, L. (1980) The incidence, causes and
secular trends of head trauma in Olmstead County, Minnesota,19351974.
Neurology, 30, 912919.
Aronow, H. (1987) Rehabilitation effectiveness with severe brain injury:
translating research into policy. Journal of Head Trauma Rehabilitation, 2,
2436.
Asogawa, S. (1992) Road traffic accidents in Nigeria. Accident Analysis and
Prevention, 24, 149155.
Australian Institute of Health and Welfare. (1992) Australias Health in 1992,
The Third Biennial Report of the Australian Institute of Health and Welfare,
AGPS, Canberra.
Baker, S., ONeill B., Haddon W. et al. (1974) The Injury Severity Score: a
method for describing patients with multiple injuries and evaluating
emergency care. Journal of Trauma, 14, 187196.
Borkenstein, R., Crowther, R., Shumate, R. et al. (1964) The role of the drinking
driver in traffic accidents. Department of Police Administration, Indiana
University, Bloomington, IN.
Bowers, S. and Marshall, L. (1980) Outcome in 200 consecutive cases of head
injury treated in San Diego County: a prospective analysis. Neurosurgery, 6,
237242.
Boyd, C., Tolson, M. and Copes, W. (1987) Trauma care: TRISS method. Journal
of Trauma, 27, 370.
Bradbury, A. and Robertson, C. (1993) Prospective audit of the pattern,
severity and circumstances of injury sustained by vehicle occupants as a
result of road traffic accidents. Archives of Emergency Medicine, 10, 1523.
22
EPIDEMIOLOGY
Brainard, B., Slauterbeck, J. and Benjamin, J. (1989) Injury profiles in

pedestrian motor vehicle trauma. Annals of Emergency Medicine, 18,
881883.
Brindle, R. (1992) Local street speed management in Australia is it traffic
calming? Injury Analysis and Prevention, 24, 2938.
Brookes, M., MacMillan, R., Cully, S. et al. (1990) Head injuries in accident and
emergency departments. How different are children from adults? Journal of
Epidemiology and Community Health, 44, 147151.
Bucklew, P., Osler, T., Eidson, J. et al. (1992) Falls and ejections from pickup
trucks. Journal of Trauma, 32, 468471.
Bull, J. (1975) The Injury Severity Score of road traffic casualties in relation to
mortality, time of death, hospital treatment and disability. Accident Analysis
and Prevention, 7, 249245.
Butler, R., Forsythe, W., Beverley, D. et al. (1993) A prospective, controlled
investigation of the cognitive effects of amateur boxing. Journal of Neurology,
Neurosurgery and Psychiatry, 66, 10551061.
Carlsson, G., Norm, H. and Ysander, L. (1989) Rearward facing child
restraints: the safest car restraints for children?, in Proceedings of the 33rd
Annual Meeting for the Association for the Advancement of Automotive Medicine,
Association for the Advancement of Automotive Medicine, Des Plaines, IL,
pp. 249265.
Champion, H., Copes, W., Sacco, W. et al. (1990) Injury severity scoring, in
Blunt Multiple Trauma, (eds J. Border, M. Allgbwer, S. Hansen et al.), Marcel
Dekker, New York, pp. 261276.
Chan, B., Walker, P. and Cass, D. (1989) Urban trauma: an analysis of 1116
paediatric cases. Journal of Trauma, 29, 15401547.
Chiu, W., Dearwater, S., McCarty, D. et al. (1993) Establishment of accurate
incidence rates for head and spinal cord injury in developing and
developed countries a capturerecapture approach. Journal of Trauma, 36,
206211.
Cope, D. and Hall, K. (1982) Head injury rehabilitation: benefits of early
intervention. Archives of Physical Medicine and Rehabilitation, 63, 433437.
Craft, A., Shaw, D. and Cartlidge, N. (1972) Head injuries in children. British
Medical Journal, 4, 200203.
Daly, K. and Thomas, P. (1992) Trauma deaths in the South West Thames
region. Injury, 23, 393396.
David, D and Simpson, D. (1995) Craniomaxillofacial Trauma, Churchill
Livingstone, Edinburgh.
Eastman, A., Lewis, F., Champion, H. et al. (1987) Regional trauma care
systems design. American Journal of Surgery, 154, 7987.
Eastman, A., Rice, C., Bishop, G. et al. (1991) An analysis of the critical
problem of trauma center reimsursement. Journal of Trauma, 31, 920926.
Edna, T. and Cappelen, J. (1985) Head injury in road traffic accidents. A
prospective study in Trondelag, Norway. Scandinavian Journal of Social
Medicine, 13, 2327.
Esposito, P., Maier, R., Rivara, F. et al. (1991) Why surgeons prefer not to care
for trauma patients. Archives of Surgery, 126, 292297.
Estwanik, J., Boitano, M. and Ari, N. (1984) Amateur boxing injuries at the
1981 and 1982 USA/ABF National Championships. Physician and Sports
Medicine, 12, 123128.
Evans, L. (1991) Traffic safety and the driver, Van Nostrand Reinhold, New York,
p. 341.
Eyester, E. and Watts, C. (1992) An update of the National Head and Spinal
Cord Injury Prevention Program of the American Association of Neurological Surgeons and the Congress of Neurological Surgeons. Clinical
Neurosurgery, 38, 252260.
Fearnside, M., McDougall, P. and Lewis, W. (1993) The Westmead Head Injury
Project. Physical and social outcomes following head injury. British Journal
of Neurosurgery, 7, 643650.
Fearnside, M., Cook, R., McDougall, P. et al. (1993) The Westmead Head Injury
Project. Outcome in severe head injury. A comparative analysis of prehospital, clinical and CT variables. British Journal of Neurosurgery, 7,
267279.
Field, J. H. (1976) Epidemiology of Head Injuries in England and Wales: With
Particular Application to Rehabilitation, HMSO, London.
Fox, M., Fabian, T., Croce, M. et al. (1991) Anatomy of the accident scene: a
prospective study of injury and mortality. American Surgeon, 57, 394397.
Friedman, G. (1987) Primer of Epidemiology, 3rd edn, McGraw Hill, New York.
Gallagher, S. and Finison, K. (1984), The incidence of injuries among 87,000
Massachusetts children and adolescents: results of the 198081 state wide
childhood injury prevention program surveillance system. American Journal
of Public Health, 74, 13401347.
Genarelli, T., Champion, H., Sacco, W. et al. (1989) Mortality of patients with
head injury and extracranial injury treated in trauma centers. Journal of
Trauma, 29, 11931202.
Glad, A. (1987) After 50 years with a per se law the drinking and driving
problem in Norway, in Alcohol, Drugs and Traffic Safety T 86, (eds P.
Noordzij and R. Rozbach), Excerpta Medica/Elsevier, Amsterdam, pp.
241244.
Gofin, R., Palti, H., Adler, B. et al. (1989) Childhood injuries: a populationbased study of emergency room visits in Jerusalem. Paediatric and Perinatal
Epidemiology, 3, 174188.
Gronwall, D. and Wrightson, P. (1974) Delayed recovery of intellectual
function after minor head injury. Lancet, ii, 605609.
Hamilton, M. and Tranmer, B. (1993) Nervous system injuries in horseback

riding. Journal of Trauma, 34, 227232.
Hass H. and Chapman-Smith, J. (1976) The role of vehicle safety glass in eye
injuries. New Zealand Medical Journal, 84, 5355.
Henry, P., Hauber, R. and Rice, M. (1992) Factors associated with closed head
injury in a pediatric population. Journal of Neuroscience Nursing, 24,
311316.
Hill, D., West, R. and Abraham, K. (1993) Impact of pedestrian injury on inner
city trauma services. Australia and New Zealand Journal of Surgery, 63, 2024.
Ho ok,
O. (1988) Head injury. Integrated trauma approach, optimal adaption
to disability. Scandinavian Journal of Rehabilitation Medicine (Supplement), 17,
6574.
Hung, C., Chiu, W., Tsai, C. et al. (1991) The epidemiology of head injury in
Hualien County, Taiwan. Journal of the Formosan Medical Association, 90,
12271233.
ISRA (1989) Safety-restraint assessment Iowa, 19781988. Journal of the
American Medical Association, 262, 28042805.
Jacobs, H. (1988) The Los Angeles Head Injury Survey: procedures and initial
findings. Archives of Physical Medicine and Rehabilitation, 69, 425431.
Jagger, J., Levine, J. and Jane, J. (1984) Epidemiologic features of head injury
in a predominantly rural population. Journal of Trauma, 24, 4044.
Jennett, B. (1975) Who cares for head injuries? British Medical Journal, 3,
267270.
Jennett, B. (1989) Some international comparisons, in Mild Head Injury, (eds H.
Levin, H. Eisenberg and A. Benton), Oxford University Press, Oxford,
pp. 2334.
Jennett, B. and MacMillan, R. (1981) Epidemiology of head injury. British
Jennett, B. and Miller, J. (1972) Infection after depressed fracture of the skull:
implications for management of non-missile injuries. Journal of Neurosurgery, 36, 333339.
Kalsbeek, W., McLaurin, R., Harris, B. et al. (1980) The National Head and
Spinal Cord Injury Survey: major findings. Journal of Neurosurgery (Supplement), 63, S19S31.
Klauber, M., Barrett-Connor, E., Marshall, L. et al. (1981) The epidemiology of
head injury: a prospective study of an entire community San Diego
County, California, 1978. American Journal of Epidemiology, 113, 500509.
Klonoff, H. and Thompson, G. (1969) Epidemiology of head injury in adults:
a pilot study. Canadian Medical Association Journal, 100, 235341.
Klopfer, J., Tielsch, J., Vitale, S. et al. (1992) Ocular trauma in the United States.
Eye injuries resulting in hospitalisation, 1984 through 1987. Archives of
Ophthalmology, 110, 838842.
Kraus, J. and Nourjah, P. (1988) The epidemiology of mild, uncomplicated
brain injury. Journal of Trauma, 28, 16371643.
Kraus, J., Black, M., Hessol, N. et al. (1984) The incidence of acute brain injury
and serious impairment in a defined population. American Journal of
Kraus, J., Anderson, C., Arzemanian, S. et al. (1993) Epidemiological aspects of
fatal and severe urban freeway crashes. Accident Analysis and Prevention, 25,
M239.
Lane, J. and McDermott, F. (1993) Do helmet wearing laws prevent head
injury? Medical Journal of Australia, 159, 719721.
Langley, J. and McLoughlin, E. (1989) Injury mortality and morbidity in New
Zealand. Accident Analysis and Prevention, 21, 243254.
Larsson, L., Melin, K., Nordstrom-Rohrberg, G. et al. (1954) Acute head
injuries in boxers clinical and electroencephalographic studies. Acta
Psychiatrica et Neurologica Scandinavica, (Supplement), 95, 142.
Lee, L., Shih, Y., Chiu, W. et al. (1992) Epidemiologic study of head injuries in
Taipei City, Taiwan. Chinese Medical Journal (Taipei), 50, 219225.
McDermott, F., Lane, J., Brazenor, G. et al. (1993) The effectiveness of bicycle
helmets: a study of 1710 casualties. Journal of Trauma, 34, 834845.
McGowan, I. (1959a) Protecting the boxer. Journal of the American Medical
Association, 169, 14091413.
McGowan, I. (1959b) Boxing injuries. American Journal of Surgery, 98,
509516.
Mackay, G., Hill, J., Parkins, S. et al. (1992) Restrained occupants on the
nonstruck side in lateral impacts. Accident Analysis and Prevention, 52,
147152.
McSwain, N. and Belles, A. (1990) Motor cycle helmets medical costs and the
law. Journal of Trauma, 30, 11891199.
Marshall, L. and Ruff, R. (1989) Neurosurgeon as victim, in Mild Head Injuries,
(eds H. Levin, H. Eisenberg and A. Benton), Oxford University Press,
Oxford, pp. 276280.
Mekky, A. (1984) Road traffic accidents in rich developing countries: the case
of Libya. Accident Analysis and Prevention, 16, 265277.
Miller, J., Sweet, R., Narayan, R. et al. (1978) Early insults to the injured brain.
Journal of the American Medical Association, 240, 439442.
Mock, C., Azdotor, K., Conklin, E. et al. (1993) Trauma outcomes in the rural
developing world: comparison with an urban level 1 trauma center. Journal
of Trauma, 35, 518523.
Moskowitz, H. and Robinson, C. (1987) Driving-related skills impairment at
low blood alcohol levels, in Alcohol, Drugs and Traffic Safety T86, (eds P.
Noordzij and R. Roszbach), Excerpta Medica/Elsevier, Amsterdam, pp.
7986.
REFERENCES
Motor Accidents Authority (1993) Large claims. Injuries and costs associated
with claims estimated to cost over half a million dollars or more. Motor
Accidents Authority of NSW, pp. 117.
Mueller, F. and Cantu, R. (1988) The annual survey of catastrophic football
injuries 19771988. Exercise and Sports Science Review, 19, 261312.
Nader, R. (1966) Unsafe at Any Speed, Pocket Books. New York.
National Center for Health Statistics (1977) Acute conditions. Incidence and
associated disability, United States, July 1974June 1975. Vital Health and
Statistics, Series 10, No. 14. National Center for Health Statistics, US
Government Printing Office, Washington, DC, pp. 167.
Newman, K., Bowman, L., Eichelberger, M. et al. (1990) The lap belt complex:
intestinal and lumbar injury in children. Journal of Trauma, 30, 11331140.
Orsay, E., Dunne, M. and Turnbull, T. (1990) Prospective study of the effect of
safety belts in motor accident crashes. Annals of Emergency Medicine, 19,
258261.
Osberg, J. and Di Scala, C. (1992) Morbidity among pediatric motor vehicle
crash victims. American Journal of Public Health, 82, 422425.
Paix, B. R., Blumbergs, P. C., Kloeden, C. N. et al. (1991) Head injury in car
occupants. Report on a pilot study. Research Report 1/92, NHMRC Road
Accident Research Unit, University of Adelaide, Adelaide, South
Australia.
Parkinson, D., Stephenson, S. and Phillips, S. (1985) Head injury: a
prospective computerised study. Canadian Journal of Surgery, 28, 7983.
Pincemaille, Y., Troseille, X., Mack. P. et al. (1989) Some new data related to
human tolerance obtained from volunteer boxers, in Proceedings of the 33rd
Strapp Car Crash Conference, Society of Automotive Engineers, Warrendale,
PA, pp. 177190.
Pitts, L., Ojemann, R. and Quest, D. (1987) Neurotrauma care and the
neurosurgeon: a statement from the Joint Section on Trauma of the AANS
and the CNS. Journal of Neurosurgery, 67, 783785.
Rimel, R., Giordani, B., Barth, J. et al. (1982) Moderate head injury: completing
the clinical spectrum of brain trauma. Neurosurgery, 11, 344351.
Ring, I., Berry, G. and Dan, N. (1986) Epidemiology and clinical outcomes of
neurotrauma in NSW. Australian and New Zealand Journal of Surgery, 56,
557566.
Roberts, A. (1969) Brain Damage in Boxers. Pitman Medical and Scientific
Publishing, London.
Roberts, I., Streat, S., Judson, J. et al. (1991) Critical injuries in paediatric
pedestrians. New Zealand Medical Journal, 104, 247248.
Romano, P. and McLoughlin, E. (1992) Unspecified injuries on death
certificates: a source of bias in injury research. American Journal of
Rompe, K., Schneider, A. and Wailrich, M. (1988) Advantages of anti-wheel
lock systems (ABS) for the average driver in difficult driving situations.
DOT HS 807 223. NHTSA Technical Report. National Highway Traffic Safety
Administration, Springfield, VA, pp. 442448.
Ross. H. (1982) Deterring the Drinking Driver, Health, Lexington, KY, p. 24.
Ross, S., OMalley, K., Stein, S. et al. (1992) Abbreviated injury scaling of head
injury as a prognostic tool for functional outcome. Injury Analysis and
Prevention, 24, 181185.
Rowbotham, G., Maciver, I., Dickson, J. et al. (1954) Analysis of 1400 cases of
acute injury to the head. British Medical Journal, i, 726730.
Russell, W. and Nathan, P. (1946) Traumatic amnesias. Brain, 49, 280300.
Rutherford, W., Greenfield, T., Hayes, H. et al. (1985) The medical effects of
seatbelt legislation in the United Kingdom. Department of Health and Security
Research Report Number 13, HMSO, London.
Sacks, J., Holmgreen, P. and Smith, S. (1991) Bicycle-associated head injuries
and deaths in the United States from 1984 through 1988. Journal of the
Salmi, L., Thomas, H., Fabry, J. et al. (1989) The effect of the 1979 French seatbelt legislation on the nature and severity of injuries to front seat occupants.
Accident Analysis and Prevention, 21, 589594.
Sandels, S. (1977) In Children, the Environment and Accidents, (ed. R. Jackson),
Pitman Medical, London pp. 2034.
Selecki, B., Ring, I., Simpson, D. et al. (1981) Injuries to the head, spine and
peripheral nerves. Report on a study, NSW Government Printer, Sydney,
NSW.
Selecki, B., Ring, I., Simpson, D. et al. (1982a) Trauma to the central and
peripheral nervous systems. Part I: An overview of mortality, morbidity
and costs; New South Wales, 1977. Australian and New Zealand Journal of
Surgery, 52, 93102.
Selecki, B., Ring, I., Simpson, D. et al. (1982b) Trauma to the central and
peripheral nervous systems. Part II: A statistical profile of surgical
treatment; New South Wales, 1977. Australian and New Zealand Journal of
Surgery, 52, 111116.
Shacksford, S., Hollingsworth-Fridlund, P., Cooper, G. et al. (1986) The effect
of regionalisation upon the quality of trauma care as assessed by concurrent
audit before and after institution of a trauma system: a preliminary report.
Journal of Trauma, 26, 812820.
Shankar, B., Ramzy, A. and Soderstrom, C. (1992) Helmet use, patterns of
injury, medical outcome and costs among motorcycle drivers in Maryland.
Accident Analysis and Prevention, 24, 385396.
23
Sharples, P., Storey, A., Aynsley-Green, A. et al. (1990) Causes of fatal

childhood accidents involving head injuries in northern region, 19791986.
British Medical Journal, 301, 11931197.
Silverberg, D., Meer, A. and Silvinger, E. (1992) Head injuries after serious
bicycle accident. European Journal of Epidemiology, 8, 826831.
Simpson, D. (ed.) (1986) Neurotrauma in Australia. Report on surveys in New
South Wales and South Australia. Australia and New Zealand Journal of
Surgery, 54, 527589.
Simpson, D., Antonio, J., North, J. et al. (1981) Fatal injuries of the head and
spine. Epidemiological studies in New South Wales and South Australia.
Medical Journal of Australia, 2, 660664.
Simpson, D., Blumbergs, P., McLean, A. et al. (1992) Head injuries in infants
and children: measures to reduce mortality and morbidity in road
accidents. World Journal of Surgery, 16, 403409.
Simpson, D. and McLean, A. (1995) Epidemiology, in Craniomaxillofacial
Trauma, (eds D. David and D. Simpson), Churchill Livingstone, Edinburgh,
pp. 8589.
Sosin, D., Sacks, J. and Smith, M. (1989) Head injury associated deaths in the
United States 1979 to 1986. Journal of the American Medical Association, 161,
22512255.
Sosin, D., Sacks, J. and Holmgreen, P. (1990) Head injury associated death
from motorcycle crashes: relationship to helmet use laws. Journal of the
Spence, L., Dykes, E., Bohn, D. et al. (1993) Fatal accidents in children: a plea
for prevention. Journal of Pediatric Surgery, 6, 214216.
States, J., Annecharico, R. and Good, R. (1990) A time comparison study of
New York State safety belt use law utilising hospital admission and police
accident report information. Accident Analysis and Prevention, 22, 509521.
Thomas, J. (1990) Road traffic accidents before and after seatbelt legislation
study in a general hospital. Journal of the Royal Society of Medicine, 83, 536.
Tiret, L., Hausherr, E., Thicoipe, M. et al. (1990) The epidemiology of head
injury in Aquitaine, (France) 1986: a community study based on hospital
admissions and deaths. International Journal of Epidemiology, 19, 133140.
Trinca, G., Johnston, B., Campbell, F. et al. (1988) Reducing traffic injury a global
challenge. Royal Australasian College of Surgeons, Melbourne, Victoria.
US Department of Transportation (1989) Development of Trauma Systems: A
State and Community Guide, US Department of Transportation, National
Highway Traffic Safety, Office of Environment and Emergency Medicine,
Washington, DC.
Van Beeck, E., Mackenbach, J., Loopman, C. et al. (1991) Determinants of
traffic accident mortality in the Netherlands: a geographical analysis.
International Journal of Epidemiology, 20, 698706.
Vazquez-Barquero, A., Vazquez-Barquero, J., Austin, O. et al. (1992) The
epidemiology of head injury in Cantabria. European Journal of Epidemiology,
8, 832837.
Walker, P. and Cass, D. (1987) Paediatric trauma: urban epidemiology and an
analysis of methods of assessing the severity of trauma in 598 injured
children. Australian and New Zealand Journal of Surgery, 57, 715722.
Waller, P., Stewart, J., Hansen, A. et al. (1986) The potentiating effects of
alcohol on driver injury. Journal of the American Medical Association, 256,
14611466.
Walton, J., Beeson, P. and Bodley-Scott, R. (eds) (1986) The Oxford Companion
to Medicine, Oxford University Press, Oxford, vol. 1, p. 351.
Wang, C., Schoenberg, B., Li, J. et al. (1986) Brain injury due to head trauma.
Epidemiology in urban areas of the Peoples Republic of China. Archives of
West, J., Williams, M. and Trunkey, D. (1988) Trauma systems current status:
future challenges. Journal of the American Medical Association, 259,
35973600.
Whitman, S., Coonley-Hoganson, R. and Desai, B. (1984) Comparative head
trauma experiences in two socioeconomically different Chicago-area
communities: a population study. American Journal of Epidemiology, 119,
570580.
WHO (1978) International Statistical Classification of Diseases and Related Health
Problems, ICD-9, 9th revision, World Health Organization, Geneva.
Problems, Clinical Modification, ICD-9-CM, 9th revision, Commission on
Professional and Hospital Activities, World Health Organization, Ann
Arbor, MI.
Problems, ICD-10, 10th revision, World Health Organization, Geneva.
Wong, T., Phoon, W. and Lee, J. (1989) Non-fatal injuries among motorcyclists
treated as in-patients in a general hospital. Annals of the Academy of Medicine,
Singapore, 18, 672674.
Woodward, A., Dorsch, M. and Simpson, D. (1984) Head injuries in country
and city, a study of hospital separations in South Australia. Medical Journal
of Australia, 141, 1317.
Wrightson, P. and Gronwall, D. (1981) Time off work and symptoms after
minor head injury. Injury, 12, 445454.
Zador, P. L. and Ciccone, M. A. (1991) Driver Fatalities in Frontal Impacts:
Comparisons Between Cars With Air Bags and Manual Belts, Insurance Institute
for Highway Safety, Arlington, VA, p. 12.
BIOMECHANICS OF CLOSED
HEAD INJURY
A. J. McLean and Robert W. G. Anderson
This chapter discusses ways in which the brain is

thought to be injured by a blunt impact to the head.
The impacting object is assumed to be unlikely to
penetrate the skull in the manner of a bullet, for
example.
The chapter is also concentrated on injuries to the
brain, rather than lacerations and abrasions to the
scalp or fractures of the skull. Obviously, if the skull is
fractured and displaced inwards, then the part of the
brain underlying the fracture will be injured. However, the brain can be very severely injured without
the skull being fractured by the impact to the head
(Gennarelli, 1980). Other intracranial injuries, such as
subdural hematomas, are referred to briefly in relation
to theories of mechanisms of primary injury to the
brain. Secondary complications of head injury also
affect the brain but they are not considered in this
chapter.
2.1
Impact to the head
2.1.1
IMPACT AND IMPULSE
Closed head injury is, in the great majority of cases, a

consequence of an impact to the head. However, there
are references in the literature to the production of
diffuse axonal injury in non-impact experiments in
which the head of an animal was accelerated in a
manner that minimized the direct contact effects of an
impact to the head (Gennarelli and Thibault, 1982;
Adams, Graham and Gennarelli, 1981). There are also
reports of brain injury resulting from acceleration of
the upper torso of an animal without any direct
impact to the head (Ommaya, Hirsch and Martinez,
1966). These reports are discussed later in this chapter.
For the present, the readers attention is drawn to the
distinction between an impact to the head and an
impulse transmitted to the head through the neck.
Both an impact and an impulse, as described above,
can accelerate a stationary head (or decelerate a
moving one) but an impact will also produce contact

effects on the head, such as skull deformation or
fracture, with an associated risk of injury to the brain.
However, in practice it appears that injury to the
human brain is almost always the result of an impact
to the head, or to a protective helmet, rather than an
impulse transmitted through the neck (Tarriere, 1981;
McLean, 1995).
An impact to a given location on the head can be
characterized by the impact velocity and the physical
properties of the struck or striking object.
2.1.2
IMPACT VELOCITY
Some forensic pathology research literature implies

that the type of brain injury differs according to
whether the head is stationary and is struck by a
moving object, or is moving and strikes a stationary
object (Yanagida, Fujiwara and Mizoi, 1989). This
distinction can be of considerable legal significance in
cases of assault in which the victim sustains a head
injury which could have been caused either by a blow
to the head or from striking the head in the resulting
fall. However, as Holbourn (1943) observed, the
moving head typically strikes an object that is considerably more massive than the head, whereas the
stationary head is more often hit by objects which are
of similar mass to the head or even lighter, such as a
club.
In physical terms the difference between the head
moving or being stationary on impact is solely in the
frame of reference. Most readers will have experienced
the paradoxical sensation of not knowing which train
is moving when the train alongside theirs in the
station starts to move. There is no physical difference
between the forces involved in a stationary head being
hit or a moving head striking a fixed object, given that
other factors such as the velocity of the impact and the
characteristics of the object contacted by the head are
the same. Throughout this chapter the terms struck
26
BIOMECHANICS OF CLOSED HEAD INJURY
and striking will therefore be used more or less

interchangeably.
In general, the head impact velocity will be greater
in, say, high-speed crashes on the road than in crashes
at low speed. However, that is not necessarily so. The
type of crash is a significant factor, with high speed
rollovers sometimes being relatively non-injurious
compared to collisions with another vehicle or a fixed
object at a much lower speed. Even in two apparently
similar crashes it is not at all uncommon for a person
in one crash to receive a severe impact to the head
when a person in the other crash may not be hit on the
head at all.
2.1.3 PHYSICAL PROPERTIES OF THE STRUCK OR

STRIKING OBJECT
(a)
Shape
As noted above, we have assumed that the impacting

object does not penetrate the skull in the manner of a
bullet, for example. The most important characteristics
of the struck or striking object are therefore its stiffness
and surface area, given that its shape is consistent with
it imparting a blunt impact to the head.
(a)
Local deformation of the skull at the point of impact

can be expected to result in direct contact injury to the
underlying brain tissue. This almost inevitably occurs
if the impact produces a displaced fracture of the
skull, but high-speed cine radiography has shown that
the skull can also be indented sufficiently in the first
few milliseconds of the impact to compress the
underlying brain and then return to its original shape
without residual evidence of such deformation in the
bone (Gurdjian, 1972; Shatsky et al., 1974).
Shatsky et al. (1974), in their studies using anesthetized monkeys, showed that in occipital impacts the
skull was not deformed and no underlying brain
lesions were observed. Impacts in the temporoparietal
region did show evidence of transient skull deformation and accompanying brain lesions. For a given
impact, the risk of underlying skull fracture will also
vary with the location of the impact on the head.
Nahum et al. (1968) estimated that for a contact area of
approximately 1 square inch (6.5 cm2) the force
required to produce a clinically significant skull
fracture in the frontal area of the cadaver skull was
twice that required in the temporoparietal area.
(b)
(b)
Stiffness
The term stiffness, as used here in the engineering

sense, is sometimes confused with hardness. The
property of stiffness is well illustrated by the compression of a spring. The less the spring deflects under a
given load the stiffer it is said to be. By comparison, a
thin sheet of glass is very hard on the surface but it
will bend, or deflect, easily when loaded. It has a low
level of stiffness, which, of course, decreases abruptly
to zero when the glass breaks.
A concrete floor is extremely stiff; almost infinitely
so in relation to the human head. A sheet metal panel
of a car, however, may be deformed several centimeters when struck by the head of a pedestrian or an
occupant of the car. Such differences in stiffness of the
object struck by the head have been shown to be
associated with differences in the type of the resulting
intracranial injury, as discussed below (Gennarelli,
1984; Willinger, Kopp and Cesari, 1991).
2.1.4
LOCATION OF THE IMPACT ON THE HEAD
The location of the impact on the head can be related

to brain injury in a number of ways: such as by local
deformation of the skull and, more importantly, by
determining the relative levels of linear and angular
acceleration of the head.
Injury adjacent to the location of the impact
Injury remote from the location of the impact
The resulting injury can also be, and very often is,
remote from the location of the impact. This is so for
injuries to both the skull and the brain. A blunt impact
to the calvarium may result in remote linear fractures
in the base of the skull. This is thought to be a
consequence of the skull cap being strong enough to
withstand the force of the impact, which is therefore
transmitted to the much thinner bone found in parts of
the base of the skull. However, once a remote linear
fracture has been initiated in such a region, it can
propagate almost instantly back into the calvarium
which is still loaded by the force of the impact (Melvin
and Evans, 1971).
The term contrecoup has long been used to
characterize an injury to the brain which is on the far
side of the head to the impact. It has been postulated
that contrecoup injury to the brain is a consequence
of rapid and localized pressure changes near the
surface of the brain tissue due to cavitation effects
arising from the brain moving relative to the cranial
cavity in response to the impact (Courville, 1942).
However, Nusholtz et al. (1984) reported that, in
occipital impacts to the head of the Rhesus monkey,
contrecoup negative pressures greater than one
atmosphere did not appear to be associated with
injury to the brain.
In the case of an occipital impact it is possible that
relative motion between the brain and the often
RESPONSE OF THE HEAD TO IMPACT
irregular bony anatomy of the anterior fossa in the

human may play a role in the causation of contrecoup
injury to the brain (Shatsky et al., 1974).
(c)
Head impact location and brain injury severity
Finally, the location of the impact to the head can

determine the nature, pattern and severity of injury
throughout the brain. More than 200 years ago
Percivall Pott, Surgeon of the City of London,
remarked upon an apparent relationship between
impact location on the head and the severity of the
resulting brain injury:
I will not assert it to be a general fact, but as far
as my own experience and observation go, I
think that I have seen more patients get well,
whose injuries have been in or under the frontal
bone, than any other bones of the cranium. If this
should be found to be generally true, may not
the reason be worth inquiring into?
27
not, the skull may be crushed to a greater or lesser

degree and the injury to the head, and the brain, will
be directly related to the location and extent of the
skull deformation. An example of such an impact
would be a masonry block falling on the head of a
person lying on a concrete floor.
As noted above, most cases of closed head injury
result from a moving head coming into contact with
a stationary object or with an object moving at a
different velocity. Injuries to the brain are generally
thought to result from the acceleration of the brain in
response to the impact to the head, with the exception of direct contact injuries resulting from skull
fracture or motion of the brain relative to the cranial
cavity. The term acceleration is used here in the
absolute sense to refer both to cases in which the
stationary head is accelerated by an impacting object
and to cases in which the moving head is decelerated
when it hits a stationary object, or one moving at a
lower velocity.
The Chirurgical Works of Percivall Pott, FRS, 1779

Some 200 years later this matter has been inquired into
and the results are consistent with Percivall Potts
experience and observation. For example, the experimental work of Gennarelli, Thibault and Tomei (1987)
on subhuman primates demonstrated that a combination of linear and angular acceleration of the head in
the coronal plane (rotation of the head about a point in
the cervical spine, somewhat analogous to the motion
resulting from a lateral impact to the head) was more
injurious to the brain than similar acceleration in the
sagittal plane (as in a frontal impact). The results of
detailed investigation of a small number of cases of
fatal and severe head injury to car occupants were also
consistent with frontal impact to the head being less
injurious to the brain than lateral impact (Simpson et
al., 1991). However, animal experiments conducted in
Japan, discussed later in this chapter, indicate that the
relationship between impacts in the sagittal and
coronal planes and injury to the brain may be
considerably more complex than is commonly supposed (Ono et al., 1980; Kikuchi, Ono and Nakamura,
1982).
2.2
Response of the head to impact
Another factor which has a marked influence on the

severity of the resulting injury is whether or not the
head is free to move in response to the impact.
2.2.1
MOVEMENT OF THE HEAD
The mechanism of injury to the head depends on

whether or not the head is free to change its velocity
when struck (Denny-Brown and Russell, 1941). If it is
2.2.2
THE FORCE OF THE IMPACT
For a given head impact velocity, an impact with a

sheet metal panel of a car will result in a much smaller
impact force, and hence lower acceleration of the
head, than will an impact with a concrete floor. This is
because the moving head is brought to rest over a
greater distance in the former case, as indicated by the
considerable dent which is often seen in the panel
following a head impact. However, the lower acceleration means that the impact force, albeit at a lower
level, is acting on the head for a longer time.
As will be seen later, there is reason to believe that
the sensitivity of the brain to injury from an impact to
the head is time-dependent. A very high level of
acceleration of the head acting for a very short time
may be less injurious than a lower level of acceleration
acting over a relatively longer time period.
2.2.3 LINEAR AND ANGULAR ACCELERATION OF

THE HEAD
If the line of action (the vector) of the impact force

passes through the center of gravity of the head then
the head will be accelerated in a straight line. In other
words, it will be subjected to a linear acceleration. This
general statement ignores any restraining effect of the
neck, which is likely to be small in the time interval
during which injury to the brain is thought to occur.
However, if the force vector does not pass through the
center of gravity then the head will be subjected to
both linear and angular acceleration, with the latter
resulting in rotation about the center of gravity. In this
chapter the terms rotational and angular are used
28
interchangeably. In some research papers a distinction

has been made between these two terms by defining
the former to mean rotation about the center of gravity
of the head and the latter to refer to rotation of the
head about some other point, such as in the cervical
spine, which results in a combination of linear and
angular motion of the center of gravity of the head.
The basic relationship between the force of the
impact and the resulting angular acceleration () of
the head is similar to that for linear acceleration except
that the offset (x) of the force vector from the center of
gravity of the head is taken into account along with
the moment of inertia (I) of the head:
F x = I .
The moment of inertia of a solid sphere, a very crude
approximation to the human head, about an axis
through the center of the sphere is:
I = 2/5 m r 2.
The force changes during the impact and so the
magnitude of the angular acceleration will change
with time during the impact.
If the human head was spherical, with its center of
gravity at the center of the sphere, then the force
vector of an impact perpendicular to the surface
would always pass through the center of gravity.
However, the human head is far from spherical and
so, for an impact perpendicular to the surface of the
skull, the distance of the force vector from the center of
gravity varies with the location of the impact on the
head. In general, regardless of differences in head
shape, the vector of an impact on the side of the
head is likely to be offset more from the center of
gravity than that of an impact on the frontal bone or
the occiput.
However, the human head does come in a wide
range of shapes and it is conceivable that the range of
this variability may be sufficient to influence the
response of the head to impact and hence the nature
and/or severity of the resulting injury to the brain
(McLean et al., 1990). For example, the force vector of
a lateral impact to the side of the frontal bone of a long
narrow head is likely to have a greater offset from the
center of gravity of the head than is a similar impact to
a rounder head, both viewed in the horizontal plane.
In practice, most impacts to the head will pocket
into the scalp to some degree. This may result in the
force vector not being at 90 to the surface of the skull,
which, in turn, will affect the distance between the
force vector and the center of gravity of the head.
However, any variation arising in this way is unlikely
to affect the general relationship between impacts on
the side of the head and greater offsets of the force
vector from the center of gravity. This means that, for
a given impact severity, lateral impacts are likely to
result in a higher level of angular acceleration of the

head than are frontal or occipital impacts (see, for
example, Vilenius et al., 1994).
If the level of angular acceleration increases with a
change in the location of the impact on the head, this
increase will be accompanied by a decrease in the level
of linear acceleration and vice versa. While it is possible
to envisage an impact to the head, such as an occipital
impact or one to the frontal bone, producing only
linear acceleration, it is most unlikely that any realistic
impact to the head will produce only angular acceleration about the center of gravity.
2.2.4
STRESS AND STRAIN
Stress (tensile or compressive) is measured in terms of

force per unit area. Strain describes the response of
the material which is being stressed. It is measured in
terms of the proportional change in length in the
direction of the tensile or compressive stress: hence
reference to, for example, a 10% strain. A compressive
strain will, of course, indicate a reduction in length,
whereas a tensile strain will indicate that the stressed
material has been stretched.
Another type of stress, and associated strain, which
is thought to be particularly relevant to injury to the
brain results from the action of a shear force. The effects
of shear stress are illustrated by the action of a pair of
scissors, or shears, cutting a thin stack of sheets of
paper. Shear stress is also measured in terms of force
per unit area but the area is measured in the plane in
which the force is acting (at right angles to the face of
the sheets of paper in the example given here).
Similarly, shear strain is the proportional displacement,
expressed in terms of the original thickness, resulting
from the action of the shear force.
(a)
Strain rate
The rate of application of a force is reflected in the rate

of the resulting strain, expressed in terms of strain per
unit of time. The response to physical loading of some
biological materials is strain-rate-dependent (Viano
and Lau, 1988).
2.3
Methods of investigation
Three types of investigative method have been used in

the study of brain injury biomechanics: experimental,
mathematical and observational.
2.3.1
EXPERIMENTAL STUDIES
Much of what is known, or postulated, about brain

injury mechanisms in living humans has come from
experimental studies. Test subjects have included
TOWARD AN UNDERSTANDING OF BRAIN INJURY MECHANISMS
human cadavers, anesthetized animals and animal

cadavers. Physical models of the head have also been
used. Experiments conducted on living humans have
defined the response of the head to non-injurious
impact.
The human cadaver head has the advantage of valid
anatomical, but not physiological, representation of
the head of the living human (although attempts have
been made to simulate vascular lesions by pressurizing the vascular system prior to impacting the head).
The anesthetized animal is, of course, a living subject
but differs anatomically from the human. Highly
sophisticated experimental techniques have been
developed in the course of the evolution of animalbased head injury studies (see, for example, Nusholtz,
Kaiker and Lehman, 1986). Although the anatomical
differences are least in the non-human primate, the
smaller size of the monkey skull and brain introduce
problems of dimensional scaling in attempts to relate
the results to the living human.
Experiments using physical models of the brain, or
skull and brain, have included measurements of
strain, recorded by measuring the distortion of an
impregnated grid or by photoelastic means, in accelerated gel-filled containers (Thibault, Gennarelli and
Margulies, 1987; Holbourn, 1943).
2.3.2
MATHEMATICAL STUDIES
Mathematical models of human and animal heads are

now at the stage which, with the ready availability of
powerful computers, justifies their use in attempts to
predict the likely impact-induced motion of the brain
relative to the skull and strains within the brain tissue
(Zhou, Khalil and King, 1994). The development of a
realistic mathematical model depends, inter alia, on
knowledge of the physical response of brain tissue to
impact loading, in addition to the response of the skull
and membranes (Melvin, Lighthall and Ueno, 1993).
Experimental data are available for use in the
validation of mathematical models of the animal
skull/brain system. Validation of mathematical models of the human head depends on the availability of
adequately detailed and accurate estimates of the
forces involved in impacts to the head of the living
human in events such as road crashes and the
resulting pathology, particularly the neuropathology.
2.3.3
OBSERVATIONAL STUDIES
The investigation of cases in which living humans

have sustained a closed head injury has the desirable
attribute that the phenomenon being investigated is
exactly that which is of interest. However, it is difficult
to obtain adequately detailed information on the
characteristics of the injury to the brain, and the
29
severity of the impact to the head can only be

estimated (Ryan et al., 1989; Gibson et al., 1985).
(a)
Neuropathology
In fatal cases the neuropathologist can provide information on injury to the brain at the microscopic level.
Even so, there are limitations imposed by the fact that
some brain lesions are not at present readily detectable
unless the fatally injured individual survives for some
hours after the incident which produced the injury. In
surviving cases, magnetic resonance imaging (MRI)
and computed tomography (CT) can be used to
identify and locate larger hemorrhagic lesions in the
brain.
(b)
Characteristics of the impact to the head
The location of the impact on the head can be

determined from the location of abrasions and contusions, depressed skull fractures and subgaleal hematomas at autopsy or in operative cases. In non-fatal cases
it can be difficult to determine the location of an
impact above the hairline.
Determining the object or objects struck by the head
usually depends on examining the setting in which
the injury occurred, such as the vehicle involved and
the crash site in the case of a road accident. The
stiffness of the struck object may be able to be deduced
from a simple description of the event, such as the
head striking a concrete floor as the result of a fall. In
a road crash the head impact is most likely to have
been with some part of the vehicle. Knowledge of the
stiffness of the struck part of the vehicle can be used to
estimate the force of the head impact, assuming that a
reasonably accurate estimate can be made of the
velocity with which the head struck the object. If a
record has been made of any residual deformation of
the struck object, an instrumented headform can be
used to measure the force required to reproduce the
dent observed in the actual impact.
2.4 Toward an understanding of brain

injury mechanisms
The study of the biomechanics of head injury has been
concentrated on the relationship between the forces
applied to the head and the resulting injury to the
brain. Some of the earlier studies used the presence or
absence of skull fracture as the outcome variable,
assuming that it would be positively related to the
severity of brain injury (see the following section on
the impact tolerance of the head). However, it was
soon recognized that the response of the whole head
to impact was likely to be the main determinant of the
nature and severity of injury to the brain.
30
2.4.1
SETTING THE SCENE
The work of Holbourn (1943), a research physicist in

the University Department of Surgery in Oxford, UK,
set the scene for what has been the most widely
accepted theory of the mechanism of injury to the
brain: that rotational motion of the head is the
predominant causal factor.
Reasoning that the brain was effectively incompressible, Holbourn hypothesized that linear acceleration of the head could not deform the brain and so was
unlikely to result in injury to the brain tissue. Angular
acceleration, however, could be expected to set up
shear strains in the brain, and the relative displacement within the brain that is implied by the creation of
such strains would be expected to be a cause of
injury.
The bowl of porridge analogy is sometimes used to
illustrate Holbourns hypothesis (Figure 2.1).
If the stationary bowl is suddenly moved sideways
(linear acceleration) there will be no appearance of
relative motion in the porridge (apart from spilling
over the side, which is not possible with a closed
vessel such as the cranial cavity). If, however, the bowl
is rotated rapidly (angular acceleration), that part of
the porridge adjacent to the bowl will tend to move
with the bowl and the porridge in the center will tend
to remain stationary. This can only happen if there is
relative motion (shear strains) within the porridge.
Holbourn tested his hypothesis using the physical
model referred to above: a gel-filled two dimensional
model of the human head. The strains produced in the
gel by acceleration of the model in that plane were
revealed by photoelastic techniques. As he predicted,
the model was relatively insensitive to linear acceleration but the pattern of strains produced by angular
acceleration could readily be demonstrated.
of the bowl produces

Figure 2.1 Angular acceleration ()
shear strains in the contents, as illustrated by the layers
sliding across each other.
The other observation which he made was that for

very short blows the duration of application of the
accelerating force was an important factor in the
production of shear strains in his model brain,
whereas for blows of long duration the levels of
shear strain were independent of the time for which
the force acts. The change from short to long
duration was estimated to occur somewhere in the
range 2200 ms (Holbourn, 1943). He also noted that
interposing a deformable object, such as a crash
helmet, between the blow and the head has the effect
of extending the duration of the impact and thereby
reducing the average level of the force transmitted to
the head (Cairns and Holbourn, 1943).
Some 20 years elapsed before Holbourns hypothesis that angular acceleration of the head was likely to
be much more injurious to the brain than linear
acceleration was followed up by other investigators.
One very practical reason for this was that it was a
relatively straightforward matter to measure linear
acceleration but techniques to measure the angular
acceleration of the head in animal experiments were
not available.
2.4.2
THE WAYNE STATE TOLERANCE CURVE
At the same time that Holbourn was conducting his

experiments on physical models in England, Gurdjian
and Webster (1943) commenced studies at Wayne State
University in Detroit on the effect of impacts administered in various ways to the head of the dog. In 1955,
together with Lissner of the College of Engineering,
they reported their Observations on the mechanism of
brain concussion, contusion, and laceration (Gurdjian, Webster and Lissner, 1955). By applying air
pressure directly to the unopened dural sac for
various time periods they were able to show that the
severity of the concussive effect depended on both the
intensity of the pressure pulse and the duration of its
application. They concluded that Concussion occurs
as the result of brain stem injury either from increased
intracranial pressure at the time of impact, direct
injury by distortion, mass movement, shearing, or
destruction by a missile. Brief reference was made to
the possibility that rotation of the head might result in
the brain being injured by abutting against bony
projections within the skull.
Gurdjian and Lissner then conducted studies on a
physical model similar to Holbourns, but with the
inclusion of a simulation of the foramen magnum and
brain stem. These studies led them to conclude that
the mechanism of concussion is shear strain in the
brain stem caused by pressure gradients due to closed
system dynamics of impact (Gurdjian and Lissner,
1961).
Lissner, Lebow and Evans (1960) also investigated

the relation between linear acceleration and intracranial pressure changes resulting from impacts to the
frontal bone of the embalmed human cadaver head.
The report on this work became notable for what was
almost a passing reference to the acceleration required
to produce a linear fracture of the frontal bone. The
plot of these results (Figure 2.2), later supplemented
by other data (Gurdjian et al., 1961), formed the basis
for the development of the Head Injury Criterion,
which is used almost universally today as the measure
of the risk of head injury in automobile crash injury
testing. This work, and the Head Injury Criterion, are
discussed at greater length in section 2.5.1.
Further investigations at Wayne State University
through the 1960s included occipital impacts to the
freely moving head of the anesthetized stumptail
monkey. In reporting on those experiments Hodgson
et al. (1969) commented that their results supported
the theory of Gurdjian and Lissner (1961). In particular, Hodgson et al. concluded that Although the
motion of the head involved both angular and
translational acceleration, the preponderance of affected cells found in the brain stem and the almost
complete absence of chromatolysis in the cortex,
makes it appear likely that translational acceleration is
the most important mechanism (Hodgson et al.,
1969).
However, in other studies of head injury mechanisms conducted in the 1960s, Ommaya and Hirsch
showed that in studies of whiplash injury the provision of a cervical collar neck support for monkeys
subjected to whole body acceleration almost eliminated the cases of concussion that were observed
when the head was free to rotate (Ommaya, Hirsch
and Martinez, 1966). In 1970, on the basis of further
Figure 2.2 The Wayne State tolerance curve. Points below

the curve are unlikely to be associated with severe brain
injury. (Source: reproduced from Gurdjian, Roberts and
Thomas, 1966, with permission.)
31
analysis of the results of these studies, they concluded

that no convincing evidence has to this date been
presented which relates brain injury and concussion to
translational motion of the head for short duration
force inputs, whether through whiplash or direct
impact (Hirsch and Ommaya, 1970).
2.4.3 FURTHER DEVELOPMENT OF EXPERIMENTS
USING HUMAN SURROGATES
In the early 1970s, Ommaya initiated work with

Gennarelli and Thibault on a series of head impact
experiments using monkeys (This work was later
continued by Gennarelli and Thibault at the University of Pennsylvania, in collaboration with Adams,
and then Graham, from the Institute for Neurological
Sciences in Glasgow.) They subjected the head of the
animal to predominately linear or angular acceleration
in a defined plane while at the same time minimizing
the direct contact effects of the impact on the head.
This was done by encasing the monkeys head in a
rigid skull cap and filling the space remaining
between the head and the cap with dental cement. The
skull cap was attached via a mechanical linkage to a
piston which, when actuated, accelerated the head
(Gennarelli and Thibault, 1982).
The rationale for attempting to minimize contact
effects of an impact to the head is clear. However, as
Gennarelli (1980) has noted, in the human there
appears to be little relationship between the presence
or absence of skull fracture and the severity of injury
to the brain. This could be interpreted to mean that the
concentrated forces of a localized impact act on the
brain in much the same way as the more uniformly
distributed forces of an impulse applied to the head as
a whole, apart from brain lesions due to local
deformation of the skull at the point of impact.
The linkage attached to the rigid skull cap constrained the motion of the head to a single plane with
the head rotating about a point in the region of the
lower part of the cervical spine. The movement of the
head was limited to an arc of 60 before the motion
was abruptly stopped. As the level of acceleration was
lower than the level of deceleration it was claimed that
the injurious event was the deceleration, although
there does seem to be reason to be concerned about the
injury potential of the acceleration phase. Physical
models subjected to the sequential accelerationdeceleration pulses showed marked distortion (strain) in
the brain tissue during the acceleration phase (Thibault, Gennarelli and Margulies, 1987; Margulies,
Thibault and Gennarelli, 1990).
Gennarelli and Thibault (1982) did find that to be
able to continue to produce subdural hematomas
when they increased the duration of the deceleration
phase they also had to increase the level of the
32
deceleration itself. This contrasted with their finding

that axonal injury and concussion could be produced
at lower deceleration levels when the duration of the
deceleration phase was increased, a result which was
consistent with the acceleration/time relationship
shown in the Wayne State tolerance curve (Figure 2.2).
Their explanation for this difference was that the
bridging veins are sensitive to the rate at which the
acceleration is applied. However, there is now evidence that the bridging veins are not strain-ratesensitive (Lee and Haut, 1989).
Lee, Melvin and Ueno (1987), working with a twodimensional finite element model of the brain of the
Rhesus monkey, concluded that the subdural hematomas may actually have been produced during the
acceleration phase of the biphasic test device developed by Thibault and Gennarelli. This is because any
increase in the duration of the deceleration phase had
to be accompanied by a corresponding decrease in the
duration of the acceleration phase, and hence an
increase in the level of the acceleration was necessary
to maintain a given level of deceleration.
2.4.4
BRAIN INJURY WITHOUT HEAD IMPACT?
In many of the publications on this very extensive

series of experiments at the University of Pennsylvania, reference has been made to the non-impact
nature of the acceleration of the head of the animal
(see, for example, Gennarelli and Thibault, 1982). The
intent may have been to draw attention to the fact that
contact phenomena, such as deformation of the skull,
were minimized by distributing the accelerative load
over a wide area of the head. A motorcyclists crash
helmet performs a similar function, while also absorbing some of the energy of the impact.
However, the term non-impact can, of course, be
interpreted to mean that the head was not subjected to
an impact. Such an interpretation can be a matter of
considerable forensic significance. For example, in
cases of alleged child abuse it is not uncommon for the
defense to allege that the infant was shaken vigorously rather than the head being struck by, or against,
some object. The probable validity of such an assertion
was investigated by Duhaime et al. (1987) who
concluded that vigorous shaking of the torso of an
infant was most unlikely to result in injury to the
childs brain in the absence of an impact to the head.
They estimated that the head acceleration level produced by such shaking was probably about one-50th
of the level resulting from an impact. This finding is
consistent with the results from our investigation of
brain injury in road crashes, although the force which
can be imparted to the torso, and hence to the head, by
an adult shaking an infant is likely to be considerably
less than can occur in a road crash.
Meaney, Thibault and Gennarelli (1994), at the

University of Pennsylvania, used a mathematical
model of the human body to investigate the likelihood
of that occurring to a car occupant subjected to a
severe lateral impact. They concluded that the acceleration of the head is unlikely to reach a level which
would be injurious to the brain. This is consistent with
McLeans finding, referred to earlier, that there were
no cases of brain injury without head impact in a
series of more than 400 fatally injured road users
(McLean, 1995).
2.4.5 THE ROLE OF LINEAR AND ANGULAR
ACCELERATION
Ommaya and Gennarelli (1974) reported that linear

acceleration of the head of the squirrel monkey in the
sagittal plane was associated with focal lesions but
was unlikely to produce cerebral concussion. However, cerebral concussion was produced in each case
when the head of the monkey was subjected to
predominantly angular acceleration. Concussion was
assessed with reference to measures of the sensory
responses at the level of the cortical neuronal pool
(Gennarelli, Thibault and Ommaya, 1972).
Later work by Gennarelli et al. (1982) emphasized
the importance of the plane in which the angular
acceleration acts. They concluded that angular acceleration of the head causes axonal injury in the brain
proportional to the degree of coronal motion.
Ono et al. (1980), from the Japan Automobile
Research Institute and Jikei University Medical
School, conducted experiments on non-human primates to examine brain injury mechanisms. Their
observations were consistent with the conclusions of
Ommaya, Hirsch and Martinez (1966) and Gennarelli,
Thibault and Ommaya (1972) that an angular acceleration component must be present to induce brain
contusion in the sagittal plane. Ono et al. further
concluded that another important mechanism for the
occurrence of contusions is deformation of the skull as
governed by the contact area of the striker.
However, the results of these tests in Japan also
showed that the occurrence of concussion, as distinct
from brain contusion, in the monkeys could not be
correlated with angular acceleration but was highly
correlated with the linear acceleration of the head.
Their definition of the severity of concussion was
based on observation of the duration of apnea, loss of
the corneal reflex and blood pressure disturbances.
The acceleration/time curve marking the threshold for
skull fracture was found to lie above the corresponding tolerance curve for concussion (Figure 2.3).
These results were scaled from the monkeys head to
that of the human, using the technique of dimensional
analysis. The scaled results were validated by compar-
Figure 2.3 JARI human head tolerance curve (JHTC).

(Source: reproduced from Ono et al., 1980, with permission.)
ing them with the acceleration/time curve for the

production of fracture in the human cadaver skull.
In a subsequent series of experiments in which the
animals head was subjected to lateral impact, Kikuchi,
Ono and Nakamura (1982) found that the acceleration/
time tolerance curve for concussion was higher than
the corresponding curve established for impacts that
accelerated the head in the sagittal plane. This finding
ran counter to the conclusion drawn by Gennarelli et al.
(1982) that the tolerance of the brain to acceleration, in
terms of duration of coma, was substantially less in the
coronal than in the sagittal plane.
Kikuchi, Ono and Nakamura acknowledged that
their results differed from the concept that lateral
impact to the head was more injurious than frontal or
occipital impact. They deduced that the difference
arose from the fact that the relative threshold levels for
skull fracture, brain concussion and various pathological brain injuries differ for impacts in the sagittal
and coronal planes.
33
Further investigations at the University of Pennsylvania by Thibault et al. led to the suggestion that a
single parameter, such as the value of the peak angular
acceleration, may not be an adequate predictor of the
deformation of the brain tissue and hence of the
severity of intracerebral injury (Thibault, Gennarelli
and Margulies, 1987). They proposed that the change
in angular velocity and possibly the total displacement may also be important parameters. This led on to
the development of a hypothesis by Margulies and
Thibault (1992) that the level of strain in the brain
tissue due to an impact to the head might be a function
of the peak change in rotational velocity, the peak
rotational acceleration and the mass of the head.
2.4.6
RELATIVE MOTION CONCEPT OF BRAIN INJURY
As mentioned previously, Holbourn (1943) argued

that rotational motion of the head was a significant
causal factor in the production of injury to the brain.
Pudenz and Sheldon (1946) were able to demonstrate
relative motion between the brain and the skull of the
monkey by means of a Lucite calvarium. They found
that the extent of such relative motion was influenced
by the direction of the impact to the head, being
greater in the sagittal than in the coronal plane.
The emphasis placed by many research workers on
rotational motion as a critical factor in the production of
injury to the brain is not universally accepted. For
example, Willinger et al. (1994) claim that knowledge of
the angular and linear components of the response of
the head to an impact is not a sufficient basis for accurate prediction of the mechanisms of any intracerebral
lesions. Applying the technique known as modal
analysis to the human head, both in vivo (Figure 2.4)
Figure 2.4 The signals measured by a load sensing hammer and an accelerometer (held against the head by the left forefinger)
are processed by modal analysis to determine the effective masses comprising the head (e.g. skull and brain) and their relative
motion for a given rate of change of acceleration of the head (which is related to the stiffness of the object struck).
34
and in vitro, Willinger et al. (1992) have noted that measurement of the acceleration response to impact of the
head of the living human indicates that in an impact
with a hard object, such as a concrete floor, the brain
appears to move relative to the skull, whereas in an
impact with a relatively soft object, such as a sheet
metal panel of a car, the brain appears to move with the
skull.
This suggests that an impact with a hard object is
likely to be accompanied by peripheral injuries, such
as ruptured bridging veins, due to relative movement
between the brain and the skull. By comparison, when
the head hits a softer object the brain tends to move
with the skull and so it is subjected to similar forces,
and the resulting accelerations, to those acting on the
skull. An impact with a soft object is therefore likely
to be characterized by damage to the tissue deep
within the brain, such as axonal injury. The results
from modal analysis of the human head suggest that
these differences in the response of the brain to impact
are independent of the severity of the impact.
Willinger et al. (1992) demonstrated that this
hypothesis was consistent with the findings from the
detailed investigations conducted by the NHMRC
Road Accident Research Unit in Adelaide of cases of
impact to the head of the living human in road crashes
and falls.
Willinger later showed that the extensive series of
experiments on non-human primates conducted by
Gennarelli and Thibault to explore the sensitivity of
the brain to linear and angular acceleration also
yielded results that were consistent with his hypothesis (Willinger et al., 1994). He has therefore suggested
that the postulated roles of these two types of
acceleration in the causation of brain injury might
simply be a correlate of underlying differences in the
characteristics of the rate of change of the acceleration
of the head. Meaney, in commenting on Willingers
hypothesis, has noted that other factors, such as the
apparent dependence of the brain on the direction of
the applied force (or, resulting acceleration) may be
equally important in predicting the occurrence of
injury to the brain (Meaney, 1994).
Viano (1987) had earlier argued that acceleration,
per se, is not the cause of injury (to the brain). Rather
rapid motion of the skull causes displacement of the
hard bony structures of the head against the soft
tissues of the brain, which lag in their motion due to
inertia and loose coupling to the skull. This argument
differs from that of Willinger et al. (1992), who claim
that there is evidence from modal analysis that in
some impact situations the brain remains closely
coupled to the skull. At Vianos suggestion, a method
was developed to facilitate the comparison of the bony
anatomy of the cranial cavity with the presence or
absence of lesions on or near the surface of the brain
(McLean et al., 1990). However, apart from lesions

adjacent to skull fractures, there was no clear evidence
of a relationship between the shape of the cranial
cavity and lesions on the surface of the brain
(unpublished).
At the time of writing, the NHMRC Road Accident
Research Unit at the University of Adelaide is collaborating with the Bioengineering Center at Wayne State
University in a comparison of the pattern of brain
lesions observed in fatal road crashes with the strains
predicted in the brain using a three-dimensional finite
element model of the skull and brain subjected to
similar estimated impact conditions. If, in due course,
such a mathematical model can be validated it will be
possible to resolve many of the conflicting theories of
the mechanisms of impact injury to the brain.
2.5
Tolerance of the head to impact
Whatever the actual mechanism or mechanisms of

injury to the brain may be in cases of blunt impact to
the head, there is a need for some quantitative
measure relating characteristics of the impact to the
risk of head injury. The designer of a crash helmet, or
of those parts of a passenger car that are likely to be
struck by the head of an occupant, needs to know
what head acceleration levels are likely to result in
severe or fatal head injuries. Without such a measure,
or criterion, the development of devices aimed at
minimizing the severity of the head injury resulting
from a given impact can be based on little more than
an assumption that some of the energy of the impact
should be absorbed before it reaches the head.
The acceleration of the head has been, and continues
to be, used as a measure of the tolerance of the head to
impact. However, as noted above, the duration of the
impact is also related to the risk of a severe head
injury. Several measures have been developed in an
attempt to quantify the tolerance of the head to impact
in terms of the magnitude of both the resulting
acceleration of the head and the duration of the
impact. Of these, the Head Injury Criterion, commonly referred to by the acronym HIC, is by far the
most widely used. For the present purpose the
derivation of HIC is outlined as a basis for consideration of some of the criticisms which have been levied
against it. The discussion concludes with a review of
the reasons why HIC continues to be used, almost
universally, as the measure of the tolerance of the
brain to blunt impact to the head.
2.5.1
HIC: THE HEAD INJURY CRITERION
The first attempt to measure the tolerance of the head

to blunt impact was carried out by researchers at
Wayne State University in Detroit, notably Gurdjian
TOLERANCE OF THE HEAD TO IMPACT
and Lissner, from the 1940s through to the 1960s (see,

for example, Gurdjian and Lissner, 1944). As noted
earlier in this chapter, they subjected cadaver heads to
a blow to the forehead and related the linear acceleration of the head to whether or not the impact
produced fractures in the frontal bone. Eight skulls
were hit and the results of six of the eight cases were
plotted on a graph having the linear (straight line)
acceleration of the head on the vertical axis and time
(measured in milliseconds) on the horizontal axis
(McElhaney, Roberts and Hilyard, 1976). These points
mostly lay on that part of the curve (shown in Figure
2.2) which lies between about 1 ms and about 7 ms
after initial contact. Additional data points from other
experimental head impact studies on animals in which
the duration of the impact was longer were added
later, together with the results of cases in which
human volunteers were subjected to non-injurious
relatively low-level accelerations acting for a comparatively long time. The slope of the extended curve
approached the horizontal asymptotically after about
10 ms (Figure 2.2).
The curve defined by the data points from the
original cadaver studies, supplemented by the additional data, became known as the Wayne State
tolerance curve (Figure 2.2). It was thought to provide
an indication of the tolerance of the brain to impact, in
terms of the time history of the acceleration imparted
to the head. This was a considerable extrapolation
from the original tests, in which the outcome measure
had been simply the presence or absence of skull
fracture. The validity of the Wayne State tolerance
curve (WSTC) depended primarily on the assumption
that, if the skull of a living human was fractured, then
that injury would probably be accompanied by
concussion.
(a)
The Gadd Severity Index
In 1966, at the Stapp Car Crash Conference, Gadd of

General Motors proposed a head injury severity
index based on the Wayne State tolerance curve
(Gadd, 1966). Gadd reasoned that some measure of
the area under the acceleration/time curve for a
given impact could form the basis for such an index.
However it was apparent that a low level of acceleration lasting for a long time was not injurious whereas
a higher level of acceleration acting for a shorter time
was much more likely to be so, even though the area
under the acceleration/time curve could be the
same.
Gadd therefore decided to weight the area measure
in favor of the acceleration component. He did this by
raising the acceleration value to the power of 2.5. He
chose this number, 2.5, because it happened to be the
absolute slope of the Wayne State curve when plotted
35
on logarithmic axes. The mathematical expression for

the Gadd Severity Index (SI) is:
SI =
a
2.5
dt
where a is the effective acceleration (thought to have

been the average linear acceleration) of the head
measured in terms of g, the acceleration of gravity, and
t is the time in milliseconds from the start of the
impact.
The Gadd Severity Index or, as it was initially called,
the Severity Index, was thought by some still not to
deal adequately with long-duration, low-acceleration
impacts. In 1971, Versace, of the Ford Motor Company,
proposed a modification of the Gadd Severity Index,
which became known as the Head Injury Criterion
(HIC). The change was proposed to focus the severity
index on that part of the impact that was likely to be
relevant to the risk of injury to the brain. This was
done by averaging the integration of the resultant
acceleration/time curve over whatever time interval
yielded the maximum value of HIC. Because this
varies from one impact to another, the expression for
Versaces modified index simply refers to times t1 and
t2. The expression for HIC is:
HIC = (t2 t1 )
t2
a/(t2 t1 ) dt 2.5
t1
where an algorithm selects t1 and t2 to yield the

maximum value.
Since then, the desirability of restricting the time
interval (t2 t1) to as low as 15 ms has been noted to
avoid the possibility of obtaining high HIC values
from long-duration, low-acceleration cases (see, for
example, Prasad and Mertz, 1985).
After the analysis of impact accelerations experienced by American football players, human volunteer
impacts with air-bags and impact tests with windscreens, Hodgson and Thomas (1972) hypothesized
that a linear acceleration/time concussion tolerance
curve may not exist and that only impacts of very
short duration (e.g. with hard surfaces) may be
important. They suggested that if the impact does not
contain a critical HIC interval of less than 15 ms, the
impact should be considered safe. As noted above,
there is observational evidence that, in fact, head
injury without head contact is so rare that it is never
seen in the clinical setting (Tarriere, 1981; McLean,
1995).
HIC has been shown to relate well to the probability
that an impact will fracture the skull of a cadaver
(Hertz, 1994), which is perhaps not surprising given
the derivation of the original points on the Wayne
State curve. However, the Head Injury Criterion bears,
at best, a crude relationship to those factors now
thought to be important in brain injury causation.
36
(b)
The JARI human head tolerance curve
Of the various other tolerance criteria which have

been proposed, the JARI human head tolerance curve
(Ono et al., 1980; Kikuchi, Ono and Nakamura, 1982)
is closest in general concept to the Wayne State curve.
The JARI tolerance curve is more soundly based than
the Wayne State tolerance curve but is nevertheless
almost identical to it. There are other head injury
criteria that have been proposed but, despite the
acknowledged inadequacies of HIC, it continues to
be by far the most widely used measure of the risk of
injury to the brain from a blunt impact to the head.
This is largely because it is specified in vehicle safety
legislation in the United States and also because there
is not yet any demonstrably superior criterion in
terms of relevance to the severity of head injury to
humans.
2.6 The state of the art of head injury

biomechanics
In conclusion, the following comment made by Goldsmith in 1981 is still a reasonable assessment of the
present situation: The state of knowledge concerning
trauma of the human head is so scant that the
community cannot agree on new and improved
criteria even though it is generally admitted that
present designations are not satisfactory.
Nevertheless, current work on mathematical models, when combined with the results of detailed
investigation of the response of the living human
brain to impact to the head, shows promise of
contributing substantially to our understanding of
brain injury mechanisms and the tolerance of the head
to impact.
2.7
References
Adams, J. H., Graham, D. I. and Gennarelli, T. A. (1981) Acceleration induced

head injuries in the monkey. II. Neuropathology, Acta Neuropathologica
(Berlin), S7, 2628.
Cairns, H. and Holbourn, H. (1943) Head injuries in motor-cyclists: with
special reference to crash helmets. British Medical Journal, 15 May,
591598.
Courville, C. B. (1942) Coup-contrecoup mechanism of cranio-cerebral
injuries. Archives of Surgery, 45(1), 1943.
Denny-Brown, D. and Russell, W. R. (1941) Experimental cerebral concussion.
Brain, 64, 93164.
Duhaime, A. C., Gennarelli, T. A., Thibault, L. E. et al. (1987 The shaken baby
syndrome. A clinical, pathological, and biomechanical study. Journal of
Neurosurgery, 66(3), 409415.
Gadd, C. M. (1966) Use of a weighted impulse criterion for estimating injury
hazard, in Proceedings of the 10th Stapp Car Crash Conference, Society of
Automotive Engineers, New York, pp. 164174.
Gennarelli, T. A. (1980) Analysis of head injury severity by AIS80, in
Proceedings of the 24th Annual Conference, American Association for
Automotive Medicine, Morton Grove, IL, pp. 147155.
Gennarelli, T. A. (1984) Clinical and experimental head injury, in The
Biomechanics of Impact Trauma, (eds B. Aldman, A. Champon and G. Lanzra),
International Centre of Transportation Studies, pp. 103115.
Gennarelli, T. A. and Thibault, L. E. (1982) Biomechanics of acute subdural
hematoma. Journal of Trauma, 22(8), 680686.
Gennarelli, T. A., Thibault, L. E. and Ommaya, A. K. (1972) Pathophysiologic
responses to rotational and translational accelerations of the head, in

Proceedings of the 16th Stapp Car Crash Conference, Society of Automotive
Engineers, New York, pp. 296308.
Gennarelli, T. A., Thibault, L. E. and Tomei, G. (1987) Directional dependence
of axonal brain injury due to centroidal and non-centroidal acceleration, in
Proceedings of the 31st Stapp Car Crash Conference, Society of Automotive
Engineers, Warrendale, PA, pp. 4953.
Gennarelli, T. A., Thibault, L. E., Adams, J. H. et al. (1982) Diffuse axonal injury
and traumatic coma in the primate. Annals of Neurology, 12, 564574.
Gibson, T. J., McCaul, K. A., McLean, A. J. and Blumbergs, P. C. (1985)
Investigation of head injury mechanisms in motor vehicle accidents a
multidisciplinary approach. Society of Automotive Engineers Technical Paper
Series 850093, Society of Automotive Engineers, Warrendale, PA.
Goldsmith, W. (1981) Current controversies in the stipulation of head injury
criteria letter to the editor. Journal of Biomechanics, 14(12), 883884.
Gurdjian, E. S. (1972) Recent advances in the study of the mechanism of
impact of the head a summary. Clinical Neurosurgery, 19, 142.
Gurdjian, E. S. and Lissner, H. R. (1944) Mechanism of head injury as studied
by the cathode ray oscilloscope, preliminary report. Journal of Neurosurgery,
1, 393399.
Gurdjian, E. S. and Lissner, H. R. (1961) Photoelastic confirmation of the
presence of shear strains at the craniospinal junction in closed head injury.
Journal of Neurosurgery, 18(1), 5860.
Gurdjian, E. S., Roberts, V. L. and Thomas, L. M. (1966) Tolerance curves of
acceleration and intracranial pressure and protective index in experimental
head injury. Journal of Trauma, 6(5), 600604.
Gurdjian, E. S. and Webster, J. E. (1943) Experimental head injury with special
reference to the mechanical factors in acute trauma. Surgery, Gynecology, and
Obstetrics, 76, 622634.
Gurdjian, E. S., Webster, J. E. and Lissner, H. R. (1955) Observations on the
mechanism of brain concussion, contusion and laceration, Surgery, Gynecology, and Obstetrics, 101, 680690.
Gurdjian, E. S., Lissner, H. R., Evans, F. G. et al. (1961) Intracranial pressure
and acceleration accompanying head impacts in human cadavers. Surgery,
Gynecology, and Obstetrics, 112, 185190.
Hertz, E. (1993) A note on the head injury criterion (HIC) as a predictor of the
risk of skull fracture, in 37th Annual Proceedings of the Association for the
Advancement of Automotive Medicine, Association for the Advancement of
Automotive Medicine, Des Plaines, IL, pp. 303312.
Hirsch, A. E. and Ommaya, A. K. (1970) Protection from brain injury: the
relative significance of translational and rotational motions of the head after
impact, in Proceedings of the 14th Stapp Car Crash Conference, Society of
Hodgson, V. R. and Thomas, L. M. (1972) Effect of long-duration impact on
the head, in Proceedings of the 16th Stapp Car Crash Conference, Society of
Hodgson, V. R., Thomas, L. M., Gurdjian, E. S. et al. (1969) Advances in
understanding of experimental concussion mechanisms, in Proceedings of
the 13th Stapp Car Crash Conference, Society of Automotive Engineers, New
York, pp. 1837.
Holbourn, A. H. S. (1943) Mechanics of head injuries. Lancet, ii, 438441.
Kikuchi, A., Ono, K. and Nakamura, N. (1982) Human head tolerance to
lateral impact deduced from experimental head injuries using primates,
Society of Automotive Engineers Technical Paper Series 826035, Society of
Automotive Engineers, Warrendale, PA.
Lee, M. C. and Haut, R. C. (1989) Insensitivity of tensile failure properties of
human bridging veins to strain rate: implications in biomechanics of
subdural haematoma. Journal of Biomechanics, 22, 537542.
Lee, M. C., Melvin, J. W. and Ueno, K. (1987) Finite element analysis of
traumatic subdural hematoma, in Proceedings of the 31st Stapp Car Crash
Conference, Society of Automotive Engineers, Warrendale, PA, pp. 6777.
Lissner, H. R., Lebow, M. and Evans, F. G. (1960) Experimental studies on the
relation between acceleration and intracranial pressure changes in man.
Surgery, Gynecology, and Obstetrics, 111, 329338.
McElhaney, J. H., Roberts, V. L. and Hilyard, F. (1976) Handbook of Human
Tolerance, Japan Automobile Research Institute, Tsukuba, Japan.
McLean, A. J. (1995) Brain injury without head impact? Journal of Neurotrauma,
12(4), 621625.
McLean, A. J., Blumbergs, P. C., Kloeden, C. N. et al. (1990) The relative motion
concept of brain injury, in Proceedings of the International Research Council on
the Biomechanics of Impacts, IRCOBI, Bron, France, pp. 181190.
Margulies, S. S. and Thibault, L. E. (1992) A proposed tolerance criterion for
diffuse axonal injury in man. Journal of Biomechanics, 25(8), 917923.
Margulies, S. S., Thibault, L. E. and Gennarelli, T. A. (1990) Physical model
simulations of brain injury in the primate. Journal of Biomechanics, 23(8),
823836.
Meaney, D. F. (1994) Discussion on Rotation-translation duality in head
trauma? Perceptive and objective evidence, in Proceedings of the International Research Council on the Biomechanics of Impacts, IRCOBI, Bron,
France, pp. 7780.
Meaney, D. F., Thibault, L. E. and Gennarelli, T. A. (1994) Rotational brain
injury tolerance criteria as a function of vehicle crash parameters, in
Proceedings of the International Research Council on the Biomechanics of Impacts,
IRCOBI, Bron, France, pp. 5162
REFERENCES
Melvin, J. W. and Evans, F. G. (1971) A strain energy approach to the
mechanics of skull fracture, in Proceedings of the 15th Stapp Car
Crash Conference, Society of Automotive Engineers, New York, pp.
666685.
Melvin, J. W., Lighthall, J. W. and Ueno, K. (1993) Brain injury biomechanics,
in Accidental Injury: Biomechanics and Prevention, (eds A. Nahum and J.
Melvin), Springer-Verlag, New York, pp. 268291.
Nahum, A. M., Gatts, J. D., Gadd, C. W. and Danforth, J. (1968) Impact
tolerance of the skull and face, in Proceedings of the 12th Stapp Car Crash
Conference, Society of Automotive Engineers, Warrendale, PA, pp.
302316.
Nusholtz, G. S., Kaiker, P. S. and Lehman, R. J. (1986) Critical limitations on
significant factors in head injury research, in Proceedings of the 30th Stapp Car
Crash Conference, Society of Automotive Engineers, Warrendale, PA, pp.
237256.
Nusholtz, G. S., Lux, P., Kaiker, P. and Janicki, M. A. (1984) Head impact
response skull deformation and angular accelerations, in Proceedings of the
28th Stapp Car Crash Conference, Society of Automotive Engineers, Warrendale, PA, pp. 4174.
Ommaya, A. K. and Gennarelli, T. A. (1974) Cerebral concussion and
traumatic unconsciousness. Brain, 97, 633654.
Ommaya, A. K., Hirsch, A. E. and Martinez, J. L. (1966) The role of whiplash
in cerebral concussion, in Proceedings of the 10th Stapp Car Crash Conference,
Society of Automotive Engineers, New York, pp. 314324.
Ono, K., Kikuchi, A., Nakamura, M. et al. (1980) Human head tolerance to
sagittal impact: reliable estimation deduced from experimental head injury
using subhuman primates and human cadaver skulls, in Proceedings of the
24th Stapp Car Crash Conference, Society of Automotive Engineers, Warrendale, PA, pp. 10160.
Prasad, P. and Mertz, H. J. (1985) The position of the United States delegation
to the ISO Working Group 6 on the use of HIC in the automotive
environment, Society of Automotive Engineers Technical Paper Series 851246,
Society of Automotive Engineers, Warrendale PA.
Pudenz, R. H. and Sheldon, C. H. (1946) The Lucite calvarium a method for
direct observation of the brain. II. Cranial trauma and brain movement.
Journal of Neurosurgery, 3, 487505.
Ryan, G. A., McLean, A. J., Vilenius, A. T. S. et al. (1989). Head impacts and
brain injury in fatally injured pedestrians, in Proceedings of the International
Research Council on the Biomechanics of Impacts, IRCOBI, Bron, France, pp.
2737.
Shatsky, S. A., Alter, W. A., Evans, D. E. et al. (1974) Traumatic dislocations of
the primate head and chest: correlation of biomechanical, radiological and
pathological data, in Proceedings of the 18th Stapp Car Crash Conference,
Society of Automotive Engineers, Warrendale PA, pp. 351381.
37
Simpson, D. A., Ryan, G. A., Paix, B. R. et al. (1991) Brain injuries in car
occupants: a correlation of impact data with neuropathological findings, in
Proceedings of the International Research Council on the Biomechanics of Impacts,
IRCOBI, Bron, France, pp. 89100.
Tarriere, C. (1981) Risk of head and neck injury if there is no direct head
impact, in Proceedings of Head and Neck Injury Criteria: A Consensus Workshop,
Session 1, National Highway Traffic Safety Administration, Washington,
DC, pp. 1315.
Thibault, L. E., Gennarelli, T. A. and Margulies, S. S. (1987) The temporal and
spatial deformation response of a brain model in inertial loading, in
Proceedings of the 31st Stapp Car Crash Conference, Society of Automotive
Engineers, Warrendale, PA, pp. 267272.
Versace, J. (1971) A review of the severity index, in Proceedings of the 15th Stapp
Car Crash Conference, Society of Automotive Engineers, New York, pp.
771796.
Viano, D. C. (1987) Biomechanics of head injury toward a theory linking
head dynamic motion, brain tissue deformation and neural trauma, Society
of Automotive Engineers Technical Paper Series 881708, Society of Automotive
Engineers, Warrendale, PA.
Viano, D. C. and Lau, I. V. (1988) A viscous tolerance criterion for soft tissue
injury assessment. Journal of Biomechanics, 21(5), 387399.
Vilenius, A. T., Ryan, G. A., Kloeden, C. et al. (1994) A method of estimating
linear and angular accelerations in head impacts to pedestrians. Accident
Analysis and Prevention, 26(5), 563570.
Willinger, R., Kopp, C. M. and Cesari, D. (1991) Brain tolerance in the
frequency field, in Proceedings of the 13th International Conference on
Experimental Safety Vehicles, National Highway Traffic Safety Administration, Washington, DC, pp. 940946.
Willinger, R., Ryan, G. A., McLean, A. J. and Kopp, C. M. (1992) Mechanisms
of brain injury derived from mathematical modelling and epidemiological
data, in Proceedings of the International Research Council on the Biomechanics of
Impacts, IRCOBI, Bron, France, pp. 179192.
Willinger, R., Taleb, L., Viguier, P. and Kopp, C. M. (1994) Rotationtranslation
duality in head trauma? Perceptive and objective evidence, in Proceedings of
the International Research Council on the Biomechanics of Impacts, IRCOBI,
Bron, France, pp. 6376.
Yanagida, Y., Fujiwara, S. and Mizoi, Y. (1989) Differences in the intracranial
pressure caused by a blow and/or a fall experimental study using
physical models of the head and neck. Forensic Science International, 41,
135145.
Zhou, C., Khalil, T. B. and King, A. I. (1994) A 3-D human finite element model
for impact injury analyses, in Proceedings of the 38th Stapp Car Crash
Conference, Society of Automotive Engineers, Warrendale, PA, pp.
137147.
PATHOLOGY
Peter C. Blumbergs
3.1
Introduction
Traumatic brain damage (non-missile injury) may be

divided into primary and secondary types of injury.
Mechanical forces acting at the moment of injury
damage the blood vessels, axons, nerve cells and glia
of the brain and initiate an evolving sequence of
secondary changes that result in complex cellular,
inflammatory, neurochemical and metabolic alterations. Numerous pathophysiological mechanisms
have been postulated to explain the progressive tissue
damage produced by secondary injuries and this has
led to various pharmacological therapies to try to limit
this type of damage.
Primary traumatic brain damage is the result of
mechanical forces producing tissue deformation at the
moment of injury. These deformations may directly
damage the blood vessels, axons, neurons and glia in
a focal, multifocal or diffuse pattern of involvement
and initiate dynamic and evolving processes which
differ for each component part (Table 3.1). There is
biomechanical evidence that the injury thresholds
(tolerance criteria) differ for the various structural
components (e.g. axons and blood vessels; Margulies
and Thibault,1992; Maxwell et al., 1988b, 1992b).
Table 3.1 Primary traumatic brain damage (neural

and/or vascular)
Diffuse
1. Diffuse axonal injury (DAI)
2. Diffuse vascular injury (DVI)
Focal
1. Vascular injury resulting in:
a) intracerebral hemorrhage
b) subdural hemorrhage
c) extradural hemorrhage (epidural hemorrhage)
2. Axonal injury
3. Contusion
4. Laceration
Secondary traumatic brain damage occurs as a

complication of the different types of primary brain
damage and includes ischemic and hypoxic damage,
cerebral swelling, the consequences of raised intracranial pressure, hydrocephalus and infection (Table
3.2).
Table 3.2 Secondary brain damage
Diffuse
1. Diffuse hypoxicischemic damage
2. Diffuse brain swelling
Focal
1. Focal hypoxicischemic injury
2. Focal brain swelling
The different types of secondary brain damage are

potentially reversible with adequate treatment and
their recognition has led to attempts to pharmacologically manipulate some of the putative factors
involved. In any given patient there may be a complex
and dynamic interplay of the different primary and
secondary types of brain damage to produce a unique
constellation of lesions both in anatomical site and
number. For example, the consequences of primary
vascular damage may be bleeding into the brain tissue
to produce an intracerebral hematoma or interference
in the perfusion of the brain tissue with resultant
ischemic damage (secondary brain damage), or a
combination of the two. Thus head injury is not a
single entity but consists of many different types of
lesion that may occur rarely in isolation or, more
commonly, in varied combinations.
The concept of focal and diffuse brain injuries
evolved as the result of routine neuroimaging of headinjured patients and is the diagnosis usually made in
the presence or absence of focal (mass) lesions on CT
head scans (Zimmerman and Bilaniuk, 1979; Gennarelli et al., 1982b).
40
PATHOLOGY
Gennarelli et al. (1982b) found significant focal

(mass) lesions (e.g. contusion, acute subdural hematoma) on CT examination in 56% of severely headinjured patients (Glasgow coma score < 8); the cases
(44%) without focal lesions were classified as having
suffered a diffuse brain injury. Marshall, Gautille and
Klauber (1991) reported a mortality rate of 23.9% in
those with diffuse injury, compared to 40.4% for
patients with focal injuries. Coma due to focal injuries
is usually related to secondary effects such as raised
intracranial pressure, and focal injuries such as contusions may be extensive without causing impaired
consciousness.
Pathological studies have shown that diffuse axonal
injury (DAI), hypoxicischemic injury and diffuse
brain swelling are the structural lesions found in
patients with diffuse brain injuries, and are usually
seen in patients who survive their head injury long
enough to be admitted to hospital. Diffuse vascular
injury (DVI) is seen in patients who die very soon after
head injury and consists of multiple small hemorrhages throughout the brain (Adams 1990, 1992).
3.2 Assessment of severity of brain

injury
The clinical assessment of the severity of head injury is
commonly based on the Glasgow Coma Scale (GCS;
Teasdale and Jennett, 1974), where an aggregate score
of 38 represents severe head injury, 912 moderate
head injury and 1315 mild head injury on examination six hours after injury. This is essentially a measure
of functional impairment of the neurological mechanisms subserving speech, motor function and eye
movement and gives no indication of the underlying
pathological or structural basis for the impairment.
Thus patients with the same GCS may have different
underlying pathological lesions producing the impairment in function, and delayed progression of these
processes may cause death, even in patients with a
GCS score above 9.
The pathological quantification of brain damage is
particularly difficult because of the large number of
possible lesions of variable severity.
3.2.1
Table 3.3

Diffuse injuries
Diffuse axonal injury (DAI)

Diffuse vascular injury (DVI)
Hypoxicischemic injury
Diffuse brain swelling
DAI and DVI are initiated by the mechanical forces

acting at the time of injury and are dynamic lesions
that evolve with time. Hypoxicischemic brain damage and diffuse brain swelling are secondary to focal
and/or diffuse vascular injury or hypoxemia, which
also evolve with time. Recently, diffuse head injury
has been subdivided into four subgroups on the basis
of CT findings. Patients with diffuse injury I are those
with no visible lesions on CT. Diffuse injury II is
present when the cisterns are not compressed, the
midline shift is less than 5 cm3 and/or there is no highor mixed-density lesion of more than 25 cm3. Diffuse
injury III is present when the cisterns are compressed
or absent and the midline shift is 05 mm with no
high- or mixed-density lesion of more than 25 cm3.
Patients with diffuse injury IV are those with a
midline shift of more than 5 mm and no high- or
mixed-density lesion of more than 25 cm3 (Marshall,
Gautille and Klauber, 1991). The mortality rate of
patients with diffuse injury I is 10% and greater than
50% for patients with diffuse injury IV (Marshall,
Gautille and Klauber, 1991). The pathological substrate of these different subgroups has not yet been
established but brain swelling and vascular injury
would appear to be especially important (Table 3.3).
QUANTITATION OF VASCULAR DAMAGE
The contusion index was derived as a system of

mapping contusions on standard diagrams and
assigning them numerical values according to surface
extent and depth (Adams et al., 1980b, 1985). The
hemorrhagic lesion score is a measure of the total
vascular damage in a traumatized brain obtained by
mapping macroscopic and microscopic evidence of
hemorrhage on diagrams of whole brain sections
divided into sectors (Ryan et al., 1994).
3.2.2
QUANTITATION OF AXONAL DAMAGE
Morphological quantitation of the severity of AI is

especially difficult because the distribution and
extent of axonal damage is not uniform or symmetrical (Strich, 1961; Adams et al., 1977) and requires
systematic microscopic study of the brain. DAI has
been divided into three grades of severity on the
basis of a combination of macroscopic and microscopic marker lesions (Adams et al., 1989; see section
3.4.3(c))
The axonal injury sector score (AISS) is obtained
by dividing standard brain sections into 116 sectors
and scoring the presence or absence of AI in each of
these sectors to provide a score ranging from 0116
(Blumbergs et al., 1995).
3.3
The mechanism of brain injury
Most head injuries are produced by contact and/or

acceleration/deceleration forces (Gennarelli and Thibault, 1985).
AXONAL INJURY
Contact injuries skull fractures, extradural hematomas, laceration/contusion-related subdural hematomas and contusions are produced by contact forces
that operate when the head strikes or is struck by
some object. Contact injuries can only be produced by
direct impacts. Coup contusions are produced by focal
vascular injury due to compressive forces operating
beneath an area of skull inbending, or tensile forces
generated by the negative pressure produced beneath
an area of skull inbending suddenly snapping back
into place.
Acceleration/deceleration injuries diffuse axonal
injury and ruptured bridging vein subdural hematomas result from violent head motion producing
compressive, tensile and shear strains in the brain
tissue irrespective of the mechanism by which the
head motion is produced.
Deep strains result in concussion syndromes and
diffuse axonal injury whereas surface strains result in
ruptured bridging vein subdural hematomas, contrecoup contusions and intermediate coup contusions
(Gennarelli and Thibault, 1985). There is experimental
evidence that rotational acceleration/deceleration in
the coronal plane is most productive of DAI (Gennarelli et al., 1982a).
Direct impacts to the head, e.g. in falls, motor
vehicle accidents and assaults where the head is free to
move, produce significant acceleration/deceleration
forces as well as the more obvious contact forces.
Indirect impacts can also produce acceleration/deceleration forces as in the violent flexionextension
movements of the head after a rear-end collision.
3.4
Axonal injury
Axonal damage interferes with axoplasmic transport

and the flow of electrical information along the nerve
fibers, resulting in failure of interneuronal communication. Axons show a limited repertoire of responses
to traumatic injury. Severe traumatic injury results in
primary axonal disruption or transection termed
primary axotomy or sets in train a series of illunderstood events culminating in secondary axonal
degeneration or secondary axotomy (Maxwell et al.,
1993, Povlishock, 1992).
3.4.1
41
This type of injury is seen in high-speed motor

vehicle accidents, where forces of large magnitude are
involved. Complete transection or avulsion of neural
tissue produces instantaneous complete axotomy of
all the nerve fibers as well as disruption of blood
vessels and glial cells.
3.4.2
SECONDARY AXOTOMY
The axonal injury (AI) in the brains of patients who

survive for some time from less severe mechanical
insults is more subtle and is characterized by an
admixture of normal and damaged axons in the same
microscopic area (Strich, 1961; Adams et al., 1982;
Figure 3.1). The axonal damage may or may not be
associated with blood vessel injury. Strich (1961)
interpreted the axonal abnormalities she found in her
fatal head injuries as similar to the retraction balls
described by Cajal on the proximal and distal ends of
experimentally severed axons and suggested that they
represented extrusions of axoplasm from torn nerve
fibers. Axonal retraction balls have also been called
axonal bulbs, axonal balloonings and axonal swellings. The experimental studies of Povlishock have
resulted in a modification of the concept that retraction balls represent extrusions of axoplasm from the
torn ends of mechanically ruptured axons. These
studies have shown that primary axonal damage
results in an evolving sequence of changes producing
axonal swellings some hours after initiation of the
injury (Povlishock, 1992). The separation of axons
some hours after injury has been termed secondary
axonal degeneration or secondary axotomy (Maxwell
et al., 1993).
The concept of a sequential evolution of axonal
damage leading to secondary axotomy is of great
clinical significance in that therapeutic intervention
may potentially minimize or prevent these changes
from occurring. Experimental studies have shown
that not all damaged axons proceed to anatomical
PRIMARY AXOTOMY
Primary axotomy is seen more commonly at the very

severe end of the spectrum of traumatic injury where
the victims do not survive to reach hospital and there
is gross physical disruption of neural tissue such as
pontomedullary disruption, transection of the corpus
callosum and dehiscence of the cerebral hemispheres,
cerebellum and brain stem.
Figure 3.1 Axonal swellings in the corpus callosum showing positive immunostaining with amyloid precursor protein (APP). 100 magnification.
42
PATHOLOGY
separation and that it is likely that some may

recover (Povlishock, 1992; Gennarelli et al., 1986). AI
may be focal, multifocal, diffuse or a combination of
these patterns of involvement.
3.4.3
DIFFUSE AXONAL INJURY (DAI)
Diffuse axonal injury (DAI) is defined as the presence

of diffuse damage to axons in the cerebral hemispheres,
corpus callosum, brain stem and cerebellum (Adams et
al., 1989). This axonal damage in short-term survivors
(days to weeks) depends on the demonstration of
axonal swellings (bulbs or retraction balls) in the white
matter of the brain (Figure 3.2). The distribution of
axonal damage is not uniform or symmetrical, with the
internal capsule, corpus callosum and superior cerebellar peduncles being most frequently involved (Ng,
Mahaliyana and Poon, 1994). There is also frequent
axonal damage in the subcortical parasagittal white
matter, fornices, dorsal peridentate white matter,
medial lemnisci, medial longitudinal fasciculi, central
tegmental tracts and corticospinal tracts (Strich, 1961;
Adams et al., 1977).
Figure 3.2 Axonal retraction ball. APP immunostaining,

400 magnification.
(a)
Microscopic diagnosis of DAI
It has been recognized since the early studies of Strich

(1961) that retraction balls take some time to develop
and the question of whether they can be recognized in
the first 12 hours post-trauma using traditional neurohistological techniques has been controversial.
Retraction balls are visible on routine hematoxylin and
eosin staining as eosinophilic round, oval or elongated
masses 540 m in size, but are better demonstrated
using silver impregnation techniques (Strich, 1961).
Adams (1992) has stated that with the standard silver
impregnation methods a survival time in excess of 18
hours is required to reliably demonstrate DAI.
However, Rand and Courville (1934), Peerless and
Rewcastle (1967) and Imajo and Roessman (1984)
identified axonal retraction balls 24 hours postmortem by light microscopy. Grady et al. (1993) found
that they could identify traumatically induced reactive

axonal change 6 hours postinjury using antibodies
against the low MW non-phosphorylated neurofilament subunit; by 12 hours, these changes had
progressed to complete axotomy. Using silver
impregnation techniques, the earliest axonal change
was only apparent at 12 hours. Recently, amyloid
precursor protein (APP) has been suggested as the
marker of choice for detecting AI (Sheriff, Bridges and
Sivaloganathan, 1994) (Figure 3.2). APP undergoes fast
axonal transport (Koo et al., 1990) and accumulates
within axonal swellings in traumatic brain injury
(Gentleman et al., 1993) as early as 1.75 hours after
injury (Blumbergs et al., 1995). APP immunostaining of
damaged axons does not stain background structurally uninvolved axons, unlike traditional silver techniques and immunocytochemical methods for phosphorylated neurofilament proteins (Sheriff, Bridges
and Sivaloganathan, 1994). It has been suggested that
the varicosities and sinusoidally enlarged axons often
seen with silver impregnation methods represent the
earliest stage in the formation of axonal bulbs (Vanezis,
Chann and Scholtz, 1987). The difficulties in interpretation of these sinusoidally enlarged varicose axons
has resulted in an attempt to define the various axonal
changes seen in brain trauma more rigorously (Crooks
et al., 1992). Axonal balloonings were defined as
argyrophil structures located at the end of disrupted
axons, axonal swellings as lobulations of axons and
axonal enlargements as axons at least twice the width
of the surrounding fibers and it was concluded that
axonal enlargements and swellings in the absence of
axonal balloonings were insufficient to diagnose
unequivocal AI (Crooks et al., 1992; Crooks, 1991).
(b)
Macroscopic markers of DAI
Strich (1961) described hemorrhages in the corpus

callosum and one or both superior cerebellar peduncles in 20/20 cases of post-traumatic diffuse white
matter degeneration. Subsequent studies confirmed
that these focal hemorrhagic lesions were usually
associated with microscopic diffuse axonal damage
(Adams et al., 1982, 1989).
Focal lesions in the corpus callosum
These are small hemorrhages in the corpus callosum,
which tend to lie to one side of the midline and may
also involve the fornices and septum pellucidum. The
hemorrhages are sometimes restricted to the splenium,
where they usually involve the lateral margins and are
frequently bilateral (Figure 3.3). If the patient survives
for some weeks, the lesion becomes granular and
brown in color, and survival of months or more results
in the formation of a shrunken cystic scar or an area of
gliosis (Figure 3.4), which may be macroscopically
AXONAL INJURY
Figure 3.3
Multiple small striate hemorrhages in the splenium of the corpus callosum in a case of DAI.
Figure 3.4
Small focal area of scarring in the corpus callosum in a case of DAI, many with several months survival.
visible or only apparent on microscopic examination.

Histologically the hemorrhages may be perivascular or
in the neuropil, and are associated with diffuse axonal
damage. Axonal retraction balls also develop at the
margins of the focal lesions and older lesions show
reactive changes of microglial, astrocytic and capillary
endothelial proliferation, followed by progressive
removal of the damaged tissue by foamy macrophages,
with the formation of small cystic scars consisting of
hemosiderin-laden macrophages in an astroglial
meshwork.
Focal lesions in the brain stem
These are represented by small hemorrhages in the
dorsolateral midbrain and rostral pons, often involv-
43
ing the superior cerebellar peduncles. When bilateral,

the lesions are usually asymmetrical (Figure 3.5). With
the passage of time these focal lesions undergo the
same sequence of histological changes described in the
corpus callosum. Occasional focal lesions show little
evidence of hemorrhage and histologically appear as
focal vacuolations of the neuropil.
These macroscopic foci of vascular injury have been
shown to be present only in the severe end of the
spectrum of DAI (Pilz, 1983; Blumbergs, Jones and
North, 1989; Adams et al., 1989) and presumably the
injury tolerance of the long perforating blood vessels
is similar, but not identical, to the injury tolerance of
axons.
DAI is present in about 30% of fatal head injuries
that survive long enough to reach hospital (Adams
44
PATHOLOGY
Figure 3.5 Bilateral hemorrhagic focal lesions in the rostral

pons in a case of DAI.
et al., 1989; Pilz, 1983; Blumbergs, Jones and North,

1989).
(c)
Grading of DAI
A grading system of DAI based on the combination of

microscopic and macroscopic markers was proposed
by Adams et al. (1989). In grade 1 DAI there is only
microscopic evidence of axonal injury. In grade 2 DAI
there is also a focal lesion in the corpus callosum and
in grade 3 DAI there is an additional focal lesion in the
dorsolateral quadrants or quadrants of the rostral
brain stem as well as in the corpus callosum. Patients
with grade 1 and 2 DAI may have complete or partial
lucid intervals (Adams et al., 1989). Such a grading
system utilizes the presence of focal vascular injury as
a marker for AI.
(d)
Clinicopathological features of DAI
It has become increasingly clear that there is a

spectrum of axonal injury and that the early studies of
marked atrophy of the white matter in vegetative state
patients (Strich, 1956, 1961) described the severe end
of the spectrum.
Different series have shown that 1430% of
patients with DAI may not be permanently unconscious from the time of injury and may experience a
lucid or partial lucid interval (Pilz, 1983; Blumbergs,
Jones and North, 1989, Adams et al., 1989).
The early studies found that DAI usually occurred
as a result of road traffic accidents and was associated
with a low incidence of skull fractures, cerebral
contusions, intracranial hematoma and raised intra-
cranial pressure (Adams et al., 1982). The acceleration/

deceleration forces in many such road traffic accidents
are believed to be of relatively gradual onset and
longer duration, similar to the mechanical loading
known to produce DAI in the laboratory (Gennarelli et
al., 1982a). More recent studies have shown that DAI
can also be found in patients with skull fractures and
focal mass lesions (Sahuquillo-Baris et al., 1988). DAI
may occur in falls greater than the persons height
(Adams et al., 1984) and may also occur as the result of
a fall from the victims own height (Imajo and Kazee,
1992). DAI has also been described in 15/50 fatal head
injuries caused by assault (Graham et al., 1992).
Some of the larger hemorrhages in the corpus
callosum and dorsolateral quadrants of the brain stem
may be imaged on CT scans and aid in the antemortem diagnosis of DAI (Zimmerman and Bilaniuk,
1979).
Quantitation of AI using the AISS has shown that
the brains of patients with mild head injuries (GCS of
1315) have lower scores than those from severe head
injuries (GCS 38) (Blumbergs et al., 1995).
It has been postulated (Katz and Alexander, 1994)
that in mild DAI unconsciousness lasts for seconds to
minutes (if at all), confusion and post-traumatic
amnesia (PTA) for minutes to hours, and residual
neurological dysfunction for days to several months;
in more severe DAI there may be days to weeks of
unconsciousness, weeks to months of confusion and
PTA, and months to years of residual dysfunction. The
most severe end of the DAI spectrum is represented by
the persistent vegetative state.
(e)
Immunocytochemical markers of DAI
Neurofilament proteins
Neurofilaments are the major cytoskeletal constituents
of mature axons and are composed of neurofilament
proteins (NFPs) made up of 68 kDa (NFP-L),
140160 kDa (NFP-M) and 200 kDa (NFP-H) subunits.
NFP immunostaining (Figure 3.6) was found to be
superior to Glees and Marsland silver impregnation
for the demonstration of axonal injury (Ng, Mahaliyana and Poon, 1994). Grady et al. (1993), using
antibodies against NFP-L and NFP-H subunits,
showed that the axonal swellings consist of focal
accumulation of the 68 kDa neurofilament protein
subunit at 6 hours postinjury and that the swellings
continue to expand and finally progress to complete
axotomy by 12 hours, supporting the experimental
studies which have demonstrated that axonal bulbs or
retraction balls are not due to immediate physical
tearing of the axons (Povlishock, 1992). Antibodies to
the low MW NFP-L were most useful in demonstrating early axonal damage (within 6 hours postinjury)
AXONAL INJURY
45
brain APP immunocytochemistry can detect AI after

only 1.75 hours survival (Blumbergs et al., 1995) and is
the method of choice for demonstrating AI (Figures 3.1
and 3.2) when compared to NAP-25, chromogranin A,
cathepsin D, NFP-L synaptophysin, ubiquitin and tau
(Sheriff, Bridges and Sivaloganathan, 1994).
(f)
Figure 3.6 Positive immunostaining of axonal swellings

and retraction balls with phosphorylated high-molecularweight neurofilament protein antibody.
without the complicating positive immunostaining of

normal axons seen with NFP-H (Grady et al., 1993).
Focal axonal increase in the 68 kDa neurofilament
subunit has been documented 1 hour after injury in
the rat (Yaghmai and Povlishock, 1992).
Ubiquitin
Ubiquitin is a low-molecular-weight stress protein
that is involved in ATP-dependent non-lysosomal
degradation of abnormal or short-lived intracellular
proteins (Finley and Varshavsky, 1985). Ubiquitinpositive immunostaining of axonal swellings after
trauma has been reported in man (Gultekin and
Smith, 1994) and in a cat model 6 hours after injury
(Schweitzer et al., 1993). However, Sheriff, Bridges and
Sivaloganathan (1994) were unable to obtain positive
ubiquitin immunostaining of axonal swellings in their
human formalin-fixed material.
-amyloid precursor protein
-amyloid precursor protein (APP) is a membranespanning glycoprotein originating from a gene on
chromosome 21 (Kang et al., 1987). APP is transported
by fast axoplasmic transport (Koo et al., 1990) and is
located at both pre- and postsynaptic sites (Schubert et
al., 1991; Shigematsu, McGeer and McGeer, 1992). APP
is an excellent marker for axonal injury in experimental animals (Nakamura et al., 1992; Otsuka, Tomonaga and Ikeda, 1991) and around cerebral infarcts
and other pathological processes in humans (Ohgami,
Kitamoto and Tateishi, 1992), including traumatic AI
(Gentleman et al., 1993). APP accumulation is thought
to occur after cytoskeletal disruption (Otsuka, Tomonaga and Ikeda, 1991) resulting in inhibition of
axoplasmic flow (Shigematsu and McGeer, 1992). In
formalin-fixed, paraffin-embedded sections of human
DAI in infancy
The Naoumenko and Feigins silver stain is necessary

to demonstrate AI in infant brains as routine silver
stains (Bodian, Glees and Marsland and Palmgren) are
claimed to be insensitive and unreliable (Vowles,
Scholtz and Cameron, 1987).
AI was present in the cerebral hemispheres and
corpus callosum, but not the brain stem, of 6/10
battered babies (Vowles, Scholtz and Cameron, 1987).
Contusional tears in early infancy are a focal marker
for DAI (Vowles, Scholtz and Cameron, 1987).
(g)
Pathogenesis of DAI
Clinicopathological studies in man (Adams et al.,

1982) and experimental studies in primates (Gennarelli et al., 1982a) have shown that DAI is caused by
angular acceleration of the head, especially in the
coronal plane.
Experimental studies in fluid percussion and controlled cortical impact models have shown that axons
are not torn by shear and tensile forces at the moment
of injury but that the axonal deformation results in a
focal impairment of axoplasmic transport and subsequent focal swelling of the axon due to abnormal
accumulations of neurofilaments and membranous
organelles, and that over the next 612 hours there is
disconnection of the proximal axonal segment from
the distal segment, which undergoes Wallerian
degeneration (Povlishock et al., 1983; Povlishock and
Kontos, 1985; Erb and Povlishock, 1988; Cheng and
Povlishock, 1988; Povlishock, 1992). The overlying
myelin sheath is stretched but not ruptured and freeze
fracture studies of the myelin sheath of affected fibers
in a non-human primate AI model also failed to show
any abnormality (Maxwell et al., 1988a). These studies
showed that the earliest structural change occurs at
the nodes of Ranvier, which demonstrate membranebound cytoplasmic blebs containing vesicles, smooth
endoplasmic reticulum and neurofilaments (Maxwell
et al., 1991). No tearing of axons was produced in the
angular acceleration model but this was achieved in a
lateral acceleration non-human primate model, where
the tensile strains acting on the axons are very much
greater. Tearing of some axons occurred without the
prior formation of axonal swellings and there was
rapid resealing of the fragmented axonal membranes
within 60 minutes. The small axons were most
46
PATHOLOGY
vulnerable to axonal fragmentation, which usually

occurred at the nodal/paranodal regions. This type of
AI was termed primary axotomy. Other axons showed
axonal swellings secondary to accumulation of cytoskeletal components and membranous organelles,
which progressed to axonal disconnection 612 hours
after injury. This second type of AI was termed
secondary axotomy (Maxwell et al., 1993).
No evidence of axonal tearing or shearing was
found in the guinea pig optic nerve stretch model but
nodal blebs consisting of small herniations of nodal
axolemma and disorganized axoplasm were found at
the nodes of Ranvier 15 minutes after injury, with
resolution 57 days later (Gennarelli et al., 1989;
Maxwell et al., 1991).
In summary the experimental studies have shown
two types of axonal injury:
Primary axotomy, where tensile forces exceed a

critical level and there is mechanical shearing or
fragmentation of the axolemma at the time of injury,
and the organization of the axonal cytoskeleton is
lost (Maxwell et al., 1993);
Secondary axotomy (secondary axonal degeneration), where the axolemma and membranous organelles are initially intact, although the axonal
cytoskeleton is altered with the formation of axonal
swellings consisting of aggregates of membranous
organelles and cytoskeletal proteins prior to axonal
discontinuity some hours after injury (Povlishock,
1992).
3.4.4
NON-TRAUMATIC AI
Axonal retraction balls or bulbs are not restricted to

traumatic brain injury but also occur adjacent to focal
lesions such as infarcts and non-traumatic hematomas

(focal AI; Adams et al., 1989; Figure 3.7).
3.4.5
LATE SEQUELAE OF AI
After a few weeks the most striking abnormality is the

presence of vast numbers of small clusters of microglia
(microglial stars) throughout the white matter of the
cerebral hemispheres, cerebellum and brain stem
(Adams, 1992). This is followed after 23 months by
Wallerian degeneration of the axons and, depending
on the severity of the axonal injury, atrophy of white
matter tracts associated with compensatory dilatation
of the ventricular system (hydrocephalus ex vacuo) in
the most severe cases.
3.5
Concussive syndromes
Concussion and its structural basis have long been

controversial issues, although the importance of acceleration forces was established by Denny-Brown and
Russell as long ago as 1941.
The neuroanatomical basis for consciousness is not
well understood but it is believed that the awake state
is a complex interaction involving the cerebral cortex,
subcortical structures including the hypothalamus,
and numerous brain stem centers.
The awake state depends on a feedback loop of
neural activity between the ascending reticular activating system of the brain stem and the cerebral cortex.
Diffuse brain injuries interfere with consciousness by
damaging the white matter pathways connecting the
cerebral hemispheres to the brain stem activating
centers; in the more severe forms of diffuse injury there
may be additional direct anatomical damage to the
cerebral cortex and brain stem as well.
Figure 3.7 Focal infarction of corpus callosum and right cingulate gyrus. Numerous axonal swellings were present at the
margins of the callosal infarct.
HYPOXICISCHEMIC DAMAGE IN HUMANS
Diffuse brain injuries form a continuum of progressively severe brain dysfunction which, at the least
severe end of the spectrum, consists of temporary
cessation of function without any structural disruption of tissue. The International Neurotraumatology
Committee defines concussion as a clinical syndrome
characterized by immediate impairment of neural
functions such as alterations of consciousness, disturbance of vision, motion and sensation due to mechanical forces (Gurdjian et al., 1979). The classical description of cerebral concussion as an essentially reversible
syndrome without detectable pathology has been
extended by Ommaya and Gennarelli (1974), who
have described a graded spectrum of concussion
syndromes, the more severe of which may be associated with structural abnormalities. Mild concussion
syndromes are those in which consciousness is preserved but there is some degree of noticeable temporary neurological dysfunction (Fisher, 1966; Yarnell
and Lynch, 1973). The mildest form (grade 1) results in
momentary confusion and disorientation unaccompanied by amnesia. In grade 2 concussion syndrome
there is initial confusion followed by amnesia for 510
minutes. This is a common injury in contact sports,
and players, although confused, continue coordinated
motor activities after the injury.
If examined immediately after the accident, these
players possess an intact recall of the events immediately prior to the injury but 510 minutes later some
degree of retrograde amnesia (forgetting of events
before the injury) will be found.
In grade 3 concussion both confusion and amnesia
are present from the time of impact, although consciousness is still retained.
Some degree of post-traumatic amnesia (forgetting
of events after injury) also occurs in addition to
retrograde amnesia.
The fact that memory mechanisms seem very
sensitive to the effects of trauma suggests that the
cerebral hemispheres are more vulnerable than the
brain stem to acceleration/deceleration forces.
In classical cerebral concussion (grade 4) there is
loss of consciousness as well as some degree of
retrograde and post-traumatic amnesia, the length of
post-traumatic amnesia being a measure of the severity of the injury. Transient systemic changes such as
bradycardia, hypertension, apnea, pupillary dilatation
or flaccidity may occur and full consciousness returns
by 24 hours.
Patients with cerebral concussion rarely come to
neuropathological examination, but there is increasing evidence that some permanent damage is present.
Recent studies, using APP as a marker for AI, have
confirmed the original observations by Oppenheimer
(1968) of axonal damage in mild head injury (Blumbergs et al., 1994).
47
Postconcussive sequelae such as subjective symptoms of lack of mental concentration, tiredness, hyperacusis, vertigo and subtle changes in personality are
common (postconcussion syndrome, post-traumatic
syndrome) after mild head injury. There is increasing
recognition that the postconcussive sequelae of even
the most minor head injuries can be of major significance in terms of ability to return to work and
competence in previous occupations (Levin, Eisenberg
and Benton, 1989). Psychological tests are impaired for
several weeks after concussion, and for longer in
patients with persisting symptoms (Gronwall and
Wrightson, 1979).
Patients who suffer a second mild injury have more
serious and more prolonged impairment of function.
This has led to the concept of cumulative concussion,
based on the hypothesis of some residual permanent
structural damage from the first injury (Gronwall and
Wrightson, 1975).
Evidence for the latter is the defective performance
on psychological testing of previously concussed
university students under conditions of artificial mild
hypoxia (Ewing et al., 1980) and the demonstration of
axonal damage in humans with mild head injury
(Oppenheimer, 1968; Blumbergs et al., 1994) and
experimentally concussed non-human primates
(Adams, Graham and Gennarelli, 1981).
It is important to remember that classical cerebral
concussion is often associated with other types of
brain injury, both focal and diffuse, the most commonly associated lesion being cerebral contusion.
Cerebral contusions do not alter consciousness, unless
large and associated with raised intracranial pressure, but may produce focal neurological deficits.
3.6
Hypoxicischemic damage in humans
In about one-third of severely head-injured patients

the cerebral blood flow is reduced to ischemic levels
(CBF < 18 ml/100 g/min) 26 hours after injury and
then subsequently returns to non-ischemic levels
(Bouma et al., 1991, 1992; Bouma and Muizelaar, 1992)
During this period the brain may be at risk from an
ischemiareperfusion type injury (Chapter 10). Cerebral blood flow studies have shown alterations in
cerebral blood flow with loss of normal PCO2 vasoconstriction following hyperventilation (Marion, Darby
and Yonas, 1991) and loss of autoregulation with
vulnerability to blood pressure extremes (Enevoldsen
and Jensen, 1978). Hypoxicischemic brain damage is
likely whenever, either singly or in combination, there
is a reduced content of oxygen in the blood (hypoxic
hypoxia) or reduced blood flow due to a reduction in
cerebral perfusion (ischemic hypoxia).
Most of the evidence supports the view that failure
of brain perfusion with resultant ischemic hypoxia is
48
PATHOLOGY
the most common type of hypoxicischemic brain

damage in man (Adams and Graham, 1984). This type
of damage is very difficult to recognize unless the
brain is properly fixed and large representative sections of the brain are histologically examined.
Ischemic changes in fatal head injury have been

classified as boundary zone (1325%), arterial territory (1719%; Figure 3.8), diffuse (2842%; Figure 3.9)
and mixed (Graham, Adams and Doyle, 1978; Graham
et al., 1989).
Ischemic damage is found in about 90% of patients
who survive for several hours after injury and is
significantly associated with a known episode of
hypoxia such as a cardiac arrest or status epilepticus
and raised intracranial pressure (Graham, Adams and
Doyle, 1978; Graham et al., 1989). Severe hypoxic
damage may occur, however, in the absence of high
intracranial pressure (Graham et al., 1988). Episodes of
systemic arterial hypotension produce ischemic damage in an arterial boundary zone pattern.
Other important factors which may contribute to
hypoxic brain damage are obstruction of the airways,
chest injuries and arterial spasm (MacPherson and
Graham, 1978).
The superoxide anion is believed to play a central
role in the vascular response to trauma and has been
linked to altered endothelial-dependent responses
(Ellison et al., 1989), alterations in blood-brain barrier
status (Wei et al., 1986) and morphological abnormalities of the endothelial cells (Kontos and Povlishock, 1986).
3.7
Figure 3.8 Multiple areas of cortical infarction in a patient

with severe head injury who survived for 28 days.
Brain swelling
Massive swelling of all or part of the brain occurring

after head injury may be due to increased tissue water
content of the brain (cerebral edema), increased
intravascular blood volume (congestive brain swelling) or a combination of the two.
Figure 3.9 Diffuse ischemic necrosis of the cortical mantle and secondary degeneration of the central white matter.
Pedestrian struck by car with cardiac arrest and subsequent vegetative state maintained on ventilator for 1 year.
BRAIN SWELLING
3.7.1
CEREBRAL EDEMA
Cerebral edema is an important but variable secondary response to trauma, the causes and consequences
of which are poorly understood. Five different types
have been described (Miller, 1993).
Vasogenic edema. Impairment of the bloodbrain

barrier leads to accumulation of protein-rich fluid in
the extracellular space. The localized swelling
related to contusions and intracerebral hematomas
is vasogenic in type.
Cytotoxic edema occurs in association with
hypoxicischemic damage where there is a disturbance of ionic gradients leading to an accumulation
of intracellular fluid.
Hydrostatic edema. The sudden increase of intravascular or transmural pressure in an intact vascular
bed results in the extracellular accumulation of
protein-poor fluid. Hydrostatic edema may follow
the sudden decompression of a mass lesion or when
autoregulation is defective following head injury.
Osmotic brain edema. Critical reductions in serum
osmolality result in increased intracellular water.
This may occur in iatrogenic hemodilution from
excessive use of intravenous dextrose water solutions or the inappropriate secretion of antidiuretic
hormone syndrome.
Interstitial brain edema. Periventricular extravasation of water may occur as a result of high-pressure
obstructive hydrocephalus. This is an uncommon
event in head injury.
3.7.2
CONGESTIVE BRAIN SWELLING
Congestive brain swelling occurs when the cerebral

blood volume is increased as a result of arterial dilatation and/or venous obstruction. This may occur very
rapidly in children and in patients with acute subdural
hematomas. Massive cerebral hemisphere swelling 20
minutes after head injury has been documented by CT
(Kobrine et al., 1977). Cerebral blood flow studies have
shown a hyperemic phase that peaks about 24 hours
after the traumatic event in patients without evidence
of significant surgical mass lesions (Marion, Darby and
Yonas, 1991; Bouma et al., 1991, 1992).
3.7.3
49
BBB disruption and vasogenic edema and ischemic

damage resulting in intracellular swelling (cytotoxic
edema). Experimental studies have confirmed that the
water content of brain tissue around cerebral contusions is increased (Tornheim, McLaurin and Thorpe,
1976).
(b)
Diffuse swelling of one cerebral hemisphere
Lobato et al. (1988) reported that massive cerebral

swelling (Figure 3.10), not related to contusions,
intracerebral hemorrhage or focal infarction, occurred
in 10.5% of 520 patients with severe head injury
studied with computed tomography. In 85% of cases
the swelling was associated with an ipsilateral subdural hematoma, in 9% with a large epidural hematoma and occurred in only 5.4% as an isolated lesion.
The rapidity of such swelling suggests that this is
congestive in type initially, although the subsequent
ICP increase may result in ischemic cytotoxic edema
as well.
(c)
Diffuse cerebral swelling of the entire brain
The reported incidence of diffuse brain swelling

following head trauma varies from 540% (Lang et al.,
1994). Diffuse brain swelling occurs more frequently
in children than in adults (Bruce et al., 1981; Aldrich et
al., 1992; Lang et al., 1994). Early studies emphasized
that diffuse cerebral swelling in children could follow
an apparently trivial injury, and that it was a common
cause of delayed neurological deterioration (Bruce et
al., 1981). Lang et al. (1994) found that 75% of children
with diffuse swelling pursued a relatively benign
course unless it was associated with severe primary
injury or a secondary hypotensive insult, whereas
diffuse swelling in adults resulted in a poor outcome
in two-thirds of cases. The pathogenesis of diffuse
BRAIN SWELLING IN HEAD INJURY
Adams (1992) has identified three main patterns of

brain swelling following head injury.
(a) Swelling around cerebral contusions and
intracerebral hemorrhages
The edema surrounding these lesions may be due to a
variable combination of vascular damage leading to
Figure 3.10 Swelling of left cerebral hemisphere subsequent to removal of an overlying acute subdural
hematoma.
50
PATHOLOGY
brain swelling is unclear but it has been postulated

that it is congestive in type and due to loss of
vasomotor tone (Bruce et al., 1981).
3.8 Neurotransmitter agonistreceptor

abnormalities
Experimental studies have suggested that excessive
neuroexcitation may result from activation of
N-methyl-D-aspartate (NMDA) and muscarinic cholinergic receptors by the release of glutamate and
acetylcholine neurotransmitters. Traumatic tissue
deformation results in membrane depolarization
(impact depolarization; Hayes, Jenkins and Lyeth,
1992) and the excessive release of these neurotransmitters. Microdialysis studies have shown glutamate elevations following head injury in humans
(Persson and Hillered, 1992) and animals (Katayama et
al., 1990; Faden et al., 1989).
Activation of NMDA receptors may result in excessive calcium entry via NMDA channels to activate
intracellular enzymes and excessive muscarinic receptor activation may alter potassium and calcium conductance via the inositol 1,4,5-triphosphate (IP3) or the
1,2-diacylglycerol protein kinase C (PKC) limbs of
G-protein-coupled phosphoinositide turnover (Hayes,
Jenkins and Lyeth, 1992). It is likely that other
neurotransmitter systems are similarly affected by
impact membrane depolarization, depending on the
region/s of the brain deformed by the mechanical
forces.
excitation secondary to the excessive release of the

excitatory neurotransmitter glutamate (Kotapka et al.,
1991). The CA1 sector of the hippocampus is rich in
glutamate receptors and receives a glutaminergic
input via the perforant pathway originating from the
entorhinal cortex. Extracellular glutamate concentration increases following fluid percussion brain injury
in the rat (Faden et al., 1989) and treatment with
phencyclidine, a glutamate receptor antagonist, prior
to the injury attenuates the hippocampal damage
(Jenkins et al., 1988).
3.10
Traumatic vascular injury
There is a large potential spectrum of traumatic

vascular injuries (Table 3.4) that may occur in isolation
or in different combinations.
Table 3.4 Traumatic vascular injury
Intraparenchymal blood vessels
Focal vascular injury
Contusions
Multifocal vascular injury
Combination of above
Petechial hemorrhage and

microhemorrhages
Extraparenchymal blood vessels

Bridging veins and
arteries
3.9 Hippocampal pathology in traumatic

brain injury
The hippocampus is frequently damaged in human
traumatic brain injury (Graham, Adams and Doyle,
1978; Graham et al., 1989) with histopathological
abnormalities present in 84% of cases (Kotapka et al.,
1992). The CA1 sector is always involved, and both
hippocampi are involved in the majority of cases
(Kotapka et al., 1992). Pathological evidence of high
ICP is present in 86% of cases (Kotapka et al., 1992)
and the hippocampal pathology in the majority of
cases has been considered secondary to decreased
perfusion pressure and/or raised intracranial pressure
(ICP) (Graham et al., 1987).
However, similar hippocampal damage has also
been noted in cases without evidence of high ICP
(Graham et al., 1988; Kotapka et al., 1992) and in a nonhuman primate acceleration injury model in which
there was no evidence of abnormal ICP or cerebral
perfusion pressure (Kotapka et al., 1991); it has been
proposed that this is due to pathological neuronal
Blood vessels in close

proximity to skull bones
Meningeal arteries and
veins
Venous sinuses
Large arteries in neck
Internal carotid arteries
Vertebral arteries
Blood vessels of the circle

of Willis
Middle, anterior and
posterior cerebral
arteries
Basilar artery
Intracranial internal
carotid and vertebral
arteries

(ASDH)
(CSDH)
Extradural (epidural)
hematoma (EDH)
Thrombosis
Dissection
Subintimal hemorrhage
Laceration
Arteriovenous fistula (AV
fistula)
Thrombosis
Dissection
Subintimal hemorrhage
Laceration
Arteriovenous fistula
TRAUMATIC VASCULAR INJURY
3.10.1
(a)
CONTUSIONS
Cerebral contusions
Cerebral contusions are focal injuries that result when

localized mechanical forces damage the small blood
vessels (capillaries, veins and/or arteries) and other
tissue components (nerve and glial cells and their
processes) of the neural parenchyma. The bleeding
from damaged blood vessels is usually the most
obvious feature on macroscopic and microscopic
examination and results in a spectrum of abnormalities ranging from microhemorrhages to confluent
hemorrhage disrupting the tissue.
Some define contusions as surface lesions of the
brain whereas others also include similar hemorrhagic
lesions in the deeper structures of the brain (Linden-
Figure 3.11 Multiple contusions involving cortex and

subcortical white matter of inferior frontal lobes and
temporal poles.
Figure 3.12
51
berg, 1971). Lindenberg (1971) classifies contusions

into contusion hemorrhages, where focal vascular
injury predominates, and contusion necrosis where
direct damage of the parenchymal elements is present.
Some contusion necroses show no evidence of hemorrhage (Lindenberg, 1971) and these may be the
pathological counterparts of the focal cortical hyperintensities detected by MR and termed non-hemorrhagic contusions (Snow et al., 1986).
In a simple contusion the overlying piaglial membrane is intact. Disruption of this membrane with
tearing of the underlying tissue constitutes a laceration. Contusions and lacerations form a continuum of
tissue injury. Surface contusions of the brain show a
wide spectrum of morphological appearances, varying from microhemorrhages only visible under the
microscope to confluent hemorrhagic necrotic lesions
extending through the cortex into the subcortical
white matter (Figure 3.11). Surface contusions typically affect the crests of gyri and often produce wedgeshaped necrosis of the involved neural parenchyma.
This is in contrast to the preservation of the subpial
molecular layer, with damage in the depths of the
sulci, seen in cerebral infarction. Contusions are
dynamic lesions that evolve with time.
The damage to the blood vessels sets in train an
intertwined cascade of events leading to hemorrhage,
delayed breakdown of the blood brain barrier and
infarction secondary to compromise of the microcirculation including thrombotic occlusion of blood
vessels. This produces a spectrum of macroscopic
changes varying from focally dilated blood vessels to
burst brain lobes (Oprescu, 1991).
Acute surface contusions are characterized by focal
vascular damage leading to punctate hemorrhages or
small linear hemorrhages aligned at right angles to the
Sulcal hematoma surrounded by compressed focally contused cerebral cortex.
52
PATHOLOGY
cortical surface due to extension of hemorrhage along

the perivascular plane. Occasionally, local subarachnoid hemorrhage adjacent to a contusion accumulates within the sulcus to form a sulcal hematoma
(Figure 3.12). This may lead to an erroneous diagnosis
of intraparenchymal hemorrhage on CT head scan.
The radiating streak-like cortical hemorrhages on
microscopy consist of perivascular accumulations of
red cells and serial sectioning may show evidence of
focal traumatic rupture or tearing of the affected blood
vessel with bleeding into the perivascular space and/
or neural parenchyma. Damaged blood vessels may
thrombose, leading to additional ischemic complications. Contusions often increase in size over hours to
days due to the evolving events related to the

interplay of hemorrhage, early ischemic necrosis and
delayed vasogenic edema (Figure 3.13). The next
phase is that of reabsorption of damaged tissue and
progressive reactive gliosis. Very small hemorrhages
may be completely reabsorbed within 23 weeks,
whereas larger hemorrhages may take many weeks or
months to reabsorb. The extravasated red blood cells
are sequentially broken down to various blood pigments, including hemosiderin. Necrotic brain tissue is
phagocytosed by macrophages derived from monocytes at sites where there has been disruption of the
bloodbrain barrier and lipid macrophages appear
25 days after the insult. The end result of these
Figure 3.13 Contusion of inferior temporal lobe showing hemorrhagic and ischemic necrosis and intraparenchymal and
subarachnoid hemorrhage.
Figure 3.14 Old traumatic lesion (plaque jaune) involving crest of gyrus consisting of a shrunken hemosiderin-stained old
contusion necrosis.
TRAUMATIC VASCULAR INJURY
processes of reabsorption is a shrunken brown cystic

lesion involving the crests of gyri and communicating
with the subarachnoid space (plaque jaune; Figure
3.14).
Contusions most frequently involve the inferior
frontal lobes (Figure 3.15) and the inferolateral temporal lobes and poles where brain tissue comes in contact
with the irregular bony surfaces of the anterior and
middle cranial fossae due to the relative motion of the
brain and skull at these sites (Adams et al., 1980b). The
occipital lobes and cerebellum are rarely contused in
the absence of skull fractures (Lindenberg, 1971;
Adams et al., 1980b).
(b)
Types of contusion
Coup contusions are contusions that occur beneath

the site of impact. Coup contusions are a type of
contact injury (Gennarelli and Thibault, 1985) and are
produced by compressive forces operating beneath an

area of skull inbending or tensile forces generated by
the negative pressure produced beneath an area of
skull inbending suddenly snapping back into place.
Contrecoup contusions occur opposite the impact
site (Courville, 1942, Gurdjian and Gurdjian, 1976).
Intermediate coup contusions are intracerebral
lesions that occur deeply within the neural parenchyma between the impact site and the opposite
side of the brain (Lindenberg, 1971).
Fracture contusions occur beneath the site of a
fracture (Lindenberg and Freytag, 1960)
Gliding contusions occur in the parasagittal
regions and are believed to be due to the rostrocaudal
movement of the brain resulting from impact or
impulsive loading forces (Lowenhielm, 1975; Voigt,
Lowenhielm and Ljung, 1977). The hemorrhages
involve the deeper layers of the cortex and the
convolutional white matter and spare the surface of
the gyrus (Figures 3.16, 3.17). Gliding contusions are
often associated with diffuse axonal injury (Adams et
al., 1986a).
Herniation contusions involve the medial temporal
lobes and the cerebellar tonsils and are produced by
movement of the brain impacting on the rigid tentorium cerebelli or the bony margins of the foramen
magnum (Figure 3.18).
(c)
Figure 3.15 Recent contusions (cortical and subcortical) of

the inferior frontal lobes.
Figure 3.16
53
Contusion patterns and head impacts
Earlier studies suggested that the contusion pattern

depends on the direction and magnitude of the
impacting force and whether the head is accelerated
by the impact (e.g. a blow to the movable head), is not
accelerated by the impact (e.g. a blow to the supported
head) or is in a state of acceleration at the moment of
Gliding contusion in parasagittal white matter and deep cortex.
54
PATHOLOGY
Figure 3.17 Bilateral gliding contusions with minimal lesion limited to deep left parasagittal cortex and a larger subcortical
hemorrhage in the right parasagittal region.
Figure 3.18
Herniation contusion of cerebellar tonsil.
impact (e.g. a fall on the head; Lindenberg, 1971;

Gurdjian and Gurdjian, 1976). Thus Gurdjian and
Gurdjian (1976) observed that a lateral impact in the
fronto-temporal area may produce a surface contusion
of the contralateral temporal lobe and contusions of
both uncinate gyri; a lateral temporo-parietal impact
may result in a contrecoup contusion of the temporal
lobe; a midline occipital impact may produce bilateral
frontal and temporal lobe contusions; an occipital
impact lateral to the midline may cause contrecoup
contusions of the frontal and temporal lobes; frontal
impacts may result in bilateral or unilateral contusions
of the frontal and temporal lobes; vertex impacts may
produce contusions of the brain stem and tears of the
corpus callosum and pituitary stalk. In contrast,
Adams et al. (1980b, 1985), using the contusion index
based on the extent and depth of contusion in various
regions of the brain, showed that surface contusions
are most severe in the frontal and temporal lobes,

irrespective of the cranial impact site, provided the
forces acting on the head are sufficient to impart
movement of the brain over the irregular bony
surfaces of the anterior and middle cranial fossae.
Both frontal and occipital impacts resulted in contusions which were most severe in the frontal lobes and
they concluded that the site of head impact could not
be extrapolated as being diametrically opposite the
area of most severe contusion (Adams et al., 1980b).
The histological aging of cerebral contusions may be
of medicolegal importance in determining the exact
time of brain injury (Oehmichen and Raff, 1980;
Loberg and Torvik, 1989). However, the details of the
time sequences for the different histological changes
vary in different published series (Lindenberg, 1971;
Oehmichen and Raff, 1980; Loberg and Torvik, 1989;
McCormick, 1985).
(d)
Experimental contusion injuries
Ommaya, Grubb and Naumann (1971), in a study of

coup and contrecoup lesions in Rhesus monkeys,
showed that blows to the back of the head produced a
contrecoup lesion in about 90% of cases whereas
frontal impacts produced a contrecoup lesion in only
10% of animals. Blows to the side of the head
produced coup and contrecoup lesions with about
equal frequency. It was suggested that these contusions were produced by distortion of the skull and
head rotation (Ommaya, Grubb and Naumann, 1971).
Gennarelli et al. (1979) produced frontal and temporal
contusions in subhuman primates using short-duration high-acceleration forces with the Penn I device
and showed that frontal lobe contusions occurred at
TRAUMATIC INTRACEREBRAL HEMORRHAGE
55
lower thresholds than temporal lobe contusions and

that the extent of the contusions was determined by
the mechanical impact whereas the depth of the
contusions increased with time, probably secondary to
progressive ischemic damage.
Dawson et al. (1980) reviewed the theories attempting to explain why, in the absence of skull fractures,
the moving head sustains contusions opposite the
point of impact but none beneath the point of impact,
and why a blow to the resting, movable head as a rule
produces contusions beneath the point of impact but
none opposite the point of impact.
(e)
Clinicopathological correlations
Neurological deficits usually correlate with the size

and location of contusions. Patients with contusions
may show progressive or sudden deterioration.
Sudden deterioration is especially a feature of
patients with severe bifrontal contusions (Vigouroux
and Guillermain, 1981; Marshall, Toole and Bowers,
1983; Statham, Johnston and Macpherson, 1989) and
temporal pole pulping (McLaurin and Helmer, 1965).
Contusions are also one of the causes of neurological
deterioration after a lucid interval (talk and die
patients), mimicking extracerebral hematomas (Reilly
et al., 1975, Rockswold, Leonard and Nagib, 1987).
However, contusions may be totally absent in
patients who have sustained severe or lethal head
injury (Adams, 1992).
Some patients with large contusions on CT head
scan may not show any alteration in conscious state
and remain in a stable clinical condition (Vigouroux,
Guillermain and Rabehanta, 1991). Surface contusions
are less severe in patients with DAI and more severe in
patients with skull fracture and those who do not
experience a lucid interval (Adams, 1992).
3.11
Lacerations
Lacerations of the brain may be defined as primary

disruptions of the neural parenchyma occurring at the
moment of injury. In some cases the damage may be
limited to the cortex but often the lacerations extend
into the deep white matter of the temporal and frontal
lobes (Figure 3.19) and occasionally into the central
gray matter. The mechanical forces required to produce
brain laceration usually also produce other lesions such
as contusions and intracerebral hemorrhage. Lacerations may be direct or indirect (Oprescu, 1991).
In direct lacerations the parenchymal disruption is
caused by a penetrating injury from various types of
missiles or an open depressed fracture of the skull
with penetration of the brain by fragments of bone
and foreign bodies. Basal skull fractures may lacerate
the dura mater and the adjacent brain tissue.
Figure 3.19 Healing old laceration/contusion involving

the left inferior frontal lobe.
In indirect lacerations the parenchymal disruption

is secondary to tissue deformation produced by
mechanical forces. The temporal lobes and/or the
inferior frontal lobes are most frequently involved and
when combined result in temporo-orbital lacerations
which may be superficial (cortico-subcortical) or
extend into the deep structures such as the temporal
horn, hippocampus and amygdaloid nucleus burst
temporal lobe.
The margins of recent lacerations show hemorrhage
and necrosis. After several days, phagocytic macrophages infiltrate the edges and there is progressive
fibroblastic and neovascular proliferative ingrowth
from the adjacent vascular meninges. Progressive
collagen deposition admixed with astrogliosis results
in a meningocerebral cicatrix which macroscopically
appears as an area of thickened meninges adherent to
the damaged cortex. The meningocerebral cicatrix
may appear yellowish-green to brown, depending on
the stage of reabsorption of the associated hemorrhage. Occasionally, meningocerebral scars extend
through the cerebral parenchyma to involve the
ventricle.
3.12
Traumatic intracerebral hemorrhages
These are defined as hematomas 2 cm or greater in size

not in contact with the surface of the brain and are
present in 15% of autopsy cases of severe head injury
(Adams, 1992). Traumatic intracerebral hemorrhages
are often multiple (Figure 3.20) and 28% are associated with subdural and 10% with epidural hematoma (Soloniuk et al., 1986); they can arise in areas
that appear normal on CT scans obtained soon after
injury (Gudeman et al., 1979; Soloniuk et al., 1986).
Lobar intracerebral hemorrhages are those that
involve a lobe of the brain and occur usually in the
temporal or frontal lobes (Jamieson and Yelland, 1972).
56
PATHOLOGY
(a)
Figure 3.20 Right basal ganglionic hemorrhage and hemorrhage in the right temporal lobe associated with laceration
and contusion.
The pathogenesis of intracranial hemorrhage (ICH) is

not always clear but it is likely that most are caused by
deformation and rupture of intrinsic blood vessels
(single or multiple) at the time of injury (DeVet, 1976).
Damage to multiple small blood vessels may result in
the coalescence of many smaller hemorrhages (Figure
3.21). One-third to one-half of patients with ICH are
unconscious on admission and up to 20% demonstrate
a classic lucid interval before the onset of coma
(Soloniuk et al., 1986). Patients who are deeply
comatose with large hematomas have a high mortality
(Teasdale et al., 1982). Large hematomas act as spaceoccupying lesions and result in intracranial hypertension and resultant transtentorial herniation. Animal studies have shown marked hypoperfusion
around hematomas as well as ipsilateral hemisphere
ischemia (Sinar et al., 1987).
Figure 3.21 Intracerebral hemorrhage in the left basal

ganglia region due to coalescence of multiple smaller
hemorrhages associated with widespread contusions of
frontal and temporal lobes.
Traumatic basal ganglia hemorrhages
With the aid of CT scanning traumatic basal ganglionic hemorrhages (TBGH) have been recognized in
some 3% of cases of severe closed head injury
(Macpherson et al., 1986; Katz et al., 1989; Colquhoun
and Rawlinson, 1989) (Figure 3.20). In a postmortem
series deep intracerebral hemorrhages (thalamus and
basal ganglia) were present in about 10% of fatal head
injuries and were associated with an increased incidence of diffuse axonal injury and gliding contusions
(Adams et al., 1986b). Most of the hematomas (43/63
cases) were small (less than 2 cm in diameter; Adams
et al., 1986b; Figure 3.22).
Their pathogenesis has been postulated to be the
result of shearing of the deep blood vessels as a result
of acceleration/deceleration forces (Macpherson et al.,
1986; Adams et al., 1986b). Why TBGH occurs in
isolation in some, or with a variety of associated
damage in others, is not known. Mosberg and Lindenberg (1959) described 20 cases of traumatic pallidal
and putamenal hemorrhage in the anterior choroidal
and lenticulo-striate vascular territories and in one
case found histological changes suggestive of traumatic tearing of a small arterial twig of the anterior
choroidal artery. They proposed that the vascular
damage was the result of shearing and stretching of
the anterior choroidal vessels against the edge of the
tentorium during tissue shifts (Mosberg and Lindenberg, 1959).
Patients with TBGHs have a generally poor prognosis (Macpherson et al., 1986) but isolated TBGHs
may have a favorable outcome (Katz et al., 1989;
Colquhoun and Rawlinson, 1989). Macpherson et al.
(1986) observed that 88% of patients with TBGH and
raised ICP had a poor outcome (severe disability,
Figure 3.22
Small hemorrhage in the right thalamus.
TRAUMATIC INTRACEREBRAL HEMORRHAGE
57
vegetative state or death) whereas TBGH with normal

ICP had a 50% poor outcome.
Hypodense lesions of the basal ganglia consistent
with infarction have been noted on CT head scans of
children following minor head injury and have been
attributed to stretching or distortion of the perforating
branches of the middle cerebral arteries (Dharker,
Mittal and Bharagava, 1993).
(b)
Traumatic intraventricular hemorrhage
A small amount of intraventricular hemorrhage (IVH)

is frequently found in head-injured patients, both in
autopsied cases and in CT scans. It is often impossible
to determine the source of hemorrhage but bleeding
may be due to small tears in the veins of the ventricle
walls, tears in the corpus callosum, septum pellucidum and fornices or tears in the choroid plexus
(Lindenberg, 1971; Grcevic, 1983). It has been suggested that the sudden dilatation of the ventricular
system at the time of impact leads to deformation and
rupture of subependymal veins (Zuccarello et al.,
1981). IVH has been found in 1.53% of all patients
with non-missile head trauma on routine CT head
scanning and in 10% of patients with severe head
injury (LeRoux et al., 1992).
In one-third of patients, IVH is secondary to
extension of frontal, temporal and basal ganglia
hematomas into the ventricular system (Figure 3.23),
often following a sagittal plane head impact (Fujitsu et
al., 1988).
The majority of the other cases of IVH (Figure 3.24)
show CT evidence of diffuse brain damage (Cordobes
et al., 1983) and have a very poor prognosis. About
50% of cases with CT evidence of DAI (hemorrhagic
lesions in the corpus callosum and dorsolateral brain
Figure 3.23
Figure 3.24 Intraventricular hemorrhage associated with

widespread (diffuse) vascular injury.
stem) also show IVH (Zimmerman, Bilaniuk and

Gennarelli, 1978; Wilberger et al., 1990). Experimental
studies of lateral head acceleration in non-human
primates have shown rapid bleeding into the ventricles, probably from the moment of injury, and the
passage of plasma and blood cells either through tears
in the choroid plexus or via the choroidal epithelial
cells (Maxwell et al., 1992a). In large isolated IVH,
occult small vascular malformations and tumors
require careful exclusion.
(c)
Delayed traumatic intracerebral hematomas
Neuroimaging has shown that many traumatic hematomas develop hours to days after the injury and may
not be visible on scanning soon after the traumatic
Burst left temporal lobe with hemorrhage into the left lateral ventricle.
58
PATHOLOGY
event. Post-traumatic hematomas have been classified

on the basis of CT abnormalities into four main types
(Fukamachi et al., 1985). Type 1 hematomas are already
evident on the initial scan; in the second type, small or
medium-sized hematomas on the initial scan increase
in size on a subsequent CT; in the third type,
hematomas develop at a site that showed no abnormality on the initial scan and in the fourth type, hematomas
develop in an area of abnormality (salt and pepper or
flecked high-density change) present on the first scan
(Fukamachi et al., 1985). The incidence of delayed
traumatic intracerebral hematomas (DTICH), or Spatapoplexie of Bollinger, varies from 0.38.7% of patients
with severe head injury (Baratham and Dennyson,
1972; Diaz et al., 1979; Gudeman et al., 1979; Ninchoji et
al., 1984; Mertol et al., 1991). DTICH occurs more
frequently in older age groups (Baratham and Dennyson, 1972) and in falls (Diaz et al., 1979). DTICH can
occur in severely brain-injured patients as well as
patients sustaining relatively mild injuries. Hematomas may be discovered within hours or days to weeks
after the injury (Baratham and Dennyson, 1972).
DTICH should be suspected in a patient who has
shown improvement or remained stable after head
injury and then rapidly deteriorates (Brown, Mullan
and Duda, 1978) or in patients that fail to improve.
Patients undergoing ICP monitoring may show sudden marked intracranial hypertension. Occasionally,
DTICHs become apparent on serial CT scans with no
change in the patients neurological status or ICP.
Suggested mechanisms of formation of DTICH
include vascular abnormalities related to diffuse
intravascular coagulation and fibrinolysis (Pretorius
and Kaufman, 1982; Kaufman et al., 1980, 1984;
Kelmowitz and Annis, 1973; Tuoho et al., 1986), vessel
wall damage secondary to pH changes as a result of
cell damage (Young et al., 1984) or vessel wall
weakness following injury (Diaz et al., 1979). Saweda
et al. (1984) found no evidence of a coagulopathy in 18
patients who developed DTICH. DTICH can also
occur after decompression of extracerebral hematomas, where the damaged vessels were tamponaded by
the pressure of the extra- or subdural hematoma
(Hirsh 1979; Modesti, Hodge and Barnwell, 1982;
Taneda and Irino, 1979).
3.13
Extradural (epidural) hemorrhage
Extradural hematomas (EDH) form when bleeding

occurs between the calvarium and the dura mater and
the normally adherent dura is stripped from the bone.
The ultimate size of the hematoma depends on the
size and nature of the vessels lacerated and how
tightly the dura is adherent to the inner table of the
skull.
EDH occurs in about 2% of all types of head injury

(Lindenberg, 1971) and up to 15% of fatal head injuries
(Freytag, 1963). The clots are most frequently found in
the temporo-parietal regions (73%) where the middle
meningeal arteries and veins have been damaged,
usually by a fracture involving the squamous temporal bone (Jamieson and Yelland, 1968).
Some 11% of clots occur in the anterior cranial fossa
(anterior meningeal artery), 9% in the parasagittal
regions (sagittal sinus) and 7% in the posterior fossa
(occipital meningeal artery, transverse and sigmoid
sinuses) (Jamieson and Yelland, 1968). Bruising of the
overlying scalp is usually a reliable guide to the site of
the hematoma. With increasing age the meningeal
vessels become embedded in bone and are at a greater
risk of being damaged with bone trauma. Extradural
hematomas are uncommon in infants under the age of
2 years and constitute about 10% of pediatric EDH
(Molloy et al., 1990). The clotted blood in an epidural
hematoma seldom liquefies and in chronic variants
(28% of surgical cases) (Iwakuma and Brunngraber,
1973) progressive organization occurs, with the formation of a thin layer of granulation tissue on the surface
of the dura adjacent to the blood clot, which after a
month or more will develop into a well-formed
neomembrane of vascular fibrous tissue encapsulating
the loculated hematoma. This membrane may occasionally become ossified.
Often, the initial injury may be apparently trivial
and the patient may experience a lucid interval (only
present in about 20% EDH) but in about one-third of
cases there are also other significant brain injuries,
such as acute subdural hematomas, contusions and
lacerations and then the patient may experience no
lucid interval and be unconscious from the time of
injury (Jamieson and Yelland, 1968). The systematic
use of CT scanning has resulted in the increased
recognition of EDH from about 4% pre-CT to 9% postCT (Guillermain, 1986) and more than half of the cases
have an arterial origin of the bleeding (middle
meningeal or anterior or posterior branches). In 40%
the origin of bleeding is either venous or osseous and
in about 10% of cases a diffuse bleeding of the dura is
noted (Guillermain, 1986). Associated fractures have
been reported with variable frequency, but Freytag
(1963) noted fractures in 205/211 (97%) postmortem
cases of EDH and Zimmerman and Bilaniuk (1982)
found fractures in 91% of cases with CT scanning. A
fracture is absent in about 28% of children with EDH
(Guillermain, 1986) because of the greater elasticity of
the childs skull. Small to moderate-sized EDHs have
been shown on serial CT scanning to increase in size
by up to 50% until about 1014 days after injury and
then progressively reabsorb to complete resolution
46 weeks after the injury (Bullock and Teasdale,
1990).
ACUTE SUBDURAL HEMATOMA
3.14
Subdural hematomas have been classified on the basis

of etiology (traumatic or non-traumatic), chronology
(the time interval between onset and clinical recognition) or operative findings (whether the hematoma is
clotted, fluid or a mixture of the two; Jones, Blumbergs
and North, 1986). The hematoma is classified as acute
when the blood is clotted (up to several days),
subacute when there is a mixture of clotted and fluid
blood (several days to 3 weeks) and chronic when the
hematoma is fluid (after 3 weeks).
There are two main types of traumatic acute
subdural hematoma (ASDH).
In traumatic ASDH related to contusions and
lacerations the subdural hematoma forms adjacent to
damaged brain, often in association with severe
diffuse primary brain damage, and these patients are
unconscious from the time of injury. Often the ASDH
is continuous through contused, lacerated brain tissue
Figure 3.25 Burst left temporal lobe with subdural hematoma, cerebral contusion and intracerebral hemorrhage.
Figure 3.26
Burst frontal lobes.
59
with an intracerebral hemorrhage. This complex of

subdural hematoma, cerebral contusion/laceration
and adjacent intracerebral hematoma is termed a
burst lobe; the temporal or frontal lobes are most
frequently involved (Figures 3.25, 3.26). Patients with
burst lobes often show delayed neurological deterioration around the third or fourth day after injury due
to swelling of the damaged brain.
The second type of ASDH is related to rupture of
bridging veins (those portions of the superficial
cerebral veins that cross the subdural space to reach
the venous sinuses) and occasionally rupture of
superficial cortical arteries or vessels within vascular
stalks bridging the subdural space (Vance, 1950).
In some cases of ruptured bridging vein ASDH
there is little or no associated brain damage and these
patients may experience a brief lucid interval before
undergoing rapid neurological deterioration similar to
that seen in typical cases of EDH. Unfortunately, the
neurological deterioration is often so rapid that these
patients fare no better than those in whom the
hematoma is merely an extension of severe primary
brain damage. Bridging vein ASDHs were found in
13% of the cases in the Glasgow series (Adams, 1991).
It has been shown that bridging veins are most
susceptible to angular acceleration forces and that 73%
of traumatic ASDH occur as the result of falls and
assaults where short-duration high-strain-rate loading
typically occurs (Gennarelli and Thibault, 1982). Only
11% of ASDHs occurred in the occupants of vehicle
crashes, where the angular acceleration is often of
longer duration and less likely to rupture the bridging
vessels (Gennarelli and Thibault, 1982).
The reported mortality rate of traumatic ASDH
varies from 3090%, with the lower mortality rates
occurring in patients who are operated on within four
hours of injury (Seelig et al., 1981).
60
PATHOLOGY
This poor outcome has been correlated with neuropathological studies showing ischemic brain damage in the hemisphere underlying the hematoma
(Graham, Adams and Doyle, 1978). An important
factor leading to this ischemic damage is raised ICP
producing impaired cerebral perfusion and a recent
study has documented that the removal of an ASDH
results in the immediate reversal of global ischemia
(Schroder, Muizelaar and Kuta, 1994). However, mass
effect and hemisphere compression are not the only
factors of importance, as hemisphere swelling beneath
an ASDH often occurs even when the hematoma is
thin (Bullock and Teasdale, 1990).
Recent studies in a rat model of ASDH, with ischemic
damage in the underlying cerebral hemisphere, have
shown massive release of excitatory glutamate and
aspartate neurotransmitters suggesting that the ischemic damage may be related to an excitotoxic mechanism (Bullock, Butcher and McCullough, 1990; Miller et
al., 1990; Bullock et al., 1991). Excessive activation of
excitatory neurotransmitter receptors, particularly the
glutamatergic N-methyl-D-aspartate receptor, can
cause neuronal damage indistinguishable from ischemic necrosis (Meldrum, 1990).
3.15
A subdural hematoma is chronic when it is discovered 3 weeks or more after the initiating injury.
The incidence of chronic subdural hematoma
(CSDH) is 12 per 100 000 people per year (Fogelholm and Waltimo, 1975). The majority of patients
are elderly or chronic alcoholics and cerebral atrophy
seems to be an important predisposition. The lesions
are bilateral in about 1520% of cases. The head
injury is often mild and in up to one half of cases is
denied altogether. The exact cause of the hemorrhage
into the subdural space is usually unknown,
although it is often attributed to rupture of a bridging vein. An atrophic brain permits the development
of a subdural hematoma without the development of
intracranial hypertension, although paradoxically,
brain distortion and atrophy is often so severe that
even when the hematoma is evacuated the brain
remains depressed beneath the dura (Figure 3.27).
Within 1 week after injury the hematoma is covered
by an outer fibroblastic neomembrane beneath the
dura, and by 3 weeks an inner membrane forms
between the hematoma and the arachnoid surface of
the brain, with complete encapsulation of the hematoma. The contents gradually liquefy during this
period, changing from solid clotted blood to an
orange-colored protein-rich fluid. The histological
aging of subdural hematoma is based on estimation of
the thickness of the outer and inner fibroblastic
neomembranes, the degree of lysis of erythrocytes and
Figure 3.27 Persistent distortion and compression deformity of the brain despite successful evacuation of left chronic
subdural hematoma.
the sequence of changes in the vasculature, from the

early appearance of giant capillaries to the formation
and subsequent hyalinization of the vascular sinusoids (Lindenberg, 1971; McCormick, 1985). The accuracy in assessing this process depends on having
adequate material available for examination. Small
biopsies of the hematoma membrane often provide
little useful information. The symptoms and signs of
CSDH are variable and non-specific.
3.16
Other subdural fluid collections
Subdural collections of serous fluid comprise a heterogeneous group of conditions which have been variously called hygromas or hydromas. The collections of
fluid may be colorless, blood-stained, xanthochromic
or bright yellow in color and under high or low
pressure. Some restrict the term subdural hygroma to
a collection of clear CSF in the subdural space
secondary to a presumed tear of the arachnoid.
Arachnoid tears have occasionally been found at
surgery (DaCosta and Adson, 1941).
Post-traumatic low-density subdural fluid collections are reported to occur in about 6% of headinjured patients (French et al., 1978; Ohno et al., 1987).
About 5% of patients with persistent subdural fluid
collections develop a chronic subdural hematoma
(Ohno et al., 1987).
3.17
Bleeding into the subarachnoid space is the most

common abnormality seen in head injury, although in
most cases it is minor and of little clinical significance. However, significant subarachnoid hemorrhage may occur and is usually associated with
BRAIN-STEM LESIONS
cortical contusions and lacerations. The accumulation of blood may become so massive that it acts as a
local space-occupying mass lesion.
Occasionally, massive basal subarachnoid hemorrhage without any evidence of other traumatic brain
damage occurs after minor head injuries, usually in
alcohol-intoxicated patients (Simonsen, 1967). In these
patients the source of hemorrhage cannot usually be
demonstrated and it has been speculated that the
bleeding has occurred from the presumed traumatic
rupture of a normal non-diseased artery at the base of
the brain (Simonsen, 1967). Others have described
total and partial tears in carotid, basilar and other
arteries at the base of the brain as the basis of
traumatic subarachnoid hemorrhage (Krauland, 1955;
Lindenberg, 1971). Recent diffuse subarachnoid
hemorrhage may interfere with the normal circulation
of the CSF, giving rise to increased intracranial
pressure and secondary brain damage from cerebral
herniation. Fibrous scarring of the subarachnoid space
secondary to reabsorption of subarachnoid blood has
been linked to the late complication of normal
pressure hydrocephalus.
3.18
The presence of numerous small hemorrhages scattered throughout the cerebral hemispheres, especially
the white matter of the frontal and temporal lobes
and the brain stem, is a very common finding in
patients who die within minutes of a closed head
injury (Tomlinson, 1970; Adams, 1990, 1992).
Microscopic examination reveals many more
hemorrhages than can be seen macroscopically and
they consist of periarterial, perivenous and pericapillary hemorrhage and small extravasations of red cells
into the neuropil. Occasional patients survive for short
61
periods of time and the vascular damage becomes

increasingly obvious (Figure 3.28). Experimental studies involving subhuman primates subjected to nondisruptive deformation from non-impact lateral acceleration have shown widespread structural changes in
the endothelium of the brain microvasculature (Maxwell et al., 1992b). The small vessels in the corpus
callosum were most vulnerable (Maxwell et al., 1988b)
and the endothelial cells showed increased pit/vesicle
activity, microvillous and crater formation which
varied in severity in different blood vessels (Maxwell
et al., 1992b).
The interendothelial tight junctions remained intact
and it was speculated that perivascular hemorrhage
occurred via diapedesis of red cells through endothelial cells by a similar mechanism to that postulated
to occur in human cerebral contusion injury (Bullock
et al., 1991). Abnormal electronlucencies of the perivascular astrocytic foot processes were also described
(Maxwell et al., 1988b, 1992b). These findings support
the hypothesis that the endothelial cell membrane
itself is directly altered by traumatic injury (Maxwell
et al., 1988b). There is evidence for brief transient
opening of the BBB after traumatic injury to the brain
(Rinder and Olsson,1968) with increased endothelial
pit/vesicle activity when the BBB is disrupted, and the
formation of endothelial balloon and crater-like
lesions when the BBB is intact (Povlishock et al., 1980;
Povlishock and Kontos, 1982). There is evidence that
the endothelial membrane damage may be secondary
to the lipid peroxidation produced by increased free
radical generation (Kontos and Povlishock, 1986).
3.19
Brain-stem lesions
Brain-stem lesions may be primary, due to mechanical

forces damaging the individual cellular components
Figure 3.28 Diffuse vascular injury with extensive striate and petechial hemorrhage of the gray and white matter of both
cerebral hemispheres.
62
PATHOLOGY
of the brain stem at the moment of injury and

initiating a dynamic series of changes which evolve
with time, or secondary, due to brain displacements
produced by raised intracranial pressure. Primary
brain-stem lesions usually occur in an undistorted
uncompressed brain stem (Crompton, 1971) and may
be focal or diffuse. Focal primary brain stem lesions
include hemorrhages, contusions, lacerations and disruptions at the pontomedullary, mesencephalicpontine and medullocervical junctions.
(a)
Primary traumatic brain-stem hemorrhages
Hemorrhage is one of the earliest recognizable signs of

injury (Hardman, 1979) and its presence in the brain
stem (macroscopic or microscopic) may be the only
evidence of a fatal brain injury (Freytag, 1963; Tomlinson, 1970). Numerous small hemorrhages in the brain
stem, usually in association with similar lesions
scattered throughout the cerebral hemispheres, are
often seen in patients who die within minutes of a
closed head injury (Tomlinson, 1970; Adams, 1990,

1992).
These are believed to be vascular markers of a type of
diffuse brain damage incompatible with life (Tomlinson, 1970). The hemorrhages may be midline or lateral
and are often prominent around the third or fourth
ventricles and the aqueduct (Tomlinson, 1970). The
bleeding may be periarterial, perivenous, pericapillary
or within the neuropil and microscopy reveals many
more hemorrhages than can be seen with the naked eye
(Tomlinson, 1970). Perivascular extravasation of red
cells may also represent a postmortem artifact and
microhemorrhages in the neuropil require careful
interpretation as they also occur in non-traumatic cases
of sudden death (Voigt, 1981). With the advent of CT
scanning primary brain stem hemorrhages have been
diagnosed with increased frequency during life
(Cooper et al., 1979; Zuccarello et al., 1983).
(b)
Contusions and lacerations of the brain stem
Brain-stem contusions and lacerations are often

directly related to adjacent basal skull fractures (coup
lesions; Simpson et al., 1989) and some may be
produced by a contrecoup mechanism as in falls on
the back of the head producing midbrain contusions
(Lindenberg, 1971).
Victims of skull-base fractures due to torsion of the
skull vault as the result of oblique lateral impacts
may show lacerations of the brain stem (Voigt and
Skold, 1974; Figure 3.29).
(c)
Pontomedullary disruption
A study of 36 cases of primary brain stem injury (3.6%

incidence of road accident fatalities) revealed that
none survived long enough to reach hospital and that
complete or partial pontomedullary tears (Figure 3.30)
Figure 3.29
fracture.
Laceration of pons secondary to skull base
Figure 3.30
Partial pontomedullary disruption associated with laceration of ventral pons.
BRAIN DAMAGE SECONDARY TO RAISED INTRACRANIAL PRESSURE
were present in two thirds of the cases (Simpson et al.,

1989). The eight cases of isolated pontomedullary tears
were due to violent cervical hyperextension resulting
in cervical fracture dislocations and axial traction and
failure at the pontomedullary junction (Simpson et al.,
1989). Although pontomedullary tears are usually
fatal, occasional cases with survival up to 26 days have
been recorded (Pilz, Strohecker and Grobovschek,
1982; Blumbergs et al., 1991).
(d)
Diffuse axonal injury and the brain stem
Diffuse axonal injury of the brain stem is diagnosed

microscopically by finding axonal swellings (axonal
retraction balls) as part of a widespread involvement
including the cerebral hemispheres. In grade 3 DAI
macroscopic hemorrhagic marker lesions are also
present in the dorsolateral quadrants of the rostral
brain stem.
(e)
Secondary brain-stem lesions
Secondary brain-stem lesions due to raised intracranial pressure often complicate severe head injuries.
In raised intracranial pressure from supratentorial
space-occupying lesions (e.g. intracerebral hematoma)
the medial temporal lobe on the side of the lesion is
squashed against the midbrain and squeezed through
the tentorial hiatus. The midbrain is compressed from
side to side and elongated anteroposteriorly, and the
perforating branches of the firmly anchored circle of
Willis are progressively stretched and torn as the brain
stem is pushed caudally. This is the generally accepted
mechanism of production of the secondary hemorrhages (Figure 3.31) and foci of ischemic and hemor-
63
rhagic necrosis seen in the brain stem in raised

intracranial pressure (Blackwood, 1963; Johnson and
Yates, 1956). Hemorrhagic infarction of the ipsilateral
calcarine cortex may occur due to compression of the
posterior cerebral artery between the edge of the
tentorium and the herniated parahippocampal gyrus
of the medial temporal lobe. Cerebellar coning (tonsillar herniation or foraminal impaction) occurs when
the downward shift of the posterior fossa contents
impacts the cerebellar tonsils against the bony margins
of the foramen magnum, leading to hemorrhagic
necrosis of the cerebral tonsils, compression of the
medulla and death from vasomotor and respiratory
center paralysis. It has been speculated that these
brain shifts represent terminal or post-mortem events
(Ropper, 1989; Fisher, 1984).
A recent MR study radiologically confirmed the
pathological findings associated with transtentorial
herniation, i.e. compression of the brain stem, anteriorposterior elongation of the midbrain, buckling of
the tectum, downward stretching of the basilar artery
and its branches and tonsillar displacement into the
foramen magnum (Reich et al., 1993).
In an undistorted brain stem, lesions localized to
one side of the tegmentum, the periaqueductal gray
matter or superior cerebellar peduncles are suggestive
of primary injury. Centrally placed lesions in the
upper brain stem are usually secondary. It may be
impossible to distinguish secondary from primary
brain-stem lesions and both types of damage may
occur together. Clinicians often erroneously use the
term primary brain-stem injury for patients who are
unconscious from the moment of injury and who do
not have a recognizable expanding intracranial lesion.
Such unconsciousness is far more commonly due to
diffuse axonal injury (DAI) or hypoxic/ischemic
injury.
Clinicopathological studies have shown that the
brain stem is not usually damaged in isolation but as
part of the syndrome of DAI in the cases that have
survived to reach a hospital (Mitchell and Adams,
1973).
3.20 Brain damage secondary to raised

Figure 3.31 Disruption of rostral pons by secondary

hemorrhages due to raised intracranial pressure.
Raised intracranial pressure was present in 75% of a

series of fatal head injuries (Graham et al., 1987).
Distortion and herniation of the brain at postmortem
does not necessarily signify an elevated ICP during
life as these changes may occur during the period of
spatial compensation prior to significant ICP
change.
Severe distortion and herniation of brain tissue may
occur in very slowly evolving space-occupying lesions
without significantly increased ICP, and in situations
64
PATHOLOGY
of a very rapid elevation of ICP, as in diffuse cerebral

swelling, there may be little distortion or herniation of
the brain (Miller and Adams, 1972).
In transtentorial herniation the medial parahippocampal gyri show wedge-shaped areas of hemorrhagic
necrosis visible with the naked eye (Figure 3.32) or
non-hemorrhagic necrosis only visible on microscopy
(Lindenberg, 1955; Adams and Graham, 1976). These
changes were invariably present when the ICP was
greater than 5.3 kPa (40 mmHg) during life, in the
majority of patients with an ICP between 2.7 and
5.3 kPa (2040 mmHg) and in no patient with an ICP
less than 2.7 kPa (20 mmHg; Adams and Graham,
1976). Scarring of these areas of pressure necrosis
serve as marker lesions of a previous episode of high
ICP with transtentorial herniation. Only 9% of
patients with severe head injury and clinical signs of
transtentorial herniation make a good functional
recovery (Andrews and Pitts, 1991).
3.21
condition was life-threatening, recovers completely. A

large multicenter study (Gennarelli et al., 1982b)
studied the outcome of patients with severe head injury
defined as a CGS of 8 or less of at least 6 hours duration.
The patients were divided into 14 separate lesion
categories (focal and diffuse) and the outcome for each
of these categories was assessed using the Glasgow
Outcome Scale. There was an overall mortality rate of
41% but this ranged from 974% among the different
lesion categories. A total of 26% had good recovery but
among the different lesion groups the range was
668%. ASDH with GCS scores of 35 showed 74%
mortality with 8% good recovery, whereas diffuse
injury coma of 624 hours with GCS scores of 68 had
9% mortality and 68% incidence of good recovery.
This study highlights the marked heterogeneity of
patients with severe head injury and demonstrates
that patients with the same GCS score have markedly
different outcomes, depending on the causative
lesion.
Long-term effects
Patients who recover from severe head injury often

have significant mental and physical handicaps producing social disability. The pathological substrate for
persisting disabilities is the aggregate of all the
different types of primary and secondary brain damage
and, because of the potentially widespread nature of
these, it is not surprising that recovery is often
incomplete. The distinction between primary brain
damage and that developing from secondary events is
helpful in explaining the apparent paradox of a patient
whose initial injury was relatively mild yet ends up
with severe brain damage, while another, whose early
3.22
Post-traumatic vegetative state
The incidence of the vegetative state, characterized by

lack of response to external stimuli and preservation
of sleepwake cycles and vegetative cardiorespiratory
functions (Jennett and Plum, 1992; ANA Committee of
Ethical Affairs, 1993), is about 1014% following
severe closed head injury (Gennarelli et al., 1982b;
Braakman, Jennett and Minderhoud, 1988; Levin et al.,
1991). This group is distinguished from severely
disabled survivors of severe head injury who retain
the capacity to respond by word or gesture to the
external environment in a meaningful manner. Fol-
Figure 3.32 Medial parahippocampal gyri showing wedge-shaped areas of hemorrhagic necrosis consistent with raised
intracranial pressure.
BRAIN INJURIES DUE TO BOXING
low-up studies have shown that 41% of vegetative

patients become conscious by 6 months, 52% by 1 year
and 58% by 3 years (Levin et al., 1991).
The more recent neuropathological studies have
stressed diffuse axonal injury and neocortical
hypoxicischemic damage, with or without secondary
brain stem damage, as the structural substrate of
severe disability and the vegetative state (Graham et
al., 1983; McLellan et al., 1986; Adams, 1992).
The earlier studies emphasized the importance of
secondary traumatic lesions, especially of the upper
brain stem (Crompton, Teare and Bowen, 1966;
Jellinger and Seitelberger, 1969, 1970; Jellinger, 1977;
Peters and Rothemund, 1977).
3.23
Post-traumatic epilepsy
Traumatic epilepsy occurs in about 5% of head-injured

patients admitted to hospital (Jennett, 1975). Patients
who have a fit within the first week of injury and
patients with compound depressed skull fractures and
intracranial hemorrhage have an increased risk of later
developing epilepsy (Jennett, 1975). About half the
patients who subsequently develop epilepsy have
their first fit within the first year after injury but in
25% it is delayed for at least 4 years. Once a late fit
occurs epilepsy is likely to persist in about 80% of
cases, although the frequency of seizures may not be
great (Jennett, 1975).
3.24
Head injury and alzheimers disease
Recent epidemiological and neuropathological studies

have supported the old concept of an association
between head injury and the subsequent development
of Alzheimers disease.
Beta A4 deposits, varying from isolated foci related
to areas of cortical traumatic damage to extensive
deposits throughout the cortex, were found in about
30% of fatal head injuries (Roberts et al., 1991, 1994).
Confounding age-related beta A4 deposition was
considered unlikely and in some cases the deposition
of beta A4 protein appeared to occur within days of
the injury (Roberts et al., 1991, 1994). It has been
hypothesized that traumatic brain injury results in the
local overexpression and mismetabolism of neuronal
APP and resultant deposition of beta A4 at synaptic
sites with disruption of synaptic transmission (Gentleman, Graham and Roberts, 1993).
3.25
Brain injuries due to boxing
3.25.1
ACUTE BRAIN DAMAGE
Blows to the head by a fist may produce violent

acceleration of the brain with peak accelerations of up
65
to 100G having been recorded (Sances et al., 1981).

This is in the same range of acceleration that the
unrestrained head may undergo in striking the dashboard in low speed car accidents (Gennarelli and
Thibault, 1982).
Rotational (angular) acceleration is produced by
most blows to the head, as in most punches the force
is not directly transmitted through the center of
gravity of the head. This type of sudden acceleration
may tear the bridging veins, resulting in acute
subdural bleeding, although the risks of intracranial
hemorrhage in boxing seem to be no greater than in
other sports (Jordan, 1987). Subdural bleeds are the
most common cause of death in boxing and account
for up to 75% of all fatal acute boxing injuries
(Unterharnscheidt, 1975; Whiteson, 1981). Acceleration/deceleration forces may also produce DAI
(Lampert and Hardman, 1984) which in milder forms
may not be clinically evident. Relative motion of the
brain over the bony ridges of the skull may produce
contusions and lacerations of the inferior frontal and
temporal lobes. A blow by a gloved fist is unlikely to
fracture the skull, but fractures and rarely extradural
hemorrhage may result from violent deceleration of
the head in falls produced by knockout punches
(Unterharnscheidt, 1975). Rotational acceleration can
also produce intracerebral hemorrhage due to tearing
of small blood vessels within the brain parenchyma.
Recently, it has been suggested that protective
headgear may increase the chance of brain injury by
providing a larger target, with the likelihood of more
off center blows being landed (Guterman and Smith,
1989).
3.25.2
CHRONIC BRAIN DAMAGE
Chronic brain damage, described as the punch-drunk

syndrome (Martland, 1928), chronic progressive
traumatic encephalopathy of boxers or dementia
pugilistica (Millspaugh, 1937), has been reported in
about 6% of professional boxers (Roberts, 1969). This
disorder is most common in boxers with long careers,
who have suffered repeated knockouts, or sluggers
notable for their capacity to absorb punishment.
Clinically these patients show speech difficulties,
clumsiness, unsteadiness of gait progressing to disabling ataxia, dementia, spasticity and extrapyramidal
disturbances of Parkinsons type (Critchley, 1957).
This syndrome is believed to be the result of repeated
blows to the head producing cumulative brain damage. The brains from punch-drunk boxers show
cerebral atrophy, enlarged ventricles and large cavum
septi pellucidi with multiple fenestrations (Figure
3.33), the latter due to tearing of the walls of the
septum pellucidum (Corsellis, Bruton and FreemanBrowne, 1973; Corsellis, 1989).
66
PATHOLOGY
Figure 3.33
Large cavum septum pellucidum with rupture of one of the walls in a patient with dementia pugilistica.
Microscopic abnormalities include degeneration of

the substantia nigra, neuronal loss and scarring of the
cerebellum, especially in the region of the foramen
magnum, and numerous neurofibrillary tangles diffusely scattered throughout the cerebral cortex and
brain stem without any concomitant senile plaque
formation as defined by routine silver stains (Corsellis, Bruton and Freeman-Browne, 1973). Re-examination of some of these cases of dementia pugilistica
using immunocytochemical methods have shown
extensive A4 amyloid protein deposition in the form
of diffuse plaques (Roberts, Allsop and Bruton, 1990;
Tokuda et al., 1991).
3.26
References
Adams, J. H. (1990) Brain damage in fatal non-missile head injury in man, in

Handbook of Clinical Neurology, vol. 13(57), (ed. R. Braakman), Elsevier
Science Publishers, Amsterdam, pp. 4363.
Adams, J. H. (1992) Head injury, in Greenfields Neuropathology, 5th edn, (eds J.
H. Adams and L. M. Ducken), Edward Arnold, London, pp. 106152.
Adams, J. H. and Graham, D. I. (1976) The relationship between ventricular
fluid pressure and the neuropathology of raised intracranial pressure.
Neuropathology and Applied Neurobiology, 2, 323332.
Adams, J. H. and Graham, D. I. (1984) Diffuse brain damage in non-missile
head injury, in Recent Advances in Histopathology 12, (eds P. P. Anthony and
R. N. M. Macsween), Churchill Livingstone, Edinburgh, pp. 241257.
Adams, J. H., Graham, D. I. and Gennarelli, T. A. (1981) Acceleration induced
head injury in the monkey. Acta Neuropathologica (Supplement), 7, 2628.
Adams, J. H., Mitchell, D. E., Graham, D. I. and Doyle, D. (1977) Diffuse brain
damage of immediate impact type. Its relationship to primary brain stem
damage in head injury. Brain, 100, 489502.
Adams, J. H., Graham, D. I., Scott, G. et al. (1980a) Brain damage in nonmissile head injury. Journal of Clinical Pathology, 33, 11321145.
Adams, J. H., Scott, G., Parker, L. S. et al. (1980b) The contusion index: a
quantitative approach to cerebral contusions in head injury. Neuropathology
and Applied Neurobiology, 6, 319324.
Adams, J. H., Graham, D. I., Murray, L. S. and Scott, G. (1982) Diffuse axonal
injury due to non-missile head injury in humans: an analysis of 45 cases.
Annals of Neurology, 12, 564574.
Adams, J. H., Doyle, D., Graham, D. I. et al. (1984) Diffuse axonal injury in
head injuries caused by a fall. Lancet, 22/29 Dec., 14201421.
Adams, J. H., Doyle, D., Graham, D. I. et al. (1985) The contusion index: a
reappraisal in man and experimental non-missile head injury. Neuropathology and Applied Neurobiology, 11, 299308.
Adams, J. H., Doyle, D., Graham, D. I. et al. (1986a) Gliding contusions in nonmissile head injury in humans. Archives of Pathology and Laboratory Medicine,
110, 485488.
Adams, J. H., Doyle, D., Graham, D. I. et al. (1986b) Deep intracerebral (basal
ganglia) haematomas in fatal non-missile head injury in man. Journal of
Neurology, Neurosurgery and Psychiatry, 9, 10381043.
Adams, J. H., Doyle, D., Ford, I. et al. (1989) Diffuse axonal injury in head
injury: definition, diagnosis and grading. Histopathology, 15, 4959.
Aldrich, E. F., Eisenberg, H. M., Saydjari, C. et al. (1992). Diffuse brain swelling
in severe head-injured children. A report from the NIH Traumatic Coma
Data Bank. Journal of Neurosurgery, 76, 450454.
ANA Committee on Ethical Affairs (1993) Persistent vegetative state: report of
the Americal Neurological Committee on Ethical Affairs. Annals of
Andrews, B. T. and Pitts, L. H. (1992) Functional recovery after traumatic
transtentorial herniation. Journal of Neurosurgery, 29, 227231.
Baratham, G. and Dennyson, W. (1972) Delayed traumatic intracerebral
haemorrhage. Journal of Neurology, Neurosurgery and Psychiatry, 3, 698706.
Blackwood, W. (1963) Vascular disease of the central nervous system, in
Greenfields Neuropathology, 2nd edn, (eds W. Blackwood, W. H. McMenemy,
A. Meyer et al.), Edward Arnold, London, pp. 71137.
Blumbergs, P. C., Jones, N. R. and North, J. B. (1989) Diffuse axonal injury in
head trauma. Journal of Neurology, Neurosurgery and Psychiatry, 52,
838841.
Blumbergs, P. C., Oatey, P. E., Sandhu, A. et al. (1991) Pontomedullary tear in
a speedboat accident. Report of a case with MRI diagnosis. Zentralblatt fur
die Neurochirurgie, 52, 8993.

Blumbergs, P. C., Scott, G., Manavis, J. et al. (1994) Staining of amyloid
precursor protein to study axonal damage in mild head injury. Lancet, 344,
10551056.
Blumbergs, P. C., Scott, G., Manavis, J. et al. (1995) Topography of axonal
injury as defined by amyloid precursor protein and the sector scoring
method in mild and severe closed head injury. Journal of Neurotrauma, 12,
565572.
Bouma, G. T. and Muizelaar, J. P. (1992) Cerebral blood flow, cerebral blood
volume, and cerebrovascular reactivity after severe head injury. Journal of
Neurotrauma, 9(Suppl.), S333S348.
Bouma, G. J., Muizelaar, J. P., Choi, S. C. et al. (1991) Cerebral circulation and
metabolism after severe traumatic brain injury: the elusive role of
ischaemia. Journal of Neurosurgery, 75, 685693.
Bouma, G. J., Muizelaar, J. P., Stringer, W. A. et al. (1992) Ultra-early evaluation
of regional cerebral blood flow in severely head-injured patients using
xenon-enhanced computerised tomography. Journal of Neurosurgery, 77,
360368.
Braakman, R., Jennett, W. B. and Minderhoud, J. M. (1988) Prognosis of the
post tramautic vegetative state. Acta Neurochirurgica (Vienna), 95, 4952.
Brown, F. D., Mullan, S. and Duda, E. E. (1978) Delayed traumatic
intracerebral haematomas. Report of three cases. Journal of Neurosurgery, 48,
10191022.
Bruce, D. A., Alavi, A., Bilaniuk, L. et al. (1981) Diffuse cerebral swelling
following head injuries in children: the syndrome of malignant brain
oedema. Journal of Neurosurgery, 54, 170178.
REFERENCES
Bullock, R., Butcher, S. and McCullough, J. (1990) Changes in extracellular
glutamate concentration after acute subdural haematoma in the rat
evidence for an excitotoxic mechanism. Acta Neurochirurgica (Supplement),
57, 274276.
Bullock, R. and Teasdale, G. (1990) Surgical management of traumatic
intracranial haematomas, in Handbook of Clinical Neurology, vol. 13(57), Head
Injury, (eds P. J. Vinken, G. W. Bruyn and R. Braakman), Elsevier,
Amsterdam, pp. 249298.
Bullock, R., Butcher, S. P., Chen, M. H. et al. (1991). Correlation of the
extracellular glutamate concentration with extent of blood flow reduction
after subdural haematoma in rat. Journal of Neurosurgery, 74, 794802.
Bullock, R., Maxwell, W. L., Graham, D. I. et al. (1991) Glial swelling following
human cerebral contusion: an ultrastructural study. Journal of Neurology,
Cheng, C. L. Y. and Povlishock, J. T. (1988) The effect of traumatic brain injury
on the visual system: A morphologic characterisation of reactive axonal
change. Journal of Neurotrauma, 5, 4760.
Colquhoun, I. R. and Rawlinson, J. (1989) The significance of haematomas of
the basal ganglia in closed head injury. Clinical Radiology, 40, 619621.
Cooper, P. R., Maravilla, K., Kirkpatrick, J. et al. (1979) Traumatically induced
brain stem haemorrhage and the computerised tomographic scan: clinical,
pathological and experimental observations. Neurosurgery, 4, 115124.
Cordobes, F., de la Fuente, M., Lobato, R. et al. (1983) Intraventricular
haemorrhage in severe head injury. Journal of Neurosurgery, 58, 217222.
Corsellis, J. A., Bruton, C. J. and Freeman-Browne, D. (1973) The aftermath of
boxing. Psychological Medicine, 3, 270303.
Corsellis, J. A. N. (1989) Boxing and the brain. British Medical Journal, 298,
105109.
Courville, C. G. (1942) Coup-contrecoup mechanism of craniocerebral
injuries: some observations. Archives of Surgery, 4, 1942.
Critchley, M. (1957) Medical aspects of boxing, particularly from a neurological stand point. British Medical Journal, i, 357362.
Crompton, M. R. (1971) Brain stem lesions due to closed head injury. Lancet,
i, 669673.
Crompton, M. R., Teare, R. D. and Bowen, D. A. (1966) Prolonged coma after
head injury. Lancet, ii, 938940.
Crooks, D. A., Scholtz, C. L., Vowles, G. et al. (1992) Axonal injury in closed
head injury by assault: a quantitative study. Medical Science and the Law, 32,
109117.
Crooks, D. A. (1991) The pathological concept of diffuse axonal injury: its
pathogenesis and the assessment of severity. Journal of Pathology, 165, 510.
DaCosta, D. G. and Adson, A. W. (1941) Subdural hygroma. Archives of
Surgery, 43, 559567.
Dawson, S. L., Hirsch, C. S. and Lucas, F. V. et al. (1980) The contrecoup
phenomenon. Human Pathology, 11, 155166.
Denny-Brown, D. and Russell, W. R. (1941) Experimental cerebral concussion.
Brain, 64, 93164.
DeVet, A. C. (1976) Traumatic intracerebral haematoma, in Handbook of Clinical
Neurology, vol. 24, (eds P. J. Vinken and G. W. Bruyn), Elsevier Science
Publishers, pp. 351368.
Dharker, S. R., Mittal, R. S. and Bharagava, N. (1993) Ischaemic lesions in basal
ganglia in children after minor head injury. Neurosurgery, 33, 863865.
Diaz, F., York, D., Larson, D. and Rockswold, G. (1979) Early diagnosis of
delayed post-traumatic intracerebral haematomas. Journal of Neurosurgery,
50, 217233.
Ellison, M. D., Erb, D. E., Kontos, H. A. and Povlishock, J. T. (1978) Recovery
of impaired endothelium-dependent relaxation after fluid-percussion brain
injury in cats. Stroke, 20, 911917.
Enevoldsen, E. M. and Jensen, F. T. (1978) Autoregulation and CO2 responses
of cerebral blood flow in patients with acute severe head injury. Journal of
Erb, D. E. and Povlishock, J. T. (1988) Axonal damage in severe traumatic brain
injury: An experimental study in the cat. Acta Neuropathologica, 76, 347358.
Ewing, R., McCarthy, D., Gronwall, D. and Rightson, P. (1980) Persisting
effects of minor head injury observable during hypoxic states. Journal of
Clinical Neuropsychology, 2, 147155.
Faden, A. I., Demeduik, P., Panter, S. S. and Vink, R. (1989) The role of
excitatory aminoacids and NMDA receptors in traumatic brain injury.
Science, 244, 798800.
Finley, D. and Varshavsky, A. (1985) The ubiquitin system: functions and
mechanisms. Trends in the Biochemical Sciences, 11, 343347.
Fisher, C. M. (1966) Concussion amnesia. Neurology, 16, 826830.
Fisher, C. M (1984) Acute brain herniation: a revised concept. Seminars in
Fogelholm, R. and Waltimo, O. (1975) Epidemiology of chronic subdural
haematoma. Acta Neurochirurgica (Vienna), 32, 247250.
French, B. N., Cobb, C. A., Corkill, G. and Youmans, J. R. (1978) Delayed
evolution of post-traumatic subdural hygroma. Surgical Neurology, 9,
145148.
Freytag, E. (1963) Autopsy findings in head injuries from blunt force.
Statistical evaluation of 1,367 cases. Archives of Pathology, 75, 402413.
Fukamachi, A., Nagaseki, Y., Kohno, K. et al. (1985) The incidence and
developmental process of delayed traumatic intracerebral haematomas.
Acta Neurochirurgica, 74, 3539.
67
Fujitsu, K., Kubawara, T., Muramoto, M et al. (1988) Traumatic intraventricular haemorrhage. Report of twenty-six cases and consideration of
the pathogenic mechanism. Neurosurgery, 23, 423430.
Gennarelli, T. A., Adams, J. H. and Graham, D. I. (1986) Diffuse axonal injury.
A new conceptual approach to an old problem, in Mechanisms of Secondary
Brain Damage, (eds A. Baethmann, K. G. Go and A. Unterberg), Plenum
Press, New York, pp. 1528.
Gennarelli, T. A. and Thibault, L. E. (1982) Biomechanics of acute subdural
haematoma. Journal of Trauma, 22, 680868.
Gennarelli, T. A. and Thibault, L. E. (1985) Biomechanics of head injury, in
Neurosurgery, (eds R. H. Wilkins and S. S. Rengachary), McGraw-Hill, New
York, vol. 2, pp. 15311535.
Gennarelli, T. A., Abel, J. M., Adams, H. and Grahams, D. (1979) Differential
tolerance of frontal and temporal lobes to contusion induced by angular
acceleration, in Proceedings of the 23rd Stapp Car Crash Conference, Society of
Gennarelli, T. A., Thibault, L. E., Adams, J. H. et al. (1982a) Diffuse axonal
injury and traumatic coma in the primate. Annals of Neurology, 12,
564574.
Gennarelli, T. A., Spelman, G. W., Langfitt, T. W. et al. (1982b) Influence of the
type of intracranial lesion on outcome from severe head injury. Journal of
Gennarelli, T. A., Thibault, L. E., Tipperman, R. et al. (1989) Axonal injury in
the optic nerve: a model simulating diffuse axonal injury in the brain.
Gentleman, S. M., Graham, D. I. and Roberts, G. W. (1993) Molecular
pathology of head trauma: altered APP. metabolism and the aetiology of
Alzheimers disease. Progress in Brain Research, 96, 237246.
Gentleman, S. M., Nash, M. J., Sweeting, C. J. et al. (1993) -Amyloid precursor
protein (APP) as a marker for axonal injury after head injury. Neuroscience
Letters, 160, 139144.
Grady, M. S., McLaughlin, M. R., Christman, C. W. et al. (1993) The use of
antibodies targeted against the neurofilament subunits for the detection of
diffuse axonal injury in humans. Journal of Neuropathology and Experimental
Graham, D. I., Adams, J. H. and Doyle, E. (1978) Ischaemic brain damage in
fatal non-missile head injuries. Journal of the Neurological Sciences, 39,
213234.
Graham, D. I., McLellan, D., Adams, J. H. et al. (1983) The neuropathology of
severe disability after head injury. Acta Neurochirurgica (Supplement), 32,
6567.
Graham, D. I., Lawrence, A. E., Adams, J. H. et al. (1987) Brain damage in nonmissile head injury secondary to high intracranial pressure. Neuropathology
and Applied Neurobiology, 13, 209217.
Graham, D. I, Lawrence, A. E., Adams, J. H. et al. (1988) Brain damage in fatal
non-missile head injury without high intracranial pressure. Journal of
Clinical Pathology, 41, 3437.
Graham, D. I., Ford, I., Adams, J. H. et al. (1989) Ischaemic brain damage is still
common in fatal non-missile head injury. Journal of Neurology, Neurosurgery
and Psychiatry, 52, 346350.
Graham, D. I., Clark, J. C., Adams, J. H. and Gennarelli, T. A. (1992) Diffuse
axonal injury caused by assault. Journal of Clinical Pathology, 45, 840841.
Grcevic, N. (1983) Traumatic tears to the tela choriodia: a hitherto unrecognised cause of post-traumatic hydrocephalus. Acta Neurochirurgica (Vienna)
(Supplement), 32, 7985.
Gronwall, D. and Wrightson, P. (1979) Delayed recovery of intellectual
function after minor head injury. Lancet, ii, 604609.
Gronwall, D. and Wrightson, P. (1975) Cumulative effect on concussion.
Lancet, ii, 995997.
Gudeman, S., Kishore, P., Miller, J. et al. (1979) The genesis and significance of
delayed traumatic intracerebral haematoma. Neurosurgery, 5, 309313.
Guillermain, P. (1986) Traumatic extradural haematomas. Advances in Neurotraumatology, 1, 150.
Gultekin, S. H. and Smith, T. W. (1994) Diffuse axonal injury in craniocerebral
trauma: a comparative histologic and immunocytochemical study. Archives
of Pathology and Laboratory Medicine, 118, 168171.
Gurdjian, E. S. and Gurdjian, E. S. (1976) Cerebral contusions: Re-evaluation
of the mechanism of their development. Journal of Trauma, 16, 3551.
Gurdjian, E. S., Brihaye, J., Christensen, J. C. et al. (1979) Glossary of
neurotraumatology. Acta Neurochirurgica (Vienna) (Supplement), 25.
Guterman, A. and Smith, R. W. (1989) Neurological sequelae of boxing. Patient
Management, April, 117141.
Hardman, J. M. (1979) The pathology of traumatic brain injuries. Advances in
Hayes, R. L., Jenkins, C. W. and Lyeth, B. G. (1992) Neurotransmittermediated mechanisms of traumatic brain injury: Acetylcholine and excitatory aminoacids. Journal of Neurotrauma, 9(Suppl. 1), S173S187.
Hirsh, L. (1979) Delayed traumatic intracerebral haematomas after surgical
decompression. Neurosurgery, 5, 653655.
Hubschmann, O. R. and Kornhauzer, D. (1983) Effects of intraparenchymal
haemorrhage on extracellular cortical potassium in experimental head
trauma. Journal of Neurosurgery, 59, 289293.
Imajo, T. and Kazee, A. M. (1992) Diffuse axonal injury by simple fall.
American Journal of Forensic Medicine and Pathology, 13, 169172.
68
PATHOLOGY
Imajo, T. and Roessman, U. (1984) Diffuse axonal injury. American Journal of

Forensic Medicine and Pathology, 5, 217222.
Iwakuma, T. and Brunngraber, C. V. (1973) Chronic extradural haematomas: a
study of 21 cases. Journal of Neurosurgery, 38, 488493.
Jamieson, K. G. and Yelland, J. D. N. (1972) Traumatic intracerebral haematoma.
Report of 63 surgically treated cases. Journal of Neurosurgery, 37, 528532.
Jamieson, K. G. and Yelland, J. D. N. (1968) Extradural haematoma: report of
167 cases. Journal of Neurosurgery, 29, 1323.
Jellinger, K. (1977) Pathology and pathogenesis of apallic syndromes
following closed head injuries, in The Apallic Syndrome, (eds G. D. Ore, F.
Gerstenbrand, C. H. Lucking et al.) Springer-Verlag, Berlin, pp. 88103.
Jellinger, K. and Seitelberger, F. (1969) Protracted post-traumatic encephalopathy, in The Late Effects of Head Injury, (eds A. E. Walker, W. F. Cavens and
M. Critchley), Charles C. Thomas, Springfield, IL, pp. 168181.
Jellinger, K. and Seitelberger, F. (1970) Protracted post-traumatic encephalopathy: pathology, pathogenesis and clinical implications. Journal of the
Neurological Sciences, 10, 5194.
Jenkins, L. W., Lyeth, B. G., Lewelt, W et al. (1988) Combined pre-trauma
scopolamine and phencyclidine attenuates post-traumatic increased sensitivity to delayed secondary ischaemia. Journal of Neurotrauma, 5, 303315.
Jennett, B. (1975) Epilepsy After Non-missile Head Injury, 2nd edn, Heinemann,
London.
Jennett, B. and Plum, F. (1992) Persistent vegetative state after brain damage.
Lancet, i, 734737.
Johnson, R. T. and Yates, P. O. (1956) Brain stem haemorrhages in expanding
supratentorial conditions. Acta Radiologica (Stockholm), 46, 250256.
Jones, N. R., Blumbergs, P. C. and North, J. B. (1986) Acute subdural
haematomas: aetiology, pathology and outcome. Australian and New Zealand
Journal of Surgery, 56, 907913.
Jordan, B. D. (1987) Neurologic aspects of boxing. Archives of Neurology, 44,
453459.
Kang, J., Lemaire, H. G., Unterbeck, A. et al. (1987) The precursor of
Alzheimers disease amyloid A4 resembles a cell surface receptor. Nature,
325, 733736.
Katayama, Y., Becker, D. P., Tamura, T. and Hovda, D. A. (1990) Massive
increase in extracellular potassium and the indiscriminate release of
glutamate following concussive brain injury. Journal of Neurosurgery, 73,
889900.
Katz, D. I., Alexander, M. P., Seliger, G. M. and Bellas, D. N. (1989) Traumatic
basal ganglia haemorrhage: clinicopathologic features and outcome. Neurology, 39, 897904.
Katz, D. I. and Alexander, M. P. (1994) Traumatic brain injury, in Handbook of
Neurorehabilitation, (eds D. C. Good and J. R. Couch), Marcel Dekker, New
York, pp. 493549.
Kaufman, H., Mouke, J., Olson, J. et al. (1980) Delayed and recurrent
intracranial haematomas related to disseminated intravascular clotting and
fibrinolysis in head injury. Neurosurgery, 7, 445449.
Kaufman, H., Hui, K., Mattson, J. et al. (1984) Clinical pathological correlations
of disseminated intravascular coagulation in patients with head injury.
Kelmowitz, R. and Annis, B. (1973) Disseminated intravascular coagulation
associated with massive brain injury. Journal of Neurosurgery, 39, 178180.
Kobrine, A. I., Timmins, E., Rajjoub, R. K. et al. (1977) Demonstration of
massive traumatic brain swelling within 20 minutes after injury. Journal of
Kontos, H. A. and Povlishock, J. T. (1986) Oxygen radicals in brain injury. CNS
Trauma, 3, 257263.
Koo, E. H., Sisoda, S. S., Archer, D. R. et al. (1990) Precursor of amyloid protein
in Alzheimers disease undergoes fast anterograde axonal transport.
Proceedings of the National Academy of Sciences of the USA, 87, 15611565.
Kotapka, M. J., Gennarelli, T. A., Graham, D. I. et al. (1991) Selective
vulnerability of hippocampal neurons in acceleration-induced experimental head injury. Journal of Neurotrauma, 8, 247258.
Kotapka, M. J., Graham, D. I., Adams, J. H. and Gennarelli, T. A. (1992)
Hippocampal pathology in fatal non-missile head injury. Acta Neuropathologica, 83, 530534.
Krauland, W. (1955) Handbuch der speziellen pathologischen Anatomie und
Histologie, Springer-Verlag, Berlin, p. 170.
Kuroda, Y., Inglis, F. M., Miller, J. D. et al. (1992) Transient glucose
hypermetabolism after acute subdural haematoma in the rat. Journal of
Lampert, P. W. and Hardman, J. M. (1984) Morphological changes in boxers.
Journal of the American Medical Association, 251, 26762678.
Lang, D. A., Teasdale, G. M., Macpherson, P. and Lawrence, A. (1994) Diffuse
brain swelling after injury: more often malignant in adults than children?
LeRoux, P. D., Haglund, M. M., Newell, D. W. et al. (1992) Intraventricular
haemorrhage in blunt head trauma: an analysis of 43 cases. Neurosurgery, 31,
678684.
Levin, H. S., Eisenberg, H. M. and Benton, A. L, (eds) (1989) Mild Head Injury,
Oxford University Press, Oxford.
Levin, H. S., Saydjari, C., Eisenberg, H. M. et al. (1991) Vegetative state after
closed head injury. A Traumatic Coma Data Bank report. Archives of
Lindenberg, R. (1971) Trauma of meninges and brain, in Pathology of the

Nervous System, vol. 2, (ed. J. Minckler), McGraw Hill, New York, pp.
17051765.
Lindenberg, R. (1955) Compression of brain arteries as a pathogenetic factor
for tissue necrosis and their areas of predilection. Journal of Neuropathology
and Experimental Neurology, 14, 223243.
Lindenberg, R. and Freytag, E. (1960) The mechanism of cerebral contusions.
Archives of Pathology, 69, 440469.
Lobato, R. D., Sarabia, R., Cordobes, F. et al. (1988) Post-traumatic cerebral
hemisphere swelling. Analysis of 55 cases studied with computerised
tomography. Journal of Neurosurgery, 68, 417423.
Loberg, E. M. and Torvik, A. (1989) Brain contusions: the time sequence of the
histological changes. Medical Science and the Law, 29, 109115.
Long, D. M. (1982) Traumatic brain oedema. Clinical Neurosurgery, 29,
174202.
Lowenhielm, P. (1975) Mathematical simulation of gliding contusions. Journal
of Biomechanics, 8, 351356.
McCormick, W. F. (1985) Pathology of closed head injury, in Neurosurgery, (eds
R. H. Wilkins and S. S. Rengachary), McGraw-Hill, New York, vol. 2, pp.
15441570.
McLaurin, R. L. and Helmer, F. (1965) The syndrome of temporal lobe
contusions. Journal of Neurosurgery, 23, 296303.
McLellan, D. R., Adams, J. H., Graham, D. I. et al. (1986) The structural basis
of the vegetative state and prolonged coma after non-missile head injury, in
Le Coma traumatique, (eds I. Papo, F. Cohadon and M. Massaroti), Liviana
Editrice, Padua, pp. 165185.
MacPherson, P. and Graham, D. I. (1978) Correlation between angiographic
findings and the ischaemia of head injury. Journal of Neurology, Neurosurgery
MacPherson, P., Teasdale, E., Dhaker, S. et al. (1986) The significance of
traumatic haematoma in the region of the basal ganglia. Journal of Neurology,
Margulies, S. S. and Thibault, L. E. (1992) A proposed tolerance criterion for
diffuse axonal injury in man. Journal of Biomechanics, 25, 917923.
Marion, D. W., Darby, J. and Yonas, H. (1991) Acute regional cerebral blood
flow changes caused by severe head injuries. Journal of Neurosurgery, 74,
407414.
Marshall, L. F., Gautille, T. and Klauber, M. R. (1991) The outcome of severe
closed head injury. Journal of Neurosurgery, 75, 528536.
Marshall, L. F., Toole, B. M. and Bowers, S. A. (1983) The National Traumatic
Coma Data Bank. Part 2: Patients who talk and deteriorate: implications for
treatment. Journal of Neurosurgery, 59, 285288.
Marshall, L. F., Marshall, S. B., Klauber, M. R. et al. (1991) A new classification
of head injury based on computerised tomography. Journal of Neurosurgery,
75, S14S20.
Martland, H. S. (1928) Punch drunk. Journal of the American Medical Association,
91, 11031107.
Maxwell, W. L., Kansagra, A. M., Graham, D. I. et al. (1988a) Freeze fracture
studies of reactive myelinated nerve fibres after diffuse axonal injury. Acta
Neuropathologica, 76, 395406.
Maxwell, W. L., Irvine, A., Adams, J. H. et al. (1988b) Response of cerebral
microvasculature to brain injury. Journal of Pathology, 155, 327335.
Maxwell, W. L., Irvine, A., Graham, D. I. et al. (1991) Focal axonal injury: the
early axonal response to stretch. Journal of Neurocytology, 20, 157164.
Maxwell, W. L., Hardy, I. G., Watt, C. et al. (1992a) Changes in the choroid
plexus, responses by intrinsic epiplexus cells and recruitment from
monocytes after experimental head acceleration injury in the non-human
primate. Acta Neuropathologica, 84, 7884.
Maxwell, W. L., Whitfield, P. C., Suzen, B. et al. (1992b) The cerebrovascular
response to experimental lateral head acceleration. Acta Neuropathologica
(Berlin), 84, 289296.
Maxwell, W. L., Watt, C., Graham, D. I. and Gennarelli, T. A. (1993)
Ultrastructural evidence of axonal shearing as a result of lateral acceleration
of the head in non-human primates. Acta Neuropathologica, 86, 136144.
Meldrum, B. (1990) Protection against ischaemic neuronal damage by drugs
acting on excitatory neurotransmission. Cerebrovascular Brain Metabolism
Review, 2, 2757.
et al. (1991) Delayed traumatic intracerebral
Mertol, T, Guner, M, Acar, U
haematoma. British Journal of Neurosurgery, 5, 491498.
Miller, J. D. (1993) Traumatic brain swelling and edema, in Head Injury, (ed. P.
R. Cooper), Williams & Wilkins, Baltimore, MD, pp. 351354.
Miller, J. D. and Adams, J. H. (1972) Physiopathology and management of
increased intracranial pressure, in Scientific Foundations of Neurology, (eds M.
Critchley, J. OLeary and B. Jennett) Heinemann Medical, London, pp.
308324.
Miller, J. D., Bullock, R. and Graham, D. I. et al. (1990) Ischaemic brain damage
in a model of acute subdural haematoma. Neurosurgery, 27, 433439.
Millspaugh, J. A. (1937) Dementia pugilistica. United States Naval Bulletin, 32,
297302.
Mitchell, D. E. and Adams, J. H. (1973) Primary focal impact damage to the
brain stem in blunt head injuries. Does it exist? Lancet, ii, 215218.
Modesti, L. M., Hodge, C. J. and Barnwell, M. (1982) Intracerebral haematoma
after evacuation of chronic extracerebral fluid collections. Neurosurgery, 10,
689693.
REFERENCES
Molloy, C. J., McCaul, K. A., McLean, A. J. et al. (1990) Extradural
haemorrhage in infancy and childhood. A review of 35 years experience in
South Australia. Childs Nervous System, 6, 383387.
Mosberg, W. H. and Lindenberg, R. (1959) Traumatic haemorrhage from the
anterior choroidal artery. Journal of Neurosurgery, 16, 209221.
Nakamura, Y., Takeda, M., Nigawa, H. et al. (1992) Amyloid beta protein
precursor deposition in rat hippocampus lesioned by ibotenic acid
injection. Neuroscience Letters, 136, 9598.
Ng, H. K., Mahaliyana, R. D. and Poon, W. S. (1994) The pathological spectrum
of the diffuse axonal injury in blunt head trauma: assessment with axon and
myelin stains. Clinical Neurology and Neurosurgery, 96, 2431.
Ninchoji, T., Vemura, K., Shimoyama, I. et al. (1984) Traumatic intracerebral
haematomas of delayed onset. Acta Neurochirurgica (Vienna), 71, 6990.
Oehmichen, M. and Raff, G. (1980) Timing of cortical contusion. Correlation
between histomorphologic alterations and post-mortem interval. Zeitschrift
fur
Rechtsmedizin, 84, 7994.
Ohagami, T., Kitamoto, T. and Tateishi, J. (1992) Alzheimers amyloid
precursor protein accumulates within axonal swellings in human brain
lesions. Neuroscience Letters, 136, 7578.
Ohno, K., Suzuki, R., Masaoka, H. et al. (1987) Chronic subdural haematoma
preceded by persistent traumatic subdural fluid collection. Journal of
Ommaya, A. K. (1979) Reintegrative action of the nervous system after
trauma, in Neural Trauma, (eds A. J. Popp et al.), Raven Press, New York, pp.
371378.
Ommaya, A. K. and Gennarelli, T. A. (1974) Cerebral concussion and
traumatic unconsciousness. Brain, 97, 633654.
Ommaya, A. K., Grubb, R. L. and Naumann, R. A. (1971) Coup and contrecoup injury: observations on the mechanics of visible brain injuries in the
rhesus monkey. Journal of Neurosurgery, 35, 503516.
Oppenheimer, D. R. (1968) Microscopic lesions in the brain following head
injury. Journal of Neurology Neurosurgery and Psychiatry, 31, 299306.
Oprescu, I. (1991) Cerebral hemispheric contusions and lacerations. Advances
in Neurotraumatology, 3, 2659.
Otsuka, N., Tomonaga, M. and Ikeda, K. (1991) Rapid appearance of betaamyloid precursor protein immunoreactivity in damaged axons and
reactive glial cells in rat brain following needle stab injury. Brain Research,
568, 335338.
Peerless, S. J. and Rewcastle, N. B. (1967) Shear injuries of the brain. Canadian
Medical Association Journal, 96, 577582.
Persson, L. and Hillered, L. (1992) Chemical monitoring of neurosurgical
intensive care patients using intracerebral microdialysis. Journal of Neurosurgery, 76, 7280.
Peters, G. and Rothemund, E. (1977) Neuropathology of the traumatic apallic
syndrome, in The Apallic Syndrome, (eds G. D. Ore, F. Gerstenbrand, C. H.
Lucking et al.) Springer-Verlag, Berlin, pp. 7887.
Pilz, P. (1983) Axonal injury in head injury. Acta Neurochirurgica (Supplement),
32, 119123.
Pilz, P., Strohecker, J. and Grobovschek, M. (1982) Survival after traumatic
pontomedullary tear. Journal of Neurology, Neurosurgery and Psychiatry, 45,
422427.
Povlishock, J. T. (1992) Traumatically induced axonal injury: pathogenesis and
pathobiological implications. Brain Pathology, 2, 112.
Povlishock, J. T. and Kontos, H. A. (1982) The pathophysiology of pial and
intraparenchymal vascular dysfunction, in Head Injury, Basic and Clinical
Aspects, (eds R. G. Grossman and P. L. Gildenberg), Raven Press, New York,
pp. 1530.
Povlishock, J. T. and Kontos, H. A. (1985) Continuing axonal and vascular
change following experimental brain trauma. Central Nervous System
Trauma, 2, 285297.
Povlishock, J. T., Kontos, H. A., Rosenblum, W. I. et al. (1980) A scanning
electron microscope analysis of the intraparenchymal brain vasculature
following experimental hypertension. Acta Neuropathologica (Berlin), 51,
203212.
Povlishock, J. T., Becker, D. P., Cheng, C. L. Y. and Vaughan, G. W. (1983)
Axonal change in minor head injury. Journal of Neuropathology and
Experimental Neurology, 42, 225242.
Pretorius, M. and Kaufman, H. (1982) Rapid onset of delayed traumatic
intracerebral haematoma and disseminated intravascular coagulation and
fibrinolysis. Acta Neurochirurgica (Vienna), 65, 103104.
Rand, C. W. and Courville, C. B. (1934) Histological change in the brain in
cases of fatal injury to the head. Archives of Neurology and Psychiatry, 31,
527555.
Reich, J. B., Sierra, J., Camp, W. et al. (1993) Magnetic resonance imaging
measurements and clinical changes accompanying transtentorial and
foramen magnum brain herniation. Annals of Neurology, 33, 159170.
Reilly, P. L., Adams, R. H., Graham, D. I. and Jennett, B. (1975) Patients with
head injury who talk and die. Lancet, ii, 375377.
Reith, K. G., Fukiwara, K., DiChiro, G. et al. (1980) Serial measurements of CT
attenuations and specific gravity in experimental cerebral oedema. Radiology, 135, 343348.
Rinder, L. and Olsson, Y. (1968) Studies on vascular permeability changes in
experimental brain concussion, part 2. Duration of altered permeability.
Acta Neuropathologica (Berlin), 11, 201209.
69
Rivano, C., Borzone, M., Carta, F. and Michelozzi, G. (1980) Traumatic

intracerebral haematomas: 72 cases surgically treated. Journal of Neurosurgical Science, 24, 7784.
Roberts, A. H. (1969) Brain Damage in Boxers. Pitman Medical, London.
Roberts, G. W., Allsop, D. and Bruton, C. (1990) The occult aftermath of
boxing. Journal of Neurology, Neurosurgery and Psychiatry, 53, 373378.
Roberts, G. W., Gentleman, S. M., Lynch, A. and Graham, D. I. (1991) A4
amyloid protein deposition in brain after head trauma. Lancet, 338,
14221423.
Roberts, G. W., Gentleman, S. M., Lynch, A. et al. (1994) Beta amyloid protein
deposition in the brain after severe head injury: implications for the
pathogenesis of Alzheimers disease. Journal of Neurology, Neurosurgery and
Psychiatry, 57, 419425.
Rockswold, G. L., Leonard, P. R. and Nagib, M. G. (1987) Analysis of
management in thirty-three closed head injury patients who talked and
deteriorated. Neurosurgery, 21, 5155.
Ropper, A. H. (1989) A preliminary MRI study of the geometry of brain
displacement and level of consciousness with acute intracranial masses.
Russell, W. R. (1971) The Traumatic Amnesias, Oxford University Press,
Oxford.
Ryan, G. A., McLean, A. J., Vilenius, A. T. S. et al. (1994) Brain injury patterns
in fatally injured pedestrians. Journal of Trauma, 36, 469476.
Sances, A., Weber, R. C., Larson, S. J. et al. (1981) Bioengineering analysis of
head and spine injuries. CRC Critical Reviews in Bioengineering, 5, 79122.
Sahuquillo-Barris, J., Maraca-Ciuro, J., Vilalta-Castan, J. et al. (1988) Acute
subdural haematoma and diffuse axonal injury after severe head trauma.
Saweda, Y., Sadomitsu, D., Sakmoto, T. et al. (1984) Lack of correlation
between delayed traumatic intracerebral haematoma and disseminated
intravascular coagulation. Journal of Neurology, Neurosurgery and Psychiatry,
77, 11251127.
Schroder, M. L., Muizelaar, J. P. and Kuta, A. P. (1994) Documented reversal of
global ischaemia immediately after removal of an acute subdural haematoma. Journal of Neurosurgery, 80, 324327.
Schubert, W., Prior, R., Weidmann, A. et al. (1991) Localisation of Alzheimer
/A4 APP. at central and peripheral synaptic sites. Brain Research, 563,
184194.
Schweitzer, J. B., Park, M. R., Einhaus, S. L. and Robertson, J. T. (1993)
Ubiquitin marks the reactive swellings of diffuse axonal injury. Acta
Seelig, J. M., Becker, D. P., Miller, J. D. et al. (1981) Traumatic acute subdural
haematoma. Major mortality reduction in comatose patients treated within
four hours. New England Journal of Medicine, 304, 15111518.
Sheriff, F. E., Bridges, L. R. and Sivaloganathan, S. (1994) Early detection of
axonal injury after human head trauma using immunocytochemistry for
-amyloid precursor protein. Acta Neuropathologica, 87, 5562.
Sheriff, F. E., Bridges, L. R., Gentleman, S. M. et al. (1994) Markers of axonal
injury in post-mortem human brains. Acta Neuropathologica, 88, 433439.
Shigematsu, K. and McGeer, P. L. (1992) Accumulation of amyloid precursor
protein in neurons after intraventricular injection of colchicine. American
Journal of Pathology, 140, 787794.
Shigematsu, K., McGeer, P. L. and McGeer, E. G. (1992) Localisation of
amyloid precursor protein in selective post-synaptic densities of rat cortical
neurons. Brain Research, 592, 353357.
Simonsen, J. (1967) Fatal subarachnoid haemorrhage in relation to minor head
injuries. Journal of Forensic Medicine, 14, 146155.
Simpson, D. A., Blumbergs, P. C., Cooter, R. D. et al. (1989) Pontomedullary
tears and other gross brain stem injuries after vehicular accidents. Journal of
Trauma, 29, 15191525.
Sinar, E., Mendelow, A., Graham, D. and Teasdale, G. (1987) Experimental
intracerebral haemorrhage: effects of a temporary mass lesion. Journal of
Snow, R. B., Zimmerman, R. D., Gandy, S. E. and Deck, M. D. F. (1986)
Comparison of magnetic resonance imaging and computed tomography in
the elevation of head injury. Neurosurgery, 18, 4552.
Soloniuk, D., Pitts, L. H., Lovely, M. et al. (1986) Traumatic intracerebral
haematomas: timing of appearance and indications for operative removal.
Statham, P. F., Johnston, R. A. and Macpherson, P. (1989) Delayed deterioration in patients with traumatic frontal contusions. Journal of Neurology,
Strich, S. J. (1956) Diffuse degeneration of the cerebral white matter in severe
dementia following head injury. Journal of Neurology, Neurosurgery and
Strich, S. J. (1961) Shearing of nerve fibres as a cause of brain damage due to
head injury. Lancet, ii, 443448.
Taneda, M. and Irino, T. (1979) Enlargement of intracerebral haematomas
following surgical removal of epidural haematomas. Acta Neurochirurgica
(Vienna), 51, 7382.
Teasdale, G. and Jennett, B. (1974) Assessment of coma and impaired
consciousness. A practical scale. Lancet, ii, 8184.
Teasdale, G., Galbraith, S., Murray, L. et al. (1982) Management of traumatic
intracranial haematoma. British Medical Journal, 285, 16951697.
70
PATHOLOGY
Tokuda, T., Ikeda, S., Yanagisawa, N. et al. (1991) Re-examination of ex-boxers

brains using immunohistochemistry with antibodies to amyloid -protein
and tau protein. Acta Neuropathologica, 82, 280285.
Tomlinson, B. E. (1970) Brain stem lesions after head injuries. Journal of Clinical
Pathology (Supplement), 4, 154165.
Tornheim, P. A., McLaurin, R. L. and Thorpe, J. F. (1976) The edema of cerebral
contusions. Surgical Neurology, 5, 171175.
Tuoho, H., Hirakawa, K., Hino et al. (1986) Relationship between abnormalities of coagulation and fibrinolysis and post-operative intracranial
haemorrhage in head injury. Neurosurgery, 19, 523531.
Unterharnscheidt, F. J. (1975) Injuries due to boxing and other sports, in
Handbook of Clinical Neurology, vol. 23, part 1, (eds P. J. Vinken and G. W.
Bruyn), North Holland, Amsterdam, pp. 527593.
Vance, B. M. (1950) Ruptures of surface blood vessels in cerebral subdural
haemorrhage. Archives of Surgery, 61, 9921006.
Vanezis, P., Chann, K. K. and Scholtz, C. L. (1987) White matter damage
following acute head injury. Forensic Science International, 35, 110.
Vigouroux, R. P. and Guillermain, P. (1981) Post-traumatic hemispheric
contusion and laceration, in Progress in Neurological Surgery 10, (ed. H.
Krayenbuhl, P. E. Maspes and W. M. Sweet) S. Karger, Basel, pp.
49163.
Vigouroux, R. P., Guillermain, P. and Rabehauta, P. (1991) Cerebral contusions and lacerations. A clinical study. Advances in Neurotraumatology, 3,
9199.
Voigt, G. E. (1981) Small haemorrhages in the brain stem. A sign of injury?
American Journal of Forensic Medicine and Pathology, 2, 115120.
Voigt, G. E., Lowenhielm, C. G. P. and Ljung, C. B. A. (1977) Rotational
cerebral injuries near the superior margin of the brain. Acta Neuropathologica
(Berlin), 39, 201209.
Voigt, G. E. and Skold, G. (1974) Ring fractures of the base of the skull. Journal
of Trauma, 14, 494505.
Vowles, G. H., Scholtz, C. L. and Cameron, J. M. (1987) Diffuse axonal injury
in early infancy. Journal of Clinical Pathology, 40, 185189.
Wei, E. P., Kontos, H. A., Ellison, M. D. and Povlishock, J. T. (1986) Oxygen

radicals in arachidonate-induced increased bloodbrain barrier permeability to proteins. American Journal of Physiology, 251, H693H699.
Whiteson, A. L. (1981) Injuries in professional boxing. Their treatment and
prevention. Practitioner, 225, 10531057.
Wilberger, J. E., Rothfus, W. E., Tabas, J. et al. (1990) Acute tissue tear
haemorrhages of the brain: computed tomography and clinicopathological
correlations. Neurosurgery, 27, 208213.
Yaghmai, A. and Povlishock, J. T. (1992) Traumatically induced reactive
change as visualised through the use of monoclonal antibodies targeted to
neurofilament subunits. Journal of Neuropathology and Experimental Neurobiology, 51, 158176.
Yarnell, P. R. and Lynch, S. (1973) The ding amnesiac states in football
trauma. Neurology, 23, 196197.
Young, H., Gleave, J., Schmidex, H. and Gregory, S. (1984) Delayed traumatic
intracerebral haematoma; a report of 15 cases operatively treated. Neurosurgery, 14, 2225.
Zimmerman, R. A. and Bilaniuk, L. (1979) Computed tomography in diffuse
traumatic cerebral injury, in Neural Trauma, (eds A. J. Popp et al.), Raven
Press, New York, pp. 253262.
Zimmerman, R. A. and Bilaniuk L. I. (1982) Computed tomographic staging of
traumatic epidural bleeding. Radiology, 144, 809812.
Zimmerman, R. A., Bilaniuk, L. T. and Gennarelli, T. A. (1978) Computed
tomography of shearing injuries of the cerebral white matter. Radiology, 127,
393396.
Zimmerman, R., Bilaniuk, L., Gennarelli, T. et al. (1978) Cranial computed
tomography in diagnosis and management of acute head trauma. American
Journal of Radiology, 13, 2934.
Zuccarello, M., Iavicoli, R., Pardatscher, K. et al. (1981) Post-traumatic
intraventricular haemorrhages. Acta Neurochirurgica (Vienna), 55, 283293.
Zuccarello, M., Fiore, D. L., Trincia, G. et al. (1983) Traumatic primary brain
stem haemorrhage. A clinical and experimental study. Acta Neurochirurgica
(Vienna), 67, 103113.
PRIMARY AND SECONDARY BRAIN

INJURY
A. David Mendelow and Peter J. Crawford
4.1
Introduction
Throughout most of the world, the majority of headinjured patients are initially managed by emergency
medical services that do not have specialized knowledge of the pathophysiology and treatment of head
injury. It is for this reason that the traditional division
into primary and secondary brain damage remains
useful; primary brain damage occurs at the time of
impact, produces its clinical effect almost immediately
and is refractory to most treatment. By contrast,
secondary brain damage occurs at some time after the
primary impact and is largely preventable and treatable. The clinicians role, therefore, is to recognize and
document the primary brain damage, then to prevent
and treat secondary brain damage. Recent research
has shown that, although primary brain damage has
been regarded as irreversible, changes in ultrastructure, the bloodbrain barrier and neuronal function
may progress over time and may provide some
potential for treatment (Povlishock 1992, 1995; Maxwell, 1995).
Understanding this concept prepares the non-specialist clinician for the main challenge in head injury
management: the prevention and treatment of secondary damage. It is therefore essential that all the causes
and consequences of secondary brain damage are
known and understood. In an ideal world, no secondary brain damage would occur! Also, this concept
paves the way for understanding how neuroprotective
strategies (hemodynamic and pharmacological) may
limit secondary brain damage. There is also merit in
classifying brain damage into focal and diffuse types
(Teasdale, 1995) but, from the clinical standpoint, the
division into primary and secondary damage remains
the most pragmatic and therapeutically useful
classification.
Secondary brain damage may begin very rapidly
after impact, so that decisions must be taken early and
correctly.
Globally, many more lives would be saved and the

morbidity from head injury would be more effectively
reduced if head injury management and services were
better organized than could ever be achieved by
improving the intensive care and pharmacological
treatments in already well-developed areas and centers. Unfortunately it is this high tech emphasis that
has dominated thinking and practice in head injury
management to date.
This concept and such organizational challenges
have been recognized for many years, and a group of
British neurosurgeons (Group of Neurosurgeons,
1984) produced guidelines for head injury management that have become a consensus document adopted throughout the UK and now in many other parts
of the world (Garibi, 1995, personal communication).
More recently, such guidelines have been modified to
include recommendations for children, and the Society of British Neurosurgeons has endorsed new
guidelines. Their introduction has resulted in an
increase in the number of hematomas being detected
in large regional centers (Miller, 1993; Treadwell and
Mendelow 1994; Figures 4.1, 4.2). A similar set of
guidelines has recently been produced in the United
States, under the auspices of the American Association
of Neurological Surgeons.
The prime aim of such guidelines is:
to reduce initial hypoxic ischemic damage using

principles of resuscitation set out in the Advanced
Trauma Life Support system (ATLS American
College of Surgeons, 1993);
to increase the early detection of hematomas so that
delay in treatment can be eliminated.
Although the ATLS recommendations play a vitally

important role in preventing hypoxia and ischemia,
the ATLS approach to head injury is not practical
throughout most of the world. The ATLS statement
All head injuries except the most minor will require a
CT scan would result in the successful diagnosis of
72
PRIMARY AND SECONDARY BRAIN INJURY
tions for CT scanning that may not be practical in

many small centers because of the limited availability
of after-hours scanning.
4.2
Figure 4.1 Effect of 1982 Edinburgh guidelines for management of head injuries on the pattern of traumatic hematomas
detected and operated on before (a) (1981) and after (b)
(1986/1989) their introduction; odds ratio with 95% confidence limits. a = detected; b = operated. (Source: reproduced from Miller et al., 1992, with permission.)
most hematomas, but could not be implemented

universally because of the lack of local 24-hour
emergency CT facilities in most countries (Hewer and
Wood, 1989). It is for this reason that guidelines for
head injury are more applicable to cities, towns and
regions without local 24-hour CT scanning facilities
for all head injuries. Since they are based on risk
factors for traumatic hematomas (Mendelow, Teasdale
and Jennett, 1983; Teasdale et al., 1990), they should
result in correct diagnosis of the majority of traumatic
hematomas by selecting for transfer those headinjured patients who are most at risk of developing
hematomas. They would be transferred to centralized
neurotrauma units where 24-hour scanning is available. Similarly, the advanced pediatric life support
(APLS) program (APLS, 1993) has made recommenda-
Primary brain damage
The pathology of brain damage has been discussed in

Chapter 3 and will not be discussed in great detail here,
other than referring to its importance in determining
the initial level of consciousness and any focal
neurological deficit. Also, the time-related ultrastructural changes that take place within minutes of injury
are discussed because of their importance in relation to
possible treatment. The clinical effects of this primary
(or impact) damage may be greatly aggravated by
secondary brain damage (Bullock, Zauner et al., 1995;
Ito, Barzo et al., 1995). Thus diffuse axonal injury
(DAI), contusions and lacerations of the brain will
produce immediate clinical effects varying from concussion, with mild DAI (Oppenheimer, 1968), to coma
and death. Focal primary damage, for example with
cerebral laceration due to a penetrating injury, may
also produce an immediate neurological deficit
depending upon the site of injury. Thereafter, any
increase in neurological deficit or deepening of the level
of consciousness will be due to secondary brain
damage.
Recent experimental research has shown that focal
axonal swelling occurs within 15 minutes of traumatic
brain injury (TBI) due to misalignment of microtubules
(Maxwell, 1995) and that their disruption leads to
nodal blebs within 2 hours. Later there is a change in
the axolemma permeability which, after 6 hours, leads
to retrograde and anterograde misalignment of neurofilaments (with compaction and mitochondrial
Figure 4.2 Effect of two editions of head injury guidelines on head injury admissions to Newcastle General Hospital in the
Northern Region of England and number of traumatic hematomas detected. (Source: data from Treadwell and Mendelow,
1994, updated for 1995 data.)
SECONDARY BRAIN DAMAGE
abnormalities). Similarly, bloodbrain barrier (BBB)

breakdown begins within 3 minutes of injury, leading
to albumin extravasation (Fukuda et al., 1995). These
early changes following DAI have been confirmed
clinically with diffusion-weighted magnetic resonance
imaging (MRI; Vink, 1995; Mikulis, 1995; Lenkinski,
Gennarelli et al., 1995). Nevertheless, it is superadded
secondary ischemic damage that results in cytotoxic
edema and elevated intracranial pressure (ICP; Ito,
Barzo et al., 1995). The combination of these early
changes with DAI and hypoxia/ischemia as a secondary event may provide some potential for pharmacological neuroprotection since increases in glutamate
concentrations are maximal after secondary ischemia
but do not occur when DAI occurs in isolation
(Bullock, Zauner et al., 1995).
Tsuji et al. (1995) have shown experimentally that
with DAI in rats there is no change in excitatory amino
acid (EAA) or extracellular calcium levels unless
secondary insults take place as well (PaO2 < 5.33 kPa;
BP < 40 mmHg).
4.3
Secondary brain damage
The classification of secondary brain damage has

traditionally been into extra- and intracranial (Table
4.1). There is merit in maintaining this classification.
4.3.1
EXTRACRANIAL SECONDARY BRAIN DAMAGE
Extracranial problems produce secondary brain damage either by hypoxia or by oligemia/ischemia (Table
73
4.1). The ultimate consequence of either is a reduction

in the availability of high-energy phosphate (adenosine triphosphate, ATP). This leads to failure of
membrane pumps so that cells either die or become
swollen (so called cytotoxic edema). The distribution
differs in that hypotension with primary oligemia and
ischemia tends to affect the arterial boundary zones
(Figure 4.3). By contrast, hypoxemia alone tends to be
more global, with neuronal loss, which leads to
cortical atrophy in survivors.
The most extreme consequences of severe and
prolonged hypoxia are the persistent vegetative state
(PVS) or death. PVS may occur with preservation of
brain-stem reflexes, but with loss of most of the cortex,
although hallmarks of DAI are found in 77% of patients
who die in PVS (Graham, Jennett et al., 1995).
Secondary brain damage due to hypoxia and hypotension also occurs in patients who have already been
admitted to intensive care units, when it is looked for
carefully: detailed monitoring studies by Jones and
Miller (Jones et al., 1993, 1994; Andrews et al., 1990) with
accurate computerized recording have revealed that
such insults occur in 32% of head-injured patients even
when there is no evidence that these have been detected
on standard nursing charts.
If the extracranial insults are associated with
intracranial lesions, then the penumbra of any region
of hemorrhage or compression becomes the primary
target of hypoxemia or ischemia. It has been clearly
shown that intracerebral hemorrhage produces an area
of ischemia around the hematoma and that this is
surrounded by a penumbra of functionally impaired
Table 4.1 Extracranial and intracranial causes of

secondary brain damage
Extracranial causes

Hypoxia
Hypotension
Hyponatremia
Hyperthermia
Hypoglycemia
Intracranial causes
Hemorrhage
Extradural
Subdural
Intracerebral
Intraventricular
Subarachnoid
Swelling
Venous congestion/hyperemia
Edema
Vasogenic
Cytotoxic
Interstitial
Infection
Meningitis
Brain abscess
Figure 4.3 CT scan showing bilateral boundary zone

infarcts from hypoxiaischemia.
74
(a)
(b)
Figure 4.4 (a) Cross-section of rat brain showing intracerebral hematoma. (b) Autoradiograph of same rat to show
larger area of ischemia.
but potentially viable tissue (Figure 4.4; Mendelow et

al., 1984).
Similarly, there is an area of ischemia that underlies
an experimental subdural hematoma which is also
bounded by a penumbra (Bullock et al., 1990). In all
these situations, hypoxemic or ischemic extracranial
insults will aggravate the secondary damage in relation
to the focal lesion. Similarly, extracranial secondary
insults will aggravate primary DAI (Bullock et al., 1995;
Ito, Barzo et al., 1995).
Regrettably, such hypoxic and hypotensive insults
have persisted in patients transferred to neurosurgery
units, despite many efforts to prevent them (Gentleman
and Jennett, 1981; Gentleman, 1992; Kohi et al., 1984).
The consequence is that ischemic brain damage
remains a very frequent finding at autopsy in patients
who die with head injury. It has fallen only very slightly
from 91% in 1978 to just over 80% in 1989 (Graham,
Adams and Doyle, 1978; Graham et al., 1989). Hopefully, better emergency rescue and pre-hospital resuscitation will reduce the incidence of ischemic brain
damage. The ATLS and APLS courses will probably
achieve this with better initial resuscitation.
The effect of hypoxia/ischemia is particularly severe
in head-injured patients where ICP may be elevated, so
that cerebral perfusion pressure (CPP) is reduced.
Levels of blood pressure that would normally be well
tolerated are insufficient to maintain normal cerebral
perfusion when ICP is elevated. It has therefore become

recognized that higher levels of CPP are associated
with better outcomes from head injury (McGraw, 1989;
Mendelow et al., 1994; Miller, 1993; Cortbus et al., 1995;
Wong et al., 1995a,b; Contant et al., 1993).
Furthermore, autoregulation that has become so well
recognized in normal physiological studies becomes
disordered following head injury (Figure 4.5).
These ischemic events may be missed when CBF is
measured because it is often measured only in the
intensive care unit after stabilization has taken place
(Mendelow et al., 1985; Obrist et al., 1979; Enevoldson
et al., 1976; Matthews et al., 1995; Sharples et al., 1994,
1995). However, studies early after head injury have
now confirmed reduced CBF with loss of autoregulation if measured within 8 hours of the injury (Schroder
et al., 1993; Chapter 11). The result is that an adult
patient with a head injury may require a CPP of 80 or
90 mmHg to maintain normal cerebral perfusion. With
increased ICP (e.g. 30 mmHg) this may require a mean
blood pressure (BP) of 120 mmHg it is interesting to
realize that the unanesthetized previously normal
patient with head injury may achieve this spontaneous rise in BP with the Cushing response, which has
been recognized for almost a century now (Cushing,
1901, 1903). However, deeply anesthetized patients
may be unable to mount a Cushing response, so
clinicians dealing with anesthetized head-injured
patients in intensive care units, resuscitation rooms
and operating theaters must be aware of the need to
maintain adequate levels of CPP. In some instances it
may be possible to determine the breakpoint or
inflection point for the lower limit of autoregulation
by measuring middle cerebral velocity with transcranial Doppler (Wong et al., 1995a,b). In this way the
Figure 4.5 Diagram showing loss of autoregulation following head injury. Cerebral blood flow (CBF) remains constant
over a range of cerebral perfusion pressure (CPP) from
50150 mmHg. Continuous line represents loss of
autoregulation.
SECONDARY BRAIN DAMAGE
CPP could be tailored to the individual patient. Early

monitoring of these parameters may lead to better
acute care of head-injured patients in future, although
at present these sophisticated monitoring techniques
are not universally available. Nevertheless, less invasive techniques like jugular venous oxygen (JVO2)
monitoring (Piper et al., 1995; Gopinath et al., 1994)
and transcranial near infrared spectroscopy (NIRS)
may become more practical in future (Germon, 1995).
In a recent study, NIRS provided a sensitive means of
detecting changes in cerebral oxygenation in 97% of
insults detected with CPP, transcranial Doppler or
laser-Doppler (Kirkpatrick et al., 1995). By contrast,
JVO2 detected only 53% of these insults.
Other causes of extracranial primary brain damage
include hypocapnia (used to treat raised ICP) and
severe hypoxemia. Although hyperventilation with
hypocapnia can reduce cerebral blood flow, there is
doubt about whether moderate hyperventilation can
produce ischemic brain damage unless cardiac output
is reduced by excessive hyperventilation. It is probably best to maintain the PaCO2 at around 32 2 mmHg
in most patients (Chapter 18).
Hyponatremia after head injury is often due to
excess ADH secretion, which results from hypovolemia caused either by fluid restriction or by hemorrhage from other injuries. The excess ADH secretion is
therefore appropriate to the hypovolemia but inappropriate to the hyponatremia. Further fluid restriction may aggravate the problem by further increasing
ADH levels (Poon et al., 1989; Jackowski, 1992). A clear
understanding of the water and electrolyte problems
following head injury may need to be determined in
each case (Nelson et al., 1981). In general, uncorrected
hyponatremia may lead to reduced levels of consciousness and even epileptic seizures. Hyponatremia
may also result from the ill-advised and sometimes
excessive use of dextrose solutions without any
sodium supplementation.
4.3.2
INTRACRANIAL SECONDARY BRAIN DAMAGE
The rate of development of secondary brain damage

depends upon the cause. It is universally recognized
that hematomas should be evacuated expeditiously.
The adverse effect of long delay times with extradural
hemorrhage and subdural hemorrhage has been documented in studies from Edinburgh, UK (Figure 4.6;
Mendelow et al., 1979) and Richmond, VA (Figure 4.7;
Seelig et al., 1981).
With extradural hemorrhage, the time from first
recorded deterioration in level of consciousness to
operation was less than 2 hours on average in
patients who made a good recovery or who were only
moderately disabled. Similarly, with acute subdural
hemorrhage the time from injury to operation strongly
75
Figure 4.6 Effect of delay (from deterioration to surgery)

on outcome from extradural hematoma. O = good recovery;
D = death. (Source: reproduced from Mendelow et al., 1979,
with permission.)
influenced outcome, delays of more than 4 hours from

injury being the most significant.
In either case the mechanism of secondary brain
damage is direct compression of the underlying cortex
causing local ischemic brain damage and brain shift,
which causes local zones of ischemia in the brain stem
and basal structures, and in the cingulate gyrus
(Chapter 18). The ischemic brain damage tends to be
focal, but if elevated ICP is unrelieved, leading to
reduced CPP, then global ischemic brain damage may
occur. Similarly, experimental studies have shown
ischemic brain damage surrounding intracerebral
hemorrhage (Jenkins et al., 1990; Kingman et al., 1988;
Mendelow et al., 1984; Nath et al., 1986, 1987; Nehls et al.,
1988; Sinar et al., 1987) and clinical studies with SPECT
have confirmed this (Figure 4.8; Wyper et al., 1995).
Ultimately the mechanism of neuronal death is
almost always a reduction in ATP with membrane
pump failure. This leads to activation of calcium
channels, with influx of calcium ions into cells resulting
in cell death. The process also leads to a release of
excitatory amino acids (EAA), including glutamate and
aspartate. These activate the receptor-operated calcium
channels, with further calcium influx (Figure 4.9).
Prevention of this calcium influx into cells has offered
scope for neuroprotective agents, such as nimodipine,
76
Figure 4.7 Effect of delay (from injury to operation) in acute subdural hematoma. (Source: reproduced from Seelig et al.,
1981, with permission.)
which block the voltage-operated calcium channels

(Mohamed et al., 1985) or the receptor-operated
channels (Hatfield et al., 1992; Chapter 7).
The potential for neuroprotective strategies in
severe head injury is likely to be maximal when
secondary brain damage is due to hemorrhage (intracerebral, subdural and subarachnoid) and ischemia
(Chapters 7 and 21).
(a) Traumatic subarachnoid hemorrhage and

secondary damage
Experimentally, SAH is associated with delayed
breakdown of the BBB (Fukuda et al., 1995), possibly
because metallic ions in blood (e.g. iron and calcium)
are potent catalysts of free radicals (Cortez et al., 1989).
Clinical outcome in traumatic SAH is worse than in
Figure 4.8 SPECT scan showing area of ischemia surrounding intracranial hemorrhage. (Source: reproduced from
Chocksey et al., 1991, with permission.)
EXTRADURAL HEMATOMA
Figure 4.9 Calcium channel (receptor-activated) to illustrate site of competitive and non-competitive antagonists.
(Source: reproduced from Davis and Barer, 1995, with
permission.)
patients without SAH on CT scan and is associated

with a higher incidence of secondary hypoxic/ischemic insults (Green et al., 1995). In traumatic SAH
Kakarieka et al. (1995) have shown that the calcium
antagonist nimodipine produced significantly fewer
unfavorable outcomes (6-month Glasgow Outcome
Scale) than placebo in a prospective randomized
controlled trial, confirming earlier suggestions from
the HIT II trauma trial with nimodipine that patients
with traumatic SAH may respond better to nimodipine than those without SAH (Braakman et al., 1994).
It should be remembered that neuroprotective
strategies can be organizational as well as pharmaceutical: beneficial effects of early operative evacuation cannot be overemphasized because the deeper the
level of consciousness at the time of operation in
patients with traumatic hematomas, the worse the
outcome (Figure 4.10; Bullock and Teasdale, 1990). For
this reason early diagnosis and treatment of hematomas will always be the most effective neuroprotective strategy in dealing with head injury.
The severity of underlying primary brain damage
with trauma can vary from almost none (in a patient
who is fully conscious, but who later deteriorates) to
severe, where the patient may be in coma from the

outset. There is more likely to be severe primary brain
damage with acute subdural and intracerebral hemorrhage than with cases of extradural hemorrhage,
where there may be very little primary damage. This
accounts for the classical lucid interval that has
become so well known with extradural hemorrhage
(Jamieson and Yelland, 1968). It puts an added burden
of responsibility on those clinicians who manage the
many patients with minor head injuries with minimal
primary brain damage, because compression from
extradural hemorrhage may result from middle
meningeal bleeding with a skull fracture in an
otherwise minor head injury. This is the basis for
recognition of the importance of skull fracture in
detecting hematomas (Chan et al., 1990; Mendelow,
Teasdale and Jennett, 1983; Teasdale et al., 1990; Santos
et al., 1995). If free access to CT scanning is available,
all head-injured patients will require a CT scan, as
assumed in the ATLS and AANS guidelines. The use
of a skull X-ray to detect a fracture then becomes
unimportant, but in the absence of universal access to
CT for head-injured patients, skull fracture remains
one of the most significant clinical and radiological
features to help detect hematomas (Table 4.2).
4.4
Extradural hematoma
Extradural hematomas (EDH) are extracerebral lesions

and thus there may be little or no primary brain
Table 4.2 Risks of traumatic hematoma with skull

fracture and altered consciousness for patients attending
Accident and Emergency (Source: from Teasdale et al.)
No. per million
per year
Absolute risk of
hematoma (1 in:)
No skull fracture
Fully conscious
Impaired consciousness
Coma
7700
550
66
7900
180
27
Skull fracture
Fully conscious
Coma
130
43
41
45
5
4
No skull fracture
Fully conscious
Coma
7200
250
23
13 000
580
65
Skull fracture
Fully conscious
Coma
100
17
11
160
25
12
Adults
Children
Figure 4.10 Effect of Glasgow Coma Score on outcome in

patients with traumatic intracranial hematoma. (Source:
reproduced from Bullock and Teasdale, 1990, with
permission.)
77
78
damage, so that, initially, consciousness may recover

or be preserved this accounts for the well known
lucid interval that occurs with EDH (although it
occurs in less than 25% of cases). The frequency of
skull fracture with extradural hematoma in children is
79.3% (Santos et al., 1995), which is as high as it is in
adults. Furthermore, in infants the large head size
relative to the body means that the volume of the
extradural space is large in relation to the blood
volume, so that hypovolemia may be the primary
presenting feature with an infantile extradural hematoma. This, coupled with elevated ICP, leads to a rapid
fall in CPP and consequent ischemic brain damage.
This is seen particularly in non-accidental injury. For
these reasons, monitoring of infants with even minor
head injury should include continuous pulse rate
recording, and a CT may be indicated if a tachycardia
develops.
Posterior fossa extradural hematomas,, though rare
(less than 5% of all EDH) may give rise to a sudden
deterioration in the level of consciousness due to
hydrocephalus, and it is important to recognize the
need to scan the posterior fossa if an occipital skull
fracture is present (Figure 4.11).
A posterior fossa extradural hematoma may extend
above the transverse sinus and compress the occipital
pole, so that great care must be taken to avoid major
blood loss from the confluens sinuum or the transverse sinus when operating on these cases. In children,
extradural haematomas are often limited by the
adherence of the dura to the sutures (Figure 4.12).
Figure 4.12 CT scan showing extradural haematoma in a

child, limited by sutures.
Chronic venous extradural hematomas are rare but

may present with late deterioration in level of consciousness (Stevenson et al., 1964). Because these lie at
the vertex, CT scanning should include high cuts
when a high skull fracture extends over the superior
longitudinal sinus.
When a patient with an EDH dies, neuropathological studies usually show massive ischemic damage
due to brain compression and/or low CPP.
4.5
Figure 4.11 CT scan showing posterior fossa extradural

hematoma with extension into the left occipital supratentorial region. Note right Sylvian subarachnoid hemorrhage.
Intradural hemorrhage
Subdural hemorrhage and intracerebral hemorrhage

were often considered as two separate conditions, and
they may well be separate. However, CT and MRI
have clearly shown that subdural hematomas often
coexist with intracerebral hemorrhage and with cerebral contusions. Acute subdural hemorrhage tends to
extend throughout the subdural space and lies mainly
on the surface. Although apparently thin, their extensive nature is associated with a larger intracranial
volume than would be expected from CT scanning
alone. They are usually unilateral and are more
frequently associated with underlying primary brain
damage, so that patients with acute subdural hematomas are more likely to present in coma, in contrast
HERNIATION
with extradural hemorrhage where primary brain

damage is often less severe. The acute subdural
hematoma associated with anticoagulants is the
exception because in these patients there is often little
or no primary brain damage so that they deteriorate,
as they bleed into the subdural space, from full
consciousness. With time, the subdural hematoma
begins to liquefy and by 1421 days it may have
become predominantly liquid so that burr hole drainage may drain it effectively, in contrast with acute
subdural hematoma where the solid clot has to be
removed via a craniotomy.
Chronic subdural hematomas represent a different
spectrum of head injury. The trauma is often minimal
or even unrecognized in one-third of cases (Cameron,
1978). Because there may be no primary brain damage
in these patients, the outcome is much better than
following the treatment of acute subdural hematomas
(Nath et al., 1985). In some patients with chronic
subdural hemorrhage there is recent acute hemorrhage so that different clot densities can be recognized
on CT scans. This recurrent hemorrhage may lead to
the development of pseudomembranes which, with
time, will result in further liquefaction of the clot.
Also, chronic subdural hemorrhage may be bilateral.
Such cases may present as dementia rather than as a
head injury.
4.6
79
damage (Nornes, 1975). The pathogenesis of the

ischemia has been extensively studied. If an equivalent volume of fluid is injected instead of blood, the
volume of ischemia is less than when blood alone is
injected (Jenkins et al., 1990) and the area of ischemia
is proportional to the volume of blood or fluid injected
(Nath et al., 1986). This ischemia and oligemia may be
due to mechanical compression of the microcirculation
(Kingman et al., 1988). The reduction in CBF and
ischemic neuronal damage is also related to the
duration of compression, since deflation of a 50 l
microballoon inflated in the rat brain and removed in
10 minutes still leaves extensive neuronal damage and
a zone of reduced CBF (Nehls et al., 1988). This
suggests that evacuation of intracerebral clots may do
little to relieve surrounding ischemic neuronal damage, although removal may reduce ICP by relieving
the mass effect. This is clinically logical, since Galbraith and Teasdale (1981) have reported that subsequent deterioration occurs if the ICP exceeds
30 mmHg in patients with traumatic intracerebral
hemorrhage. This suggests that surgical treatment is
effective for reducing ICP but not for reducing local
ischemic damage.
Acute intraventricular hemorrhage and subarachnoid hemorrhage may result in acute hydrocephalus
necessitating ventricular drainage although this is
rare.
Traumatic intracerebral hemorrhage may occur in

isolation or be part of a complex intradural hemorrhage. Isolated intracerebral hemorrhage is much
more common in the elderly and may at times be
difficult to distinguish from spontaneous intracerebral
hemorrhage. This is because a primary spontaneous
hemorrhage may result in a fall, which causes a
secondary head injury. The mechanism of development of traumatic intracerebral hemorrhage is similar
to spontaneous intracerebral hemorrhage: an artery or
arteriole is disrupted by shearing forces, or ruptures
spontaneously, allowing blood under arterial pressure
to expand into the brain parenchyma. Bleeding stops
when the tissue pressure around the clot reaches
arterial pressure. Experimental studies have shown
that this type of intracerebral hemorrhage leads to
acute ischemia in the immediately adjacent brain. The
clot may remain contained within the parenchyma or
burst into the ventricle, the subdural space or the
subarachnoid space. In contained hemorrhage, there is
a ring of ischemia around the hematoma, which in
turn is surrounded by a penumbra of functionally
impaired but potentially recoverable tissue (Mendelow et al., 1984). The uncontained type of hemorrhage,
like subarachnoid hemorrhage, leads to a global fall in
CPP with much more widespread ischemic neuronal
4.7
Herniation
4.7.1
TENTORIAL HERNIATION
Apart from elevating ICP and reducing CPP (global

effects) and causing local compression and vasoconstriction, hematomas may also cause herniation, the
commonest form of which is tentorial herniation. The
uncus of the temporal lobe is pushed through the
tentorial edge, compressing the third cranial nerve
and the posterior cerebral artery. It is common clinical
knowledge that a IIIrd nerve palsy following a head
injury is a sign of tentorial herniation due to a
hematoma, but compression of the posterior cerebral
artery is less well recognized. This will produce a
medial occipital infarct, which may be the only
residual long-term consequence if the clot is removed.
These patients may have an homonymous hemianopia
on recovery and the infarct may be demonstrable on
magnetic resonance imaging (Figure 4.13).
4.7.2
SUBFALCINE HERNIATION
Subfalcine herniation may occur as a result of displacement by hematomas. This may cause ischemia
due to compression of branches of the anterior
cerebral artery (Figure 4.14).
80
In conclusion, secondary brain damage from hematomas can have widespread manifestations and consequences but the principle of the diagnosis of
secondary brain damage due to a hematoma must be
recognized so that this rare complication of head injury
(less than 1% of all head injuries who attend Emergency
rooms) can be recognized and treated rapidly.
4.8
Figure 4.13 MRI showing right medial occipital infarct

months after posterior cerebral artery compression from
delayed evacuation of an extradural hematoma with tentorial herniation.
4.7.3
TONSILLAR HERNIATION (MEDULLARY CONE)
Herniation of the cerebellar tonsils is seen particularly

with posterior fossa hematomas. The cerebellar tonsils
become displaced through the foramen magnum and
may be found below the level of C2. This is associated
clinically with medullary compression and ultimately
brain-stem death.
Brain swelling may be due to vascular engorgement,

or to brain edema. The recognition of venous congestion as a cause of elevated ICP, especially in children
(Bruce et al., 1981) led to the widespread use of
hyperventilation and hypocapnia in the treatment of
raised ICP. More recently, the dangers of excessive
hyperventilation have been recognized (Sharples et al.,
1994). Venous congestion secondary to cerebrovascular dilatation is also thought to be the mechanism
for the generation of A-(plateau) waves of raised ICP
(Rosner, 1993).
Brain edema is due to an increase in the water
content of brain interstitial space, the neurons or the
glia. It has been divided into vasogenic, cytotoxic and
interstitial types (Klatzo, 1979). Some have suggested
that better terms would be open barrier edema for
vasogenic edema and closed barrier edema for
cytotoxic edema (Betz et al., 1989). Recent studies in
animal models (Fukuda, Tanno et al., 1995) have shown
that bloodbrain barrier (BBB) breakdown may occur
within a few minutes of injury, challenging the earlier
view that vasogenic (open barrier) edema occurred
only late after head injury. These workers, however, did
show that late BBB breakdown often occurred following SAH, perhaps because the metallic ions in blood are
potent catalysts of free radicals (Cortez et al., 1989).
Swelling of axons may occur soon after primary DAI
because of changes in axolemma permeability (Pettus
and Povlishock, 1995) but it is doubtful that this mild
axonal swelling would be sufficient to cause whole
brain swelling, although it may be associated with
transient axonal malfunction. After 6 hours, retrograde
and anterograde malalignment of neurofilaments takes
place in axons with compaction and mitochondrial
swellings. That such mitochondrial abnormalities may
occur following traumatic brain injury has been
suggested by recent clinical studies in severe head
injury in children (Matthews et al., 1996). These
mitochondrial changes are rapidly reversible in mild
injuries, and may constitute an anatomical basis for
concussion (Petrus and Povlishock, 1995).
4.8.1
Figure 4.14 MRI to show subdural hematoma with subfalcine shift.
Brain swelling
VASOGENIC EDEMA (OPEN BARRIER EDEMA)
As its name implies, open barrier edema is caused by

a change to BBB so that protein leaks into the
BRAIN SWELLING
interstitial space (Milhorat, 1992). The increased

oncotic pressure draws water with it, so that water
accumulates between cells. The interstitial space is the
primary pathway for clearing extracellular edema
proteins. The primary mechanism is diffusion that is
independent of pressure gradients (Ohata and Marmarou, 1992). In the direct infusion edema model in
the rat, Ohata and Marmarou showed that high- and
low-molecular-weight dextrans, as edema markers,
moved preferentially towards the cortical surfaces and
the subarachnoid CSF.
This type of edema may also develop later around
contusions and intracerebral hemorrhage after several
days (Marmarou, 1994; Bullock, 1985; Teasdale, 1995;
Figure 4.15). In general, vasogenic edema is more
common in relation to tumors and abscesses than in
head injury. Dexamethasone is effective in reducing
vasogenic edema (Todd and Teasdale, 1989). Dexamethasone has not been shown to be effective in head
injury although, in one controlled trial of high dose
steroids, a subgroup of patients with traumatic intracerebral hemorrhage and contusions seemed to have a
better outcome when steroids were given (Grumme et
al., 1995).
4.8.2
CYTOTOXIC EDEMA (CLOSED BARRIER EDEMA)
In this type of brain swelling, the cells themselves

swell, as the name cytotoxic edema implies. The
cause is usually hypoxia/ischemia (either local or
global) resulting in a loss of high-energy phosphates
and malfunction of the sodium/potassium pumps.
Water enters the cells, and calcium channels open,
with influx of these ions into neurons. The influx of

calcium is itself cytotoxic and results in cell death
(Rappaport et al., 1987). Further ischemia may result in
the release of free radicals, which change the phospholipid membrane between the ion channels. This aspect
is discussed more fully in Chapter 21. Damage to the
phospholipid membrane may have a knock-on effect,
so that the whole cell membrane starts to leak with
further influx of calcium and water. Ischemic neurons
also liberate excitatory amino acids (EAA) which
activate receptor-operated calcium channels in the cell
membrane. These include n-methyl-D-aspartate
(NMDA) receptors and AMPA (amino-3-hydroxy5-methyl-4-isoxazole propionic acid) receptors which,
when open, also allow the influx of calcium into the
cells (Chapters 7 and 21).
Therapeutic initiatives have focused on the use of
calcium-channel-blocking drugs and free-radical scavengers as well as NMDA-receptor antagonists (competitive and non-competitive), both experimentally
and, more recently, clinically.
The distribution of cytotoxic edema will depend
upon the distribution of ischemia, and it may be
detected by diffusion weighted MRI (Chapter14),
T2-weighted MRI (if more severe) or CT, in its most
severe forms (e.g. stroke), by 812 hours after onset. If
there is global ischemia due to reduced CPP, infarction
and edema will occur in the arterial boundary zones
(Figure 4.3). Around or beneath hematomas, cytotoxic
edema may remain focal. If large end-arteries are
damaged or compressed, there may be an infarct in
the appropriate arterial territory (middle, anterior or
posterior cerebral artery territory). In each of these
situations, there will be a surrounding penumbra of
oligemic tissue, which has the potential to benefit from
pharmacological and intensive care treatment to optimize CBF and O2 delivery.
4.8.3
Figure 4.15 CT scan demonstrating small left parietal

contusion with surrounding edema.
81
INTERSTITIAL EDEMA
Hydrocephalus results in increased intraventricular

cerebrospinal fluid pressure, with transependymal
exudation of CSF through the ependyma into the
brain tissue. This may manifest itself on CT or MRI as
periventricular lucency (PVL). Interstitial edema can
be relieved by ventricular drainage.
All three types of brain edema may coexist in a
head-injured patient, although cytotoxic edema is
usually the earliest to appear. The degree of edema in
patients with head injury can be measured with
gravimetric columns (Shigeno et al., 1982) and this
correlates well with the Hounsfield numbers on CT
scan (Bullock et al., 1985) and with MRI proton density
figures (Marmarou et al., 1990). It is therefore possible
to measure the degree of brain edema in patients using
CT and MR imaging. These techniques have clearly
82
demonstrated the different patterns of distribution of

edema in various patients (Figure 4.3). For example, in
global ischemia there may be boundary zone ischemia
with edema visible on both CT and MRI. Focal edema
may occur with infarction as a result of vasospasm
from subarachnoid blood (Wilkins and Odom, 1970)
or of vessel occlusion (Figure 4.16). Edema may also
occur around contusions (Figure 4.15) and intracerebral hemorrhage (Figure 4.17).
4.8.4
PATHOPHYSIOLOGY OF EDEMA
Magnetic resonance imaging with diffusion weighting (DWI) has made it possible to characterize
different types of edema with apparent diffusion
coefficients (ADC), which are reduced in cytotoxic
edema and increased in vasogenic edema (Ito et al.,
1995). Following DAI in the rat acceleration model,
cytotoxic edema has been shown to follow secondary
insults. It is likely that these diffusion-weighted
techniques will detect changes in tissue water diffusion in the early stages of head injury (Vink, 1995;
Chapter 14). Functional imaging, in particular MR
spectroscopy, (e.g. N-acetyl aspartate), is likely to
reveal lesions that will correlate with clinical and
functional deficits.
The plethora of second messenger molecules that
produce ischemic neuronal damage and increased
water content experimentally is also being shown to
be involved clinically, using a variety of techniques,
including MR spectroscopy, microdialysis and measurement of CSF and blood/serum levels in patients.
Figure 4.17 CT scan showing acute subdural hematoma,

contralateral contusions with surrounding edema and subarachnoid hemorrhage beneath the tentorium and in the
perimesencephalic cistern.
The variability of the nature and distribution of

edema makes imaging and more accurate time sequence
monitoring desirable if effective treatments are to be
found in head injury. Using new MRI modalities, it
should be possible in the future to differentiate the
different causes and time courses of focal and global
swelling, each of which may require different treatments mannitol, steroids, calcium antagonists,
NMDA-receptor antagonists, ventricular drainage,
free radical scavengers or decompressive surgery.
Perhaps only when we are able to divide the
pathology in head-injured patients into relevant categories will we be able to improve outcome. To classify
all severely head-injured patients into one disease
group (head injury) is unlikely to provide a single
effective therapeutic option. This has been well demonstrated in the several large multicenter trials that
have failed to identify an individual strategy effective
in large heterogeneous groups of head-injured
patients (Bailey et al., 1991; Todd and Teasdale, 1989;
Ward et al., 1985).
4.8.5
Figure 4.16 CT scan showing focal ischemia due to

traumatic internal carotid artery occlusion (note contralateral subdural hematoma).
BRAIN SWELLING IN CHILDREN
It has long been recognized that brain swelling may

occur after even relatively minor head injuries in
children (Bruce et al., 1981). Onset may be rapid but
sometimes may not occur until days later. Two
explanations exist for this rapid onset swelling.
INFECTION
There is very little space available for expansion

within the skull of a child, so that a minor degree of
brain swelling produces a marked increase in ICP.
Children may have a cerebral circulation which
responds more actively to trauma than that of
adults. Bruce et al. (1981) favor the latter, but studies
in Newcastle upon Tyne using receiver-operatingcharacteristic (ROC) curves were able to demonstrate that ICP and CPP have a much greater effect
on outcome in children and young adult patients
under the age of 40 than in older patients, where
there may be more space available within the
cranium (Chambers et al., 1995). This type of
secondary brain damage should therefore be anticipated in children and there should therefore be a
lower threshold for the institution of ICP monitoring, and (if necessary) ventilation in children than in
adults.
It is generally considered that hyperemia is more

common in children and, therefore, if there is a role for
hyperventilation in the treatment of any patients with
head injury, it is likely to be more effective in children
and adolescents. Recent studies in children with head
injury in intensive care units have shown that cerebral
blood flow may fall with long-term hyperventilation
used to treat hyperemia (Sharples et al., 1994) so that it
may be as harmful as in adults.
The reason for the dangerous delayed brain swelling
that sometimes occurs in children with head injury
remains uncertain. In 36 children with severe head
injury, CBF, CMRO2 and CMR lactate was measured
and it was demonstrated that CMRO2 fell significantly between 12 and 48 hours while lactate
production increased, suggesting a shift to anaerobic
metabolism of glucose (Matthews et al., 1996a).
Improvement in outcome from severe head injury in
children may therefore occur only when continuous
monitoring of many biochemical and functional parameters is undertaken and when such results are
immediately available to the medical and nursing
staff. In addition, monitoring to detect such secondary
insults should be undertaken with equipment programmed to provide an alarm when abnormalities
take place, because of the unreliability of routine
nursing data (Jones et al., 1994).
4.8.6 POTENTIAL THERAPEUTIC INITIATIVES TO

LIMIT BRAIN EDEMA AND CELL DAMAGE
The stage in the ischemic cascade at which edema

begins to occur is not yet clear and may vary from
patient to patient. In the early stages, maintaining CPP
may be most important. Later, preventing and treating
brain edema may reduce focal damage and elevated
ICP. Neuroprotective strategies may limit the size of
83
the penumbra around areas of ischemia. Strategies

may include the use of free radical scavengers to
prevent lipid peroxidation, NMDA- and AMPA-receptor antagonists to limit damage from excitatory amino
acids such as glutamate, and calcium antagonists to
prevent influx of calcium into cells.
Finally, progressive edema may result from the
release of leukotrienes from white cells and platelets
(Hallenback et al., 1986). It is of interest to note that
profound degrees of ischemia and oligemia can be
tolerated in hibernating ground squirrels where CBF
falls to 8 ml/100 g/min (Frerichs et al., 1993). In that
study it was shown that the hemoglobin level remains
normal during hibernation, but a severe leukopenia
and thrombocytopenia occur. It is thus possible that
depletion of platelets and leukocytes may be a
natural form of neuroprotection during hypothermia.
Leukopenia and thrombocytopenia may therefore
offer some form of protection by limiting the progressive edema that follows ischemia. Studies in the
intracerebral hemorrhage and middle cerebral artery
occlusion models in the rat have shown that immunosuppression with whole-body irradiation reduces the
white cell and platelet count and limits brain edema
(Kane et al., 1992; Strachan et al., 1992). Further studies
are required to discover whether or not influencing
the function of white cells and platelets may be of
some therapeutic value.
4.9
Infection
Infection complicates 58% of severely head-injured

patients. It will usually become evident a few days
after the injury and will take the form of either a brain
abscess or meningitis. It is an important cause of
secondary deterioration and brain damage. The organism is almost always bacterial and enters the nervous
system via a breach in the dura associated with a
fracture either of the vault or of the skull base.
4.9.1
(a)
INFECTION AFTER SKULL FRACTURE
Fracture of the vault
It is very unusual for linear fractures without skin

disruption to lead to secondary infection. By contrast,
compound linear fractures may allow entry of contaminating bacteria. Compound depressed fractures
are far more likely to lead to secondary infection,
although the risk depends to a great extent upon the
degree of depression, the extent of any contamination
and whether or not there has been any penetration.
There is also a difference between higher-velocity
missile injuries, which impart kinetic energy to the
tissues, causing necrosis and extensive destruction of
tissue, and the low-velocity (non-missile) type of injury
84
more common in civilian experience. The latter is less

likely to cause such severe tissue destruction, so
recovery of vital tissue is more likely to occur. This may
explain why it was always the custom to debride and
elevate compound depressed fractures in wartime, and
why this teaching influenced the management of nonmissile head injuries for many years in civilian practice.
Non-missile compound depressed fractures that are
not contaminated may be associated with less extensive
tissue destruction and may not therefore require
extensive debridement and elevation of the fracture
(van der Heever and van de Merwe, 1986).
The simple linear and depressed fractures of the
vault seen with closed head injury are rarely associated with infection.
(b)
Basal skull fracture
This is more common than linear vault fracture after

severe closed head injury and may involve the anterior
and/or middle fossa. Anterior fossa fractures may
initially result in bleeding into paranasal sinuses with
epistaxis and later with CSF rhinorrhea. However,
anterior fossa fractures may result in free communication between the paranasal sinuses and a dural tear
without overt CSF rhinorrhea. The patient may therefore present days, weeks, or even years after head
injury with meningitis without ever developing a CSF
leak. Therefore a high index of suspicion is necessary in
patients with severe closed head injury where epistaxis, periorbital bruising (so called raccoon/panda
eyes) and extensive subconjunctival hemorrhage (Figure 4.18) occurs. Similarly, middle fossa fractures may
or may not be associated with CSF otorrhea. Initially,
the only clinical sign of fracture may be bleeding from
the ear or bruising over the mastoid process (Battles
sign; Figure 4.19). Fractures of the petrous bone may
also cause deafness or a lower motor neuron facial
nerve palsy. Again, a high index of suspicion of a CSF
fistula is needed because, if present, these patients may
later present with meningitis.
Figure 4.19
4.9.2
MENINGITIS
Whichever the portal of entry, with basal skull

fractures organisms tend to be commensals from the
paranasal sinuses or middle ear. Pneumococcal infection is therefore commonest, unless patients receive
prophylactic antibiotics, which are more likely to
encourage the overgrowth of resistant organisms.
With vault fractures, staphylococci are the most likely
organisms to cross the dura. Staphylococcal meningitis is also more likely to follow postoperative wound
infection when a craniotomy has been performed, for
example, for evacuation of a hematoma. Secondary
infection may also occur with prolonged ventricular
drainage. Meningitis may give rise to hydrocephalus,
so that late deterioration in the level of consciousness
with elevated ICP may occur some days after injury in
a patient with very little primary brain damage. This
again reinforces the usefulness of the concept of
primary and secondary brain damage in the management of patients with head injury.
4.9.3
Figure 4.18 Anterior fossa fracture with bilateral panda

eyes (note subconjunctival hemorrhage).
Middle fossa fracture with Battles sign.
BRAIN ABSCESS
Post-traumatic brain abscess formation is generally

rare and is almost always associated with a penetrating injury, particularly if there is in-driven foreign
PYREXIA FOLLOWING HEAD INJURY
material. It is extremely rare to develop a brain abscess

following severe closed head injury unless there is
fracturing of the posterior wall of the frontal sinus,
although it may sometimes be impossible to detect
that a penetrating injury has occurred. Such a case was
seen several years after an apparently closed head
injury. A frontal brain abscess was drained via
craniotomy, and a non-radio-opaque wooden arrowhead was found within the abscess cavity. The
fragment had entered unrecognized at the time of the
original head injury. Although rare, such cases demonstrate that brain abscess may be a late complication of
an apparently closed head injury.
4.10
Post-traumatic vascular damage
4.10.1 CAROTID AND MIDDLE CEREBRAL ARTERY

(MCA) OCCLUSIONS
The internal carotid artery (ICA) may be damaged in

the neck by acute flexion (Zelenock et al., 1982),
extension (Stringer and Kelly, 1980) or directly by a
safety belt, in which case a cutaneous abrasion over
the cervical carotid artery is usually seen. These blunt
injuries cause dissection of the intima with stenosis
(Ueda et al., 1986), occlusion (Schultz et al., 1984) or
thrombosis with distal embolization (Janon, 1970).
Total occlusion of the internal carotid artery may lead
to major hemisphere infarction that is visible on CT
scan. The absence of flow in the internal carotid or the
ipsilateral middle cerebral artery (MCA) may be
detected with transcranial Doppler (TCD; Schneider et
al., 1988). These patients may have relatively slight
primary brain damage but a dense focal deficit
(mimicking a stroke).
The ICA may also be damaged at the skull base by
basal skull fractures (Aarabi and McQueen, 1978), in
which case the patient may have associated severe
primary brain damage and may therefore be comatose. The diagnosis of traumatic ICA occlusion is
therefore more difficult in these cases, although it
should be suggested by the characteristic infarct on CT
or absent MCA flow on TCD.
Occlusion of intracranial arteries may also result
from the herniations referred to above. These compress and occlude the anterior or posterior arteries.
Survivors of such herniation may display a classical
infarct in the vascular territory of that vessel (Figure
4.16). The intracranial ICA or MCA may become
damaged with rapid acceleration or deceleration
against the sharp sphenoid wing. Primary traumatic
MCA occlusions are well recognized in young patients
with closed head injury and may be associated with
dissecting aneurysms (Kunze and Schiefer, 1971; Sato
et al., 1971) and were more easily recognized in the
85
pre-CT era, when angiography was used commonly in

closed head injury.
4.10.2
TRAUMATIC SUBARACHNOID HEMORRHAGE
Traumatic subarachnoid hemorrhage (SAH) is known

to be associated with cerebral vasospasm in more than
40% of cases (Macpherson and Graham, 1973) and
with cerebral blood flow and TCD monitoring techniques the incidence of vasospasm was found to be
27% (Martin et al., 1992). Traumatic vasospasm following closed head injury is therefore an important cause
of cerebral ischemia which is analogous to that found
in spontaneous SAH.
4.10.3
ANEURYSMS
It may be difficult to differentiate a pre-existing

aneurysm that ruptures as a result of trauma from a
true post-traumatic aneurysm. Undoubtedly both
types exist but it may be impossible to distinguish one
from the other, although true post-traumatic aneurysms tend to be more common in the distal MCA
(Fleischer et al., 1975). The true trauma-induced
aneurysm may also enlarge (Bendoit and Wortzman,
1973) and is at risk of rupturing.
4.10.4
FISTULAE
Blunt injuries may lead to arteriovenous fistulae,

which are usually associated with a cranial bruit. The
commonest site is the carotid cavernous fistula, which
often develops several days after the injury. It is
characterized by proptosis with vasodilation of the
sclera and may go on to cause SAH (Dohrmann et al.,
1985). Fistulae may occur at other sites related to skull
fractures (Feldman et al., 1980). They may also give
rise to late SAH or even subdural hemorrhage in
patients who otherwise have little or no primary brain
damage.
4.11
Pyrexia following head injury
The development of pyrexia following head injury

may be a sign of brain-stem damage. However the
occurrence of pyrexia was one of the most significant
predictors of mortality and morbidity in a series of 124
adult head-injured patients who were monitored
continuously (minute by minute) in Edinburgh (Jones
et al., 1994). There is also a revived interest in
hypothermia as form of neuroprotection in headinjured patients: a multicenter prospective randomized controlled trial is currently under way in the
USA, funded by the National Institute of Health.
86
4.12
Conclusion
While many classifications of brain damage are based

on pathological findings, biochemical and radiological
features, the importance of classifying brain damage
clinically into primary and secondary types is that an
understanding of the pathology will lead to more
appropriate management. Most importantly, it helps
non-specialists to understand the role of the prevention and treatment of brain swelling, hemorrhage and
infection. It is important to teach medical students,
nurses and doctors how to recognize primary and
secondary brain damage. They may then be better able
to prevent it from occurring in the first place or to
minimize the effects once the cycle of secondary brain
damage is initiated. In this way, through team cooperation, the gold standard of head injury management
eliminating secondary brain damage may be
reached.
4.13
References
Aarabi, B. and McQueen, J. (1978) Traumatic internal carotid occlusion at the

base of the skull. Surgical Neurology, 10, 233.
American College of Surgeons (1993) Advanced Trauma Life Support Course
Manual, American College of Surgeons, Washington, DC, p. 168.
Andrews, P., Piper, I. et al. (1990) Secondary insults during intrahospital
transport of head injured patients. Lancet, 335, 327.
APLS (1993) Advanced Paediatric Life Support Manual. British Medical Journal
Publications, London, p. 155.
Bailey, I., Bell, A., Gray, J. et al. (1991) A trial of the effect of nimodipine on
outcome after head injury. Acta Neurochirurgica (Vienna), 110, 97105.
Benoit, B. and Wortzman, G. (1973) Traumatic cerebral aneurysms. Clinical
features and natural history. Journal of Neurology, Neurosurgery and Psychiatry, 36, 127.
Betz, A., Iannotti, F. et al. (1989) Brain oedema: a classification based on bloodbrain barrier integrity. Cerebrovascular Brain Metabolism Review, 1, 133.
Braakman, T. and the European study group on Nimodipine in severe head
injury (1994) A multicentre trial of the efficacy of Nimodipine on outcome
after severe head injury. Journal of Neurosurgery, 80, 797.
Bullock, R. and Teasdale, G. (1990) Surgical management of traumatic
intracranial haematomas, in Handbook of Clinical Neurology, vol. 13(57), Head
Injury, (eds P. J. Vinken, G. W. Bruyn and R. Braakman), Elsevier,
Bullock, M., Smith, R. et al. (1985) Brain specific gravity and CT scan density
measurements after human head injury. Journal of Neurosurgery, 63, 64.
Bullock, R., McCulloch, J. et al. (1990) Focal ischemic damage is reduced by
CPP-ene. Studies in two animal models. Stroke, 21(Suppl. III), 32.
Bullock, R., Zauner, A. et al. (1995) Excitatory amino acid release patterns after
severe human head injury experience with microdialysis in 30 patients.
Journal of Neurotrauma, 12, 372.
Cameron, M. (1978) Chronic subdural haematoma: a review of 114 cases.
Journal of Neurology, Neurosurgery and Psychiatry, 41, 834.
Chambers, I., Choksey, M. et al. (1995) Receiver operator characteristic (ROC)
curves: a new approach to the analysis of cerebral perfusion pressure and
intracranial pressure in relation to outcome in severe head injury. Journal of
Chan, K.-H., Mann, K. et al. (1990) The significance of skull fracture in acute
traumatic intracranial haematomas in adolescents. Journal of Neurosurgery,
72, 189.
Choksey, M. S., Costa, D. C., Iannotti, F. et al. (1991) 99Tc-M-HMPAO SPECT
studies in traumatic intracerebral haematoma. Journal of Neurology, Neurosurgery and Psychiatry, 54, 611.
Contant, C. F., Robertson, C. S., Gopinath, S. P. et al. (1993) Determination of
clinically important thresholds in continuously monitored patients with
head injury (abstract). Journal of Neurotrauma, 10, S57.
Cortbus, F., Jones, P. A., Miller, J. D. et al. (1995) Cause distribution and
significance of episodes of reduced cerebral perfusion pressure following
head injury (abstract). Journal of Neurotrauma, 12, 368.
Cortez, S., McIntosh, T. et al. (1989) Experimental fluid percussion brain

injury: vascular disruption and neuronal and glial alterations. Brain
Research, 482, 271.
Cushing, H. (1901) Concerning a definite regulatory mechanism of the
vasomotor centre which controls blood pressure during cerebral compression. Bulletin of the Johns Hopkins Hospital, 12, 290.
Cushing, H. (1903) The blood pressure reaction of acute cerebral compression,
illustrated by cases of intracranial haemorrhage. American Journal of Medical
Science, 125, 1017.
Davis, M. and Barer, D. (1995) Theoretic aspects of neuroprotection in acute
ischaemic stroke. 1: Basic mechanisms. Vascular Medicine Review, 6,
283298.
Dohrmann, P., Batjer, H. et al. (1985) Recurrent subarachnoid haemorrhage
complicating a traumatic carotid cavernous fistula. Neurosurgery, 17, 480.
Enevoldsen, E. M., Cold, G. E., Jensen, F. T. et al. (1976) Dynamic changes in
regional cerebral blood flow, intraventricular pressure, cerebrospinal fluid
pH and lactate levels during the acute phase of head injury. Journal of
Feldman, R., Hieshima, G. et al. (1980) Traumatic dural arteriovenous fistula
supplied by scalp, meningeal, and cortical vessels: case report. Neurosurgery,
6, 670.
Fleischer, A., Patton, J. et al. (1975) Cerebral aneurysms of traumatic origin.
Surgical Neurology, 4, 233.
Frerichs, K., Kennedy, C. et al. (1993) Local cerebral blood flow and glucose
metabolism during hibernation, a model for natural tolerance to cerebral
ischaemia. Journal of Cerebral Blood Flow and Metabolism, 13(Suppl. 1),
S422.
Fukuda, K., Tanno, H. et al. (1995) The BloodBrain Barrier disruption to
circulating proteins in the early period after fluid percussion brain injury in
rats. Journal of Neurotrauma, 12, 315.
Galbraith, S. and Teasdale, G. (1981) Predicting the need for operation in the
patient with an occult traumatic intracranial haematoma. Journal of
Gentleman, D. (1992) Causes and effects of systemic complications among
severely head injured patients transferred to a neurosurgical unit. International Surgery, 77, 297.
Gentleman, D. and Jennett, B. (1981) Hazards of interhospital transfer of
comatose head injured patients. Lancet, ii, 853.
Germon, T. J., Picydell-Pearce, C. W. et al. (1995) Can near infrared
spectroscopy detect changes in regional oxyhaemoglobin due to functional
activation? Journal of Neurology, Neurosurgery and Psychiatry, 60, 2421.
Gopinath, S. P,. Robertson, C. S., Contant, C. F. et al. (1994) Jugular venous
desaturation and outcome after head injury. Journal of Neurology, Neurosurgery and Psychiatry, 57, 717723.
Graham, D. I., Adams, J. H. and Doyle, E. (1978) Ischaemic brain damage in
fatal non-missile head injuries. Journal of the Neurological Sciences, 39,
213234.
Graham, D., Ford, I. et al. (1989) Ischaemic brain damage is still common in
fatal non-missile head injury. Journal of Neurology, Neurosurgery and
Graham, D., Jennett, B. et al. (1995) The structural basis of the vegetative state
after acute head injury. Journal of Neurotrauma, 12, 345.
Green, M., Marciano, F. et al. (1995) Traumatic subarachnoid haemorrhage:
impact on clinical course and outcome during acute hospitalisation in nonpenetrating head injury. Journal of Neurotrauma, 12, 476.
Group of Neurosurgeons (1984) Guidelines for initial management after head
injury in adults. British Medical Journal, 288, 983.
Grumme, T., Baethmann, A. Kolodziejcvyk, D. et al. (1995) Treatment of
patients with severe head injury by triamcinolone a prospective,
controlled multicenter clinical trial of 396 cases. Research in Experimental
Medicine, 195, 217229.
Hallenback, J., Dutka, A. et al. (1986) Polymorphonuclear leucocyte accumulation in brain regions with low blood flow during the early postischaemic
period. Stroke, 17, 246.
Hatfield, R., Gill, R. et al. (1992) Dose response relationship and therapeutic
window for Dizocilpine (MK801) in a rat focal ischaemia model. European
Journal of Pharmacology, 216, 17.
Hewer, R. and Wood, V. (1989) Availability of computed tomography of the
brain in the United Kingdom. British Medical Journal, 298, 1219.
Ito, J., Barzo, P., Fatouros, P. et al. (1995) Characterization of edema by
diffusion weighted imaging following closed head injury and secondary
insult in the rat. Journal of Neurotrauma, 12, 475.
Jackowski, A. (1992) Disordered sodium and water in neurosurgery. British
Jamieson, K. and Yelland, J. (1968) Extradural hematoma: report of 167 cases.
Janon, E. (1970) Traumatic changes in the internal carotid artery with basal
skull fractures. Radiology, 96, 55.
Jenkins, A., Mendelow, A. D., Graham, D. I. et al. (1990) Experimental
intracerebral haematoma: the role of blood constituents in early ischaemia.
British Journal of Neurosurgery, 4, 4552.
Jones, P., Piper, I. et al. (1993) Microcomputer based detection of secondary
insults and 12 month outcome after head injury. Journal of Neurotrauma,
10(Suppl. 1), S102.
REFERENCES
Jones, P. A., Andrews, P. J. D., Midgley, S. et al. (1994) Measuring the burden
of secondary insults in head-injured patients during intensive care. Journal
of Neurosurgery and Anaesthetics, 6, 414.
Kakarieka, A., Harders, A. et al. (1995) Traumatic subarachnoid haemorrhage:
a clinical entity and its treatment with nimodipine. Journal of Neurotrauma,
12, 375.
Kane, P., Modha, P. et al. (1992) The effect of immunosuppression on the
development of cerebral oedema in an experimental model of intracerebral
haemorrhage: whole body and regional irradiation. Journal of Neurology,
Kane, P., Cook, S. et al. (1994) Cerebral oedema following intracerebral
haemorrhage: the effect of the NMDA receptor antagonists MK-801 and
D-CPPene. Acta Neurochirurgica, 60(Suppl.), 561.
Kingman, A., Mendelow, A. et al. (1988) Experimental intracerebral mass:
description of model, intracranial pressure changes and neuropathology.
Journal of Neuropathology and Experimental Neurology, 47, 128.
Kirkpatrick, P., Smielewski, P. et al. (1995) Detection of cerebral hypoxic events
using near infrared spectroscopy in head injured patients. Journal of
Neurotrauma, 12, 374.
Klatzo, I. (1979) Cerebral oedema and ischaemia, in Recent Advances in
Neuropathology, (ed. W. T. Smith and J. B. Cavanagh) Churchill Livingstone,
Edinburgh, p. 27.
Kohi, Y., Mendelow, A. et al. (1984) Extracranial insults and outcome in
patients with acute head injury relationship to the Glasgow Coma Scale.
Injury, 16, 25.
Kossmann, T., Morganti-Kossmann, M. et al. (1995) Production of complement
proteins in human brain injury correlates to intraventricular cerebrospinal
fluid interleukin-6 levels. Journal of Neurotrauma, 12, 461.
Kunze, S. and Schiefer, W. (1971) Angiographic demonstration of a dissecting
aneurysm of the middle cerebral artery. Neuroradiology, 2, 201.
Lenkinski, R. E., Gennarelli, T. A., McIntosh, T. K. et al. (1995) Magnetic
resonance spectroscopy of the injured brain. Journal of Neurotrauma, 12, 351.
McGraw, C. (1989) A cerebral perfusion pressure of greater than 80 mmHg is
more beneficial, in Intracranial Pressure VII, (eds J. T. Hoff and A. L. Betz),
Springer-Verlag, Berlin.
Macpherson, P. and Graham, D. (1973) Arterial spasm and slowing of the
cerebral circulation in the ischaemia of head injury. Journal of Neurology,
Marmarou, A., Fatouros, P. et al. (1990) In vivo measurement of brain water by
MRI. Acta Nerurochirurgica (Supplement), 51, 123.
Marmarou, A. (1994) Traumatic brain oedema an overview. Acta Neurochirurgica, S60, 421424.
Martin, N. A., Doberstein, C., Zane, C. et al. (1992) Postraumatic cerebral
arterial spasm: transcranial Doppler ultrasound, cerebral blood flow, and
angiographic findings. Journal of Neurosurgery, 77, 575583.
Matthews, D., Matthews, J. et al. (1995) Changes in cerebral oxygen
consumption are independent of changes in body oxygen consumption
after severe head injury in childhood. Journal of Neurology, Neurosurgery and
Matthews, D., Aynsley-Green, A. et al. (1996a) Modified hormonal effects on
fat metabolism following severe head injury in children. Paediatric
Research.
Matthews, D., Sharples, P. et al. (1996) Evidence for progressive changes in
cerebral mitochondrial function in severely head-injured children. Journal of
Maxwell, W. (1995) Microtubular changes in axons after stretch injury. Journal
of Neurotrauma, 12, 363.
Mendelow, A., Karmi, M. et al. (1979) Extradural haematoma: effect of delayed
treatment. British Medical Journal, i, 1240.
Mendelow, A. D., Teasdale, G. M., Jennett, B. et al. (1983) Risks of intracranial
haematoma in head injured patients. British Medical Journal, 287,
11731176.
Mendelow, A., Bullock, R. et al. (1984) Intracranial haemorrhage induced at
arterial pressure in the rat: Part 2. Short term changes in local cerebral blood
flow measured by autoradiography. Neurology Research, 6, 189.
Mendelow, A. D., Teasdale, G. M., Russell, T. et al. (1985) Effect of mannitol on
cerebral blood flow and cerebral perfusion pressure in human head injury.
Mendelow, A., Chambers, I. et al. (1994) Cerebral perfusion pressure: the 60
minute average minimum value and outcome from head injury. Journal of
Neurotrauma, 10(Suppl 1).
Mikulis, D. (1995) Functional imaging in neurotrauma. Journal of Neurotrauma,
12, 351.
Milhorat, T. (1992) Classification of the cerebral oedemas with reference to
hydrocephalus and pseudotumour cerebri. Childs Nervous System, 8, 301.
Miller, J. (1993) Monitoring the brain of patients with head injury: transcranial
Doppler sonography. Journal of Neurotrauma, 10(Suppl. 1), S109.
Mohamed, A., Gotoh, O. et al. (1985) Effect of pretreatment with the calcium
antagonist, Nimodipine on local cerebral blood flow and histopathology
after middle cerebral artery occlusion. Annals of Neurology, 18, 705.
Nath, F., Mendelow, A. et al. (1985) Chronic subdural haematoma in the CT
scan era. Scottish Medical Journal, 30, 152.
Nath, F., Jenkins, A. et al. (1986) Early haemodynamic changes in experimental
intracerebral haemorrhage. Journal of Neurosurgery 65, 697.
87
Nath, F., Kelly, P. et al. (1987) Experimental intracerebral haemorrhage: effects

on blood flow, capillary permeability and histochemistry. Journal of
Nehls, D., Mendelow, A. et al. (1988) Experimental intracerebral haemorrhage:
progression of haemodynamic changes after production of a spontaneous
mass lesion. Neurosurgery, 23, 439.
Nelson, P. B., Seif, S. M., Maroon, J. C. and Robinson, A. G. (1981)
Hyponatraemia in intracranial disease. Perhaps not the syndrome of
inappropriate secretion of antidiuretic hormone (SIADH). Journal of
Nornes, H. (1975) Monitoring patients with intracranial aneurysms. Clinical
Obrist, W. D., Gennarelli, T. A., Segawa, H. et al. (1979) Relation of cerebral
blood flow to neurological status and outcome in head-injured patients.
Ohata, K. and Marmarou, A. (1992) Clearance of brain oedema and
macromolecules through the cortical extracellular space. Journal of Neurosurgery, 77, 387.
Oppenheimer, D. (1968) Microscopic lesions in the brain following head
injury. Journal of Neurology, Neurosurgery and Psychiatry, 31, 299.
Pettus, E. and Povlishock, J. (1995) Evidence for prolonged alterations in
axolemmal permeability following traumatic brain injury. Journal of
Piper, I., Midgely, S. et al. (1995) Intracranial pressure waveform indices
correlate with cerebral perfusion pressure but fail to predict efficacy of
mannitol therapy for raised ICP: a prospective study in head injured
patients. Journal of Neurology, Neurosurgery and Psychiatry, 58, 387.
Poon, W., Mendelow, A. et al. (1989) Secretion of antidiuretic hormone in
neurosurgical patients: appropriate or inappropriate? Australian and New
Zealand Journal of Surgery, 59, 173.
Povlishok, J. (1992) Traumatically induced axonal injury: pathogenesis and
Povlishock, J. (1995) Traumatically induced delayed axotomy triggered by
varied cytoskeletal and axolemmal abnormalities. Journal of Neurotrauma,
12, 363.
Rappaport, Z., Young, W. et al. (1987) Regional brain calcium changes in the
rat middle cerebral artery occlusion model of ischaemia. Stroke, 18, 760.
Rosner, M. (1993) Cyclic CSF pressure waves causally related to systemic
arterial blood pressure, in Intracranial Pressure VIII, (eds C. J. J. Avezaat, J. H.
M. van Eindhoven, A. I. R. Maas and J. T. J. Tans), Springer-Verlag,
Berlin.
Santos, A., Shu, E. et al. (1995) Extradural haematoma in children: analysis of
74 cases. Journal of Neurotrauma, 12, 389.
Sato, O., Bascom, J. et al. (1971) Intracranial dissecting aneurysm: case report.
Schneider, P., Rossman, M. et al. (1988) Effect of internal carotid artery
occlusion on intracranial haemodynamics: transcranial Doppler evaluation
and clinical correlation. Stroke, 19, 589.
Schroder, M., Muizelaar, J. et al. (1993) Regional ischemia and regional
cerebral blood volume (CBV) after severe head injury in man. Journal of
Neurotrauma, 10(Suppl. 1), S57.
Schultz, U., Kutemeyer, et al. (1984) Traumatic occlusion of both internal
carotid arteries. Journal of Neurology, 231, 233.
Seelig, J. M., Becker, D. P., Miller, J. D. et al. (1981) Traumatic acute subdural
haematoma. Major mortality reduction in comatose patients treated within
four hours. New England Journal of Medicine, 304, 15111518.
Sharples, P. M., Stuart, A. G., Matheras, D. S. et al. (1994) Cerebral blood flow
and metabolism in children with severe head injuries. Part I: Relation to
age, Glasgow Coma Score, outcome, intracranial pressure and time after
injury. Journal of Neurology, Neurosurgery and Psychiatry, 57, 145152.
Sharples, P., Matthews, D. et al. (1995) Cerebral blood flow and metabolism in
children with severe head injuries. Part 2: Cerebrovascular resistance and its
determinants. Journal of Neurology, Neurosurgery and Psychiatry, 58, 153159.
Shigeno, T., Brock, M. et al. (1982) The determination of brain water content:
microgravimetry versus dryingweighing method. Journal of Neurosurgery,
57, 99.
Sinar, E., Mendelow, A., Graham, D. and Teasdale, G. (1987) Experimental
intracerebral haemorrhage: effects of a temporary mass lesion. Journal of
Stevenson, G., Brown, H. et al. (1964) Chronic venous epidural haematoma at
the vertex. Journal of Neurosurgery, 21, 887.
Strachan, R., Kane, P. et al. (1992) Immunosuppression by whole-body
irradiation and its effect on oedema in experimental cerebral ischaemia.
Acta Neurologica Scandinavica, 86, 256.
Stringer, W. and Kelly, D. (1980) Traumatic dissection of the extracranial
internal carotid artery. Journal of Neurosurgery, 6, 123.
Teasdale, G. (1995) Head injury. Journal of Neurology, Neurosurgery and
Teasdale, G., Murray, G. et al. (1990) Risks of acute traumatic intracranial
haematoma in children and adults: implications for the management of
head injuries. British Medical Journal, 300, 363.
Todd, N. V. and Teasdale, G. M. (1989) Steroids in human head injury. Clinical
studies, in Steroids in Diseases of the Central Nervous System, (ed. R. Capildio),
John Wiley & Sons, New York, pp. 151161.
88
Treadwell, l. and Mendelow, A. (1994) Audit of head injury management in

the Northern Region. British Journal of Nursing, 3, 136.
Tsuji, O., Marmarou, A. et al. (1995) Electrolyte and excitatory amino acid
surge following closed head injury and secondary insult in the rat. Journal
Ueda, T., Kikuchi, H. et al. (1986) Traumatic stenosis of the internal carotid
artery in children. Surgical Neurology, 26, 368.
Van der Heever, C. and van de Merwe, D. (1986) Management of depressed
skull fractures: selective conservative management of non-missile injuries.
Vink, R. (1995) Functional magnetic resonance characterization of the injured
brain: can MR provide clinically relevant information beyond conventional
imaging? Journal of Neurotrauma, 12, 351.
Ward, J. D., Becker, D. P., Miller, J. D. et al. (1985) Failure of prophylactic
barbiturate coma in the treatment of severe head injury. Journal of

Wilkins, R. and Odom, G. (1970) Intracranial arterial spasm associated with
craniocerebral trauma. Journal of Neurosurgery, 32, 626.
Wong, F., Piper, I. et al. (1995a) Determination of cerebral perfusion pressure
thresholds in head injury patients. Journal of Neurology, Neurosurgery and
Wong, F., Piper, I. et al. (1995b) Differences in treatment thresholds of cerebral
perfusion pressure in head injuries. Journal of Neurotrauma, 12, 413.
Wyper, D., Hadley, D. et al. (1995) Recovery after closed head injury:
functional image correlates. Journal of Neurotrauma, 12, 472.
Zelenock, G. B., Kazmers, A. et al. (1982) Extracranial internal carotid artery
dissections. Non-iatrogenic traumatic lesions. Archives of Surgery, 117,
425.
Brain metabolism and cerebral

blood flow
5.1 The basic principles of brain

metabolism and blood flow
5.1.1
(a)
CEREBRAL HEMODYNAMICS
Anatomical and physiological considerations
Cerebral blood flow is influenced and regulated by a

number of factors, including arterial blood pressure,
intracranial pressure, venous outflow, blood viscosity,
PaCO2 and PaO2, collateral flow, vasoreactivity and the
status of cerebral autoregulation. The regulation of
vascular resistance lies mainly in the arterioles and
precapillary segments (Auer and Loew, 1983). It is
important to note that cerebral metabolism is the
major determinant of regional blood flow.
Anatomy
Each carotid artery contributes approximately 40% to
the total cerebral perfusion, the remaining 20% coming from the two vertebral arteries, which fuse to form
the basilar artery. Autopsy studies, however, show
that there are many variations in the vascular anatomy
and a normal circle of Willis was seen in only 52% of
brains (Alpers, Berry and Paddison, 1959). Collateral
channels occur on the pial surface or within the brain
itself. These channels cross the arterial boundary
zones (sometimes termed watershed zones) and the
severity of focal ischemia after systemic hypotension
or occlusion of a major vessel is determined by these
collaterals.
The major fraction of venous blood draining from
the brain is collected in the transverse sinuses which,
together with the inferior petrosal sinuses, form the
internal jugular veins. The cerebral venous drainage
territories are not sharply delineated and overlap. It is
estimated that approximately 70% of blood in each
internal jugular vein originates in the ipsilateral and
30% in the contralateral hemisphere. Approximately
3% of blood in the internal jugular vein is from
extracranial sources, mainly via the superior petrosal

sinus (Jennett, Miller and Harper, 1976).
The distribution of capillaries is functionally
organized throughout the central nervous system. The
capillary density may provide an anatomical indicator
of oxidative metabolism. Brain areas with high basal
levels of glucose metabolism contain a high density of
capillaries. Some 90% of cerebral capillaries are continuously perfused and 10% are subject to recruitment
in response to metabolic and other stimuli (Sokoloff et
al., 1977). A close relationship exists between aerobic
metabolism and the microvasculature.
(b) Cerebral perfusion pressure and
cerebrovascular resistance
The net driving force for the cerebral circulation is
defined as the cerebral perfusion pressure (CPP),
which is the mean arterial blood pressure (MABP)
minus the cerebral venous pressure. If the pressure
within the thin-walled cortical veins and the bridging
veins in the subarachnoid space is equal to or less than
the external (CSF) pressure, the veins may collapse
and increase the resistance to flow. Under most
circumstances, however, the pressure in these veins is
slightly above extravascular, (that is intracranial pressure), in order to permit continuous flow. The venous
sinuses do not collapse completely because of their
stiff walls. Because of the close relationship between
cerebral venous pressure and ICP, CPP is generally
defined as the difference between arterial pressure
and ICP. In a normal brain, changes in CPP between 50
and130 mmHg produce only minimal changes in CBF.
A constant flow in this range of CPP is maintained by
an increase or decrease in vascular resistance (Harper,
1966). Outside this range, cerebral vasodilatation or
vasoconstriction cannot maintain a normal cerebral
blood flow (CBF). Following ischemia, trauma or
subarachnoid hemorrhage, the normal autoregulatory
relationship may be lost or its range altered such that
a CPP of about 70 mmHg may be needed to prevent
90
BRAIN METABOLISM AND CEREBRAL BLOOD FLOW
ischemic brain damage. A mean systemic blood

pressure of 60 mmHg, which is satisfactory in the
normal individual, may result in a profound fall in
CBF, such that ischemic brain damage may occur after
only 1530 minutes.
Thus, CPP = MABP ICP and CBF = CPP/CVR
where CVR is cerebrovascular resistance.
Blood is a non-Newtonian fluid and viscosity is
velocity-dependent. This, with the pressure loss across
cerebral circulation and the complex geometry of the
brain circulation, makes it difficult to calculate CVR.
However, although the law of Poiseuille describes
steady laminar flow of Newtonian fluids in cylindrical
tubes, it can be used to describe CBF and the CVR.
F = P / [(8 l)/(r4)]
(P = pressure; F = flow; = viscosity; l = length; r =
radius).
Since vascular resistance is inversely proportional to
the fourth power of the vessel diameter, a modest
change in cerebral vessel diameter will produce a
marked change in flow resistance
Most resistance to the cerebral circulation lies within
the arterioles and precapillary sphincters. There is a
pressure drop from 90 mmHg in the conductance
arteries to 2030 mmHg in the capillaries across these
resistance vessels. The resistance arterioles are able to
regulate their own diameter through their vasomotor
smooth muscle. They are supplied with a complex
peptidergic nerve fiber network with cell bodies in
various brain-stem nuclei such as the dorsal raphe
nuclei.
(c)
Effects of hematocrit on CBF
Blood viscosity is determined by a number of factors,

including erythrocyte size and concentration, shear
rate, temperature, pH, plasma protein level and plasma
lipid concentration. Clinical studies show an inverse
relationship between CBF and hematocrit. A hematocrit of 47% is associated with a significantly reduced
CBF, whereas lowering the hematocrit within the
physiological range, e.g. by volume expansion, increases CBF. (Kuschinsky et al., 1972; Humphrey et al.,
1979) A similar inverse relationship can be seen
between red cell concentration and CBF (Thomas et al.,
1977). On the other hand, reduced CBF increases blood
viscosity and stasis may occur. A fall in PO2 and pH
increases red cell stiffness and may decreases CBF
further. In normal brain, autoregulatory mechanisms
can compensate for changes in blood viscosity by
changing the cerebrovascular resistance. These changes take place in the resistance, not the conductance
vessels (Hudak et al., 1989; Giller et al., 1993; Muizelaar
et al., 1992). If viscosity increases, vasodilation occurs to
overcome the increase in cerebrovascular resistance.
5.1.2
(a)
REGULATION OF CBF
Arterial gas tension and cerebral vasoreactivity
Minimal changes in arterial PCO2 lead to variations of

the cerebral perfusion. CO2 molecules are highly
diffusible and pass the bloodbrain barrier easily. The
glycolytic pathway, the major energy-producing
source in the brain, produces six moles of CO2 per
mole of glucose. Hypercapnia relaxes cerebrovascular
smooth muscles, whereas hypocapnia produces vasoconstriction. An arterial PCO2 of 2025 mmHg may
reduce the CBF by 4050%, and conversely an increase
in PCO2 over 50 mmHg increases CBF by more than
50%.(Olesen, Paulson and Lassen, 1971) These changes in CBF due to alterations in CO2 occur almost
immediately in healthy brain, but the response may be
altered after head injury (Bouma and Muizelaar,
1992).
Carbon dioxide induces changes in the hydrogen ion
concentration (modulating extracellular pH), and it is
most likely this which acts on the smooth muscle,
rather than PCO2 (Kontos, Raper and Patterson, 1977;
Wahl et al., 1970). Experimental studies have shown
that HCO3 solutions decrease CBF whereas low
HCO3 solutions increase CBF, supporting the importance of tissue pH (Cameron and Caronna, 1976).
Various brain structures have different densities of
perfused capillaries, but the number of perfused
capillaries is unchanged during hypercapnia
(although blood flow is increased) compared to
normocapnia (Gobel et al., 1989). Extracellular pH may
also modulate the action of vasomotor agents such as
norepinephrine (noradrenalin). In isolated pial vessels,
the contractile response induced by catecholamines
was reduced by 50% at a pH of 7.0 (Edvinsson and
Sercombe, 1976). Certain pharmacological agents and
therapeutic techniques influence cerebrovascular reactivity via changes in gas tension. The carbonic anhydrase inhibitor acetazolamide decreases extracellular
brain pH at a constant arterial PCO2, and increases
normocapnic CBF. However this effect lasts only for a
short time (Ehrenreich et al., 1961; Heuser et al., 1975).
Barbiturates depress cerebrovascular reactivity but the
mechanisms involved are not well understood. Barbiturates lower both cerebral metabolism and CBF and
may reduce CO2 reactivity (Kassell et al., 1981;
Edvinsson and McCulloch, 1981). In clinical studies,
diminished CO2 reactivity is associated with a worse
prognosis and with decreased effectiveness of barbiturates (Cold, 1989; Nordstrom et al., 1988; Bouma and
Muizelaar, 1992; Muizelaar et al., 1991; Giller et al.,
1993). If vasoreactivity is preserved, patients with
high ICP may respond to barbiturate therapy with
further reduction in energy consumption and reduced
CBF and volume. In these patients with preserved
reactivity, prolonged hyperventilation may diminish
BASIC PRINCIPLES OF BRAIN METABOLISM AND BLOOD FLOW
brain oxygen delivery and it must be used with caution

(Muizelaar et al., 1991).
(b)
Autoregulation and CBF
In the normal brain CBF remains relatively constant

despite changes in the systemic blood pressure and
CPP. These vascular responses are also true for changes
in CPP due to ICP increase. This phenomenon, called
cerebral autoregulation, has a range of between 40 and
160 mmHg. The limits of autoregulation are not fixed
and are influenced by both pharmacological agents and
pathological conditions (systemic as well as intracranial). Autoregulation describes a pressureflow
homeostatic relationship of the cerebral circulation and
several hypotheses have been formulated to explain it.
However, it is most probably controlled and regulated
by the interplay of several mechanisms.
The metabolic hypothesis
This proposes that a reduction in CBF results in the
release of chemical mediators. Regional CBF is then
regulated by constriction and dilation of cerebral
arterioles and opening and closing of precapillary
sphincters in the metarterioles, which are responsible
for distribution of blood through the cerebrovascular
bed. But a coupling of CBF autoregulation to vasoactive metabolites or molecules is not well defined. CO2,
O2, K+, Ca2+, H+ and adenosine have all been
proposed as playing a role (Kuschinsky and Wahl,
1978). Hypercapnia and hypoxia are powerful stimulants for CBF increase. Adenosine, which is formed
from ATP breakdown during energy consumption, is a
potent vasodilator. It increases with reduction in blood
pressure, but the role of adenosine and adenosine
phosphorylated derivatives in controlling CBF is still
unsolved (Winn et al., 1980; Kontos, 1985). Tissue
oxygen pressure may also be a metabolic component
of the control of cerebral autoregulation. This
O2-sensitive mechanism will cause cerebral arteriolar
dilation in response to increased venous pressure, an
effect which can be reversed by local hyperoxia (Wei
and Kontos, 1988).
The myogenic hypothesis
The short latency of the autoregulatory response is
considered to be an argument in favor of a myogenic
mechanism, whereby smooth muscle cells of small
arteries and arterioles constrict or dilate in response to
changes in transmural pressure. A rapid change in
intravascular pressure alters the state of the actin and
myosin filaments within the smooth muscle cells.
Experimentally, changes in arterial transmural pressure correlate positively with changes in smooth
91
muscle action potentials and membrane potentials

which lead to spontaneous activation (Harder et al.,
1989; Kontos et al., 1978).
Endothelial-cell-related factors have also been
implicated in autoregulation. Endothelium-derived
relaxing factor (EDRF) has been identified as nitric
oxide (NO) or a NO-containing substance produced in
cerebral blood vessels. EDRF action is mediated by the
activation of guanylate cyclase and formation of cyclic
GMP. Ca2+/calmodulin activates NO synthetase,
which generates NO from L-arginine. Stimulated
neurons and glia also produce NO and may mediate
local vasodilation (Furchgott and Vanhoutte, 1989;
Faraci, 1993; Kim and Vanhoutte, 1990). An impaired
endothelium-dependent response with decreased production of EDRF but increased release of endothelium-derived contracting factors, such as endothelin
and arachidonic acid, has been found in subarachnoid
hemorrhage. Endothelin is an extremely potent and
long-lasting vasoconstrictor peptide. In experimental
focal cerebral ischemia, activation of nitric oxide
synthetase with a burst of NO and activation of
guanylate cyclase has been found (Kader et al.,
1993).
5.1.3
(a)
CEREBRAL ENERGY METABOLISM
Cerebral energy generation
The brain comprises only 2% of body weight, but

receives 15% of cardiac output and uses 20% of total
body oxygen and 25% of total body glucose. Within
normal levels of global CBF the brain extracts about
50% of the oxygen and 10% of the glucose from
arterial blood. High-energy phosphates, predominantly adenosine triphosphate (ATP), are the most
important energy source for the brain. ATP is produced almost entirely by the oxidative metabolism of
glucose.
Complete aerobic oxidation of one molecule of
glucose yields 38 molecules of ATP (via the glycolytic
and tricarboxylic acid pathway), whereas anaerobic
breakdown of one molecule of glucose produces only
two molecules of ATP. Carbon dioxide, the endproduct of aerobic oxidation, is easily eliminated from
the brain and crosses the bloodbrain barrier (BBB).
On the other hand lactic acid, the end-product of
anaerobic glycolysis, is toxic to neurons and decreases
the pH. A severalfold increase in cerebral tissue lactate
concentration can be seen in ischemic animal models,
where a low level of cerebral perfusion provides some
glucose but insufficient oxygen. Experimental data
show that gray matter uses about three times more
glucose than white matter, although there are considerable inter-regional variations (Abrams et al., 1984;
Phelps, Mazziotta and Schelbert, 1986).
92
It is speculated that neurons consume about 75% of

all oxygen in the CNS. About 80% of all energy
generated may be necessary for the maintenance of
ionic gradients. Glial cells, which make up almost 50%
of the brain volume, have a much lower metabolic rate
and account for less than 10% of total cerebral
metabolism (Siesjo,
1984).
(b) The effect of the bloodbrain barrier on brain
metabolism
The bloodbrain barrier comprises morphological and
functional mechanisms that restrict or facilitate the
passage of substances from blood to brain. Endothelial cells of cerebral capillaries are sealed together by
tight junctions. They contain a large number of
mitochondria and pinocytic vesicles that facilitate
the transport of substances into the brain. Astrocytes
also play an important role in the normal functioning
of the BBB. The uptake of a substance by the brain
depends on its polarity, the presence of carrier
mechanisms, pH, enzymes and the size of the molecules. The endothelial cells are not simply a barrier
between cerebrovascular smooth muscle and the
blood. Endothelium produces vasoactive substances
(see above), such as endothelin-1 which is a potent
vasoconstrictor of cerebral arteries, and vasodilators,
NO and acetylcholine (Faraci, 1993; Robinson and
McCulloch, 1990). Endothelial cells are also part of the
mechanism of hemodynamic regulation via their
responses to changes in transmural pressure (Harder
et al., 1989). Oxygen passes from the blood to the brain
by diffusion, whereas glucose crosses the BBB via a
saturable mechanism that is activated by hypoxia or
hypoglycemia. Bloodbrain barrier mechanisms may
be disturbed by head injury but little is known about
the long-term effects of BBB dysfunction in human
head injury (Sokrab et al., 1988).
(c)
tissue. Because the movement of oxygen depends on

partial pressure gradients, the cells furthest away from
capillaries are most vulnerable to hypoxia. Oxygen
delivery is also dependent on CBF. Leniger-Follert et
al. (1975) found, in cats, that brain PO2 remained
constant if the blood pressure was kept between
50150 mmHg. However, below a blood pressure of
50 mmHg there was a dramatic drop in brain PO2.
Local PO2 may be regulated by vasoconstriction
during increased arterial oxygen supply. However,
local regulation is abolished by tissue anoxia or by
adding CO2 to the inspired air (Leniger-Follert, Gronczewski and Danz, 1984; Leusen, Weyne and Demeester, 1980; Lubbers,
1968).
Several research groups have investigated continuous oxygen measurements in brain tissue. Major
drawbacks in the past were the considerable drifting of
oxygen sensors over time and the labor-intensive
calibration procedures (Fleckenstein et al., 1990). In
their animal studies, Maas et al. (1993) found a brain
tissue oxygen of 28 mmHg under normocapnia and
normal CPP, but when CPP was lowered to 40 mmHg
or less there was a sharp decline in brain tissue oxygen
tension. Recently a multiparameter probe has become
available for measuring brain PO2, PCO2, pH and
temperature in one combined sensor. Studies at the
Medical College of Virginia have demonstrated stable
tissue PO2 (42 9 mmHg), PCO2 (58 14 mmHg) and pH
(7.0 0.2; Zauner et al., 1995). Hypoxemia, hypocapnia
and hypercapnia significantly changed tissue PCO2,
whereas PO2 was markedly changed during hypoxemia and hypocapnia. On occluding the middle
cerebral artery to produce focal ischemia, PO2 fell to
19 6 mmHg (53%) immediately and stayed low
throughout the experiment, whereas PCO2 increased to
71 11 mmHg (+ 20%; Zauner et al., 1995; Figure 5.1).
Brain tissue oxygen and carbon dioxide
Brain function and tissue integrity are highly dependent on a continuous supply of oxygen and clearance of
CO2. Tissue oxygenation is the product of the blood
flow and arterial O2 content. The percentage of
hemoglobin carrying oxygen depends on several factors, the most important of which is the partial
pressure of oxygen (PO2). The affinity of oxygen for
hemoglobin is increased by decrease in temperature,
hydrogen ion concentration (increased pH) or PCO2,
shifting the oxyhemoglobin curve to the left so that
release of oxygen into tissue is decreased. Conversely,
the release of oxygen into tissue is increased when the
curve is shifted to the right. However hypoxemia
shifts the dissociation curve to the right, decreases O2
loading and consequently reduces O2 unloading in the
Figure 5.1 Example of changes in brain oxygen, PCO2 and

pH in a feline model of focal ischemia (middle cerebral
artery occlusion), using a multiparameter probe. The onset
of middle cerebral artery occlusion is shown by the arrow.
BASIC PRINCIPLES OF BRAIN METABOLISM AND BLOOD FLOW
This technique was applied to 18 severely head-injured

patients. In patients who made good recoveries
following severe head injury, brain PO2 was
35 8 mmHg, brain PCO2 was 50 9 mmHg, brain pH
was 7.12 0.15 and brain temperature was 37.8 0.8C.
93
However, patients with poor outcome (died or vegetative state) had a brain tissue oxygen of 15 6 mmHg,
a brain PCO2 of 67 23 mmHg, a brain pH of 6.95 0.4
and a brain temperature of 36.3 2.4C (Zauner,
Bullock and Young, 1995a; Figure 5.2).
Brain P CO2 and P O2 (mmHG)
(a)
Br P CO2
Br P O2
Glucose
Lactate
(b)
Figure 5.2 (a) Brain PO2, PCO2, lactate, glucose and ICP plotted against time in a patient with good outcome. The brain PO2
increases in the first 24 hours after injury. Note the drop in brain lactate in the first 36 hours after injury. (b) Brain PO2, PCO2,
and dialysate glucose and lactate in a patient with arterial hypoxemia (PaO2-55 mmHg on admission to ICU). Note that brain
CO2 was very high (pulmonary contusions). Hyperventilation at FiO2 1.0 failed to improve brain PO2 to > 20 mmHg, and
dialysate lactate remains very high. (The patient died 9.5 hours after monitoring commenced.)
94
5.1.4
(a)
NEUROTRANSMITTERS AND CBF
Metabolic effects of neurotransmitters
The close relationship between regional CBF and

glucose utilization is generally accepted. CBF increases with increased neuronal activity and also during
tissue hypoxia. It is thought that PCO2, pH and
adenosine serve as coupling factors, but the mechanisms involved cannot be fully explained. Extracellular K+ affects membrane function and cerebrovascular tone (Kuschinsky et al., 1972). Astrocytes may
play a role by raising K+ concentration around
cerebral arterioles.
The role of neurotransmitters with vasomotor
effects on cerebral vessels and cerebral metabolism is
less well understood. The presence of dopamine
receptors (D2) in neurons and cerebral blood vessels
suggests that they can increase CBF and glucose
metabolism (Sharkey and McCulloch, 1995). Different
cerebral regions have a variable amount of - and adrenoreceptors. The literature reports a diversity of
roles for catecholamines and both vasoconstriction
and vasodilation have been reported. It seems that the
systemic application of catecholamines has little
direct effect on cerebral metabolism. However, if the
bloodbrain barrier is disrupted, CBF and cerebral
metabolic rate of oxygen are increased (Edvinsson,
1982). The effect of gamma-aminobutyric acid
(GABA), one of the most widely distributed neurotransmitters in the human brain, on CBF and metabolism is not known. Both increased and decreased
regional CBF have been reported after systemic
administration (Alborch et al., 1984; Kelly and McCulloch, 1983).
An intrinsic neuronal pathway that may influence
and partially regulate the cerebral circulation and
metabolism has received attention, although the
results of experimental studies are diverse. For example, electrical stimulation of the cerebellar fastigial
nucleus increased cerebral cortical blood flow, but no
change in local metabolism and glucose utilization
was observed (Nakai et al., 1983).
5.2 Normal values for CBF and

metabolism
The most widely used clinical techniques for measuring CBF are described in detail in Chapter 11. In
summary, quantitative blood flow measurement is
available using:

stable xenon-enhanced computer tomography;

inhalational i.v. xenon-133;
PET scanning;
MRI angiography and flow measurement by oxyhemoglobin spectral shifts, and contrast indication

transit times this is currently being investigated

for use in patients with acute severe head injuries;
SPECT (Section 9.5);
laser Doppler flowmetry and the thermal diffusion
techniques these are useful for monitoring progressive changes but they are essentially
qualitative;
transcranial Doppler sonography;
near-infrared spectroscopy the last two methods
are non-invasive and are being used more frequently at the bedside, but they do not measure
CBF directly, and their value in the intensive care
management needs further evaluation;
continuous monitoring of jugular venous oxygen
saturation to give an assessment of global cerebral
oxygenation and an indirect indication of CBF;
brain tissue sensors for oxygen and carbon dioxide
these allow direct substrate delivery measurements but are still undergoing experimental
evaluation.
The normal values for xenon CBF and jugular bulb

oximetry are summarized below. Other techniques,
which measure relative quantitative or qualitative
changes, are described in Chapter 11.
5.2.1
XENON CBF MEASUREMENTS
A wide range in CBF values has been reported for

different brain locations measured by the stable (nonradioactive) xenon and the xenon-133 methods (Risberg et al., 1975; Austine et al., 1972; Gur, Yonas and
Good, 1989; Gur, Good and Yonas, 1992; Nakano et al.,
1992). Dettmers found a mean hemispheric CBF value
of 44.4 5.7 ml/100 g/min in 17 normal volunteers
(mean age 29 years; Dettmers et al., 1992). In a recent
series of nine healthy volunteers of similar age, we
found that the mean hemispheric CBF was 54.9 9.0.
Different clinical approaches for CBF evaluation
necessitate the differentiation of various brain compartments. Meyer et al. reported a CBF of 52.8 10.9,
57.7 15.2 and 15.8 3.3 ml/100 g/min for frontal cortex, thalamus and frontal white matter respectively in
22 neurologically normal volunteers. They also noted
a decline in cortical gray matter blood flow with
advancing age (Meyer et al., 1992). The variations in
CBF using stable xenon in different studies are most
likely due to the limitations in the spatial resolution of
computed tomography, the use of different xenon
concentrations and different durations of xenon inhalation, as well as biological variation. Most of the
current stable xenon equipment and calculations are
for adult patients, and less is known about CBF in
infants and children. However, Suzuki et al. used the
intravenous xenon-133 clearance method to study CBF
in 80 children, aged from 1 week to 15 years (Suzuki,
1990). Within the first weeks of life CBF was
BLOOD FLOW AND METABOLISM FOLLOWING HEAD INJURY
2060 ml/100 g/min. Between 6 months and 1 year

mean CBF was 74 10 and between 1 and 2 years
101 12. There was a peak of 108 5 between 2 and 4
years, followed by a slow decline to 90 10 between 4
and 9 years, and a further decline to 71 12 ml/
100 g/min between the ages of 9 and 15.
As noted earlier, the brain has little storage capabilities for ATP and glucose and virtually no reserves of
oxygen. Therefore CBF must be tightly coupled to
cerebral metabolism. At a normal serum glucose
concentration, the influx of glucose to the brain is
twice the glucose utilization. However, if serum
glucose falls below 2.5 mol/ml the availability of
glucose to the brain is reduced (Robinson and Rapoport, 1986). The average cerebral metabolic rate for
glucose (CMRG) in adults is 0.325 mol/g/min (Hatazawa et al., 1988). In normal brain more than 90% of
glucose usage is met by aerobic metabolism and the
remaining by anaerobic glycolysis, which yields lactate as the end product (Hatazawa et al., 1988). The
cerebral metabolic rate of lactate (CMRL) can be
positive or negative depending on whether there is
uptake or excretion of lactate by the brain. Normally
under aerobic conditions, the CMRL is 0.02 mol/
g/min (Ritter and Robertson, 1994). The normal
cerebral metabolic rate for oxygen (CMRO2) in healthy
adults ranges from 1.83.9 ml of oxygen/100 g/min
(mean 3.4 = 1.5 mol/g/min; Muizelaar and Schroder,
1994). After head injury CMRO2 is often reduced to an

average of 0.9 mol/g/min. Less than 0.6 mol/g/min
is insufficient to maintain normal cellular function
(Ritter and Robertson, 1994). The cerebral arteriovenous oxygen difference (AVDO2), representing the
amount of oxygen extracted by the brain, ranges from
4.58.5 ml (mean 6.5) O2/100 ml of blood. In a normal
coupled relationship between AVDO2 and CBF
(AVDO2 = CMRO2/CBF), AVDO2 remains unchanged
when the CMRO2 changes. However if CMRO2
remains constant, changes in AVDO2 reflect uncoupled changes in CBF (Robertson et al., 1989). If CBF
decreases following head injury, AVDO2 will increase
as the brain compensates by extracting a greater
amount of oxygen. A further uncompensated decline in
CBF leads to ischemia and a fall in CMRO2 and an
increase in the cerebral lactate production.
The average jugular venous saturation (SjvO2) in
normal adults is 65%, ranging from 5571%. However
in comatose patients, values greater than 50% are
considered normal (Cruz, 1993a, b; Sheinberg et al.,
1992). Whenever SjvO2 drops below 50% for more than
15 minutes, the cause must be sought (Sheinberg et al.,
1992). If technical errors are excluded, a decline in
SjvO2 may be caused by a low PaO2, prolonged
hyperventilation, decreased CPP (< 70 mmHg), vasospasm or, very occasionally, low hemoglobin (Gopinath et al., 1994).
95
5.3 CBF and metabolism following head

injury
The severity of injury to the brain resulting from
mechanical trauma depends not only on the structural
damage inflicted by the impact, but also on complex
pathophysiological events occurring in the first hours
or days. In the past a distinction has been made between
primary and secondary injury. Current research is
seeking ways of intervening before or during the
secondary events to correct metabolic derangements.
A classification of head injury based on initial CT
findings was introduced by the Traumatic Coma Data
Bank (TCDB; Marshall et al., 1991). This classification
shows a strong relationship between CT appearance,
mortality and the frequency of increased ICP in the
early phase of injury. It is also helpful in the early
prediction of outcome in these patients. The stable
xenon technique can be performed at the same time as
a diagnostic CT scan, thus allowing a better interpretation of the CBF results.
In keeping with the known heterogeneity of traumatic brain injury, considerable variations in CBF
values, ranging from abnormally low to abnormally
high flow, are found. There is little information on the
time course and significance of CBF changes in the
early hours following brain injury. Both very low and
very high blood flow values have been associated with
poor outcome. Low CBF represents either a state of
ischemia or reduced brain metabolism and is seen
especially in patients with severe head injury. On the
other hand, hyperemia or luxury perfusion due to
uncontrolled vasodilation contributes to brain swelling and high ICP and may also be associated with
poor outcome (Obrist et al., 1984; Bouma and Muizelaar, 1992).
5.3.1 CEREBRAL ISCHEMIA AFTER SEVERE HEAD
INJURY
Decreased CBF is frequently seen following severe

head injury. It may be secondary to depressed metabolism rather than a sign of cerebral ischemia (Obrist et al.,
1984; Adams and Graham, 1972). Many of the earlier
studies of CBF and AVDO2 were performed several
hours after injury and in several of these reports, CBF
values below the threshold of ischemia (18 ml/
100 g/min) and critically high values of AVDO2,
indicative of ischemia, were extremely rare (Obrist et
al., 1984; Bouma and Muizelaar, 1992; Adams and
Graham, 1972; Jones et al., 1981; Bruce et al., 1973;
Muizelaar and Schroder,
1994). Autopsy studies,

however, have indicated that cerebral ischemic damage
is present in 80% of patients dying from severe head
injury (Adams and Graham, 1972; Jennett et al., 1973).
Furthermore, experimental work suggests that head
96
injury enhances the vulnerability of brain tissue to

ischemia and that this vulnerability persists for at least
24 hours following the injury (Jenkins et al., 1989).
Using the stable xenon CT technique with the initial
diagnostic CT scan, Schroder
et al. (1995) confirmed

that the threshold for global ischemia of 18 ml/
100 g/min was accurate and that most patients with
CBF values below this value will develop low-density
changes on CT scans and die within 48 hours. They
assumed that these very low CBF values last for a few
hours only and are therefore missed in studies
performed beyond 4 hours after injury. It is generally
assumed that if CBF stays above 18 ml/100 g/min but
below 2025 ml/100 g/min neurons will survive but
may not function. However, recent research suggests
that infarction will ensue with a CBF of 5 ml/
100 g/min sustained for more than 1.5 hours, 10 ml/
100 g/min for more than 3 hours, 15 ml/100 g/min for
more than 3.5 hours or 18 ml/100 g/min for more than
4 hours (Obrist et al., 1984; Bouma et al., 1991, 1992;
Heiss, Hayakawa and Waltz, 1976; Schroder,
Muizelaar and Kota, 1994).

CMRO2 is typically reduced after severe head injury
to 0.61.2 mol/g/min. This decrease in metabolic
requirements may protect against cerebral ischemia
caused by the injury (Ritter and Robertson, 1994).
Salvant and Muizelaar suggested that a parallel
reduction in CBF and CMRO2 without an increase in
AVDO2 is consistent with diminished cerebral metabolic requirements or reduced capacity for oxidative
metabolism (Salvant and Muizelaar, 1993). If CBF falls
below the ischemic level of 18 ml/100 g/min, CMRO2
is low and AVDO2 is increased, usually to over
48 ml/100 ml blood, as the brain compensates by
extracting a greater amount of oxygen (Ritter and
Robertson, 1994; Obrist et al., 1984; Robertson, Narayan and Gokaslan, 1989). If ischemia continues, an
increase in CSF lactate and CMRL is observed. Very
high cerebral lactate levels can be seen in patients who
die early after severe head injury (Robertson, Narayan
and Gokaslan, 1989). In theory, it is possible to
differentiate viable brain at severe risk for ischemia
(low CBF, high AVDO2, low CMRO2), from compensating and low-risk brain (low CBF, normal
AVDO2, low CMRO2), and from irreversible brain
damage (low CBF, low AVDO2, very low CMRO2;
Muizelaar and Schroder,
1994).
With the ability to measure global cerebral oxygen
saturation via a jugular bulb catheter either continuously or by intermittent blood samples, cerebral
desaturation (defined as a sustained jugular bulb
oxygen saturation less than 55% for more than 15
minutes) might theoretically be detected early (Sheinberg et al., 1992). However in some studies, more than
50% of the fiberoptic catheter readings are not reliable
and major improvements in the equipment are needed
to make this technique more widely applicable. Our

studies with continuous brain tissue PO2, PCO2, pH
and temperature monitoring via a single fiber suggest
that this may be a better way to continuously monitor
cerebral oxygenation and metabolism (Zauner et al.,
1995; Zauner, Bullock and Young, 1995a, b, c).
5.3.2 CEREBRAL HYPEREMIA AFTER SEVERE HEAD
INJURY
Hyperemia is defined as CBF in excess of metabolic

demands and is frequently seen in severely headinjured patients. When CBF is reduced, CBF and
CMRO2 often stay coupled, whereas in hyperemia
uncoupling of CBF and metabolism is more likely
(Obrist et al., 1984; Bruce et al., 1973; Enevoldsen et al.,
1976). Although the etiology is unknown, hyperemia
implies one of:

metabolic derangement;
vasodilation due to loss of vasomotor tone;
severe tissue acidosis (Obrist et al., 1984; Bouma
and Muizelaar, 1992; Lassen, 1966; Langfitt, Weinstein and Kassell, 1965).
Hyperemia after severe brain injury usually follows

a period of reduced CBF or increased ICP. It may start
approximately 48 hours after injury and last from a
few hours to 710 days (Bouma and Muizelaar, 1992).
Hyperemia is frequently seen shortly after removal of
an epidural or acute subdural hematoma (Schroder,
Muizelaar and Kota, 1994). Pressurevolume index

(PVI) and cerebral blood volume (CBV) studies have
suggested that the association between hyperemia
and ICP is probably due to an increase in cerebral
blood volume rather than CBF (Bouma and Muizelaar, 1992; Marmarou et al., 1987).
These observations suggest the usefulness of combined studies of CBV and CBF with the xenonenhanced computed tomography (Chapter 11) in
acute severely head-injured patients.
According to Obrist et al. (1984), in patients with
head injury CBF below 33 ml/100 g/min at a PaCO2 of
34 mmHg should be considered to be reduced, and all
values above this level are hyperemic. CBF between 33
and 55 ml/100 g/min represents relative hyperemia and
above 55 ml/100 g/min, absolute hyperemia.
Previous studies have reported hyperemia in more
than 50% of patients after severe head injury, generally
associated with high ICP and uncoupling of CBF and
CMRO2 (Obrist et al., 1984; Bouma and Muizelaar,
1992). In children hyperemia was found even more
frequently in some studies, but others found no
increased incidence for hyperemia in children aged
318, nor a relationship between ICP and CBF (Bouma
and Muizelaar, 1992; Salvant and Muizelaar, 1993;
Muizelaar, Marmarou and Desalles, 1989; Muizelaar et
REFERENCES
al., 1989a, b; Sharples et al., 1995). Children normally

have a higher CBF depending on their age (see above),
and this must be taken into consideration.
5.3.3 CEREBROVASCULAR REACTIVITY AFTER
SEVERE HEAD INJURY
Under normal physiological conditions there is a

36% change in CBF per mmHg PaCO2 (Obrist et al.,
1984; Bouma and Muizelaar, 1992; Enevoldsen et al.,
1976). In most patients with severe head injury, some
depression of CO2 reactivity is seen for some time
(Bouma and Muizelaar, 1992). Diminished CO2 reactivity is associated with a poor outcome and is most
probably a reflection of the severity of the injury
rather than a causative mechanism. Reduced CO2
reactivity is also associated with decreased effectiveness of barbiturate therapy in the control of raised ICP
(Messeter et al., 1986). A xenon-enhanced CBF study in
17 severely head-injured patients found that hemispheric CO2 reactivity ranged from 1.38.5%
per mmHg change in PaCO2 (Bouma and Muizelaar,
1992).
Prolonged hyperventilation may decrease brain
oxygenation and must be used with caution and only
for a short period of time (Muizelaar et al., 1991). In
the future, measuring brain tissue PO2 and PCO2 may
allow PaCO2 to be adjusted more accurately and safely
(Zauner, Bullock and Young, 1995a; Zauner et al., 1995;
Figure 5.3).
Current knowledge on the status of cerebral autoregulation following severe head injury is inconclusive.
In general, pressure autoregulation is usually intact
for the first hours after injury but may deteriorate in
many patients (Obrist et al., 1984; Bouma and Muizelaar, 1992; Enevoldsen et al., 1976; Lassen, 1966;
Muizelaar, Marmarou and Desalles, 1989). However,
conclusions regarding the outcome of patients according to the status of autoregulation should be made
with great caution. Bouma and Muizelaar performed
158 autoregulation tests, using CBF measurements
and an alpha-adrenergic agonist, on 117 severely
head-injured patients. They found autoregulation to
be intact in 51% and defective in 49% (Bouma and
Muizelaar, 1992). They concluded that the presence or
absence of pressure autoregulation had no specific
temporal profile, nor a clear relationship to clinical
status or outcome. This has also been the finding in
head-injured children (Muizelaar et al., 1989b; Sharples et al., 1995). However, the status of autoregulation
is important for determining the most appropriate
blood pressure and CPP in a particular patient, and
may influence the choice of treatment for raised ICP.
For example, mannitol may be more effective if
autoregulation is intact (Bouma and Muizelaar,
1992).
5.3.4 POST-TRAUMATIC CEREBRAL
HYPERMETABOLISM FOR GLUCOSE
Positron emission tomography has been used to

measure brain metabolism after human head injury.
Uniformly reduced glucose metabolism has been
reported, although marked regional variations were
present in two studies performed more than 1 week
after injury (Alavi et al., 1986; Langfitt et al., 1986).
Recently Hovda et al. (1995) have found striking
increases in regional glucose utilization in posttraumatic studies performed within the first few days
of injury. This was seen especially in the underlying
brain after hematoma removal. This early posttraumatic hyperglycolysis may result from cellular
efforts to re-establish ionic gradients. Thus the
hyperemic zones seen in some patients may represent
an appropriate response to metabolic needs.
5.4
Figure 5.3 Changes in brain PO2, PCO2 and pH in the

human brain during hyperventilation after severe head
injury. Brain PO2 is lowered during hyperventilation, resulting in decreased substrate delivery, although this remains
above theshold levels ( 20 mmHg). Brain PCO2 follows the
washout of arterial CO2, resulting in an increase in the pH of
the brain tissue.
97
References
Abrams, R. M., Ito, M., Frisinger, J. E. et al. (1984) Local cerebral glucose
utilisation in fetal and neonatal sheep. American Journal of Physiology, 246,
R608R618.
Adams, J. H. and Graham, D. I. (1972) The pathology of blunt head injury, in
Scientific Foundations of Neurology, (eds M. Critchley, J. L. OLeary and B.
Jennett), Heinemann, London, pp. 488491.
Alavi, A., Dann, R., Chawluk, J. et al. (1986) Positron emission tomography
imaging of regional cerebral glucose metabolism. Seminars in Nuclear
Medicine, 16, 234.
Alborch, E., Torregrosa, G., Terrasa, J. C. et al. (1984) GABA receptors mediate
cerebral vasodilation in the anaesthetised goat. Brain Research, 321,
103110.
Alpers, B. J., Berry, R. O. and Paddison, R. M. (1959) Anatomical studies in the
circle of Willis in normal brains. Archives of Neurology and Psychiatry, 81,
409418.
Auer, L. M. and Loew, F. (1983) The Cerebral Veins. An Experimental and Critical
Update, Springer-Verlag, Vienna.
98
Austine, G., Horn, N., Rouhe, S. et al. (1972) Description and early results of
an intravenous radioisotope technique for measuring regional cerebral
blood flow in man. European Neurology, 4351.
Bouma, G. J. and Muizelaar, J. P. (1992) Cerebral blood flow, cerebral blood
Neurotrauma, 9(Suppl. 1), S333S348.
metabolism after severe traumatic brain injury: the elusive role of
ischaemia. Journal of Neurosurgery, 75, 685693.
xenon-enhanced computed tomography. Journal of Neurosurgery, 77,
360368.
Bruce, D. A., Langfitt, T. W., Miller, J. D. et al. (1973) Regional cerebral blood
flow, intracranial pressure and brain metabolism in comatose patients.
Cameron, I. R. and Caronna, J. (1976) The effect of local changes in potassium
and bicarbonate concentration on hypothalamic blood flow in the rabbit.
Journal of Physiology, 262, 415430.
Cold, G. E. (1989) Measurements of CO2 reactivity and barbiturate reactivity
in patients with severe head injury. Acta Neurochirurgica, 98, 153163.
Cruz, J. (1993a) Cerebral oxygenation monitoring and management. Acta
Neurochirurgica (Supplement), 59, 8690.
Cruz, J. (1993b) On-line monitoring of global cerebral hypoxia in acute brain
injury: relationship to intracranial hypertension. Journal of Neurosurgery, 79,
228233.
Dettmers, C., Broich, K., Hartmann, A. et al. (1992) Estimated values in normal
subjects, in Cerebral Blood Flow Measurements with Stable Xenon-enhanced
Computer Tomography, (ed. H. Yonas), Raven Press, New York.
Edvinsson, L. (1982) Sympathetic control of cerebral circulation. Trends in
Neuroscience, 8, 425429.
Edvinsson, L. and McCulloch, J. (1981) Effects of phenobarbital on contractile
responses of feline cerebral arteries. Journal of Cerebral Blood Flow and
Metabolism, 1, 437440.
Edvinsson, L. and Sercombe, R. (1976) Influence of pH and pCO2 on alphareceptor mediated contraction in brain vessels. Acta Physiologica Scandinavica, 97, 325331.
Ehrenreich, D. L., Burns, R. A., Alman, F. N. et al. (1961) Influence of
acetazolamid on cerebral blood flow. Archives of Neurology, 5, 227232.
Enevoldsen, E. M., Cold, G. E., Jensen, F. T. et al. (1976) Dynamic changes in
regional cerebral blood flow, intraventricular pressure, cerebrospinal fluid
pH and lactate levels during the acute phase of head injury. Journal of
Neurosurgery, 44, 191214,
Faraci, F. M. (1993) Endothelium-derived vasoactive factors and regulation of
the cerebral circulation. Neurosurgery, 33, 648658.
Fleckenstein, W., Maas, A. I., Nollert, G. and de Long, D. A. (1990) Oxygen
pressure in cerebrospinal fluid, in Clinical Oxygen Pressure Measurement 11,
(eds A. M. Ehrly, W. Fleckenstein and H. Landgraf), Blackwell, Berlin, pp.
368393.
Furchgott, R. F. and Vanhoutte, P. M. (1989) Endothelium-derived relaxing
and contracting factors. FASEB Journal, 3, 20072018.
Giller, C. A., Bowman, G., Dyer, H. et al. (1993) Cerebral arterial diameters
during changes in blood pressure and carbon dioxide during craniotomy.
Gobel, U., Theilen, H. and Kuschinsky, W. (1990) Congruence of total and
perfused capillary network in rat brains. Circulation Research, 66, 271281.
Gobel, U., Klein, B., Schrock, H. and Kuschinsky, W. (1989) Lack of capillary
recruitment in the brains of awake rats during hypercapnia. Journal of
Cerebral Blood Flow and Metabolism, 9, 491499.
Good, W. F., Gur, D. and Yonas, H. (1992) Technical aspects, in Cerebral Blood
Flow Measurements with Stable Xenon-enhanced Computer Tomography, (ed. H.
Yonas), Raven Press, New York.
Gopinath, S. P,. Robertson, C. S., Contant, C. F. et al. (1994) Jugular venous
Gur, D., Yonas, H. and Good, W. F. (1989) Local cerebral blood flow by xenonenhanced CT: current status, potential improvements, and future directions.
Cerebrovascular Brain Metabolism Review, 1, 6896.
Harder, D. R., Kauser, K., Roman, R. J. et al. (1989) Mechanism of pressureinduced myogenic activation of cerebral and renal arteries: role of the
endothelium. Journal of Hypertension, 7, S11S15.
Harper, A. M. (1966) Autoregulation of cerebral blood flow: influence of the
arterial blood pressure on the blood flow through the cerebral cortex.
Hatazawa, J., Ito, M., Matusuzawa, T. et al. (1988) Measurement of the ratio of
cerebral oxygen consumption to glucose utilisation by positron emission
tomography: its consistency with the values determined by the Kety
Schmidt method in normal volunteers. Journal of Cerebral Blood Flow and
Metabolism, 8, 426432.
Heiss, W. D., Hayakawa, T. and Waltz, A. G. (1976) Cortical neuronal function
during ischaemia. Archives of Neurology, 33, 813820.
Heuser, D., Astrup, J., Lassen, N. A. et al. (1975) Brain carbonic acid acidosis
after acetazolamide. Acta Physiologica Scandinavica, 93, 385390.
Hovda, D. A., Lee, S. M., Smith, M. L. et al. (1995) The neurochemical and
metabolic cascade following brain injury: moving from animal models to

man. Journal of Neurotrauma, 12, 903906.
Hudak, M. L., Jones, M. D., Popel, A. S. et al. (1989) Hemodilution causes sizedependent constriction of pial arterioles in the cat. American Journal of
Physiology, 257, H912H917.
Humphrey, P. R. D., Marshall, J., Russel, R. W. R. et al. (1979) Cerebral blood
flow and viscosity in relative polycythemia. Lancet, i, 873876.
Jenkins, L. W., Moszynski, K., Lyeth, B. et al. (1989) Increased vulnerability of
the mildly traumatized rat brain to cerebral ischaemia: the use of controlled
secondary ischaemia as a research tool to identify common or different
mechanisms contributing to mechanical and ischaemic brain injury. Brain
Research, 477, 211224.
Jennett, B., Miller, J. D. and Harper, A. M. (1976) Effect of Carotid Artery Surgery
on Cerebral Blood Flow, Excerpta Medica, Amsterdam.
Jennett, B., Graham, D. I., Adams, H. P. et al. (1973) Ischemic brain damage after
fatal blunt head injury, in Cerebrovascular Diseases, 8th Conference, (eds F. H.
McDowall and R. W. Brennan), Grune & Stratton, New York, pp. 163176.
Jones, T. H., Morawetz, R. B., Crowell, R. M. et al. (1981) Thresholds of focal
cerebral ischaemia in awake monkeys. Journal of Neurosurgery, 36, 463470.
Kader, A., Frazzini, V., Solomon, R. A. and Trifiletti, R. R. (1993) Nitric oxide
production during focal cerebral ischaemia in rats. Stroke, 24, 17091716.
Kassell, N. F., Hitchon, P. W., Gerk, M. K. et al. (1981) Influence of changes in
arterial pCO2 on cerebral blood flow and metabolism during high-dose
barbiturate therapy in dogs. Journal of Neurosurgery, 54, 615619.
Kelly, P. A. and McCulloch, J. (1983) The effect of the GABA agonist muscimol
upon the relationship between local cerebral blood flow and glucose
utilisation. Brain Research, 258, 338342.
Kim, P. and Vanhoutte, P. M. (1990) Endothelium dependent relaxations and
chronic vasospasm after subarachnoid hemorrhage. Blood Vessel, 27,
263268.
Kontos, H. A. (1985) Oxygen radicals in cerebral vascular injury. Circulation
Research, H57, 508516.
Kontos, H. A., Raper, A. J and Patterson, J. L. (1977) Analysis of vasoactivity
and local pH, pCO2 and bicarbonate on pial vessels. Stroke, 8, 358360.
Kontos, H. A., Wei, E. P., Navari, R. M. et al. (1978) Responses of cerebral
arteries and arterioles to acute hypotension and hypertension. American
Journal of Physiology, 234, H371H383.
Kuschinsky, W. and Wahl, M. (1978) Local chemical and neurogenic regulation
of cerebral vascular resistance. Physiology Reviews, 58, 656689.
Kuschinsky, W., Wahl, M., Bosse, O. et al. (1972) Perivascular potassium and
pH as determinants of local pial arterial diameter in cats: a microapplication
study. Circulation Research, 31, 240247.
Langfitt, T. W., Weinstein, J. D. and Kassell, N. F. (1965) Cerebral vasomotor
paralysis produced by intracranial hypertension. Neurology, 15, 622641.
Langfitt, T. W., Obrist, W. D., Alavi, A. et al. (1986) Computerized tomography,
magnetic resonance imaging, and positron emission tomography in the
study of brain trauma. Journal of Neurosurgery, 64, 760767.
Lassen, N. A. (1966) The luxury perfusion syndrome and its possible relation to
acute metabolic acidosis localised within the brain. Lancet, ii, 11131115.
Leniger-Follert, E., Gronczewski, J. and Danz, C. (1984) Regulation of
microflow in the cat brain during insulin induced hypoglycemia. Advances
in Experimental Medical Biology, 169, 297303.
Leniger-Follert, E., Lubbers,
D. W. et al. (1975) Regulation of local tissue PO2 of

the brain cortex at different arterial O2 pressures. Pflugers Archiv, 359,
8195.
Leusen, I. R., Weyne, J. J. and Demeester, G. M. (1980) Regulation of acid-base
equilibrium of cerebrospinal fluid, in Neurobiology of Cerebrospinal Fluid, (ed.
J. H. Wood), pp. 2542.
Lubbers,
D. W. (1968) The oxygen pressure field in the brain, in Oxygen

Transport in Blood and Tissue, (eds D. W. Lubbers,
U. C. Loft, G. Thews and

E. Witzleb), Springer-Verlag, Berlin, pp. 6792.
Maas, A. I. R., Fleckenstein, W., de Long, D. A. and van Santbrink, H. (1993)
Monitoring cerebral oxygenation: experimental studies and preliminary
clinical results of continuous monitoring of cerebrospinal fluid and brain
tissue oxygen tension. Acta Neurochirurgica (Supplement), 59, 5057.
Marmarou, A., Maset, A. I., Ward, J. D. et al. (1987) Contribution of CSF and
vascular factors to elevation of ICP in severely head injured patients. Journal
Marshall, L. F., Gautille, T., Klauber, M. R. et al. (1991) The outcome of severe
closed head injury. Journal of Neurosurgery, 75, S28S36.
Messeter, K., Nordstrom, C. H., Sundbarg, H. M. et al. (1986) Cerebral
hemodynamics in patients with acute severe head trauma. Journal of
Meyer, I. S., Imai, A., lchijo, M. et al. (1992) Local cerebral blood flow and local
lambda values change with normal advancing age, in Cerebral Blood Flow
Measurements with Stable Xenon-enhanced Computer Tomography, (ed. H.
Muizelaar, J. P., Marmarou, A. and Desalles, A. A. (1989a) ICP and PVI with
blood pressure alterations and relation with CBF autoregulation, in
Intracranial Pressure VII, (eds J. T. Hoff and A. L. Betz), Springer-Verlag,
Berlin, pp. 825828.
Muizelaar, J. P. and Schroder,
M. L. (1994) Overview of monitoring of cerebral

blood flow and metabolism after severe head injury. Canadian Journal of the
Neurological Sciences, 21, S6S11.
REFERENCES
Muizelaar, J. P., Marmarou, A., Desalles, A. A. et al. (1989b) Cerebral blood
flow and metabolism in severely head-injured children. Part I: Relation
with GCS, outcome, ICP and PVI. Journal of Neurosurgery, 71, 6371.
Muizelaar, J. P., Ward, J. D., Marmarou, A. et al. (1989) Cerebral blood flow and
metabolism in severely head injured children. Part II: Autoregulation.
Muizelaar, J. P., Marmarou, A., Ward, J. D. et al. (1991) Adverse effects of
prolonged hyperventilation in patients with severe head injury: a randomized clinical trial. Journal of Neurosurgery, 75, 731739.
Muizelaar, J. P., Bouma, G. J., Levasseur, J. E et al. (1992) Effect of hematocrit
variations on cerebral blood flow and basilar artery diameter in vivo.
American Journal of Physiology, 262, H949H954.
Nakai, M., Iadecola, C., Ruggiero, D. et al. (1983) Electrical stimulation of
cerebellar fastigial nucleus increases cortical blood flow without change in
local metabolism: evidence for an intrinsic system in brain for primary
vasodilation. Brain Research, 260, 3549.
Nakano, S., Yamashita, T., Kashiwagi, S. et al. (1992) in Cerebral Blood Flow
Measurements with Stable Xenon-enhanced Computer Tomography, (ed. H.
Nordstrom, C. H., Messeter, K., Sundbarg, G. et al. (1988) Cerebral blood flow,
vasoreactivity, and oxygen consumption during barbiturate therapy in
severe traumatic brain lesions. Journal of Neurosurgery, 68, 424431.
Obrist, W. D., Langfitt, T. W., Jaggi, J. L. et al. (1984) Cerebral blood flow and
metabolism in comatose patients with acute head injury. Journal of
Olesen, J., Paulson, O. B., Lassen, N. A. (1971) Regional cerebral blood flow in
man determined by the initial slope of the clearance of intra-arterially
injected 133Xe. Stroke, 2, 519540.
Phelps, M. E., Mazziotta, J. C., Schelbert, H. R. (1986) Positron Emission
Tomography and Autoradiography, Principles and Applications for the Brain and
Heart, Raven Press, New York.
Risberg, J., Ali, Z., Wilson, E. M. et al. (1975) Regional cerebral blood flow by
133
xenon inhalation: preliminary evaluation of an initial slope index in
patients with unstable flow compartments. Stroke, 6, 142148.
Ritter, A. M. and Robertson, C. S. (1994) Cerebral metabolism. Neurosurgical
Intensive Care, 5, 633645.
Robertson, C., Grossman, R., Goodman, C. et al. (1987) The predictive value of
cerebral anerobic metabolism with cerebral infarction after head injury.
Robertson, C. S., Narayan, R. K., Gokaslan, Z. L. et al. (1989) Cerebral
arteriovenous oxygen difference as an estimate of cerebral blood flow in
comatose patients. Journal of Neurosurgery, 70, 222230.
Robinson, M. J. and McCulloch, J. (1990) Contractile responses to endothelin
in feline cortical vessels in situ. Journal of Cerebral Blood Flow and Metabolism,
10, 285289.
Robinson, P. and Rapoport, S. (1986) Glucose transport and metabolism in the
brain. American Journal of Physiology, 250, R127RI36.
Salvant, J. and Muizelaar, J. (1993) Changes in cerebral blood flow and
metabolism related to the presence of subdural hematoma. Neurosurgery, 33,
387391.
99
Schroder,
M. L., Muizelaar, J. P. and Kota, A. J. (1994) Documented reversal of

global ischaemia immediately after removal of an acute subdural hematoma. Journal of Neurosurgery, 80, 324327.
Schroder,
M. L., Muizelaar, J. P., Kota, A. J. et al. (1995) Thresholds for cerebral

ischaemia after severe head injury: relationship with late CT findings and
outcome. Journal of Neurotrauma.
Sharkey, J. and McCulloch, J. (1995) Relationship between local cerebral blood
flow and glucose utilisation with selective dopamine receptor agonists.
Journal of Cerebral Blood Flow and Metabolism, 5(Suppl.), 537538.
Sharples, P. M., Stuart, A. G., Matheras, D. S. et al. (1995) Cerebral blood flow
and metabolism in severely head injured children. Part I: Relationship to
age, GCS, outcome, ICP and time after injury. Journal of Neurology,
Sheinberg, M., Kanter, M. J., Robertson, C. S. et al. (1992) Continuous
monitoring of jugular venous oxygen saturation in head-injured patients.
Siesjo,
B. K. (1984) Cerebral circulation and metabolism. Journal of Neurosurgery, 60, 883908.
Sokoloff, L., Reivich, M., Kennedy, C. et al. (1977) The 14C deoxyglucose
method for measurement of local cerebral glucose utilisation: theory,
procedure and normal values in the conscious and anesthetized albino rat.
Journal of Neurochemistry, 28: 897916.
Sokrab, T., Johansson, B., Kalimo, H. et al. (1988) A transient hypertensive
opening of the bloodbrain barrier can lead to brain damage. Acta
Suzuki, K. (1990) The changes of regional cerebral blood flow with advancing
age in children. Nagoya Medical Journal, 34, 159170.
Thomas, D. J., DuBoulay, G. H., Marshall, J. et al. (1977) Effect of hematocrit on
cerebral blood flow in man. Lancet, ii, 941943.
Wahl, M., Deetjen, P., Thurau, K. et al. (1970) Micropuncture evaluation of the
importance of perivascular pH for the arteriolar diameter on the brain
surface. Pflugers Archiv, 316, 152163.
Wei, E. P. and Kontos, H. A. (1988) Increased venous pressure causes
myogenic constriction of cerebral arterioles during local hyperoxia.
Circulation Research, 55, 249252.
Winn, H. R., Welsh, J. E., Rubio, R. et al. (1980) Brain adenosine production in
rat during sustained alteration in systemic blood pressure. American Journal
of Physiology, 239, H636H641.
Zauner, A., Bullock, R. and Young, H. F. (1995a) Continuous brain oxygen,
CO2, pH and temperature monitoring in neurosurgical patients. Neurosurgery, 37, 570 (abstract).
Zauner, A., Bullock, R. and Young, H. F. (1995b) Continuous monitoring of
oxygen, CO2, pH and temperature in brain tissue, using a single sensor.
Journal of Neurotrauma, 12, 468 (abstract).
Zauner, A., Bullock, R. and Young, H. F. (1995c) Brain oxygen, CO2, and
cerebral blood flow measurements in focal ischaemia. Journal of Cerebral
Blood Flow and Metabolism, 15(Suppl. 1), S297 (abstract).
Zauner, A., Bullock, R., Di, X. et al. (1995) Brain oxygen, CO2, pH, and
temperature monitoring: evaluation in the feline brain. Neurosurgery, 37,
11681177.
INTRACRANIAL PRESSURE
AND ELASTANCE
Ian Piper
6.1 The problem: raised intracranial

pressure after head injury
6.1.1
INTRODUCTION
Head injury is a common form of trauma. For example,

in the UK head injury occurs in more than 500 000
persons per annum of which about 10% are diagnosed
as severe, 15% moderate and the remainder as minor
head injury (Miller et al., 1992; Pickard and Czosnyka,
1993). Head trauma is a significant cause of death and
disability, especially in young males (median age < 30)
and is associated with raised intracranial pressure
(ICP). Raised ICP is defined as pressure greater than
20 mmHg and appears most commonly in about
5075% of patients with severe head injury who
remain comatose after resuscitation.
Over the past 50 years there has been an active and
wide-ranging research into the causes and consequences of raised ICP which, to date, has been the
subject of nine international symposia embracing such
diverse disciplines as neurosurgery, anesthesia, radiology, biophysics, electronic and mechanical engineering, mathematics and computer science.
In particular, the introduction during the 1970s of
the continuous monitoring of ICP has led to renewed
activity in both clinical and experimental research into
the physiology and pathophysiology of maintaining
craniospinal volume and pressure. This interest has
not just been in monitoring pressure alone but also in
using information derived from pressure monitoring
to help both predict raised ICP and determine the
underlying cause.
ICP is a reflection of the relationship between
alterations in craniospinal volume and the ability of
the craniospinal axis to accommodate added volume.
The craniospinal axis is essentially a partially closed
box with container properties including both viscous
and elastic elements. The elastic or its inverse, the
compliant, properties of the container will determine
what added volume can be absorbed before intra-
cranial pressure begins to rise. So an understanding of

raised ICP encompasses an analysis of both intracranial volume and craniospinal compliance.
This chapter reviews the relationship of raised ICP
to outcome and its significance as part of the development of the primary injury and as a superimposed
secondary insult. This is followed by a review of both
the historical and current concepts underlying our
present understanding of the physiology and pathophysiology of maintaining intracranial pressure and
volume.
6.1.2 RAISED INTRACRANIAL PRESSURE:
RELATIONSHIP TO OUTCOME
Raised ICP has in the past been found to be associated

with a poorer outcome from injury with the higher the
level of ICP, particularly the peak ICP level, correlating with the expected prognosis for mortality and
morbidity (Becker et al., 1977; Marshall et al., 1979;
Miller et al., 1977, 1981; Pitts et al., 1980). There has,
however, been controversy over the usefulness of
monitoring raised ICP with some groups, with a no
ICP monitoring policy, finding in their studies of head
injury mortality and morbidity that outcome is similar
to other groups that do monitor ICP (Stuart et al.,
1983). Reported differences in the utility of ICP
monitoring could be due to variability in both management and monitoring protocols between different
neurosurgical centers. Variation in type of ICP pressure monitor, site of placement, treatment thresholds,
patient referral characteristics and outcome measures
can all combine to produce a large variability in both
measured ICP and outcome, irrespective of whether
ICP is monitored or how it is treated. Another source
of variation in terms of raised ICP is the inherent
variability of the head-injured population, outcome
being dependent on a number of other factors. For
example, mass lesions are generally accompanied by
elevations in ICP of greater than 40 mmHg and are
associated with poorer outcome, while diffuse injuries
102
INTRACRANIAL PRESSURE AND ELASTANCE
tend to have lower ICP levels associated with a

similarly poor outcome (Miller et al., 1977, 1981). Age
is also an important factor, with an age-dependent
distribution of ICP for both type of injury and
outcome. This is particularly so for pediatric cases
(Alberico et al., 1987; Choi et al., 1991; Luerssen,
Klauber and Marshall, 1988; Volmer et al., 1991). ICP
can even be raised in the absence of overt signs of
swelling or mass lesions on CT. In a small study of
severely head-injured patients, OSullivan et al.
(1994) demonstrated that some comatose head-injured
patients whose initial CT scan was normal, with no
mass lesion, midline shift or abnormal basal cisterns,
developed raised ICP greater than 20 mmHg that
lasted longer than 5 minutes. This included a subset of
patients showing pronounced raised ICP of greater
than 30 mmHg.
Data from large prospective trials carried out from
single centers and from well-controlled multicenter
studies have provided the most convincing evidence
for a direct relationship between ICP and outcome
(Narayan et al., 1981; Saul and Ducker, 1982; Marmarou et al., 1991; Jones et al., 1994). Narayan et al.
(1981), in a prospective study in 133 severely headinjured patients, demonstrated that the outcome prediction rate was increased when the standard clinical
data, such as age, Glasgow Coma Score on admission
(GCS) and pupillary response with extraocular and
motor activity, was combined with ICP monitoring
data. Marmarou et al. (1991), reporting on 428 patients
data from the National Institute of Healths Traumatic
Coma Data Bank, showed that, following the usual
clinical signs of age, admission motor score and
abnormal pupils, the proportion of hourly ICP recordings greater than 20 mmHg was the next most significant predictor of outcome. Outcome was classified
by the Glasgow Outcome Score (GOS) at 6 months
follow-up. They also found, using stepwise logistic
regression, that, following ICP, arterial pressure below
80 mmHg was also a significant predictor of outcome.
Jones et al. (1994) studied prospectively 124 adult
head-injured patients during intensive care using a
computerized data collection system capable of minute-by-minute monitoring of up to 14 clinically indicated physiological variables. They found that ICP
above 30 mmHg, arterial pressure below 90 mmHg
and cerebral perfusion pressure below 50 mmHg
significantly affected patient morbidity.
Although in the past there have been differing
opinions about the contribution of continuous monitoring of ICP to reduction in mortality and morbidity
following head injury, there is now sufficient evidence
to remove doubt about the value of ICP monitoring
towards improving the prediction of outcome and
allowing more informed decisions to be made about
patient management.
6.1.3 RAISED INTRACRANIAL PRESSURE:

RELATIONSHIP TO PRIMARY AND SECONDARY
INJURY
Both experimental and clinical studies have clearly

shown that, following traumatic brain injury, normal
physiological mechanisms for maintaining cerebral
perfusion can become impaired (Lewelt, Jenkins and
Miller, 1980, 1982; Povlishock and Kontos, 1985;
Nordstrom et al., 1988; Miller and Adams, 1992).
These studies demonstrate that brain injury can
cause impairment or loss of autoregulation defined
as the ability of the cerebral vessels to respond to
changes in arterial gases or to arterial pressure. As a
result of these changes there can, at times, be a
decrease in cerebrovascular resistance, which can
lead to raised ICP in both adults and children
(DeSalles, Muizelaar and Young, 1987; Jaggi et al.,
1990; Muizelaar et al., 1989a, b; Uzzell et al., 1986).
While brain injured patients are being managed in
intensive care there are, superimposed on the primary injury, periods of reduced PaO2 or episodes of
arterial hypotension often as a result of other injuries
or treatment by hypnotic drugs (Rose, Valtonen and
Jennett, 1977; Gentleman and Jennett, 1981; Miller et
al., 1981; Miller and Becker, 1982; Marmarou et al.,
1991; Jones et al., 1994). With an impaired physiological mechanism unable to respond adequately to
these adverse changes in physiological parameters
(or secondary insults), ischemic brain damage can
occur. These secondary, chiefly ischemic brain insults
are common with Graham, Hume Adams and Doyle
(1978) reporting, in a series of 151 fatal cases of
severe head injury, a 91% incidence of ischemic brain
damage found on autopsy. A second study carried
out by the same group over 10 years later found a
similar high incidence (> 80%) of ischemic brain
damage despite subsequent improvements in intensive care of head-injured patients (Graham et al.,
1989). Other studies have shown that patients whose
primary injury was judged not to have been severe
(patients who talk and die) go on subsequently to
deteriorate as a result of secondary largely avoidable
events (Reilly et al., 1975; Rose, Valtonen and Jennett,
1977; Sharples et al., 1990). These studies were
responsible for the concept of avoidable brain injury
and stimulated considerable clinical and experimental research into both the pathophysiology of
secondary brain damage and methods aimed at
detecting and predicting adverse physiological
conditions.
Grossman et al. (1975), studying the relationship
between cortical electrical activity and cerebral blood
flow (CBF) and CPP during CSF infusion in nonhead-injured primates, showed the relevance of maintaining adequate cerebral perfusion. Provided CBF
RAISED INTRACRANIAL PRESSURE AFTER HEAD INJURY
and CPP were preserved, electrical activity was

maintained despite ICP increases above 50 mmHg.
However, the cortical activity rapidly disappeared
when CBF fell markedly, to levels below approximately 20 ml/100 g/min. Narayan et al. (1981) confirmed the importance of brain electrical activity as a
marker for adequate cerebral perfusion in headinjured patients. In a prospective study of 133 patients
he demonstrated that multimodality evoked potentials when taken together with both basic clinical
signs and ICP formed the best predictive model of
outcome. More recent studies of brain electrical
activity have again reaffirmed the association
between outcome, ICP and brain electrical activity.
Park et al. (1993) studied a small group of patients
classified as having minor, moderate or severe diffuse
axonal injury (DAI) with serial monitoring of multimodality evoked potentials over 4 weeks. Patients
were followed up at 3 months and assessed using the
Glasgow Outcome Score. Patients with both moderate
and severe DAI showed marked changes in somatosensory and visual evoked potentials, which were
correlated with outcome. Increasingly, motor evoked
potentials elicited by transcranial electrical or magnetic stimulation are also being studied as potential
measures of neurological function in neurotrauma
patients. Kawai et al. (1993), in a feline model of
diffuse brain compression, monitored the central
motor pathways by transcranial magnetic stimulation
of the motor cortex. They found that the N4 component (4.99 ms latency) of the spinal motor evoked
103
potential (L1L2) was progressively prolonged and

depressed as ICP rose above 40 mmHg.
Apart from cerebral electrical measurements as an
indication of neurological function, there has been
much interest in the relationship between ICP, CPP
and CBF. The landmark study of Miller, Stanek and
Langfitt (1972) produced some of the first experimental evidence confirming the concept that changes
in intracranial pressure affect cerebral blood flow not
directly but through changes in cerebral perfusion
pressure (CPP), where CPP is defined as the difference
between mean arterial pressure and intracranial pressure. Strictly speaking, the actual cerebral perfusion
outflow pressure would be cerebral venous pressure,
although it is, in most situations, impractical to
measure this pressure routinely. However it has been
established that over a wide range of pressures
cerebral venous pressure is well approximated (within
34 mmHg) by ICP (Johnson and Rowan, 1974; Yada,
Nakagawa and Tsuru, 1973; Nakagawa, Tsuru and
Yada, 1974). In an experimental study of CBF as
determined by the venous outflow technique in dogs,
Miller also demonstrated that, when MAP and ICP
rise in parallel so that CPP remains constant at
60 mmHg, CBF increases with MAP in animals found
to be non-autoregulating. It was further shown that, as
CPP drops in autoregulating animals, the breakpoint
at which CBF starts to decrease is at a higher level if
CPP is reduced through hemorrhagic arterial hypotension than if it is reduced through intracranial hypertension (Figure 6.1). This work suggests that cerebral
Figure 6.1 Plot showing that the breakpoint at which cerebral blood flow (CBF) starts to decrease is at a higher level if
cerebral perfusion pressure (CPP) is reduced through hemorrhagic arterial hypotension (dMAP) than if it is reduced through
intracranial hypertension (dICP).
104
perfusion is more sensitive to arterial hypotension

than to intracranial hypertension.
The clinical significance of this information is that in
the management of head injury it is often necessary to
employ therapy to lower raised ICP. Therapeutic
agents for reducing raised ICP often do so at the
expense of reduced MAP and as a consequence CPP
may not improve. If autoregulation is preserved, CBF
should remain unchanged despite parallel changes in
MAP and ICP. However, clinically, autoregulation is
likely to be impaired in those conditions in which ICP
is increased such as head injury or subarachnoid
hemorrhage (Harper, 1966; Muizelaar, Lutz and
Becker, 1984; Muizelaar and Becker, 1986; Muizelaar et
al., 1989a, b; Obrist et al., 1984; Bouma and Muizelaar,
1990). Under these circumstances, it is important that
reduction in ICP should not be achieved at the
expense of lowering CBF and provoking brain
ischemia.
This earlier work of Miller was later extended by
Chan, Miller and Piper (1992) to include CPP ranges of
60, 50 and 40 mmHg. At CPP levels of 50 and
60 mmHg, when autoregulation was intact, CBF
remained unchanged. However, with loss of autoregulation, there was a trend for CBF to increase as
MAP and ICP were increased in parallel at a CPP of 50
and 60 mmHg. Absolute CBF levels were significantly
different between the autoregulating and non-autoregulating groups. At a CPP of 40 mmHg CBF showed
a linear correlation with BP. This work demonstrates
that, when autoregulation is impaired, there is a
functional difference between autoregulating and
non-autoregulating cerebral vessels despite similar
MAP and CPP and that, when autoregulation is
impaired, CBF depends more on arterial driving
pressure than on cerebral perfusion pressure.
The importance of arterial pressure as the prime
factor governing CPP-related secondary insults has
been well demonstrated by the recent work by Jones,
Miller and colleagues (1994) who carried out a
prospective study over 4 years of the frequency and
severity with which secondary insults occur to headinjured patients while being managed in intensive
care. They developed a microcomputer-based data
collection system that allows the acquisition of data
from up to 15 monitored variables minute by minute
(Piper et al., 1991). At each bedspace data collection
was under the control of a microcomputer; serial
links between the patient monitors and the microcomputer allow the controlled transfer of multiple
channels of physiological data once per minute. The
controlling software allowed medical staff to add
comments to the current active computer file at any
time, precisely annotating significant events. The
software performs artifact detection, calculates
derived data and highlights valid data that falls
outside normal physiological levels. Collected data

was stored to disk and can be printed either locally
or remotely.
From the valid physiological data produced, manual processing of data was used to identify secondary
insults, which are defined at one of three grades of
severity and which must last for 5 minutes or longer to
be recorded as an insult. This permits calculation of
the frequency, severity and total duration of insults,
measured in minutes.
An analysis was made of 124 adult head-injured
patients who were monitored during intensive care
using the computerized data collection system. Information was logged at 1 minute intervals and scanned
to identify insults when values fell outside threshold
limits for 5 minutes or longer. Three grades of insult
were defined for each variable. The duration of insults
has been analyzed in relation to the Glasgow Outcome
Score of these patients at 12 months after injury. The
monitored patients included 68 with severe head
injury (Glasgow Coma Score of 8 or less with no eyeopening), 36 with moderate head injury (GCS 912)
and 20 with minor head injury (GCS 1315 but with
multiple injuries, scoring 16 or more on the Injury
Severity Scale).
Insults were found in 91% of patients at all degrees
of severity of head injury. 10% of patients had insults
that were only at the lowest Grade 1 level, 31% had
insults at both Grade 1 and Grade 2 levels and 50% of
patients had at least one insult at Grade 3 level in
addition to Grade 1 and 2 insults. Overall, the majority
(77%) of all insults detected in the ITU were at Grade
1 level and these represented 85% of the total duration
in minutes of insult measured.
In a subset of 71 patients (51 severe, 18 moderate
and two minor), all of whom had six channels
monitored concurrently (ICP, BP, CPP, SaO2, body
temperature, heart rate) insult and outcome data were
analyzed using stepwise logistic regression to determine the effects of age, admission GCS score, pupil
responses, injury severity score and insult duration on
outcome scored on the Glasgow Outcome Scale at 12
months. Duration of hypotension, hypoxemia and
pyrexia were found to be significant predictors of
mortality. When good versus bad outcome was considered (good recovery and moderate disability versus
severe disability and death), the logistic regression
analysis showed duration of hypotension and bilateral
loss of pupil light response to be the most important
predictors.
Differences in the duration of insult between outcome groups 12 months postinjury were compared
with each grade of insult, using KruskallWallis oneway analysis of variance and by MannWhitney U
tests. Significant differences in the distribution of
hypotensive insults was found between the outcome
PHYSIOLOGY AND PATHOPHYSIOLOGY OF INTRACRANIAL PRESSURE
grades at all levels of severity of insult. Similar results

were found for cerebral perfusion pressure insult
duration. These data confirm the important adverse
effect of even moderate reductions in arterial pressure
(systolic BP less than 90 mmHg or mean BP less than
70 mmHg).
Miller and coworkers further considered whether
the estimate of the burden of secondary insults was
excessive, as a result of selection of an atypical group
of head-injured patients. This was not the case, as the
mortality rates for severe, moderate and minor head
injury were 31%, 11% and 5% respectively. The other
possibility is that the computerized system is more
efficient in the detection of abnormal values in
monitored variables than the nursing staff in the
intensive care unit. This was examined by an additional study (Corrie et al., 1993), where the number,
duration and severity of computer-identified insults
were compared with the values recorded by nursing
staff on the standard intensive care unit bedside
24-hour chart. Data from computer recordings and
nursing charts from 20 head-injured patients were
compared for 37 periods, each of a full 24 hours,
during which data from four selected channels were
obtainable without interruption. Insults detected
using the computer system were divided into those
completely identified, completely missed or partially
identified on the bedside nursing chart. Of 216 periods
of insult of 5 minutes or longer that were identified
using the computer system, 69 (32%) were missed on
the standard bedside nursing chart. The majority of
these missed insults were at Grade 1. Insults at Grade
2 or Grade 3 level were more likely to be recorded on
the nursing chart. However, of all the insult types
studied, Grade 1 arterial hypotension was most often
missed (94%). Charting accuracy improved to 87%
detection of computer-identified insults when those
insults that crossed the designated hour and half hour
recording slots in the nursing chart were considered.
Insults that lasted less than 30 minutes were significantly more likely to be missed than those that
were longer. While this study is reassuring in that the
bedside nursing chart was reliable in the detection of
most insults that involved raised ICP, it is a major
concern that periods of mild arterial hypotension
were frequently missed when outcome data indicates
that even these relatively minor episodes have an
adverse effect on patient outcome.
Thus secondary insults are common, result in
mainly ischemic brain damage and are a major
contribution to disablement. They are important
because they are common and yet so potentially
avoidable. Clearly a critical challenge facing us is to
develop patient monitoring systems and protocols
that will lead to rapid detection and resolution of
secondary insults.
105
6.2 The principles: physiology and

pathophysiology of intracranial pressure
6.2.1 CONTROL OF INTRACRANIAL VOLUME AND
PRESSURE: HISTORICAL CONCEPTS
The history of the subject of intracranial pressure has

been well reviewed (Masserman, 1935a, Stern, 1963;
Langfitt, 1969) and starts from the doctrine named after
Monro (1783) and Kellie (1824), which proposed that
the brain and its contained blood were incompressible,
enclosed in a rigid and inextensible skull, of which the
total volume remained constant. In its original form
the MonroKellie doctrine did not take into account
the CSF as a component of the cranial compartment.
The concept of reciprocal volume changes between
blood and CSF was introduced in 1846 by Burrows and
was later extended in the early 20th century by Weed
and McKibben (1919; Weed, 1929) to allow for reciprocal changes in all the craniospinal constituents.
Kocher in 1901 translated into clinical terms the four
stages of cerebral compression proposed almost 25
years earlier by the experimental studies of Duret
(1878). Kocher described four stages of cerebral
compression related to the expansion of intracranial
brain tumors. In stage 1, the initial increase in tumor
volume is compensated by a reduction in volume of
the other intracranial components, chiefly CSF and
venous blood. This spatial compensation results in no
net increase in intracranial volume or pressure and
hence no clinical symptoms. In stage 2 the compensatory mechanisms are exhausted, ICP increases and the
patient becomes drowsy, with headache. Stage 3 is
characterized by a considerable increase in ICP, an
associated deterioration in conscious level with intermittent elevations of blood pressure (BP) accompanied
by bradycardia. In the fourth and final stage, the
patient is unconscious, with bilateral fixed dilated
pupils and falling BP, usually leading to death.
Cushing (1901, 1902, 1903), then a research worker
for Kocher, described in both experimental and clinical
studies the close relationship between increases in ICP
and BP and proposed that the BP rose in order to
maintain adequate blood supply to the hindbrain, the
stimulus to this vasopressor response believed to be
medullary ischemia (Jennett, 1961; Johnson and
Rowan, 1974).
At about this time a false confidence developed in
the lumbar CSF pressure technique (lumbar puncture)
which caused Cushings findings to be challenged.
Reports emerged (Browder and Meyers, 1936; Smyth
and Henderson, 1938; Evans et al., 1951) that some
patients showing clinical signs of brain compression
had normal lumbar CSF pressures and that in other
patients elevations in BP were found at times when
ICP was well below the level of BP.
106
Partly because of this apparent dissociation between

ICP and clinical symptoms, emphasis switched away
from ICP measurement towards the relationship
between craniospinal volume and pressure, particularly the importance of the elastic properties of the
craniospinal system. Ayala (1923, 1925) studied the fall
in lumbar pressure that occurred when CSF was
removed from patients, describing the degree of
decline in terms of the volume of CSF removed and
the elasticity of the meninges. Ayalas index developed from this work and is defined as the fall in
pressure divided by the volume of fluid removed. This
index was found to be low in patients diagnosed with
benign intracranial hypertension and high in patients
with cerebral tumors.
Weed and Flexner (1932; Weed, Flexner and Clark,
1932; Flexner, Clark and Weed, 1932; Flexner and
Weed, 1933) systematically studied the effect of
hydrostatic columns on the elastic properties of the
craniospinal system by observing the pressure and
volume changes during up/down head-tilting experiments in animals. They defined a coefficient of
elasticity based on Hooks law (Eo = stress/strain),
which failed to show any change under a variety of
experimental conditions. Their work was critically
reviewed by Massermann (1934, 1935a), who carried
out similar studies in patients. Ryder et al. (1951) were
the first to characterize the craniospinal volume
pressure relationship as a non-linear quantity,
describing it as a hyperbolic function, which implies
an increase in elastance as pressure increases. This
was in conflict with the work by Weed and coworkers,
although the latter group only studied the elastic
properties over a limited physiological pressure range.
Furthermore, it was also partly the work of Ryder et al.
(1953) that restored confidence in intracranial pressure
measurement by demonstrating a differential pressure
between intraventricular and lumbar CSF pressure
recording. This phenomenon was reported as early as
1895 by Bayliss, Hill and Gulland, who noted that it
was impossible to obtain valid ICP measurements
below the tentorium during later stages of progressive
supratentorial brain compression.
It was not until the 1960s when Lundberg (1960)
published his now classic monograph, that interest in
clinical ICP measurement was rekindled. Using ventricular fluid pressure recording in brain tumor
patients, Lundberg was the first to delineate the
frequency with which raised ICP occurs clinically, at
times reaching pressures as high as 100 mmHg. Lundberg also described three types of spontaneous pressure wave fluctuations: A waves or plateau waves of
large amplitude (50100 mmHg) with a variable duration (520 min), B waves, which are smaller (up to
50 mmHg), sharper waves with a dominant frequency
of 0.52/min, and finally C waves, which are small
(up to 20 mmHg), rhythmic oscillations with a frequency of 48/min.

This work was then extended to include head
injuries (Lundberg, Troup and Lorin 1965; Johnston,
Johnston and Jennett, 1970), intracranial hemorrhage
(Richardson, Hide and Eversden, 1970), posthypoxic
brain damage (Langfitt et al., 1974) and benign
intracranial hypertension (Johnston and Paterson,
1972). ICP can therefore increase under an assortment
of experimental and clinical circumstances, the frequency often being underestimated by the lumbar
pressure recording technique. This phenomenon of
pressure underestimation was fully defined by Langfitt and coworkers (1964a, b) in experimental studies
of extradural brain compression, where progressive
loss of transmission of ICP across the tentorial hiatus
occurred, with the pressure in the posterior fossa and
lumbar subarachnoid space progressively under-reading the ventricular pressure and eventually returning
to normal pressure.
Some of the most important work at this time was
also carried out by Langfitts group (Langfitt, Weinstein and Kassel, 1965) who redefined Kochers four
stages of cerebral compression under controlled
experimental conditions in Rhesus monkeys with
simultaneous measurement of arterial and intracranial
pressure, jugular or sagittal sinus pressure, cerebral
blood flow (CBF) and measures of brain metabolism.
They defined stage 1 as the period of spatial compensation, with very little increase in ICP despite slow
inflation of an extradural balloon. Electroencephalogram (EEG), CBF and brain oxygenation were normal
and stable at this time. Stage 2 occurred at the end of
spatial compensation and was characterized by an
exponential increase in ICP with a steady extradural
balloon inflation rate. Towards the end of this stage,
ICP increased by more than 15 mmHg with 0.1 ml
injections into the extradural balloon, and spontaneous increases in BP occurred that initiated further
increases in ICP. Further waves of increased ICP could
be triggered at this time by hypercapnia and hypoxia.
In stage 3, ICP was approaching the level of BP, with
the vasomotor reflexes becoming less effective in
driving BP up above the ICP. EEG slowed and became
flat as ICP reached the level of BP. At this stage,
altering the arterial concentration of carbon dioxide
(PaCO2 ) had no response, an effect which Langfitt
termed vasomotor paralysis. Also at this stage,
induced changes in BP produced almost identical
changes in ICP. Deflation of the balloon at this stage
could cause a return of ICP to normal levels, with a
partial return of EEG. If balloon inflation continued,
stage 4 was entered, where decompensation was
irreversible, BP dropped and death followed. Deflation of the balloon at this time resulted in only a
temporary fall in ICP.
6.2.2 CONTROL OF INTRACRANIAL VOLUME AND

PRESSURE: CURRENT CONCEPTS
Following on from this earlier work, the research

carried out in the 1970s and early 1980s provides
much of the basis for our current concepts of intracranial pressure and craniospinal compliance.
Marmarou, interested in CSF dynamics in relation
to the pathological state of hydrocephalus, was the
first to provide a full mathematical description of the
craniospinal volumepressure relationship. Marmarou (1973) developed a mathematical model of the
CSF system that produced a general solution for the
CSF pressure. The model parameters were subsequently verified experimentally in an animal model of
hydrocephalus. As a corollary from this study, Marmarou demonstrated that the non-linear craniospinal
volumepressure relationship could be described as a
straight line segment relating the logarithm of pressure to volume, which implies a monoexponential
relationship between volume and pressure (Figure 6.2).
The slope of this relationship Marmarou termed the
pressurevolume index (PVI) which is the notional
volume required to raise ICP tenfold. Unlike elastance
(change in pressure per unit change in volume dP/
dV) or its inverse, compliance (change in volume per
unit change in pressure dV/dP), the PVI characterizes
the craniospinal volumepressure relationship over
the whole physiological range of ICP.
The PVI is calculated from the pressure change
resulting from a rapid injection or withdrawal of fluid
107
from the CSF space (Figure 6.3), and has found

widespread use both clinically and experimentally as
a measure of lumped craniospinal compliance (Marmarou, Shulman and LaMorgese, 1975; Sullivan et al.,
1977; Takagi et al., 1980; Kosteljanetz, 1985; Shapiro et
al., 1985; Takizawa, Gabra-Sanders and Miller, 1986a;
Maset et al., 1987). Any factor increasing in volume
within the craniospinal axis will deplete available
compensatory exchange space (decompensation),
reduce compliance and eventually lead to increased
intracranial pressure. Shapiro and Marmarou (1982)
have found a PVI reduced by 80% of control values to
be predictive of raised ICP in pediatric head injury.
Tans and Poortvliet (1983), also using the PVI in
patients, state that the values of 10 ml and 13 ml are
key values, with lower values indicating that active
ICP reduction and improvement in compliance are
required.
Marmarous mathematical model developed an
improved understanding not only of lumped intracranial compliance but also of the inter-relationships of
the static and dynamic processes of formation, storage
and absorption mechanisms of CSF. Clearly, the
balance between formation and storage is critical and if
the absorption of CSF is hindered, perhaps as a result of
increased CSF outflow resistance, this will result, once
the storage capacity of CSF becomes exhausted, in
raised ICP. There is a clear relationship between CSF
pressure and cerebral venous pressure and Davson
(1967) has shown that, by withdrawing CSF at the
estimated rate of CSF production (approximately
Figure 6.2 Log intracranial pressure (ICP) versus intracranial volume relationship defined by Marmarou (1973). The
pressure volume index (PVI) is the notional volume (ml) which, when added to the craniospinal volume, causes a tenfold
rise in ICP.
108
Figure 6.3 Formulae for deriving the pressure volume index (PVI), volume pressure response (VPR) and the CSF outflow
resistance (Ro ) where Po is the baseline CSF pressure, Pp is the peak pressure resulting from a bolus volume injection Vo and
P2 refers to the pressure point on the return trajectory at time T2 (usually selected at 2 min postinjection).
0.3 ml/min), it is possible to determine the baseline

cerebral venous pressure. This value can then be
substituted into the steady-state ICP equation:
ICP = formation rate outflow resistance
+ venous pressure.
Marmarou has extended Davsons work and his
general solution for intracranial pressure allowed the
derivation of an equation for CSF outflow resistance
based on a bolus injection technique (Marmarou, 1973;
Marmarou, Shulman and LaMorgese, 1975). Through
a single volume injection (Vo ) and noting the starting
pressure (Po ), peak pressure resulting from the volume injection (Pp ) and the pressure (P2 ) on the return
trajectory at a time (t2 usually 2 minutes), the
outflow resistance (Ro ) can be calculated (Figure 6.3).
In head injury management, the usefulness of knowing CSF outflow resistance stems from the premise
that increased CSF outflow resistance is one possible
non-vascular cause of raised ICP. In general terms,
causes of raised ICP can be categorized into vascular
and non-vascular mechanisms. Vascular mechanisms would include active cerebral vasodilation due
to stimuli such as increased CO2 levels or decreased
arterial inflow pressure (assuming intact pressure
autoregulation) or passive distention of cerebral
vessels in the absence of autoregulation or by venous
outflow obstruction. Non-vascular mechanisms
would include increases in brain bulk due to increased
brain water content (edema) or to an increasing
intracerebral, extradural or subdural mass. A further
non-vascular mechanism would be an increase in CSF

outflow resistance, possibly due to obstruction in the
normal CSF pathway, which results in dilation of the
channels proximal to the site of the obstruction.
The development of a simple bolus method for
measurement of CSF outflow resistance has led to
work validating the bolus method against both constant infusion and perfusion methods. Takizawa,
Gabra-Sanders and Miller (1985) confirmed the earlier
data of Sullivan et al. (1979), indicating that, under
baseline conditions, there was a significant correlation
between steady-state and dynamic measurements of
CSF outflow resistance, but the slope of the regression
line was less than unity so that the bolus method
progressively underestimated CSF outflow resistance
compared to the steady-state method. However, a
separate set of measurements was obtained after the
CSF system was loaded by infusion of artificial CSF so
that the baseline feline PVI decreased from 0.720.56.
This had no effect on the linear relationship between
dynamic and steady-state measurements, but did
produce a regression line with a much better fit to
unity. Thus, under those clinically relevant conditions
where volume buffering is likely to be decreased, the
bolus method provides an adequate measure of CSF
outflow resistance.
CSF outflow resistance measurement is used less
often in head injury research but is generally accepted
as valuable in the diagnosis of diseases associated
with disturbances in the CSF dynamics. Thus, techniques for measuring CSF outflow resistance have
found widespread application in research into hydrocephalus (Katzmann and Hussey, 1970; Ekstedt, 1978;
Borgeson and Gjerris, 1982; Tans and Poortvliet, 1984;
Borgesen et al., 1989; Maksymowicz et al., 1993), the
origin of B-waves and as a means of timing shunt
placement (Dirnagl et al., 1989; Tans and Poortvliet,
1984; Tanaka and Nishimura, 1989; Goderski and
Graff-Radford, 1993). Using these bolus techniques,
Marmarou has extended their utility in head injury by
demonstrating that, through measurement of the PVI
and CSF outflow resistance, it is possible to calculate
the percentage contribution of CSF and vascular
factors to the elevation of ICP (Marmarou et al.,
1987).
This important study has shown that the CSF
contribution to ICP in severely head-injured patients
accounts for only about 30% of the ICP rise while the
majority of ICP is attributable to vascular mechanisms. Recently, one of the basic assumptions underlying ICP dynamics tests has been called into question:
that venous outflow pressure, as estimated by sagittal
sinus pressure, remains constant. Marmarou et al.
(1993), measuring jugular bulb pressure as an estimate
of sagittal sinus pressure, showed that sagittal sinus
pressure may be elevated in approximately 40% of
severely head-injured patients. Also, in those patients
with a significant correlation between jugular bulb
pressure and intracranial pressure, there was a significantly higher percentage vascular contribution to
ICP elevation. This work shows that elevation of
venous outflow pressure does contribute to ICP
elevation and that assessment of CSF outflow resistance, PVI and jugular bulb pressure may, in selected
patients, be useful to measure when targeting therapy
for raised ICP.
At about the same time that Marmarou introduced
the PVI technique, Miller and co-workers (Miller and
Garibi, 1972; Miller, Garibi and Pickard, 1973) defined
a further measure of the craniospinal volumepressure relationship, the volume pressure response
(VPR). The VPR, calculated from the intracranial
pressure response resulting from a rapid bolus injection of saline into the CSF space, was a direct
measure, not of compliance, but of its inverse: elastance. The VPR technique was in several ways
preferable to the PVI technique in that it was a simpler
measure of craniospinal volume depletion, involving
none of the assumptions about the monoexponential
nature of the pressure versus volume relationship
inherent in Marmarous technique. Furthermore, the
VPR increases in value as the patients condition
worsens, which makes it easier to understand
clinically.
Miller pointed out that if there were only a single
volumepressure curve then no new information
would be gained by measuring compliance or ela-
109
stance, and a knowledge of absolute ICP alone would

suffice in determining the state of a patients craniospinal volume decompensation. However, Miller and
coworkers (Miller and Pickard, 1974; Miller, 1975;
Miller, Leach and Pickard, 1975) have shown that the
shape of the volumepressure relationship changes
under a variety of conditions between patients and
within patients at different times and under different
circumstances. In head-injured patients, the VPR
correlated better to the degree of brain midline shift,
as imaged on CT scan, than it did to absolute ICP
alone. The VPR served as an indicator for surgical
decompression, critical levels being between
35 mmHg/ml (Miller, Garibi and Pickard, 1973; Hase
et al., 1978).
Lofgren
(1973; Lofgren,
von Essen and Zwetnow,

1973; Lofgren
and Zwetnow, 1973, 1976) extended the

ICP range over which the volumepressure relationship was studied, including a negative pressure range
(relative to atmospheric pressure). In experimental
studies in dogs using spinal block techniques, he
showed the volumepressure curve to be the sum of
two separate curves representing high compliance
related to the spinal portion of the dural sac and a
low compliance curve, at elevated ICP, related mostly
to the cranial portion (Figure 6.4). At the most
elevated ICP, there was a sudden decrease in elastance
as ICP approached diastolic pressure, possibly due to
shifting of blood from the vascular bed when CBF
ceased (Lofgren,
von Essen and Zwetnow, 1973).

The importance of vascular factors as a determinant
of lumped craniospinal compliance was demonstrated
clearly by the work of Gray and Rosner (1987a, b),
who showed that, when CBF autoregulation was
intact with cerebral perfusion pressure (CPP) levels
greater than 50 mmHg, there was a linear increase in
PVI with increasing CPP. However, with CPPs below
the autoregulatory range, CBF fell progressively followed this time by increases in the PVI again. This
work demonstrates that the PVI is a complex function
of CPP, the direction of the CPPPVI relationship
being dependent on whether CPP is above or below
the autoregulatory range for CBF (Figure 6.5).
Not only is craniospinal compliance critically
dependent on vascular factors but Anile, Portnoy and
Branch (1987) have demonstrated that compliance is
also time-dependent. They showed that the VPR
calculated from slow, medium and rapid bolus injections yields different values. They conclude that
lumped craniospinal compliance can be divided into
two components based on the rate of injection of the
volume bolus: physical compliance and physiological
compliance. Physical compliance is a measure of such
factors as expansion of spinal dura matter and of any
minute amount of brain compression and skull expansion that may occur (Heifetz and Weiss, 1981).
110
Figure 6.4 Extended craniospinal volumepressure relationship defined by Lofgren
(1973), demonstrating low and high

elastance regions.
Figure 6.5 Plot showing the pressurevolume index (PVI), as a measure of compliance, versus cerebral perfusion pressure
(CPP), both within (squares) and beyond (triangles) an autoregulating range of CPP. Autoregulation was defined as intact
if cerebral blood flow measurements taken at the time of the PVI measurements were within 15% of baseline values. Note
that the lines of best fit through the PVI/CPP relationships have slopes of different direction depending on whether the CPP
range is within or beyond the autoregulating range. (Source: adapted from Gray and Rosner, 1987.)
Physiological compliance of the intracranial system is

related to cerebrovascular alterations, particularly
venous outflow resistance (Chopp and Portnoy, 1983).
These data show that, to understand craniospinal
pressurevolume relationships, the dynamic and the
viscoelastic properties of CSF, nervous tissue and
vascular factors must all be considered. Zee and
Shapiro (1989), using a gas-bearing electrodynamometer and a linear variable displacement transducer,
measured the relationship between brain compression
force and displacement to study the viscoelastic
properties of the brain. They demonstrated that dogs,
made hydrocephalic with intracisternal injection of
kaolin, developed brains that became less stiff (more
compliant) and more viscous than normal brain. They
propose that this weakening of brain tissue may
account for the increase in brain compliance associated with hydrocephalus, although the time course
of these changes is still unknown.
Walsh and Schettini (1989) have shown that brain
tissue elasticity bears no relation to lumped craniospinal elastance as measured by the VPR. The brain
elastic response was measured extradurally with a
coplanar transducer recording the brain displacement
simultaneously with the pressure required to cause
that brain displacement. The resulting pressure versus
brain displacement relationship is similar to the
pressure versus injection volume relationship previously described. The tangent of this brain elastic
response curve is a parameter G0 which is a measure
of brain tissue elasticity. They have shown in ten dogs,
in an extradural balloon-inflation model of raised ICP,
that the VPR increases with increasing ICP but G0
remains unchanged. Upon cardiac arrest, however, the
VPR decreases and G0 increases. They propose that G0
increases as a result of an ion shift from the extracellular to the intracellular compartment, leading to
increased intracellular water and hence increased
cellular tension (Van Harreveld and Ochs, 1956).
From this work it is clear that a knowledge of a
patients craniospinal volumepressure relationship is
an adjunct to ICP measurement for predicting states of
raised ICP. However, the use of the PVI or VPR
methods is not without disadvantages. With these
techniques there is an increased risk of infection,
usually due to Staphylococcus epidermidis, with reported infection rates ranging from 0.5% up to 9%
(Lundberg, 1960; Wyler and Kely, 1972; Troupp and
McDowell, 1976). Infection is a particular complication
of ventriculostomy and relates to the duration of
monitoring. Narayan et al. (1982) found an 8.5%
incidence of ventriculostomy-related infection in
patients monitored for over 5 days but no similar
infections in patients monitored for 3 days or less. This
relationship of duration of monitoring to risk of
infection has not been confirmed by others, and the
111
rate of access to the CSF system to obtain samples or to

recalibrate may be important. Other disadvantages
include a risk of provoking secondary pressure rises
with rapid volume injection through activation of
secondary vasodilation (Avezaat and Van Eijndhoven,
1984; Langfitt et al., 1974). Furthermore, variability
between measurements is high as it is difficult to
manually inject consistent volumes of fluid rapidly at
a constant rate of injection. As a result an average of
repeated measures is usually required.
As a consequence of these limitations the PVI or VPR
tests are not routinely used in neurosurgical practice. In
an effort to find a less invasive means to obtain this data
Avezaat and Van Eijndhoven (1979; Avezaat, Van
Eijndhoven and Wyper, 1984) systematically studied
the ICP waveform pulse amplitude (ICPplse ) as a
measure of craniospinal elastance. The rationale
behind this concept is that with each heartbeat there is
a pulsatile increase in cerebral blood volume, the
equivalent of a small intracranial volume injection,
and the ICPplse is the intracranial pressure response to
that volume increment and should therefore be directly
related to the craniospinal elastance (dP/dV). That is,
as craniospinal elastance increases (compliance decreases) the ICPplse should increase, provided that the
volume increment remain constant. The observation
that as ICP increases so does the amplitude of the
intracranial pressure pulsations is not a new one,
having been first described in 1866 by Leyden.
Avezaat and Van Eijndhoven first extended the
mathematical description of the exponential craniospinal volumepressure relationship by introducing a
constant term P0 into the pressurevolume equation
(Figure 6.6). Primarily for mathematical convenience
this term shifts the volumepressure curve as a whole
up or down its axis, which allows correction for
pressure transducer reference position and postural
changes. Mathematically, P0 is the pressure at zero
elastance (see the equation in Figure 6.6) and must
therefore have physiological significance as a determinant of the normal intracranial equilibrium pressure
and Zwetnow (1973) showed that

(Peq ). Lofgren
alterations in CVP can shift the pressurevolume
curve up or down its axis, which would suggest CVP
may be a factor determining P0.
Avezaat and Van Eijndhoven described the mathematical relationship between ICPplse and ICP by
substituting the ICPplse for the elastance (dP/dV) and
pulsatile blood volume for the volume injection. This
relationship was verified in both clinical and experimental studies. They found that the ICPplse increased
linearly with ICP up to a pressure of 60 mmHg,
whereupon a breakpoint occurred (Figure 6.7). Above
60 mmHg the ICPplse increased more rapidly with
rising ICP. They argue that the breakpoint is a marker
for loss of CBF autoregulation, postulating that onset of
112
P = Peq eE1 dVe + Po

Figure 6.6 Craniospinal volumepressure relationship demonstrating that for the same increase in craniospinal volume
(dVe ) the ICP pulse amplitude (dP) increases when total craniospinal volume (Ve ) increases. This is due to the exponential
nature of the curve, which is described mathematically by the equation. P = ICP; Peq = equilibrium ICP; E1 = elastic
coefficient; Ve = elastic volume (addition to total volume); P0 = constant term.
vasomotor paralysis causes a decreased arteriolar

inflow resistance, which results in an increased phase
shift between the inflow and outflow pulsatile blood
volume. This translates to an overall increased intracranial pulsatile blood volume and will tend to increase
the slope of the ICPplse versus ICP relationship.
It is assumed that the pulsatile blood volume (dV),
the input function to the elastance calculation (dP/
dV), is unchanging. This is a tenuous assumption in
severely injured patients, some of whom may have

compromised or fluctuating cardiovascular function.
As a consequence of the dependence of the ICPplse
versus ICP relationship on the pulsatile blood volume,
the clinical utility of this technique as a measure of
lumped craniospinal elastance is limited unless a
measure of the pulsatile blood volume can be monitored simultaneously and controlled for in patients.
Despite this limitation, analysis of the ICP pulse
Figure 6.7 ICP pulse amplitude (ICPplse ) versus intracranial pressure relationship defined by Avezaat and Van Eijndhoven
(1979), demonstrating a direct linear dependence of ICPplse on mean ICP. A breakpoint occurs at an ICP of approximately
60 mmHg whereupon the slope of the relationship increases.
amplitude is still widely studied clinically and may, in

the future, prove useful as an estimate of cerebrovascular autoregulatory reserve (Laniewski, Czosnyka
and Maksymowicz, 1993).
Portnoy and Chopp recognized the importance of
measurement of the input function in an analysis of
the ICPplse and were the first to apply a systems
analysis approach to the ICP waveform (Chopp and
Portnoy, 1980). Systems analysis is a technique
whereby an attempt is made to define the physical
characteristics of a system using only the system
input and output signals (Marmareliz and Marmareliz, 1987). Portnoy and Chopps method assumes that
the BP waveform is the chief input signal to the
cerebrovascular system and the ICP waveform is the
output response to that stimulus. Both BP and ICP
waveforms are converted into the frequency domain
by Fourier analysis, and the resulting frequency
spectra are used in the calculation of the system
transfer function (Figure 6.8). The system transfer
function consists of amplitude and phase components.
The amplitude transfer function is a measure of how
much pressure is transmitted through the cerebrovascular bed at a given frequency. The phase transfer
function is a measure of how much a pressure signal is
phase-shifted as it is transmitted through the cerebrovascular bed at a given frequency.
113
Using these methods, Portnoy and Chopp (1981)

found, in an experimental model of raised ICP in cats,
that arterial hypercarbia and hypoxia produced an
increase in ICPplse, and an increase mainly in the
fundamental of the amplitude transfer function. The
changes induced were greater than those caused by
intraventricular infusion of saline to the same ICP
level. The VPR was less during hypercapnia than
during intraventricular infusion at the same ICP level,
which suggests that the increase in ICPplse is related
more to arteriolar vasodilation than to steepening of
the craniospinal volumepressure relationship.
Extending their model by including analysis of the
sagittal sinus pressure (SSP) waveform in dogs,
Portnoy et al. (1982) found that the ICP waveform and
SSP waveform were almost identical, indicating that
the ICPplse is derived from the cerebral venous bed.
With the animals breathing pure oxygen it was
observed that an attenuation of the amplitude transfer
function fundamental occurred in the transmission of
the arterial pulse through to the CSF space under
conditions of low ICP (ICP < 7 mmHg). They attributed this attenuation to functional autoregulatory tone
of the precapillary cerebral resistance vessels, and
further demonstrated that a flat amplitude transfer
function (equal transmission of all harmonics) can be
experimentally induced by intraventricular infusion
Figure 6.8 Systems analysis. The input blood pressure (f(t)) and output intracranial (g(t)) waveforms recorded from
locations across the cerebrovascular bed (CVB) can be described by spectral analysis in terms of their harmonic components
(F() and G()). Spectral analysis of the BP and ICP signals resolved each waveform as a series of sine waves consisting of
a fundamental component and five harmonics of the fundamental. The transfer function (H()) defines how the input signal
is transformed into the output signal and consists of amplitude and a phase components. The amplitude curve describes
how much pressure is transmitted through the CVB at each harmonic frequency. The phase curve describes how much each
pressure sine wave is shifted in its cycle as it is transmitted through the CVB.
114
of mock CSF or arterial hypercarbia. They propose

that the conversion from an attenuated low frequency
transmission to a flat amplitude transfer function is
evidence for reduced arteriolar vasomotor tone.
Applying these techniques to hydrocephalic dogs,
Portnoy, Branch and Chopp (1985) found that, when
ICP was less than 9 mmHg, there was an attenuated
low-frequency transmission from BP to ICP; however,
when ICP was greater than 12 mmHg a flat amplitude
transfer function was present. These findings in
hydrocephalic dogs were identical to non-hydrocephalic controls, and they concluded that pressure transmission from BP to ICP in hydrocephalic dogs was
determined by the cerebrovascular bed and not by the
hydrocephalic process.
Branch, Chopp and Portnoy (1989), recording the
pressure waveform from a small cortical vein, provided further evidence that the ICP waveform was
derived from the cerebral venous bed by demonstrating that the ICP and cortical venous amplitude spectra
were identical across a variety of experimental
conditions.
Using similar techniques to Portnoy and Chopp,
Takizawa, Gabra-Sanders and Miller (1987) have
shown that the first four harmonics of the ICP
waveform and the amplitude transfer function all
show a positive correlation to raised ICP and an
inverse correlation to CPP. A distortion factor k was
used to show that, as ICP increased towards 50 mmHg
the ICP waveform became more like a sine wave,
changing less as ICP continued to increase. Takizawa,
Gabra-Sanders and Miller (1986b) also found that
cerebrovascular pressure transmission increased with
saline infusion and arterial hypercarbia equally when
ICP was recorded either in the lateral ventricle or in
the cisterna magna, but transmission was attenuated
in the lumbar space. This attenuation in the lumbar
space was decreased by saline loading of the craniospinal axis. They propose that the spinal sac functions
as a low-pass filter to the conduction of the ICPplse.
Also using a systems analysis approach to studying
cerebrovascular pressure transmission, but using different methods, Kasuga et al. (1987) have demonstrated resonance within the intracranial cavity in
dogs. They randomized pressure pulse transmission
into the cranial cavity through the control of an
implanted cardiac pacemaker. Using the carotid pulse
waveform as an input function and the extradural
pressure waveform as the output function, they
calculated the transfer function from the autocorrelation of the input function and the cross-correlation of
the input and output functions by means of a least
squares method. They showed that the amplitude
transfer function decreased between the frequencies of
1 and 7 Hz, then suddenly increased to form a marked
peak at about 1015 Hz, whereupon the phase transfer
function also changed from positive to negative. This

showed that the lower frequencies of the pulse wave
were suppressed during transmission through the
intracranial cavity and that resonance was present
under normal intracranial conditions. Kasuga, Nagai
and Hasegawa (1989) subsequently showed that, with
both extradural balloon inflation and intraventricular
infusion models of raised ICP in dogs, the resonant
frequency increased above the control value. Associated with this increase in resonant frequency, there
was an increased transfer of the low frequency
components. However, with arterial hypercarbia, ICP
increased but with no significant change in the
resonant frequency, although low-frequency pressure
transmission increased in a similar fashion to both
groups.
Bray and Robertson (Bray et al., 1986), also using
Fourier analysis of the ICP waveform in patients,
identified two main frequency bands in the ICP
waveform power spectrum. The centroid (powerweighted average frequency) of the low frequency
band (0.22.6 Hz) they correlated to cerebral blood
flow using the nitrous oxide method, while the highfrequency band centroid (415 Hz) they found
inversely correlated to the PVI as a measure of
craniospinal compliance. Further clinical experience
(Robertson et al., 1989) with the high-frequency centroid showed that the percentage of time spent with a
high-frequency centroid greater than 9 Hz bore no
relation to ICP but that the centroid frequency
correlated exponentially to increased mortality. Case
reports showed that the high-frequency centroid was a
better measure of the clinical state of the patient than
was the absolute ICP alone. However, subsequent
work shows that the high-frequency centroid is also
affected by heart rate thus diminishing its predictive
reliability (Contant et al., 1993).
Adapting the systems analysis method of Portnoy
and Chopp to a clinical study of cerebrovascular
pressure transmission, Piper et al. (1990a), in an
observational study of 1500 pressure records in 30
severely head-injured patients, identified four patterns of amplitude transfer function. Both forms that
showed an elevated fundamental pressure transmission from BP to ICP were associated with raised ICP,
whereas the remaining forms with a normal fundamental amplitude transfer function were associated
with ICP less than 15 mmHg. Following on from this
work a further explanatory experimental study was
performed in cats, demonstrating that the fundamental amplitude transfer function can be increased
by active arteriolar vasodilation, by loss of autoregulatory vascular tone, or through reduced cerebrovascular transmural pressure (Piper et al., 1993). It
may be possible to distinguish these mechanisms,
working from the observed phase shift between the
fundamental of the BP and ICP waveform. In this

experimental model, active arteriolar vasodilation was
followed by an increasingly negative phase shift and
decreased transmural pressure resulted in no overall
phase shift whereas impaired autoregulation showed
that an increased fundamental amplitude was accompanied by a positive phase shift. Further studies are
needed that correlate these ICP waveform measures
with CBF and pressure autoregulation in head-injured
patients.
Another area of research showing promise as a
means of studying the effect of intracranial hypertension on craniospinal compliance and autoregulatory reserve concerns the continuous measure of
transcranial middle cerebral artery (MCA) flow velocity and its correlation with CPP. Chan et al. (1993)
demonstrated that, in continuously monitored headinjured patients, the MCA Doppler pulsatility index
(PI; systolicdiastolic/mean flow velocity), when
plotted against CPP, showed a breakpoint at
70 mmHg below which the PI increases. Simultaneous
measurement of jugular venous oxygen saturation in
the same patients demonstrated a fall in jugular
venous saturation towards ischemic levels from the
same CPP breakpoint. This data would indicate that
below a CPP threshold of 70 mmHg autoregulation in
these patients was becoming exhausted (Chapter 14).
This information is useful as it provides a means of
determining the optimal CPP threshold for treating
raised ICP at any time during the management of
head-injured patients. If there was only one critical
115
CPP threshold it would be a simple matter to treat if

CPP fell below this threshold. However, there is now
increasing evidence, from both clinical and experimental studies, that as a result of the varying severity
of head injury and the development of the injury
process with time, the critical CPP threshold changes
both between patients and within patients on different
days (Price et al., 1994; Wong et al., 1995; Lewis et al.,
1995). The development of analysis methods for
detecting changes in CPP breakpoint may have a
significant impact on the future management of
cerebral perfusion.
A similar relationship between CPP and exhaustion
of autoregulation may also be identified through
analysis of the ICP waveform (Figure 6.9). Although
this relationship is currently under study in headinjured patients, it confirms the earlier experimental
study in cats of Takizawa, Gabra-Sanders and Miller
(1987), who demonstrated that the fundamental
amplitude transfer function showed a positive correlation to raised ICP and an inverse correlation to CPP,
with the latter demonstrating a breakpoint phenomenon as CPP exceeded 60 mmHg.
Much of the work just described shows promise in
elucidating the status of cerebral autoregulation in
head-injured patients, but we are still no closer to
improving upon the methods developed by Marmarou and Miller for assessment of the craniospinal
volumepressure status. However, some of the limitations of these manual volumepressure techniques are
now being overcome as a result of innovative applica-
Figure 6.9 Plot of the ICP-waveform-derived parameter (ICP pulse to BP pulse ratio) versus cerebral perfusion pressure
(CPP). At a CPP of less than 70 mmHg, a breakpoint occurs where there is an increasing transfer of the BP pulse through
to the CSF.
116
tion of computer technology. For example, Smielewski

et al. (1993) reported a new method of measuring
craniospinal compliance and CSF outflow resistance
in hydrocephalic patients, based on controlled CSF
drainage. This system uses an electromagnetically
driven clamp, which opens or closes the outlet of the
lumbar drain under computer control, permitting online controlled drainage of CSF, measurement of CSF
pressure and volume. This automated drainage
method does not raise ICP and so overcomes the risk
of provoking uncontrolled rises in ICP associated with
the continuous infusion or bolus techniques. Similarly,
Piper et al. (1990b) reported on an automated method
for measuring craniospinal compliance based on an
electronic square wave pressure generator triggered
under computer control and able to produce small
(0.05 ml) volume injection/withdrawal sequences into
the CSF space. The compliance is calculated with this
method from the amplitude of the intracranial pressure response to this small volume increase. The
resulting pressure response itself is small (12 mmHg)
and is isolated from background noise using computer-controlled signal averaging. Work is under way
developing a practical clinical device based on this

technique for measuring craniospinal compliance in
patients at risk of raised ICP.
Also as a result of improvements in both hardware
and software computer technology, it is becoming
more practical to develop increasingly complex mathematical models of the craniospinal system. Some
investigators have modeled the craniospinal and
cerebrovascular systems as second- or higher-order
systems containing a series of distributed resistive,
inductive and capacitive components (Takemae et al.,
1987; Hoffmann, 1987; Sorek, Bear and Karni, 1988;
Ursino, 1988). More recently, finite element analysis
has been applied to model the viscoelastic properties
of the brain (Nagashima, Shirakumi and Rapoport,
1990; Hamano et al., 1993). Such models are becoming
increasingly useful as methods for the controlled
testing of hypotheses and simulation of physiological
conditions that would otherwise be difficult to reproduce in animal models. For example, Czosnyka et al.
(1993) developed a mathematical model of the cerebrovascular bed and craniospinal compartment (Figure 6.10). The electrical equivalent circuit of this model
Figure 6.10 Electrical equivalent circuit of a cerebral blood flow (CBF) and CSF circulation model developed by Czosnyka
et al. (1993). The upper figure defines the model parameters and the lower figure shows the autoregulatory relationship
between vascular resistance (CVR) and CPP. ABP = arterial blood pressure; Pa = arterial pressure in basal cerebral arteries;
Pv = cerebral venous pressure; Pi = intracranial pressure; Pss = sagittal sinus pressure; Ca = compliance of arterial bed; Cv
= compliance of venous bed; Ci = complicance of CSF lumbar compartment; If = CSF formation rate; RCSF = CSF
reabsorption; Rb = resistance of bridging veins; Ra = resistance of basal arteries; CVR = main cerebrovascular resistance.
REFERENCES
comprises arterial and venous resistances, capacitances (or compliance), CSF formation rate, a nonlinear craniospinal compliance, CSF outflow resistance and arterial and venous pressure sources. What is
particularly useful in this model is the non-linear
arterial resistance characteristic (Figure 6.10), which
allows representation of the autoregulatory process to
changes both in cerebral perfusion pressure and
arterial CO2. Using this model it was possible analytically to define the CPP-dependent ICP pulse amplitude and transcranial Doppler flowvelocity relationship. This confirmed the earlier experimental and
clinical data of Avezaat, Van Eijnhoven and Wyper
(1979), Takizawa, Gabra-Sanders and Miller (1987),
Czosnyka et al. (1990) and Chan et al. (1992).
In conclusion, current research using multimodality
monitoring and applied computer technology is a
most promising approach to the study of raised ICP
and may prove to be a powerful aid in the investigation of cerebrovascular pathophysiology and craniospinal volumepressure relationships.
6.3
References
Alberico, A. M., Ward, J. D., Choi, S. C. et al. (1987) Outcome after severe head
injury: relationship to mass lesions, diffuse injury and ICP course in
pediatric and adult patients. Journal of Neurosurgery, 67, 648656.
Anile, C., Portnoy, H. D. and Branch, C. (1987) Intracranial compliance is timedependent. Neurosurgery, 20, 389395.
Avezaat, C. J. J. and Van Eijndhoven, J. H. M. (1984) Cerebrospinal fluid pulse
pressure and craniospinal dynamics: a theoretical, clinical and experimental
study. Thesis, Erasmus University, Rotterdam.
Avezaat, C. J. J., Van Eijndhoven, J. H. M. and Wyper, D. J. (1979)
Cerebrospinal fluid pulse pressure and intracranial volumepressure
relationships. Journal of Neurology, Neurosurgery and Psychiatry 42,
687700.
Ayala, G. (1923) Uber

den diagnostischen Wert des Liquordruckes und einen
Apparat zu seiner Messung. Zeitschirft fur
die gesamte Neurologie und
Psychiatrie, 84, 4295.
Ayala, G. (1925) Die Physiopathologie der Mechanik des Liquor cerebrospinalis und der Rachidealquotient. Monatschrift fur
die Psychiatrie und
Neurologie, 58, 65105.
Bayliss, W. M., Hill, L. and Gulland, G. L. (1895) On intracranial pressure and
the cerebral circulation. Journal of Physiology, 18, 334362.
Becker, D. P., Miller, J. D., Ward, J. D. et al. (1977) The outcome from severe
head injury with early diagnosis and intensive management. Journal of
Neurosurgery 47, 491502.
Bouma, G. J. and Muizelaar, J. P. (1990) Relationship between cardiac output
and cerebral blood flow in patients with intact and with impaired
autoregulation. Journal of Neurosurgery, 73, 368374.
Borgeson, S. E. and Gjerris, F. (1982) The predictive value of conductance to
outflow of CSF in normal pressure hydrocephalus. Brain, 105, 6586.
Borgeson, S. E., Gjerris, F., Fedders, O. et al. (1989) Measurement of resistance
to CSF outflow clinical experience in 333 cases, in Intracranial Pressure VII,
(eds J. T. Hoff and A. L. Betz), Springer-Verlag, Berlin, pp. 353359.
Branch, C. A., Chopp, M. and Portnoy, H. D. (1989) Fourier analysis of
intracranial pressures during experimental intracranial hypertension, in
Berlin, pp. 175180.
Bray, R. S., Sherwood, A. M., Halter, J. A. et al. (1986) Development of a clinical
monitoring system by means of ICP waveform analysis, in Intracranial
Pressure VI, (eds J. D. Miller, G. M. Teasdale and J. O. Rowan), SpringerVerlag, Berlin, pp. 260264.
Browder, J. and Meyers, R. (1936) Observations on behaviour of the systemic
blood pressure, pulse and spinal fluid pressure following craniocerebral
injury. American Journal of Surgery, 31, 403427.
Burrows, G. (1846) On Disorders of the Cerebral Circulation and on the Connection
Between Affections of the Brain and Diseases of the Heart, Longman, London.
Chan, K. H., Miller, J. D. and Piper, I. R. (1992) Cerebral blood flow at constant
cerebral perfusion pressure but changing arterial and intracranial pressure:
relationship to autoregulation. Journal of Neurosurgery and Anaesthetics, 4,
188193.
117
Chan, K. H., Miller, J. D., Dearden, N. M. et al. (1992) The effect of changes in
cerebral perfusion pressure upon middle cerebral artery blood flow velocity
and jugular bulb venous oxygen saturation after severe brain injury. Journal
Chan, K. H., Dearden, N. M., Miller, J. D. et al. (1993) Multimodality
monitoring as a guide to treatment of intracranial hypertension after severe
head injury. Neurosurgery, 32, 547553.
Choi, S. C., Muizelaar, J. P., Barnes, T. Y. et al. (1991) Prediction tree for severely
head injured patients. Journal of Neurosurgery, 75, 251255.
Chopp, M. and Portnoy, H. (1980) Systems analysis of intracranial pressure.
Comparison with volumepressure test and CSF-pulse amplitude analysis.
Chopp, M. and Portnoy, H. D. (1983) Hydraulic model of the cerebrovascular
bed: an aid to understanding the volumepressure test. Neurosurgery, 13,
511.
Contant, C. F., Robertson, C. S., Narayan, R. K. et al. (1993) Effects of heart rate
on the shape of the intracranial pressure wave and related parameters, in
Intracranial Pressure VIII, (eds C. J. J. Avezaat, J. H. M. van Eindhoven, A. I.
R. Maas and J. T. J. Tans), Springer-Verlag, Berlin, pp. 356362.
Corrie, J., Piper, I. R., Housley, A. et al. (1993) Microcomputer based data
recording improves identification of secondary insults during intensive
care management of head injured patients. British Journal of Intensive Care, 3,
225233.
Czosnyka, M., Piechnik, S., Koszewski, W. et al. (1993) The dynamics of
cerebral blood flow, perfusion pressure and CSF circulation a modeling
study, in Intracranial Pressure VIII, (eds C. J. J. Avezaat, J. H. M. van
Eindhoven, A. I. R. Maas and J. T. J. Tans), Springer-Verlag, Berlin, pp.
699706.
Czosnyka,M., Batorski, L., Laniewski, P. et al. (1990) A computer system for
the identification of the cerebrospinal compensatory model. Acta Neurochirurgica (Vienna), 105, 112116.
Cushing, H. (1901) Concerning a definite regulatory mechanism of the
vasomotor centre which controls blood pressure during cerebral compression. Bulletin of the Johns Hopkins Hospital, 12, 290292.
Cushing, H. (1902) Some experimental and clinical observations concerning
states of increased intracranial tension. American Journal of Medical Science,
124, 375400.
Cushing, H. (1903) The blood pressure reaction of acute cerebral compression,
illustrated by cases of intracranial haemorrhage. American Journal of Medical
Science, 125, 10171044.
Davson, H. (1967) Physiology of the Cerebrospinal Fluid, J. & A. Churchill,
London.
DeSalles, A. F., Muizelaar, J. P. and Young, H. F. (1987) Hyperglycemia,
cerebrospinal fluid lactic acidosis and cerebral blood flow in severely headinjured patients. Neurosurgery, 21, 4550.
Dirnagl, U., Garner, C., Haberl, R. et al. (1989) Correlation between B-Waves
and intracranial pressurevolume relationships, in Intracranial Pressure VII,
Duret, H. (1878) Etudes experimentales et cliniques sur les traumatismes cerebraux,
Paris.
Ekstedt, J. (1978) CSF hydrodynamic studies in man: 2. Normal hydrodynamic variables related to CSF pressure and flow. Journal of Neurology,
Evans, J. P., Espey, F. F., Kristoff, F. V. et al. (1951) Experimental and clinical
observations on rising intracranial pressure. Archives of Surgery, 63,
107114.
Flexner, L. B., Clark, J. H. and Weed, L. H. (1932) The elasticity of the dural sac
and its contents. American Journal of Physiology, 101, 292303.
Flexner, L. B. and Weed, L. H. (1933) Factors concerned in positional
alterations of intracranial pressure. American Journal of Physiology, 104,
681692.
Gentleman, D. and Jennett, B. (1981) Hazards of interhospital transfer of
comatose head injured patients. Lancet, ii, 835855.
Godersky, J. C. and Graff-Radford, N. R. (1993) Evaluation of pressure
monitoring and CSF conductance in NPH, in Intracranial Pressure VIII, (eds
C. J. J. Avezaat, J. H. M. van Eindhoven, A. I. R. Maas and J. T. J. Tans),
Springer-Verlag, Berlin, pp. 811815.
Graham, D. I., Hume Adams, J. and Doyle, D. (1978) Ischaemic brain damage
in fatal non-missile head injuries. Journal of the Neurological Sciences, 39,
213234.
Graham, D. I., Ford, I., Hume Adams, J. H. et al. (1989) Ischaemic brain
damage is still common in fatal non-missile head injury. Journal of
Gray, W. J. and Rosner, M. J. (1987a) Pressurevolume index as a function of
cerebral perfusion pressure. Part 1: The effects of cerebral perfusion
pressure changes and anaesthesia. Journal of Neurosurgery, 67, 369376.
Gray, W. J. and Rosner, M. J. (1987b) Pressurevolume index as a function of
cerebral perfusion pressure. Part 2: The effects of low cerebral pressure and
Grossman, R. G., Turner, J. W., Miller, J. D. and Rowan, J. D. (1975) The
relationship between cortical electrical activity, cerebral perfusion pressure
and cerebral blood flow during increased intracranial pressure, in Cerebral
Circulation and Metabolism, (eds T. W. Langfitt, L. C. McHenry, M. Reivich
and H. Wollman), Springer-Verlag, Berlin, pp.232234.
118
Hamano, S., Nagashima, T., Kojima, N. et al. (1993) A finite element analysis
of intracerebral stress distribution with viscoelastic model, in Intracranial
Pressure VIII, (eds C. J. J. Avezaat, J. H. M. van Eindhoven, A. I. R. Maas and
J. T. J. Tans), Springer-Verlag, Berlin, pp. 716719.
Harper, A. M. (1966) Autoregulation of cerebral blood flow: influence of the
arterial pressure on blood flow through the cerebral cortex. Journal of
Hase, U., Reulen, H. J., Meinig, G. and Schurmann,
K. (1978) The influence of

the decompressive operation on the intracranial pressure and the pressure
volume relation in patients with severe head injuries. Acta Neurochirurgica,
45, 113.
Heifetz, M. D. and Weiss, M. (1981) Detection of skull expansion with
increased intracranial pressure. Journal of Neurosurgery, 55, 811812.
Hoffmann, O. (1987) Biomathematics of intracranial CSF and haemodynamics. Simulation and analysis with the aid of a mathematical model. Acta
Jaggi, L. J., Obrist, W. D., Gennarelli, T. A. and Langfitt, T. W. (1990)
Relationship of early cerebral blood flow and metabolism to outcome in
acute head injury. Journal of Neurosurgery, 72, 176182.
Jennet, W. B. (1961) Experimental brain compression. Archives of Neurology, 4,
599607.
Johnston, I. H., Johnston, J. J. and Jennett, W. B. (1970) Intracranial pressures
following head injury. Lancet, ii, 433436.
Johnston, I. H. and Paterson, A. (1972) Benign intracranial hypertension:
aspects of diagnosis and treatment, in Optic Nerve, (ed. J. S. Cant), Himpton,
London, pp. 155165.
Johnston, I. H. and Rowan, J. O. (1974) Raised intracranial pressure and
cerebral blood flow: 3. Venous outflow tract pressures and vascular
resistances in experimental intracranial hypertension. Journal of Neurology,
Kasuga, Y., Nagai, H. and Hasegawa, Y. (1989) Transmission characteristics of
pulse waves in the intracranial cavity of dogs during normal intracranial
condition, intracranial hypertension, hypercapnia and hydrocephalus, in
Berlin, pp. 196200.
Kasuga, Y., Nagai, H., Hasegawa, Y. and Nitta, M. (1987) Transmission
characteristics of pulse waves in the intracranial cavity of dogs. Journal of
Katzmann, R. and Hussey, F. (1970) A simple constant infusion manometric
test for measurement of CSF absorption. Neurology, 20, 534545.
Kawai, N., Nagao, S., Fujisawa, M. et al. (1993) Analysis of pyramidal tract
function in experimental intracranial hypertension by magnetic stimulation,
in Intracranial Pressure VIII, (eds C. J. J. Avezaat, J. H. M. van Eindhoven, A. I.
Kellie, G. (1824) An account of the appearances observed in the dissection of
two of three individuals presumed to have perished in the storm of the
third and whose bodies were discovered in the vicinity of Leith on the
morning of the 4th, November 1821, with some reflections on the pathology
of the brain. Transactions of the Medico-Chirurgical Society of Edinburgh, 1,
84169.
Kocher, T. (1901) Hirnerschutterung, Hirndruck und chirurgische Eingriffe
beim Hirnerkrankungen, in Nothnagel: Spezielle Pathologie und Therapie,
Vienna.
Kosteljanetz, M. (1985) Pressurevolume conditions in patients with subarachnoid and or intraventricular haemorrhage. Journal of Neurosurgery, 63,
398403.
Langfitt, T. W. (1969) Increased intracranial pressure. Clinical Neurosurgery, 16,
436471.
Langfitt, T. W. (1975) Intracranial volumepressure relationship, in Intracranial Pressure II, (eds N. Lundberg, U. Ponten and M. Brock), SpringerVerlag, Berlin, pp. 6976.
Langfitt, T. W., Weinstein, J. D. and Kassel, N. F. (1965) Cerebral vasomotor
paralysis produced by intracranial hypertension. Neurology, 15, 622641.
Langfitt, T. W., Weinstein, J. D., Kassell, N. F. and Simeone, F. A. (1964a)
Transmission of increased intracranial pressure I: within the craniospinal
axis. Journal of Neurosurgery, 21, 989997.
Langfitt, T. W., Weinstein, J. D., Kassell, N. F. and Gagliardi, L. J. (1964b)
Transmission of increased intracranial pressure II: within the supratentorial
space. Journal of Neurosurgery, 21, 9981005.
Langfitt, T. W., Kumar, V. S., James, H. E. and Miller, J. D. (1974) Continuous
recording of intracranial pressure in patients with hypoxic brain damage, in
Brain Hypoxia, (eds J. S. Brierley and B. S. Meldrum), Heinemann, London,
pp. 118135.
Laniewski, P., Czosnyka, M. and Maksymowicz, W. (1993) Continuous
analysis of the intracranial pressure waveform as a method of autoregulatory reserve assessment, in Intracranial Pressure VIII, (eds. C. J. J.
Avezaat, J. H. M. van Eindhoven, A. I. R. Maas and J. T. J. Tans), SpringerVerlag, Berlin, pp. 376381.
Leyden, E. (1866) Beitrage und Untersuchungen zur Physiologie und
Pathologie des Gehirns. Uber

Hirndurck und Hirnbewegungen. Virchows
Archiv, 37, 519559.
Lewelt, W., Jenkins, L. W. and Miller, J. D. (1980). Autoregulation of cerebral

blood flow after experimental fluid percussion injury of the brain. Journal of
Lewelt, W., Jenkins, L. W. and Miller, J. D. (1982) Effects of experimental fluid
percussion injury of the brain as cerebrovascular reactivity to hypoxia and
hypercapnia. Journal of Neurosurgery, 56, 332338.
Lewis, S. B., Wong, M. L. H., Bannan, P. E. et al. The relationship of cerebral
blood flow to transcranial Doppler indices and cerebrovascular waveform
analysis in a sheep model. Journal of Neurotrauma, 12, 414.
Lofgren,
J. (1973) Pressurevolume relationships of the cerebrospinal fluid

system. Thesis, University of Goteborg.
Lofgren,
J., von Essen, C. and Zwetnow, N. N. (1973) The pressurevolume

curve of the cerebrospinal fluid space in dogs. Acta Neurologica Scandinavica,
49, 557574.
Lofgren,
J. and Zwetnow, N. N. (1973) Cranial and spinal components of the

cerebrospinal fluid pressurevolume curve. Acta Neurologica Scandinavica,
49, 575585.
Lofgren,
J. and Zwetnow, N. N. (1976) Intracranial blood volume and its

variation with changes in intracranial pressure, in Intracranial Pressure III,
(eds J. W. F. Beks, D. A. Bosch and M. Brock), Springer-Verlag, Berlin, pp.
2528.
Luerssen, T. G., Klauber, M. R. and Marshall, L. F. (1988) Outcome from head
injury related to patients age: a longitudinal prospective study of adult and
paediatric head injury. Journal of Neurosurgery, 68, 409416.
Lundberg, N. (1960) Continuous recording and control of ventricular fluid
pressure in neurosurgical practice. Acta Psychiatrica et Neurologica Scandinavica, 36(Suppl.), 149.
Lundberg, N., Troupp, H. and Lorin, H. (1965) Continuous recording of the
ventricular fluid pressure in patients with severe acute traumatic brain
injury. Journal of Neurosurgery, 22, 581590.
Maksymowicz, W., Czosnyka, M., Koszewski, W. et al. (1993) Post-shunting
improvement in hydrocephalic patients described by cerebrospinal compensatory parameters, in Intracranial Pressure VIII, (eds C. J. J. Avezaat, J. H.
M. van Eindhoven, A. I. R. Maas and J. T. J. Tans), Springer-Verlag, Berlin,
pp. 829832.
Marmareliz, P. Z. and Marmareliz, V. Z. (1987) Analysis of Physiological Systems:
The White Noise Approach, Plenum Press, New York, pp. 1128.
Marmarou, A. (1973) A theoretical and experimental evaluation of the
cerebrospinal fluid system, PhD Thesis, Drexel University.
Marmarou, A., Shulman, K. and LaMorgese, J. (1975) Compartmental analysis
of compliance and outflow resistance of the cerebrospinal fluid system.
Marmarou, A., Maset, A. L., Ward, J. D. et al. (1987) Contribution of CSF and
vascular factors to elevation of ICP in severely head-injured patients.
Marmarou, A., Anderson, R. L., Ward, J. D. et al. (1991) Impact of ICP
instability on outcome in patients with severe head trauma. Journal of
Neurosurgery, 75, S59S66.
Marmarou, A., Foda, M. A., Bandoh, K. et al. (1993) Elevated venous outflow
pressure in head injured patients, in Intracranial Pressure VIII, (eds C. J. J.
Marshall, L. F., Smith, R. W. and Shapiro, H. M. (1979) The outcome with
aggressive treatment in severe head injuries. Part 1: The significance of
intracranial pressure monitoring. Journal of Neurosurgery, 50, 2025.
Maset, A. L., Marmarou, A., Ward, J. D. et al. (1987) Pressurevolume index in
Masserman, J. H. (1934) Cerebrospinal hydrodynamics IV: Clinical experimental studies. Archives of Neurology and Psychiatry, 32, 523553.
Masserman, J. H. (1935) Intracranial hydrodynamics: central nervous system
shock and edema following rapid fluid decompression of ventriculosubarachnoid spaces. Journal of Nervous and Mental Diseases, 80, 138158.
Miller, J. D. (1975) Volume and pressure in the craniospinal axis. Clinical
Miller, J. D. and Adams J. H. (1992) The pathophysiology of raised intracranial
pressure, in Greenfields Neuropathology, 5th edn, (eds J. H. Adams and L. W.
Duchen), Edward Arnold, Sevenoaks, pp. 69105.
Miller, J. D. and Becker, D. P. (1982) Secondary insults to the injured brain.
Journal of the Royal College of Surgeons of Edinburgh, 27, 292298.
Miller, J. D., Becker, D. P. and Ward, J. D. et al. (1977) Significance of
intracranial hypertension in severe head injury. Journal of Neurosurgery, 47,
503516.
Miller, J. D. and Garibi, J (1972) Intracranial volume/pressure relationships
during continuous monitoring of ventricular fluid pressure, in Intracranial
Pressure, (eds M. Brock and H. Dietz), Springer-Verlag, Berlin, pp.
270274.
Miller, J. D., Garibi, J. and Pickard, J. D. (1973) Induced changes of
cerebrospinal fluid volume: effects during continuous monitoring of
ventricular fluid pressure. Archives of Neurology, 28, 265269.
Miller, J. D., Leech, P. J. and Pickard, J. D. (1975) Volume pressure response in
various experimental and clinical conditions, in Intracranial Pressure II, (eds
N. Lundberg, U. Ponten and M. Brock), Springer-Verlag, Berlin, pp. 9799.
Miller, J. D. and Pickard, J. D. (1974) Intracranial volume/pressure studies in
patients with head injury. Injury, 5, 265268.
REFERENCES
Miller, J. D., Stanek, A. E. and Langfitt, T. W. (1972) Concepts of cerebral
perfusion pressure and vascular compression during intracranial hypertension, in Progress in Brain Research, vol. 35, Cerebral Blood Flow, (eds J. S.
Meyer and J. P. Schade), Elsevier, Amsterdam, pp. 411432.
Miller, J. D. Butterworth, J. E., Gudeman, S. K. et al. (1981) Further experience
in the management of severe head injury. Journal of Neurosurgery, 54,
289299.
Miller, J. D, Jones, P. A., Dearden, N. M. and Tocher, J. L. (1992) Progress in the
management of head injury. British Journal of Surgery, 79, 6064.
Monro, A. (1783) Observations on the Structure and Function of the Nervous
System, Creech & Johnston, Edinburgh.
Muizelaar, J. P. and Becker, D. P. (1986) Induced hypertension for the
treatment of cerebral ischemia after subarachnoid hemorrhage. Direct effect
on cerebral blood flow. Surgical Neurology, 25, 317325.
Muizelaar, J. P., Lutz, H. A. and Becker, D. P. (1984) Effect of mannitol on ICP
and CBF and correlation with pressure autoregulation in severely headinjured patients. Journal of Neurosurgery, 61, 700706.
Muizelaar, J. P. Marmarou, A., DeSalles, A. A. F. et al. (1989) Cerebral blood
flow and metabolism in severely head-injured children. Part 1: Relationship
with GCS score, outcome, ICP and PVI. Journal of Neurosurgery, 71, 6371.
Muizelaar, J. P. Ward, J. D., Marmarou, A. et al. (1989) Cerebral blood flow and
metabolism in severely head-injured children. Part 2: Autoregulation.
Nagashima, T., Shirakuni, T. and Rapoport, S. I. (1990) A two dimensional
finite element analysis of vasogenic brain edema. Neurologia Medicochirurgica, 30, 19.
Nakagawa, Y., Tsuru, M. and Yada, K. (1974) Site and mechanism for
compression of the venous system during experimental intracranial
hypertension. Journal of Neurosurgery, 41, 427434.
Narayan, R. K., Greenberg, R. P., Miller, J. D. et al. (1981) Improved confidence
of outcome prediction in severe head injury: a comparative analysis of the
clinical examination, multimodality evoked potentials, CT Scanning and
intracranial pressure. Journal of Neurosurgery, 54, 751762.
Narayan, R. K., Kishore, P. R. S., Becker, D. P. et al. (1982) Intracranial pressure:
to monitor or not to monitor: a review of our experience with severe head
Nordstrom, C. H., Messeter, K, Sundbarg, G. et al. (1988) Cerebral blood flow,
vasoreactivity and oxygen consumption during barbiturate therapy in
metabolism in comatose patients with acute head injury. Relationship to
intracranial hypertension. Journal of Neurosurgery, 61, 241253.
OSullivan, M. G., Statham, P. F., Jones, P. A. et al. (1994) Role of intracranial
pressure monitoring in severely head-injured patients without signs of
intracranial hypertension on initial computerised tomography. Journal of
Park, C. K., Hong, Y. K., Cho, K. S. et al. (1993) Diffuse axonal injury: changes
of cerebral blood flow, intracranial pressure and evoked potentials, in
Pickard, J. D. and Czosnyka M. (1993) Management of raised intracranial
pressure. Journal of Neurology, Neurosurgery and Psychiatry, 56, 845858.
Piper, I. R., Miller, J. D., Dearden, N. M. et al. (1990a) Systems analysis of
cerebrovascular pressure transmission: an observational study in head
injured patients. Journal of Neurosurgery, 73, 871880.
Piper, I. R., Miller, J. D., Whittle, I. R. and Lawson, A. (1990b) Automated timeaveraged analysis of craniospinal compliance. Acta Neurochirurgica (Suppl.),
51, 387390.
Piper, I. R., Lawson, A., Dearden, N. M. and Miller, J. D. (1991) Computerised
data collection: a microcomputer data collection system in head injury
intensive care. British Journal of Intensive Care, 1, 7378.
Piper, I. R., Chan, K. H., Whittle, I. R. and Miller, J. D. (1993) An experimental
study of cerebrovascular resistance, pressure transmission and craniospinal
compliance. Neurosurgery, 32, 805816.
Pitts, L. H., Kaktis, J. V., Juster, R. and Heilbrun, D. (1980) ICP and outcome
in patients with severe head injury, in Intracranial Pressure IV, (eds K.
Shulman, A. Marmarou, J. D. Miller et al.), Springer-Verlag, Berlin, pp.
59.
Portnoy, H. D., Branch, C. and Chopp, M. (1985) The CSF pulse wave in
hydrocephalus. Childs Nervous System, 1, 248254.
Portnoy, H. D. and Chopp, M. (1981) Cerebrospinal fluid pulse wave form
analysis during hypercapnia and hypoxia. Neurosurgery, 9, 1427.
Portnoy, H. D., Chopp, M., Branch, C. and Shannon, M. B. (1982) Cerebrospinal fluid pulse waveform as an indicator of cerebral autoregulation.
Povlishock, J. T. and Kontos, H. A. (1985) Continuing axonal and vascular
change following experimental brain trauma. Central Nervous System
Trauma, 2, 285298.
Price D. J,. Czosnyka, M., Czosnyka, Z. and Price, J. R. (1994) Correlation of
continuous measures of autoregulation and compensatory reserve with
cerebral perfusion pressure and ICP, in Intracranial Pressure IX, (eds H.
Nagai, K. Kamiya and S. Ishii), Springer-Verlag, Berlin, pp. 6063.
Reilly, P. L., Graham, D. I., Adams, J. H. et al. (1975) Patients with head injury
who talk and die. Lancet, ii, 34650.
119
Richardson, A., Hide, T. A. H. and Eversden, I. D. (1970) Long-term

continuous intracranial pressure monitoring by means of a modified
subdural pressure transducer. Lancet, ii, 687689.
Robertson, C. S., Narayan, R. K., Contant, C. F. et al. (1989) Clinical experience
with a continuous monitor of intracranial compliance. Journal of Neurosurgery, 71, 673680.
Rose, J., Valtonen, S. and Jennett, B. (1977) Avoidable factors contributing to
death after head injury. British Medical Journal, ii, 615618.
Ryder, H. W., Espey, F. F., Kristoff, F. V. and Evans, P. P. (1951) Observations on
the interrelationships of intracranial pressure and cerebral blood flow.
Ryder, H. W., Espey, F. F., Kimbell, F. D. et al. (1953) The elasticity of the
craniospinal venous bed. Journal of Laboratory and Clinical Medicine, 42,
944.
Saul, T. G. and Ducker, T. B. (1982) Effect of intracranial pressure monitoring
and aggressive treatment on mortality in severe head injury. Journal of
Sharples, P. M., Storey, A., Aynsley-Green, A. et al. (1990) Avoidable factors
contributing to death of children with head injury. British Medical Journal,
300, 8791.
Shapiro, K. and Marmarou, A. (1982) Clinical applications of the pressure
volume index in treatment of pediatric head injuries. Journal of Neurosurgery, 56, 819825.
Shapiro, K., Fried, A., Takai, F. and Kohn, I. (1985) Effect of the skull and dura
on neural axis pressurevolume relationships and CSF hydrodynamics.
Smielewski, P., Czosnyka, M., Maksymowicz, W. et al. (1993) Identification of
the cerebrospinal compensatory mechanisms via computer-controlled
drainage of the cerebrospinal fluid, in Intracranial Pressure VIII, (eds C. J. J.
Smyth,C. E. and Henderson, W. R. (1938) Observations on the cerebrospinal
fluid pressure on simultaneous ventricular and lumbar punctures. Journal of
Neurology and Psychiatry, 1, 226237.
Sorek, S., Bear, J. and Karni, Z. (1988) A non-steady compartmental flow
model of the cerebrovascular system. Journal of Biomechanics, 21, 695704.
Stern, W. E. (1963) Intracranial fluid dynamics. The relationship of intracranial
pressure to the Monro-Kellie doctrine and the reliability of pressure
assessment. Journal of the Royal College of Surgeons of Edinburgh, 9, 1836.
Stuart, G., Merry, G., Smith, J. A. and Yelland J. D. M. (1983) Severe head
injury managed without intracranial pressure monitoring. Journal of
Sullivan, H. G., Miller, J. D., Becker, D. P. et al. (1977) The physiological basis
of intracranial pressure change with progressive epidural brain compression. Journal of Neurosurgery, 47, 532550.
Sullivan, H. G., Miller, J. D., Griffith, R. L. et al. (1979) Bolus vs steady-state
infusion for determination of CSF outflow resistance. Annals of Neurology, 5,
22838.
Takagi, H., Walstra, G., Marmarou, A. and Shulman, K. (1980) The effect of
blood pressure and PaCO2 upon bulk compliance (PVI), in Intracranial
Pressure IV, (eds K. Shulman, A. Marmarou, J. D. Miller et al.), SpringerVerlag, Berlin, pp. 163166.
Takemae, T., Kosugi, Y., Ikebe, J. et al. (1987) A simulation study of intracranial
pressure increment using an electrical circuit model of cerebral circulation.
IEEE Transactions in Biomedical Engineering, 34, 958962.
Takizawa, H., Gabra-Sanders, T. and Miller, J. D. (1985) Validity of
measurements of cerebrospinal fluid outflow resistance estimated by the
bolus injection method. Neurosurgery, 17, 6367.
Takizawa, H., Gabra-Sanders, T. and Miller, J. D. (1986a) Analysis of changes
in intracranial pressure and pressurevolume index at different locations in
the craniospinal axis during supratentorial epidural balloon inflation.
Takizawa, H., Gabra-Sanders, T. and Miller, J. D. (1986b) Variations in
pressurevolume index and CSF outflow resistance at different locations in
the feline craniospinal axis. Journal of Neurosurgery, 64, 298303.
Takizawa, H., Gabra-Sanders, T. and Miller, J. D. (1987) Changes in the
cerebrospinal fluid pulse wave spectrum associated with raised intracranial
pressure. Neurosurgery, 20, 355361.
Tanaka, K. and Nishimura, S. (1989) The importance of outflow resistance of
the shunt system for elimination of B-waves, in Intracranial Pressure VII,
Tans, J. T. and Poortvliet, D. C. (1983) Intracranial volumepressure
relationship in man. Part 2: Clinical significance of the pressurevolume
index. Journal of Neurosurgery, 59, 810816.
Tans, J. T. and Poortvliet, D. C. (1984) Comparison of ventricular steadystate infusion with bolus infusion and pressure recording for differentiating arrested and non-arrested hydrocephalus. Acta Neurochirurgica,
72, 1529.
Troupp, H. and McDowell, D. G. (1976) Summary of session on patient
management, in Intracranial Pressure III, (eds J. W. F. Beks, D. A. Bosch and
M. Brock), Springer-Verlag, Berlin, pp. 279280.
Ursino, M. (1988) A mathematical study of human intracranial hydrodynamics. Part 1. The cerebrospinal fluid system pulse pressure. Annals of
Biomedical Engineering, 16, 379401.
120
Uzzell, B. P., Obrist, W. D., Dolinskas, C. A. and Langfitt, T. W. (1986)

Relationship of acute CBF and ICP findings to neuropsychological outcome
in severe head injury. Journal of Neurosurgery, 65, 630635.
Van Harreveld, A. and Ochs, S. (1956) Cerebral impedance changes after
circulatory arrest. American Journal of Physiology, 187, 180192.
Vollmer, D. G., Torner, J. C., Jane, J. A. et al. (1991) Age and outcome following
traumatic coma: why do older patients fare worse? Journal of Neurosurgery,
75, S37S49.
Walsh, E. K. and Schettini, A. (1989) Brain tissue elasticity and CSF elastance,
in Intracranial Pressure VII, (eds J. T. Hoff and A. L. Betz), Springer-Verlag,
Berlin, pp. 271274.
Weed, L. H. (1929) Some limitations of the MonroKellie hypothesis. Archives
of Surgery, 18, 10491068.
Weed, L. H. and Flexner, L. B. (1932) Cerebrospinal elasticity in the cat and
macaque. American Journal of Physiology, 101, 668677.
Weed, L. H., Flexner, L. B. and Clark, J. H. (1932) The effect of dislocation of
cerebrospinal fluid upon its pressure. American Journal of Physiology, 100,

246261.
Weed, L. H. and McKibben, P. S. (1919) Pressure changes in cerebrospinal
fluid following intravenous injection of solutions of various concentrations.
American Journal of Physiology, 48, 512530.
Wong, F. C., Signorini, D., Piper, I. R. and Miller, J. D. Differences in treatment
thresholds of cerebral perfusion pressure in head injuries. Journal of
Wyler, G. R. and Kely, W. (1972) Use of antibiotics with external ventriculostomies. Journal of Neurosurgery, 37, 185187.
Yada, K., Nakagawa, Y. and Tsuru, M. (1973) Circulatory disturbance of the
venous system during experimental intracranial hypertension. Journal of
Zee, C. M. and Shapiro, K. (1989) The viscoelasticity of normal and
hydrocephalic brain tissue, in Intracranial Pressure VII, (eds J. T. Hoff and A.
L. Betz), Springer-Verlag, Berlin, pp. 263266.
INJURY AND CELL FUNCTION

Ross Bullock
7.1
Introduction
This chapter reviews the mechanisms by which

primary impact acceleration forces applied to the
neuraxis cause damage to the cranial contents and the
cellular basis by which secondary insults, particularly
ischemia, exacerbate primary damage to produce the
composite clinical picture usually seen.
There are now data available from in vitro and in
vivo models and from human studies that allow
conclusions to be drawn about the effects of trauma
upon neuronal and astrocytic function. Trauma has
effects on cell membranes, ion channels of axons and
neurons and astrocytes, and also on whole brain
systems affecting substrate delivery, blood flow, brain
metabolism and neurological function. Although
many aspects of the pathophysiology of neurotrauma
remain poorly understood, the enormous advances
made over recent years have provided promising new
avenues for therapy which will be clinically tested
over the next few years.
do not lose consciousness, because axonal injury is

absent and the reticular activating system and projection fibers are not disturbed by shear forces.
Thibault and Gennarelli, and coworkers, have used
the Penn 1 and Penn 2 primate impact acceleration
injury model to characterize the relationship between
the magnitude of acceleration/deceleration force, the
time duration over which it is applied, and the
consequences for the intracranial contents (Figure 7.1;
Gennarelli, Thibault and Adams, 1985). A brief, highintensity decelerational force, as seen when the head
strikes a solid floor at the end of a standing fall, will
tear parasagittal bridging veins, causing an acute
subdural hematoma. When the deceleration force is of
higher magnitude and longer in duration, as in vehicle
7.2 Biomechanical characteristics of

neurotrauma
The manner in which kinetic energy is applied to the
cranium during injury is extremely variable. At one
extreme, the helmeted head of a restrained aircraft
pilot may decelerate from a velocity of several
hundred miles per hour to zero over tenths of a second
during a crash. The cranium may never contact a solid
object, yet the brain is irreversibly damaged. Such
pure deceleration injury maximally damages axons
and most frequently occurs as a result of motor vehicle
accidents, where it is always compounded by an
additional impact component. At the opposite end of
the spectrum are those unusual injuries in which the
stationary head (e.g. of a machine operator) is slowly
crushed by slow-moving machinery. Such injuries
classically produce massive fractures, extra-axial
hematomas and contusions, yet these patients usually
Figure 7.1 Relationship between the magnitude of shear

force and duration of application, and clinical effects, after
severe brain injury. Note that above a certain magnitude of
acceleration, survival is no longer possible. Long durations
of acceleration/deceleration tend to produce diffuse axonal
injury, as seen in motor vehicle accidents, while briefduration shear force tends to produce subdural hematoma
and/or concussion.
122
Figure 7.2 Mathematical modeling of pressure waves in a simulated cranium with and without neck, following frontal
impact, 1.8 ms after impact. Note the lines of differential pressure radiating out from the inferior frontal region, which
corresponds closely to the distribution of contusion in severely injured humans. (Source: modified from Kuijpers, Claessens
and Sauren, 1995 with permission.)
accidents, diffuse axonal injury is caused. When both

the magnitude and the duration of the decelerational
force is less, transient unconsciousness, concussion,
is caused but few structural effects are seen when the
brain is examined either ultrastructurally or by light
microscopy (Chapters 2 and 3).
7.2.1
BRAIN MOVEMENT DURING IMPACT
Studies performed more than 20 years ago using

various biomechanical modeling techniques, demonstrated that the brain moves within the cranial cavity
in response to decelerational forces (Holbourne, 1943;
Gurdjian, Lissner and Hodgson, 1966). The gelatinlike brain reverberates and swirls within the cranial
cavity for many milliseconds after impact in response
to lines of force (Kuijpers, Claessens and Sauren, 1995;
Ommaya et al., 1966; Peerless and Rewcastle, 1967).
The brain is anchored within the cranial cavity only by
the parasagittal bridging veins, parasinusoidal granulations, cranial nerves and tentorium. Movement of
the lobes of the brain forward towards the anterior
cranial basal structures, particularly the sphenoidal
ridges, concentrates force at the bases of the frontal
lobes and the tips of the temporal lobes (Figure 7.2;
Gurdjian, Lissner and Hodgson, 1966). Surface contusions are thus very much more frequent at these sites
than elsewhere (Chapters 2 and 3). It is generally held
that shearing forces transmitted through the brain
stem and the reticular activating system are responsible for immediate loss of consciousness, although
there is evidence in the human brain that shearing
force also concentrates in the deep white matter
structures such as the corona radiata, explaining the
frequent finding of parasagittal gliding contusions
(Stritch, 1956; Chapter 3).
7.2.2
FOCAL INJURY
Focal brain injuries are usually associated with a

breach of the cranial coverings, such as a compound
depressed fracture. By definition, they produce focal
cortical contusions with or without associated intracerebral hematoma formation (Figure 7.3). Such injuries may occur in association with a more significant
diffuse injury of the type described above, e.g. when
the rapidly decelerating cranium strikes a sharp
pointed surface within a motor vehicle, but in most
instances they are the result of the stationary cranium
being struck by moving objects with relatively small
mass such as sticks, baseball bats or golf clubs.
Usually, impacts of this type do not cause prolonged
unconsciousness but they may cause permanent focal
Figure 7.3 Distribution of contusions in 72 fatally headinjured brains, irrespective of the site of cranial impact
(Source: modified from Gurdjian, Lissner and Hodgson,
1966, with permission.)
BIOMECHANICAL EFFECTS AND AGE
neurological deficit due to the immediate effects of the

penetrating/focal injury, or even death due to the
delayed consequences of cerebral contusion or intracranial hematoma (see below and Chapter 19).
7.2.3
PENETRATING INJURY
Penetrating injuries may be caused by missiles or

projectiles, which have low mass but strike the
cranium with high or very high velocity, or by stab
wounds, in which sharp objects moving at low
velocity are driven into the cranial cavity. Stab
wounds damage vascular structures, cranial nerves
and white matter fiber tracts. When the wounding
instrument remains imbedded within the skull and
brain the prognosis is remarkably good with appropriate management, but if the weapon is levered free
by the assailant, the resulting arc of blade movement
within the brain may be devastating (Du Trevou and
Van Dellen, 1992). Angiography is usually needed to
exclude vascular injury.
7.2.4
MISSILE INJURIES
Missile injuries and their effects on the brain are

highly variable and unpredictable. Cranial wounding
by high-velocity military firearms (> 2000 m/s) is
almost always fatal because the kinetic energy transmitted to the cranium is enormous. When a bullet
traverses the intracranial contents, its effects are
twofold: a pressure wave reverberates through the
brain, transmitting extremely high pressure as the
bullet enters followed by low pressure of equal
magnitude behind the missiles path. These pressure
waves occur in proportion to the velocity of the
projectile, and largely determine survival. The second consequence of missile injuries is direct vascular
and neural disruption which results from the tearing
effects of the bullet itself and in-driven skull fragments. These effects are similar to other forms of
contusion, although the magnitude of vascular damage and consequent hematoma formation may be
much greater. Patients in persistent deep coma,
(Glasgow Coma Scale 35) after bullet injury of high
or low velocity have a 9095% likelihood of a bad
outcome (death or severe disability) unless an associated extra-axial hematoma can be removed (Graham et al., 1990). Although cranial gunshot wounds
and their management are not a primary focus of
this book, most aspects of their pathophysiology and
management are common to blunt or closed cranial
injury. (For an excellent review, the reader is referred
to Chapters 60 and 61 in Narayan, Willberger and
Povlishock, 1996.)
The two factors common to all types of severe
cranial inquiry are:
123
the transmission of kinetic energy through the brain

which passes as pressure waves of varying magnitude and duration (Figure 7.2);
direct impact damage to small and large vascular
structures and neurons close to the cortical
surface.
From Figure 7.1 it may be seen that the magnitude of

shearing force needed to transiently impair neurological function is many times less than the magnitude
of shear force needed to disrupt axons. It has been
estimated that the magnitude of shear that is required
to damage the pial vasculature may be five times
greater than that needed to damage axons (Gennarelli,
Thibault and Adams, 1985; Hayashi et al., 1980).
Although this gradation of force suggests that vascular structures should be damaged less frequently
than axons and membranes, this may not be the case,
because focal forces, as seen in cerebral contusion are
often concentrated within the first few millimeters of
the cortex under the pia, especially at the frontal and
temporal tips, thus producing damage to the microvasculature and distribution arterioles in the majority
of patients who sustain moderate and severe brain
trauma (Figure 7.3).
7.3
Biomechanical effects and age
At both extremes of the age spectrum, the brain is

more vulnerable to vascular damage in response to
shearing forces. In the premature neonate, for example, relative absence of myelination and reduced
astrocyte maturity are probably responsible for the
high incidence of periventricular white matter hemorrhage resulting from the shearing forces sustained
during birth trauma.
In the elderly, brain atrophy may result in reduced
neuronal and astrocyte density with poorer support of
vascular structures, such that progressive pericontusional hemorrhage and edema are greatly facilitated.
The section below will review the effects of these
different patterns of biomechanical forces upon the
hierarchy of structures in the neuraxis.
7.3.1
CELL MEMBRANES AND ION CHANNELS
Neurons and their dendritic processes represent an

enormous membrane surface area. Although dendritic spines, synapses, gap junctions and myelinated
axons constitute specialized regions within the neuronal membrane, ion channels are by far the most
frequent structures embedded in neuronal membranes. To date over 15 varieties of neuronal ion
channels have been identified. The most common of
these, voltage gated channels, are closely linked
functionally with the sodium/potassium ATPase
124
pump (Sachs, 1991; Schwartz and Kandell, 1991;

Siegelbaum and Koester, 1991). Many types of ion
channel are linked to the specific agonist-gated
receptors, and others are linked within the cell with
second messenger systems, such as adenylate cyclase
and G proteins (for review see Siegelbaum and
Koester, 1991).
Although the surface area and complexity of
bimolecular leaflet cell membrane may be less for
astrocytes, there is now clear evidence that astrocytes
are excitable, possess ion channels, and may be
depolarized, though to a much lesser extent than
neurons (Bowman et al., 1992; Kimelberg and Norenberg, 1989; Orkand, Nicholls and Kuffler, 1966).
Astrocytic membranes also constitute an important
component of the bloodbrain barrier and there is
now extensive evidence to show that this barrier
function is disturbed transiently by mechanical
trauma (see below).
The effects of traumatic mechanical shearing upon
neuronal membranes is not well understood. Bimolecular leaflet membranes themselves appear to be
stretchy and relatively resistant to deformation and
movement, even at high shear rates, and indeed
some degree of motility and elasticity occurs normally with arterial pulse pressure waves and with
extremes of sagittal plane neck movement which
stretches the spinal cord and medulla several
centimeters.
Recent studies using patch clamp techniques and in
vitro tissue culture of neurons growing on deformable
surfaces have shown that ion channel function may be
radically altered by mechanical deformation at certain shear magnitudes and rates (Bowman et al., 1992;
Sachs, 1991; Tavalin, Ellis and Satin, 1995).
Specific classes of mechanotransducing ion channels have been identified using patch clamp techniques in both neurons and glia (Bowman et al., 1992;
Sachs, 1991). Some of these ion channels remain
leaky for several hours after mechanical deformation (Tavalin, Ellis and Satin, 1995). We speculate that
the majority of voltage-sensitive and agonist-gated
ion channels are also sensitive to transient mechanical deformation by shearing forces. Recent in vitro
experiments have shown that monolayer neuron and
astrocyte cultures growing on a flexible plastic membrane, which are rapidly deformed by a brief air jet
impulse, undergo rapid calcium entry and subsequent neuronal death, with efflux of lactate and
potassium into the culture medium (Tavalin, Ellis
and Satin, 1995).
Data from in vivo trauma models such as fluid
percussion injury and contusional impact models have
shown massive rapid transient efflux of potassium
into the extracellular fluid (ECF), associated with a fall
in sodium content in ECF (Bullock et al., 1995; Di et al.,
1996b; Katayama et al., 1990; Nilsson et al., 1993). These

changes may be explained by rapid alterations in the
pump leak relationship that exists between voltagedependent ion channels and the sodiumpotassiumdependent ATPase pump, and the opening of agonistgated channels.
We speculate that the structural complexity and
allosteric relationships that characterize agonistdependent ion channels and their receptors means that
they may be even more sensitive to mechanical
deformation with consequent increased leakiness.
Astrocytes are known to function as potassium
uptake buffers, having the capacity to rapidly take up
potassium from the extracellular space (Bullock et al.,
1991, 1994; Kimelberg and Nordenberg, 1989; Newman,
1986). However, this results in rapid astrocyte swelling, sometimes to enormous proportions. Such astrocyte swelling is the ultrastructural hallmark of both
acute cerebral ischemia (see below) and focal cerebral
contusion. It is almost always seen in animal trauma
models and in humans after trauma (Figure 7.4;
Bullock et al., 1991; Schroeder et al., 1994).
Figure 7.4 Electron micrograph made from human gray

matter taken from the periphery of a resected contusion in a
patient with increased intracranial pressure after severe
head injury (magnification 1500). Note the microvessels
(V) with narrowed lumen and an entrapped red cell. The
asterisks denote massively swollen astrocytic cytoplasm,
above due to swelling of the astrocyte end feet processes
abutting on a blood vessel and below surrounding the
astrocyte nucleus.
BIOMECHANICAL EFFECTS AND AGE
7.3.2 EFFECT OF SHEAR FORCES ON SYNAPSES AND

SYNAPTIC FUNCTION
Direct investigation of synaptic function is difficult

immediately after trauma. Recent microdialysis studies have investigated the time course of changes in
neurotransmitters within the extracellular space after
fluid percussion injury and brief transient surges in
release of excitatory amino acids and acetylcholine
have been demonstrated (Figure 7.5). Moreover, a
three- to fourfold surge in extracellular potassium was
also demonstrated, using the same techniques
(Katayama et al., 1990; Newman, 1986). About onethird of this potassium release could be blocked using
Tetrodo toxin, suggesting that two-thirds of the potassium release was occurring through agonist-operated
channels. Recent studies from our laboratory have
similarly shown that blockade of voltage-operated ion
channels prior to traumatic brain injury (TBI) failed to
ameliorate the negative neurological and behavioral
effects of the trauma, and produced only a modest
effect on K+ flux in the ECF, suggesting that agonistoperated ion channels are more important after TBI in
mediating ionic events (Di et al., 1996b).
There are now data available from microdialysis
studies in patients who have sustained severe head
injury and in patients with ischemic events superimposed upon their primary trauma, which show that
ECF excitatory amino acids (EAAs) rise to levels 5060
times higher than normal values when a secondary
ischemic event occurs superimposed on the trauma
Figure 7.5 The rapid and transient increase in potassium

release and extracellular glutamate, as detected by microdialysis after fluid percussion injury in the rat. (Source:
reproduced from Katayama et al., 1990, with permission.)
125
(Bullock et al., 1995; Zauner et al., 1996a). Excitatory

neurotransmitters released from damaged cells and
neurite processes may be responsible for these increases, and EAAs may also come from the intravascular
compartment. This conclusion is supported by the
finding that levels of structural amino acids in these
patients were also raised and appeared to fluctuate in
parallel with EAAs (Di et al., 1996a).
The behavioral changes that persist up to weeks or
months after TBI, even in animals without any
evidence of structural damage, have been taken as
evidence to support functional changes at the synaptic
level or in relation to second messenger systems (see
below). Recent neurochemical studies have shown
evidence of synaptic alterations and G-protein coupling relationships in the cell membrane that manifest
as prolonged increased in protein synthesis in
response to muscarinic cholinergic receptor activation
and certain categories of catecholaminergic receptors
(Delahunty et al., 1995; Miyazaki et al., 1992; Prasad et
al., 1994).
These changes may translate into effects on longterm potentiation in the hippocampus which have
been demonstrated in the absence of structural changes after trauma and may be an important mechanism
underlying the traumatic effects on learning and
memory.
7.3.3
EFFECT OF SHEAR INJURY UPON AXONS
About 50 years ago, neuropathological studies first

demonstrated accumulation of axoplasmic retraction
balls at sites of axonal discontinuity (Strich, 1956;
Peerless and Rewcastle, 1967).
These were chiefly found on large myelinated fibers
in patients who were unconscious from the time of
injury and subsequently died. These retraction balls
were found in high density in white matter tracts in
about 25% of severely head-injured patients (Adams,
Doyle and Ford, 1989).
Recent ultrastructural studies have shown, surprisingly, that the development of retraction balls is a
gradual process requiring more than 12 hours in lower
mammals such as rodents and around 24 hours in
humans to become fully developed (Povlishock, 1992).
This raises the intriguing possibility that diffuse
axonal injury may be amenable to therapeutic intervention. In accordance with these findings, Blumbergs
et al. have recently shown that a number of patients
who showed the features of diffuse axonal injury
(DAI) on examination of the brain at postmortem were
actually lucid and conscious during part of their
clinical course (Chapter 3). This suggests that neurons
and axons that subsequently show the changes of
diffuse axonal injury may function after impact prior
to degeneration, or that other, less affected, axon tracts
126
do not progress to DAI. Povlishock et al. have recently

shown that retraction balls and axonal degeneration
are particularly common at the points where axons
must deviate around small vessels within white
matter, suggesting that the axons may be fixed at these
points (Povlishock, 1992).
Carefully sequenced ultrastructural examinations
beginning a few minutes after diffuse cerebral impact
have now demonstrated a fascinating sequence of
events that incriminates primarily the microtubular
intra-axonal cytoplasmic system (Pettus et al., 1994).
Adjacent to nodes of Ranvier, the outer myelin layer
and axoplasmic membrane is first seen to balloon and
distend, compatible with the hypothesis that ion flux
through conformationally distorted ion channels may
be occurring. Subsequently, normal axonal transport
ceases because the intracytoplasmic microtubular system begins to collapse and concatenate upon itself like
a deck of cards collapsing (Pettus and Povlishock,
1995). The factors that cause this cytoskeletal collapse
are not fully understood. However, this process is
thought to result in axoplasmic stasis at the site of
these changes with subsequent accumulation of axoplasm proximally to form the reactive axonal swellings previously termed retraction balls (Grady et al.,
1993).
These changes have far-reaching consequences for
neuronal function. Interruption of the axon causes
proximal Wallerian degeneration of the affected neuron. Distally, the axon degenerates, fragments and
disappears, resulting in deafferentation of the affected
neuronal fields. The functional consequences of this
process may include seizures because of lack of
inhibitory effects, spasticity, intellectual decline and
unmodulated behavior patterns. When this is widespread, and Wallerian degeneration destroys many
neurons, the whole brain becomes atrophic, with
ventriculomegaly and, in the worst cases, a persistent
vegetative state (Adams et al., 1989; McLellan et al.,
1986).
7.3.4 ANIMAL MODELS AND THERAPEUTIC
IMPLICATIONS
Until recently, it has been difficult to produce diffuse

axonal injury in rodent models because of the biomechanical characteristics of the rodent cranium and
brain, and most animal studies have been performed
using cats and mini-pigs. The Marmarou weight-drop
model, in which a 500 g weight is dropped through
2 m on to the unrestrained helmeted cranium of an
anesthetized rodent has produced diffuse axonal
injury in relatively high density (Marmarou et al.,
1994). This model, therefore, offers the possibility of
testing therapeutic hypotheses relatively cheaply in
large numbers of animals.
7.3.5 EFFECTS OF SHEAR FORCE UPON

MICROVASCULATURE
The cerebral microvasculature is more resistant to

shear damage than axons. In the majority of significant head injuries, however, focal concentrations
of force develop at the tips of the frontal and temporal
poles that are sufficient to disrupt these pial vessels
causing a focal contusion. In other words, focal injury
is superimposed on diffuse injury (Kuijpers, Claessens
and Sauren, 1995).
Recent ultrastructural studies in both head-injured
humans and in appropriate animal models have
demonstrated major anatomical changes in the injured
microvasculature. These changes include the following (Bullock et al., 1994; Bullock et al., 1991):
swelling of perivascular astrocytic end feet such

that apparent narrowing and distortion of the
vascular lumen is frequently seen (Figure 7.4);
increased endothelial microvacuolation and micropseudopodial activity, suggesting increased transendothelial flux of intravascular components such
as water, ions or protein-rich fluid;
perivascular hemorrhage and transvascular diapedesis of red cells these hemorrhages may coalesce
to form a frank intracerebral hematoma or hemorrhagic contusion;
increased intravascular leukocyte adherence this
process has been observed in both trauma and
ischemia, and may be caused by cytokine activation
due to free radical release in response to shearing
injury in the endothelial cell walls.
Frank vascular disruption has been seen surprisingly

infrequently in human pericontusional biopsy material, suggesting that small vessels stretch and leak
much more frequently than they tear or burst. These
microvascular changes have profound functional consequences, chiefly reduction of local cerebral blood
flow (CBF) and development of vasogenic and cytotoxic edema, with increased intracranial pressure
(ICP) (Schroeder et al., 1994; see below).
7.4 Major vascular damage secondary to

shear injury
Acute subdural hematoma complicates about 20% of
severe head injuries and carries the worst outcome of
any of the subgroups of severe head injury patients.
This complication is almost always caused by rupture
of any of three types of surface vessel.
7.4.1
RUPTURE OF BRIDGING VEINS
Biomechanical studies with the Penn 1 and Penn 2

models have shown that reproducible acute subdural
hematomas can occur as a result of rupturing of the
MAJOR VASCULAR DAMAGE SECONDARY TO SHEAR INJURY
(a)
127
(c)
(b)
(d)
Figure 7.6 Excitatory amino acids in two patients with severe head injury. (a), (b) Patient with no focal lesion, and
uncomplicated recovery. Note that dialysate glutamate and aspartate are close to the normal level in extracellular fluids
(2 mol). (c), (d) Patient with large acute subdural hematoma, brain shift and hemispheric swelling (the patient died). Note
the massive increase in excitatory amino acids. Cerebral blood flow in the hemisphere beneath the subdura was 16 ml/
100 g/min.
parasagittal bridging veins when the cranium is

rapidly decelerated with a relatively low magnitude of
shear force applied (Gennarelli, Thibault and Adams,
1985). Such circumstances are clinically produced by a
fall from standing height on to a solid surface such as
a floor. Boxing injuries, particularly rotational accelerations of the head, also classically produce these
injuries. Avulsion of parasagittal and Sylvian bridging
veins is usually accompanied by a degree of diffuse
axonal injury and polar contusion (Adams, Doyle and
Ford, 1989; Chapter 3). What remains unclear, however, is the mechanism by which low-pressure venous
bleeding can accumulate to form a hematoma of
sufficient size to compress the brain. We speculate that
episodes of coughing, straining, or vomiting may

pump out sufficient blood to progressively tamponade the brain as clotting occurs. However, such
subdurals also occur in patients unconscious from the
time of injury, who have none of these episodes. That
the blood vessels are attached to the sagittal sinus,
which is held open rather than compressed by the
hematomas, may be important.
7.4.2
SUBDURAL HEMATOMAS OF ARTERIAL ORIGIN
When polar contusions are extensive, they may burst

through the pia to accumulate in the subdural space.
This results in the classical burst lobe injury charac-
128
terized by subdural hematoma, polar contusion, intracerebral hematoma and hemispheric swelling.
7.4.3 COALESCENCE AND RUPTURE OF
PARENCHYMAL SMALL VESSELS BLEEDING FROM A
CONTUSION
This frequently occurs when a coagulation defect

develops as a result of consumption of clotting factors
or anticoagulant therapy (Bullock et al., 1990a). Such
subdurals are usually associated with sizable intraparenchymal clots.
7.5
Metabolic consequences of TBI
Because the brain is dependent upon aerobic metabolism for substrate delivery (oxygen and glucose), and
because of the frequent impairment of oxygenation
and perfusion that occurs following severe head
injury, metabolic derangement is an extremely frequent and important consequence of TBI. Metabolic
changes may be global, involving the whole brain, or
focal, developing in the region of intracerebral and
subdural hematomas and contusions. Evidence demonstrating metabolic derangement after TBI has come
from several sources, including:

animal studies using the 2-deoxyglucose technique

to measure glucose metabolism in global models
such as fluid percussion injury and focal models
such as contusion and subdural hematoma;
positron emission tomography studies using fluorodeoxyglucose in humans;
measurement of jugular/arterial differences of oxygen and lactate to yield global measures of oxygen
consumption and lactate production in humans
(AVDO2 and CMRO2);
measurements of whole brain lactate, ATP and
other metabolites using MR spectroscopy in animal
models;
measurement of extracellular fluid lactate, glucose
and oxygen content by microdialysis and oxygen
electrodes in humans.
Data from these studies allow the following synthesis

to be constructed regarding the metabolic consequences of TBI.
Immediately following impact, the shearing forces
applied to neuronal tissues result in massive ion
fluxes across neuronal membranes, widespread loss of
resting membrane potential and release of neurotransmitters into the extracellular space. Within
minutes of these events, the brain attempts to restore
ionic homeostasis by reuptake of neurotransmitters
and ion pumping. These processes are intensely
energy-dependent and result in an abrupt increase in
glucose utilization. Studies with the fluid percussion
model in rats have shown that this increase in glucose

metabolism, to facilitate ATP generation, is brief and
maximally localized to those parts of the brain that
are maximally deformed by the shearing forces
(Kawamata et al., 1995). Evidence suggests that ionic
pumping, and glutamate surges in astrocytes both
preferentially activate anaerobic glycolysis, thus producing lactate, especially in astrocytes (Pellerin and
Magistretti, 1994). This depletes ECF glucose. When
focal lesions such as subdural hematoma, focal
infarction or cerebral contusion are present, then
glucose use increases for a longer period in the
penumbral border zone around the densely ischemic
core of these lesions (Kuroda and Bullock, 1992;
Sutton et al., 1994). This increase may persist for two
to four hours in the rat, and 57 days, in humans
(Kuroda et al., 1992; Bergsneider et al., 1996). In both
animal models and humans, glucose use is depressed
when measured days after the injury, and remains so
for weeks after impact, which is consistent with the
reduced metabolic needs of the comatose brain (Yoshino et al., 1991).
In humans, recent PET studies have shown that
these increases in glucose are maximal in the penumbral zone of contusions and in the hemisphere
underlying hematomas when this brain is viable
(Bergsneider et al., 1996; Figure 7.7). Such tissues are
often adjacent to low-density cytotoxic edema areas
on CT scan. Studies performed with PET at later time
points have shown uniformly decreased metabolism,
both for glucose and oxygen in humans, for 14 weeks
after impact.
7.5.1
LACTATE
Numerous studies in animals and humans have shown

that TBI is characterized by increased brain lactate
production, which normalizes gradually after the first
few days in those humans who survive, but remains at
levels five to ten times normal in those who die
(Crockard and Taylor, 1972; Inao et al., 1988). Measurements have been made in extracellular fluid, CSF and
whole brain using different techniques but the results
are consistent (Kawamata et al., 1995; Figure 7.8).
Recent microdialysis studies have demonstrated that
extracellular fluid glucose declines to extremely low
levels sometimes zero when lactate is increasing
(Myseros et al., 1995; Zauner et al., 1996b; Figure 7.8).
Marmarou et al. originally postulated that these data
could be explained best by a shift to anaerobic
metabolism within cerebral tissue such that the glycolytic pathway would consume glucose and generate
lactate while the Krebs cycle and mitochondrial
NADPH phosphorylation had ceased to occur (Inao et
al., 1988), otherwise lactate would be consumed. Such
situations generally occur in circumstances of restricted
METABOLIC CONSEQUENCES OF TBI
(a)
(b)
(c)
(d)
129
Figure 7.7 Increased focal glucose use in humans and two animal models after neurotrauma. (a) CT scan and
(b) fluorodeoxyglucose-11 PET study in a patient with frontal contusions 4 days after injury. Note the dark areas of increased
glucose use adjacent to the contusions, and markedly reduced glucose use in the contused cortex (light areas on PET study)
(c) Increased glucose use and decreased regional cerebral blood flow in a rat model of focal cerebral contusion immediately
after trauma. Glucose use was around 150 mol/g/min and CBF was as low as 5 ml/100 g/min in hippocampus and
contusion periphery (Sutton et al., 1994). (d) Decreased glucose use and slightly decreased cerebral blood flow 24 hours after
injury in the same rat contusion model as shown in (c). (Source: Reproduced by courtesy of Dr David Hovda, UCLA Brain
Injury Research Center.)
oxygen delivery to tissues. Although the oxygen

tension at which aerobic Krebs cycle metabolism ceases
to occur in human cerebral tissue is not well known, it is
probable that this develops only when brain oxygen
tension levels fall below about 20 mmHg (Zauner et al.,
1996b). Marmarou and others have shown in animal
experiments that anaerobic cerebral metabolism with
generation of lactate appears to occur even in the

absence of blood flow limitation, suggesting that
trauma incapacitates mitochondrial phosphorylation,
thus causing a shift towards anaerobic metabolism
(Inao et al., 1988). Recently, however, an interesting link
between increased ECF glutamate and increased
anaerobic glycolysis has been shown in astrocytes, such
130
that astrocytes consume glucose and produce large

amounts of lactate, which may be used as a preferential
energy substrate by neurons. Pellerin and Magistretti
hypothesize that this allows astrocytes to preferentially
pump ions and maintain ion homeostasis when
glutamate is high, as in neuroexcitation (Pellerin and

Magistretti, 1994). This would explain many of the
above findings in neurotrauma.
Recently we have studied severely head-injured
patients with an oxygen-sensitive microelectrode
(a)
(b)
Figure 7.8 Multiparameter monitoring in two patients with severe head injury. (a) Patient died 38 hours after establishment
of monitoring. Note that blood flow in the region of the multiparameter monitor was 15 ml/100 g/min 5 hours after injury.
Note that dialysate lactate increases to levels three times normal. Brain oxygen concomitantly decreases from the threshold
level of around 20 mmHg, and brain glucose steadily declines (normal dialysate glucose 800 mmol/l). (b) Patient with massive
brain injury progresses rapidly from uncontrollable ICP to brain death. Note that blood flow measured at time zero is 9.3 ml/
100 g/min and brain oxygen declines rapidly to anaerobic levels, indicating cessation of blood flow. Cerebral lactate climbs to
massively increased levels (5000 mmol/l). Brain glucose is close to zero throughout the monitoring period.
METABOLIC CONSEQUENCES OF TBI
implanted in the brain, capable of measuring brain

oxygenation, CO2 generation, pH and temperature in
parallel with extracellular fluid lactate and glucose
levels measured by microdialysis (Zauner et al.,
1996b). These studies have confirmed the finding that
lactate generation is increased in about 65% of
measurements, even in the presence of adequate
cerebral blood flow and brain oxygen levels, thus
confirming the hypothesis that lactate may be a
normal metabolite during neuroexcitation (Doppenberg et al., 1996; Figure 7.9).
It remains to be determined whether high levels of
ECF lactate are intrinsically harmful to the injured
brain. However marked cerebral acidosis may exacerbate calcium-mediated damage to intracellular
enzyme systems and may also interfere with ion
channel function (King, McLaurin and Knowles, 1994;
Siesjo,
1992(a)). High tissue lactate levels could foster
a decline in brain pH, as has been shown in numerous
post-traumatic studies, both in animal models and in
our recent human studies.
7.5.2 FLOWMETABOLISM UNCOUPLING AFTER
TRAUMA
Several studies have shown that cerebral blood flow

may be markedly reduced within the first few hours
after severe brain injury in both humans and animal
models (Bouma, Muizelaar and Stringer, 1992; Obrist
et al., 1984). In zones of focal cerebral contusion and
beneath intracranial hematomas, flow may fall to
131
levels close to the thresholds for ischemic brain

damage (Kuroda and Bullock, 1992; Schroeder et al.,
1994; Jones et al., 1981). When there is a concomitant
increase in glucose metabolism, then cerebral tissue is
placed at an increased risk of damage to intracellular
structures dependent upon continuous oxygen delivery, such as mitrochondrial and various enzyme
systems. These include, in particular, the enzyme
systems that break down free radicals, thus leading to
delayed damage in the hours that follow, especially
during the reperfusion phase (Kontos, 1985; Siesjo,
1992a, b, Schroeder et al., 1994b).

Therefore, we hypothesize that the most severely
damaged tissue, that which sustains the greatest
magnitude of shearing injury, will be unable to restore
ionic homeostasis in spite of maximally increasing
glycolytic activity. If tissue blood flow is reduced
during this time of maximal metabolic need, tissue
glucose and oxygen levels will fall to subthreshold
levels. Tissue swelling will be exacerbated and ischemic necrosis will occur. The vulnerability of brain
regions to ischemia varies; hence the process is not
uniform. CBF may be further reduced at the tissue
level by such processes as astrocytic swelling (Figure
7.10), and generally by low blood pressure, high ICP
(itself generated by cytotoxic swelling) or intracranial
hematomas causing distortion. Probably the effects of
all these insults may be cumulative and occur to a
varying extent in the majority of patients with severe
head injury.
Figure 7.9 Schematic diagram (modified from Pellerin and Magistretti, 1994) to show proposed mechanisms whereby
increased ECF glutamate generated during neuroexcitation (e.g. after trauma) causes an increase in glucose uptake by
astrocytes, and subsequent generation of lactate, which is then used as an energy substrate by neurons, e.g. to maintain ionic
homeostasis. During this process, astrocytes may swell in response to ion shifts.
132
Figure 7.10 Hypothetical scheme to depict post-traumatic and postischemic events, with opening of ion channels and
uptake of potassium by astrocytes, which jeopardizes the microcirculation.
INTRACELLULAR MECHANISMS
(a)
Implications for therapy
If this hypothesis is true, it would explain the relative

success of therapies such as metabolic suppression
using barbiturates or hypothermia, or raising cerebral
perfusion and CBF by the use of pressors. Diuretics
and rheological agents, such as mannitol, may help to
improve tissue perfusion during these crucial early
periods (Muizelaar et al., 1983). Ion channel blockade
using agents directed at both voltage-dependent and
agonist-operated channels may be important avenues

for future therapy, as may be the use of hemoglobin
substitutes, which augment the oxygen carrying
capacity of the microcirculation to damaged tissue (Di
et al., 1996a; Figure 7.8).
7.5.3 BIOCHEMICAL CONSEQUENCES OF IMPACT
INJURY
In addition to its effect on ion channels and membranes, which are described above, traumatic shearing
forces also have profound effects upon intracellular
biochemical events and also membrane and cytoplasmic second messenger systems.
(a)
(a)
(b)
(c)
Figure 7.11 (a) Diagrammatic representation of an agonistoperated ion channel, such as the three channels for
glutamate within the cell membrane: NMDA, kianate and
AMPA receptors. Note that different agonists and coagonists
such as glutamate and glycine may activate the ion channel.
(b) A second-messenger-linked receptor system, such as the
metabotropic glutamate receptor, which induces more complex changes within the cell membrane. Second-messengerlinked systems are vulnerable to upregulation or deactivation
after trauma and may underlie the mechanistic basis for the
changes in memory and higher function after neurotrauma.
(c) Molecular schematic representation of the G-proteincoupled human 2-adrenergic receptor. Note the complexity
of the transmembrane protein domains and their potential
vulnerability to shear forces after trauma. (Source: reproduced from Schwartz and Kandell, 1991, with permission.)
133
Second messenger systems and neurotrauma
Second messengers are large molecules, situated

usually within the neuronal membrane or adjacent to
its inner surface within the cytoplasm, which have the
capability of modulating or amplifying external signals brought to the cell via neurotransmitters and
mediators, such as glutamate (metabotropic receptor),
adenosine, steroids and acetylcholine (Figure 7.11). A
number of recent studies have shown that second
messenger systems, probably because of their large
molecular size and the complexity of their stearic
interactions, are vulnerable to the shear forces of
neurotrauma. In some circumstances, second messenger systems may be amplified (up to 200-fold or
more) while other types of second messengers are
downregulated or deactivated by neurotrauma (Delahunty et al., 1995; Kuroda, Dewar and Bullock, 1993). It
is thought that such systems may play an important
role in complex neurological processes such as encoding of memory and so these change in second
messenger systems could constitute a mechanism for
the behavioral and memory changes that are seen in
both animals and humans after neurotrauma (Bortolotto et al., 1994). In the fluid percussion model, no
anatomical changes are seen in the presence of these
long-lasting neurobehavioral deficits (Prasad et al.,
1994; Miyazaki et al., 1992).
7.6
Intracellular mechanisms
7.6.1
THE ROLE OF CALCIUM
Several studies have now demonstrated rapid and

massive intracellular increase in free calcium ions
within minutes after trauma (Fineman et al., 1993;
Nilsson et al., 1993; Kawamata et al., 1995). This is
similar to the situation following severe acute ischemia, in which calcium increases about tenfold within
seconds of the onset of severe ischemia (Siesjo,
1992a).
134
In both these clinical situations, calcium is thought to

enter cells through many different channels which
may be opened by several mechanisms, including:
voltage-dependent channel opening, induced by

mechanical deformation of membrane and ion
channels, as described;
agonist-dependent channel opening, mediated by
neurotransmitter substances released in excess into
extracellular fluid (see below).
Intracellular vesicular calcium may be released into

the cytosol by exocytosis induced by shear trauma or
cytoskeletal disruption.
The specific T-, L- and N-type calcium channels,
which are voltage-dependent, may also be activated
by trauma, ischemia and other intracellular events.
These channels are of interest because calcium antagonists e.g. nimodipine, which blocks L channels, and
the omega conotoxins, which block N channels are
undergoing clinical evaluation (Chapter 18). Calcium
may be particularly important in mediating the
delayed ischemic neuronal death that may be superimposed upon the effects of primary trauma in the
majority of severely head-injured patients (see below;
Figure 7.10).
Recently, the dihydropyridine calcium antagonist
nimodipine has shown a significant beneficial effect in
two clinical trials specifically focused upon headinjured patients with subarachnoid hemorrhage
(Harders et al., 1996; European Study Group on
Nimodipine in Severe Head Injury, 1994). This effect is
probably mediated via these neuroprotective effects
rather than by restoration of blood flow alone.
7.6.2
THE ROLE OF FREE RADICAL GENERATION
Free radicals are highly reactive ionic molecules

bearing an unpaired electron in their outer electron
shell, which confers extremely high chemical reactivity. These compounds are the normal byproduct of
oxidative metabolism within mitochondria, and they
fulfill important physiological roles within various
tissues, such as polymorphonuclear-leukocyte-mediated destruction of bacteria (Kontos, 1985; Siesjo
1992b). Because these highly charged molecular species may react with various structures within the cell
membranes, proteins and the genome, potent enzyme
systems exist within all cells to break them down.
Intracellular calcium inactivates some of these mechanisms, such as peroxidase and xanthine oxidase
(Kontos, 1985). It is speculated that free radicals are
generated especially in circumstances of postischemic
reperfusion when the tissue is replete with oxygen
(from which most free radicals are derived) and after
ischemic events when the protective enzymes may be
inactivated. Free radical generation is also favored by

the presence of free ferrous iron, which acts as a
catalyst to the HaberWeiss reaction depicted below
(Siesjo,
1992b):
Generation of hydroxyl radicals (OH):
O2 + Fe3+ O2 + Fe2+
H2O2 + Fe2+ OH + OH + Fe3+
O2 + H2O2 O2 + OH + OH
Free iron may be found in traumatized brain within a
few hours of injury as repair mechanisms cleave iron
from hemoglobin in breaking down red cells, which is
an extremely frequent sequel of severe brain injury. To
date, however, there is no technique for demonstrating
the existence of free radical species in the human brain
after trauma, although studies using a number of
indirect techniques have shown an increase in free
radical activity following trauma and ischemia in
various animal models (Siesjo,
1992a, b; Hall and
Braughler, 1989). There is evidence to suggest that
free-radical-mediated damage may be particularly
marked within the endothelium of cerebral capillaries
after trauma, and it is likely that the large neuroprotective molecules directed against the free radical
mechanism (Tirilazad (amino steroid) and polyethylene-glycol-conjugated superoxide dismutase (PEGSOD)) do not penetrate extensively into brain tissue.
This may in part explain their failure to benefit
patients in clinical trials (Chapter 18).
7.6.3
THE ROLE OF HYDROGEN IONS
Although hydrogen ions in the extracellular space are

powerful cerebral vasodilators, high concentrations of
hydrogen ions within cells appear to be harmful
because they alter the function of intracellular
enzymes (Siesjo,
1992a). An interesting feedback system exists whereby low pH causes conformational
changes in the NMDA ion channel, thus preventing
further ingress of sodium and calcium and egress of
potassium during cellular acidosis.
7.6.4
POLYAMINES
The polyamines spermine, spermidine, putrescine and

cadaverine are metabolic breakdown products of
purine catabolism within the cell. They accumulate in
toxic concentrations within cells following both trauma
and ischemia because the breakdown enzymes for their
destruction are inactivated, probably by calciummediated mechanisms within the cell. The presence of
high concentrations of these substances within cells is
cytotoxic both in tissue culture and in vivo.
7.6.5 EFFECT OF NEUROTRAUMA ON PROTEIN

SYNTHESIS AND GENE SIGNALING
Recently, molecular techniques such as northern and

Western blotting, in situ hybridization and immunohistochemistry have allowed the effects of neurotrauma upon protein synthesis and gene expression to
be explored. Understanding for these processes is,
however, far from complete. In general, early after
impact, protein synthesis is suppressed for a few
hours but at about 56 hours gene expression begins to
occur and in those neurons and astrocytes that are
destined to survive impact, and which are capable of
manifesting a regenerative response, a phase of
intense protein synthesis occurs that lasts for many
days. There is now intense interest focused upon the
idea that identification of potentially harmful intrinsic
mediators that are produced by neurons or astrocytes
early after injury could allow blocking drugs to be
devised, which could then be potentially therapeutic.
Among the earliest substances produced in large
amounts during this early post-traumatic period are
the vasodilator neurotransmitter nitric oxide (NO),
which may be directly toxic to neurons in excessive
amounts (Dawson, 1994). Nitrate oxide synthase, the
enzyme responsible for its production, is induced
shortly after injury. Blockade of the enzyme has now
been shown to be neuroprotective after brain injury
(Mesenge et al., 1996).
7.6.6
GENE REGULATION IN THE NORMAL NEURON
Because neurons do not reproduce, the vast majority

of their genetic material remains repressed by operon
genes and regulator proteins. Only those genes concerned with structural protein synthesis function at a
low level, and those concerned with neurotransmitter
synthesis and maintenance of energy metabolism
continue to be highly active during normal neuronal
life. There is evidence that, during embryonal development, enormous numbers of redundant neurons
develop in the central nervous system (CNS), and only
those that make active functioning synaptic connections with their target organs persist into the adult
form of the CNS. The process responsible for this dieback phenomenon among redundant neurons has
been shown to be apoptosis or programmed cell
death which is under genetic control and now
appears demonstrable using molecular biological
techniques (see below).
135
HSP70 begins to be synthesized and its gene becomes

activated about 6 hours after sublethal, stressful
events such as transient cerebral ischemia and neurotrauma (Massa, Swanson and Sharp, 1996). Its synthesis maximizes at about 4872 hours in rodents and
then declines. Other genes with the same temporal
and mechanistic characteristics have also been identified these include c-fos, c-jun and the heme oxygenase gene, the last of which has particular importance
after
subarachnoid
and
intracerebral
hemorrhage because heme oxygenase is the enzyme
responsible for breaking down hemoglobin into free
iron and biliverdin, and iron may catalyze free radical
production (Rehcrona, Hauge and Siesjo,
1989; Siesjo,
1992a). The third of the immediate early (IE) triad is

the glucose transporter gene, which is activated
massively in neurons that usually do not go on to die
after a stressful event. It is uncertain whether these
neurons will continue to function normally during the
time these IE genes are expressed. Thus the concept
emerges that these genes are a marker of populations
of neurons that may be vulnerable to subsequent
second insults that may occur during the period when
these IE genes are regulated. Alternatively, they may
be a manifestation of protective mechanisms whereby
the neurons are protecting themselves against a
transient, but potentially lethal, stress. Some studies
have suggested that IE gene expression may be
protective and, therefore, if these genes could be
expressed by a gene manipulation therapeutic technique, this may be neuroprotective (Figure 7.12). There
is speculation that calcium entry may be the trigger
for this IE gene expression and, thus, that it may be
ion-channel-dependent (Marsa et al., 1996).
7.6.8
APOPTOSIS AND NEUROTRAUMA
In non-neural tissues and in neoplasia, death of

redundant cells without any inflammatory response,
has long been recognized by pathologists, and its
histological features are well known. Recently, new in
situ and molecular techniques (e.g. TUNEL staining,
and blotting for the c-myc gene) have suggested that
this form of delayed, genetically programmed cell
death also occurs in neurons after brain ischemia and
trauma (Islam et al., 1995; Rink et al., 1994). Its
importance, its potential for therapy and its relationship to delayed neuronal death (see below) remain to
be elucidated.
7.6.7 IMMEDIATE EARLY GENE EXPRESSION

STRESS GENES AFTER TRAUMA
7.6.9 SECONDARY ISCHEMIC NEUROLOGICAL

DAMAGE
Recently, the heat shock protein HSP70 has been

identified, and its gene has become easily identifiable
by both Western blotting and in situ hybridization.
The incidence of ischemic brain damage seen at

postmortem in severely head-injured patients who die
is extremely high, with estimates ranging between
136
(a)
amenable to any form of postevent treatment. Indeed,

to date, no laboratory study has yet shown that the
development of retraction balls can be delayed or
retarded by therapy. Those processes that are set in
train at the time of the primary impact but then may
be magnified in subsequent days to exacerbate brain
damage are therefore of great interest. The observation
that about 70% of severely head-injured patients
manifest high intracranial pressure during their clinical course in the intensive care unit accords with this
concept (Becker et al., 1977).
7.6.10 THE GENESIS OF ISCHEMIC BRAIN DAMAGE
AFTER SEVERE HUMAN HEAD INJURY
(b)
Figure 7.12 (a, b) The time scale of early blood flow and
gene expression changes after neurotrauma.
60% and 90% of patients (Graham, 1985; Chapter 3).

During life, most of these patients do not manifest the
long periods of low cerebral perfusion pressure (e.g.
< 30 mmHg for 30 min or more) that are known to be
necessary for the generation of ischemic damage.
Likewise, in animal models of impact-type head
injuries such as fluid percussion, weight drop and
contusional impact, widespread ischemic damage is
not seen other than around the periphery of focal
contusions. Thus, there is a fundamental paradox and
the high incidence of ischemic brain damage is not
easily explained. This concept of delayed secondary
neurological damage after head injury is also supported by the lucid interval statistics. Between 30%
and 40% of severely head-injured patients who die
will, at some time, have demonstrated a period of
lucidity sufficient to obey commands or speak
(Adams, Doyle and Ford, 1989; Graham, 1995). This
implies that primary impact events were not sufficiently severe to damage the brain beyond the
capacity for function, therefore emphasizing the
importance of secondary damage (Reilly et al., 1975;
Chapter 4). The genesis of these types of secondary
brain damage after severe head injury has become the
focus of major research endeavors in those academic
centers that care for head-injury patients. This is
because primary impact damage of the diffuse axonal
injury type has hitherto been considered not to be
Until recently, numerous studies using various cerebral blood flow measurement techniques had failed to
demonstrate levels of blood flow sufficiently low to
cause ischemic neuronal damage. However, tomographic regional blood flow measurements early after
severe injury have now clearly demonstrated flow
levels < 18 ml/100 g/min sufficient to generate neuronal ischemic necrosis in about 34% of severely
injured patients (Bouma, Muizelaar and Stringer, 1992;
Schroeder et al., 1994; Schroeder, Muizelaar and Kuta,
1994). These were predominantly patients with fixed
dilated pupils, acute subdural hematoma or early
acute brain swelling (Chapter 5). Other studies using
these same techniques have revealed profound
regional flow reductions around intraparenchymal
lesions such as contusions and intracerebral hematomas where blood flow is about 18 ml/100 g/min
(Schroeder et al., 1994; Schroeder, Muizelaar and Kuta,
1994; Zauner et al., 1996a). This is in accordance with
the uniform neuropathological observation in humans
that pyknotic neuronal degeneration and astrocytic
swelling is seen in the tissues surrounding focal
contusions (Adams, Doyle and Ford, 1989; Figure 7.4;
Chapter 5).
7.6.11 MECHANISM BY WHICH REDUCED CEREBRAL
BLOOD FLOW CAUSES TISSUE DAMAGE
The landmark studies by Symons group and Jones et al.

have demonstrated a time-dependent hierarchy of
neuronal events in response to progressive reduction of
CBF (Astrup, Siesjo and Symon, 1981; Branston et al.,
1974; Jones et al., 1981; Figure 7.13). In the healthy,
normally autoregulating brain, cortical flow reduction
down to levels around 20 ml/100 g/min may be
tolerated without functional consequences, although
the EEG may begin to slow and the subject may
develop anxiety and drowsiness. Abruptly, at around
20 ml/100 g/min, consciousness is lost and the brain
loses the capacity to make neurotransmitter substances
so that coma ensues (Branston et al., 1974; Siesjo,

Figure 7.13 The relationship between cerebral blood flow

and tissue infarction, modified from Jones et al. 1981 data
obtained from studies in awake primates. Note that after 30
minutes of severe ischemia, reversible neuronal changes
have occurred but recovery is possible. However, when the
ischemic period is prolonged to 23 hours, irreversible
infarction takes place. In head-injured humans there is
probably a left shift of this curve.
1992a,b). When flow falls below 18 ml/100 g/min,

ionic homeostasis becomes jeopardized because the
energy-dependent Na/K-ATPase pump system, which
maintains ionic gradient across the cell wall, cannot
function. At this level, neurons move to anaerobic
metabolism and lactate begins to be generated in large
amounts. When flow falls further to levels around
10 ml/100 g/min, membrane integrity is lost, massive
calcium influx begins and the biochemical cascade of
neuronal destruction becomes irreversible (see below).
The ultrastructural hallmarks of this process are
mitochondrial swelling and perineuronal astrocytic
process vacuolation, followed by swelling of the Golgi
apparatus and intracellular cytoplasmic vesicles. Eventually, nuclear definition is lost (karyorrhexis; Graham,
1985; for review see Siesjo,
1992a, b). Many of these
postischemic events are synergistic with the loss of
ionic homeostasis seen after trauma.
7.6.12
LOSS
137
hippocampal neurons of the molecular layer, CA1

and CA3 sectors;
cerebellar granular cells;
cortical neurons, particularly the larger cells, in
areas such as the cuneate visual cortex.
This may occur when flow is globally reduced

throughout the whole brain for various periods of
time. Within the context of head injury, this type of
neuronal loss is especially important in patients with
raised intracranial pressure, where cerebral perfusion
pressure may be marginal (around 3040mmHg) for
many hours, or even days. In such patients, recent
studies have demonstrated an extremely high frequency of ischemic neuronal loss especially in the
hippocampus (Graham, 1985; Chapter 3). Such bilateral hippocampal loss and cerebellar damage may
explain the high frequency of memory disorders and
coordination difficulty seen in the majority of severely
head-injured survivors. This concept also accords with
the almost universal finding of marked cerebral
atrophy in those patients who survive severe head
injuries.
7.6.13
DELAYED NEURONAL DEATH
First described by Kirino et al. about 15 years ago,

delayed death of large neurons, especially in hippocampal CA1 and CA3 sectors, 510 days after an
ischemic insult, has been clearly shown in rodent
models, and probably in humans, after cardiac arrest
episodes (Kirino, 1982; Kirino, Yamure and Sano,
1984). When histology is studied 2448 hours after
ischemia, no abnormality is seen in these rodent global
ischemia models, and gross behavior is also normal
initially. This suggests a sublethal insult that later
amplifies to cause cell death. High levels of glutamate
have been implicated in these phenomena, and glutamate antagonists prevent the delayed cell death.
Apoptosis may be an alternate explanation for this
phenomenon.
INFARCTION VERSUS SELECTIVE NEURONAL
When flow is profoundly reduced (i.e. less than

510 ml/100 g/min) within the distribution of one
cerebral end artery for more than 6090 minutes,
infarction ensues. That is, there is immediate necrosis
of all cell types within a zone of the brain (Symon,
Pasztor and Branston, 1974). However, when the flow
reduction is less marked (e.g. to levels around
1518 ml/100 g/min) and when this persists for periods of more than about 30 minutes, then selective
neuronal loss may occur (DeGirolami, Crowell and
Marcoux, 1984; Pulsinelli, Brierley and Plum, 1982).
Neuronal types that are the most vulnerable within
the mammalian brain are:
7.6.14 DOUBLE INSULT MODELS AND SYNERGISTIC

DAMAGE MECHANISMS AFTER NEUROTRAUMA
Jenkins et al. have shown that a mild global ischemic

insult insufficient to cause neuronal death alone, when
combined with mild trauma (fluid percussion injury
(FPI) insufficient to cause neuronal death or major
behavioral sequelae), causes massive hippocampal
neuronal necrosis (Jenkins et al., 1989). This occurs
when trauma and ischemia occur up to 24 hours apart.
When the ischemia occurred more than 7 days after FPI,
however, neuronal necrosis did not occur. This synergistic damage phenomenon could also be abolished by
treatment with scopolamine, a muscarinic cholinergic
antagonist, and phencyclidine, a glutamate antagonist
138
(Jenkins et al., 1988). Both these phenomena are of

considerable clinical importance and although not fully
understood both are amenable to drug therapy.
7.7 Brain swelling and cellular events

after neurotrauma
Brain swelling occurs in almost all patients with
severe brain injury, and to a lesser extent in 510% of
those with moderate injuries (Miller et al., 1981;
Marmarou, 1994; Chapters 3 and 4). This high incidence is a major reason for neurointensive care
management for these patients, and this is a major
reason why aggressive surgical management of mass
lesions improves outcome. The causes of brain swelling after severe head injury are multifactorial and
poorly understood (Figure 7.14). The widespread
misconception that the increase in brain volume that
occurs after neurotrauma is due to vasogenic edema
which should be treated with corticosteroids is still
responsible for the unnecessary deaths of head-injury
patients; despite the lack of evidence supporting the
use of corticosteroids in numerous clinical trials
(Chapter 18). These misconceptions were fueled by
laboratory studies using models, such as the cold
injury model, that do not mimic the edema pattern of
head injury but rather that of meningiomas and
gliomas (Long, 1984).
There is now compelling evidence from animal
models such as weight-drop injury and fluid percussion that, immediately after impact, extracellular fluid
volume may increase briefly for less than 30 minutes
and that this may be associated with a transient
opening of the bloodbrain barrier to mediummolecular-weight markers 5070 kDa (Ito et al.,
1996). Some authorities feel that this brief bloodbrain
Figure 7.14
barrier opening develops because of the hypertensive

surge associated with impact, particularly in fluid
percussion injury. There is, of course, no data for these
early postinjury events in humans. More likely, there
is a net shift of small ions, chiefly Na/K and Cl,
together with obligated water, from the intravascular
compartment to the extracellular fluid. This occurs
down a concentration gradient, because ions have
begun to shift into neurons and astrocytes that are
grossly swollen (Figure 7.4). Recent MRI diffusionweighted imaging studies show that, after about 1
hour, the extracellular space rapidly shrinks as water
moves into cells because of the ionic shifts already
referred to above (Ito et al., 1996). Diffusion-weighted
imaging techniques will show a reduction in the
diffusion index consistent with constrainment of
water molecules within cells: this situation is exacerbated by superimposed ischemia (Ito et al., 1996).
At the worst end of the spectrum brain tissue that is
most severely injured is unable to restore ionic
homeostasis because insufficient glucose can be
delivered to the tissue via the microcirculation and a
vicious cycle may be set up whereby the microcirculation is squeezed by astrocyte end feet that
swell because of potassium uptake. This tissue
appears dark because of cytotoxic edema on the CT
scan, and there is loss of graywhite definition or
the ground-glass appearance on the CT scan (Chapter 9). Under these circumstances, intracranial pressure rises to further jeopardize global brain perfusion
and cause death. This same sequence of events occurs
within the penumbra of focally contused tissue
round localized contusions (Schroeder et al., 1994).
There is now clear evidence that the majority of
early brain edema, both global and focal, is cytotoxic
rather than vasogenic. In humans studied with both
Massive intraoperative brain swelling following removal of an acute subdural hematoma.
CONCLUSION
gadolinium-enhanced MRI and pertechnetateenhanced SPECT scans, vasogenic edema with opening of the bloodbrain barrier is only seen at later
time points around contusions, and not at all in
patients with diffuse non-focal injuries (Bullock et al.,
1990; Lang et al., 1991; Todd and Graham, 1990).
Vasogenic edema probably becomes important around
focal contusions on the second through the 10th15th
day.
Brain engorgement studies using MRI- and CTbased techniques to estimate cerebral blood volume
have shown that blood volume is uniformly reduced
initially after acute brain injury, although many
patients will demonstrate a phase of hyperemic CBF,
with increased flow values demonstrable from the
second through the seventh day after injury, most
strongly seen after removal of intracranial hematomas
(Chapter 5; Fatouros et al., 1985).
Marmarou et al. have recently used mathematical
modeling techniques to estimate that the vascular
component of brain swelling after severe brain injury
probably represents about 25% of the overall increased
in brain bulk, with the remainder being due predominantly to cytotoxic edema (Marmarou et al., 1993).
Clearly, however, all three of these components of
(a)
139
swelling may occur in the same patient, and they may

fluctuate in magnitude within the same patient at
different times after severe brain injury (Figure 7.15).
We hypothesize that the erectile brain swelling
seen in severely injured patients, chiefly after removal
of acute subdural hematoma, is caused primarily by
redistribution within the swollen brain tissue, primarily due to cytotoxic edema (Figure 7.14). However,
postischemic reperfusion is obviously an important
component of this rapid-onset brain swelling, and
indeed this can often be seen with the naked eye, as
small blood vessels over the pia dilate and reperfuse
after the subdural has been removed. Blood flow
studies have confirmed this (Chapter 5).
7.7.1
RESOLUTION OF BRAIN SWELLING
Kimelberg and others have hypothesized that astrocytes function as a syncytium or wick to conduct
potassium away from neurons, particularly in injured
brain, and thereby aid in the establishment of ionic
homeostasis (Kimelberg and Norenberg, 1989). Thus
there is a net loss of potassium from injured tissue into
the microvasculature which begins hours after onset.
In mildly affected tissue, astrocyte swelling will begin
to resolve after about 12 hours. In our own human
ultrastructural studies, astrocytes around contusions
appear to be shrinking by about the fifth day after
injury (Bullock et al., 1991). Clearly, when the microcirculation is competent and CBF remains above about
20 ml/100 g/min, recovery of brain swelling will be
much more rapid, and it is unlikely to occur at all
when blood flow in the microcirculation is below
these threshold levels.
7.7.2
IMPLICATIONS FOR THERAPY
Mannitol may exert its tremendously beneficial effects

in head-injured patients by opposing ionic flux into
damaged tissue by keeping the intravascular space
hyperosmolar (Nath and Galbraith, 1986).
Clearly, however, its rheological effects are also
extremely important, and mannitol has been shown to
reduce cerebral blood volume, through vasoconstriction, and it also increases CBF (Muizelaar et al., 1983).
A number of therapeutic strategies that block ion
channels are currently in clinical trials, and the effect
of measures that reduce the metabolic demand of the
damaged tissue, such as barbiturates and hypothermia, are being evaluated (Chapters 18, 19).
7.8
(b)
Figure 7.15 (a, b) Schematic to show the time course and
potential mechanisms causing post-traumatic brain edema.
Conclusion
Mortality rates for severe brain injury have fallen

about 10% per decade over the last 25 years, yet
during this period no single drug therapy has been
140
shown to be effective in clinical trials. This reduction

in mortality rates has been achieved by improvements
in prehospital care, early diagnosis, removal of intracranial hematomas and better neurointensive care
therapy. Each of these measures acts by optimizing the
conditions for substrate delivery (oxygen and glucose)
to the injured brain, and their importance cannot be
overemphasized. The focus for the next decade in
head injury care, therefore, needs to be twofold:
to disseminate knowledge regarding the principles

of optimal care that have been worked out over the
last 25 years, so that all head-injured patients may
benefit;
to translate the enormous advances in pharmacological therapy for brain damage, and monitoring
and detection of brain damage, from the laboratory
into the clinical arena.
7.9
Acknowledgments
Many of the studies referred to above were supported

by NIH grant NS12587.
7.10
References
Adams, J. K., Doyle, D. and Ford, l. (1989) Diffuse axonal injury in head
injury: definition, diagnosis, and grading. Histopathology, 15, 4959.
Astrup, J., Siesjo,
P. K. and Symon, L. (1981) Thresholds in cerebral ischemia
the ischemic penumbra. Stroke, 12, 723725.
Bergsneider, K., Shalmon, E., Kelly, D. F. et al. (1996) Hyperglycolysis acutely
following severe human traumatic brain injury: assessment and implication. Journal of Neurosurgery.
Bortolotto, Z. A., Bashir, Z. L., Davies, C. K. and Collingridge, G. L. (1994) A
molecular switch activated by metabotropic glutamate receptors regulates
induction of long term potentiation. Nature, 368, 740743.
Bouma, G. J., Muizelaar, J. P. and Stringer, W. A. (1992) Ultra-early evaluation
of regional cerebral blood flow in severely head injured patients using
xenon enhanced computed tomography. Journal of Neurosurgery, 77,
360368.
Bowman, C. L., Ding, J. P., Sachs, F. and Sokabe, M. (1992) Mechanotransducing ion channel in astrocytes. Brain Research, 584, 272286.
Branston, N. K., Symon, L., Crockard, H. A. et al. (1974) Relationships between
the cortical evoked potential and local cortical blood flow following acute
middle cerebral artery occlusion in the baboon. Experimental Neurology, 45,
195208.
Bullock, R., Hannemann, C. O., Murray, L. and Teasdale, G. M. (1990)
Recurrent hematomas following craniotomy for traumatic intracranial
mass. Journal of Neurosurgery, 72, 914.
Bullock, R., Landolt, H., Maxwell, W. L. and Fujisawa, H. (1994) Massive
astrocytic swelling in response to extracellular glutamate a possible
mechanism for post traumatic brain swelling? Acta Neurochirurgica, 60,
465467.
Bullock, R., Statham, P., Patterson, J. et al. (1990) The time course of vasogenic
edema after focal human head injury evidence from SPECT mapping of
blood brain barrier deficit. Acta Neurochirurgica, 51(Suppl.), 286288.
Bullock, R., Maxwell, W. L., Graham, D. I. et al. (1991) Glial swelling following
human cerebral contusion: an utrastructural study. Journal of Neurology,
Bullock, R., Zauner, A., Tsuji, O. et al. (1995) Patterns of excitatory amino acids
release and ionic flux after severe human head trauma, in Proceedings of the
9th International Symposium on ICP and Brain Damage, Springer-Verlag,
Tokyo, pp. 6471.
Crockard, H. A. and Taylor, A. R. (1972) Serial CSF lactate/pyruvate levels as
a guide to prognosis in head injury. European Neurology, 8, 151157.
Dawson, D. A. (1994) Nitric oxide and focal cerebral ischemia: multiplicity of
actions and diverse outcome. Cerebrovascular Brain Metabolism Review, 6,
299324.
DeGirolami, U., Crowell, R. M. and Marcoux, S. W. (1984) Selective necrosis

and total necrosis in focal cerebral ischemia. Neuropathologic observations
on experimental middle cerebral artery occlusion in the macaque monkey.
Neuropathology and Experimental Neurology, 43, 5764.
Delahunty, T. K., Jiang, J.-Y., Black, R. T. and Lyeth, B. G. (1995) Differential
modulation of carbachol and trans-ACPB-stimulated phosphoinositide
turnover following traumatic brain injury. Neurochemical Research, 20,
405411.
Di, X., Harpold, T., Watson, J. C. and Bullock, R. (1996a) Excitotoxic damage
in neurotrauma: fact or fiction? Restorative Neurology and Neuroscience, 9,
231241.
Di, X., Lyeth, B., Hamm, R. and Bullock, R. (1996b) Voltage dependent Na/K
ion channel blockade fails to ameliorate behavioral deficits after traumatic
brain injury in the rat. Journal of Neurotrauma.
Doppenberg, E., Zauner, A., Bullock, R. et al. (1996) Massively increased extracellular fluid lactate in severe human head injury. Evidence for preferential
glycolysis? Journal of Neurosurgery, (abstract).
DuTrevou, M. and Van Dellen, J. R. (1992) Penetrating stab wound to the
brain: the timing of angiography in patients presenting with the weapon
already removed. Neurosurgery, 3l, 905911.
European Study Group on Nimodipine In Severe Head Injury (1994) A multicentered trial of the efficacy of nimodipine on outcome after severe head
Fatouros, P., Schroeder, M. L., Muizelaar, P. et al. (1985) Dynamic CT and
cerebral blood volume measurement, in Quantitative CBF Measurement
Using Stable Xenon CT Clinical Application, (eds M. Tomonaga, A. Tanaka
and H. Yonas), Fatoura, New York, pp. 95110.
Fineman, I., Hovda, D. A., Smith, K. et al. (1993) Concussive brain injury is
associated with a prolonged accumulation of calcium: a 45 calcium
autoradiographic study. Brain Research, 624, 94102.
Gennarelli, T. A., Thibault, L. E. and Adams, J. H. (1985) Diffuse axonal injury
and traumatic coma in the primate, in Trauma of the Central Nervous System,
(eds R.-G. Dacey et al.), Raven Press, New York, pp. 169193,
Grady, M. S., McLaughlin, M. R., Christman, C. W. et al. (1993) The use of
antibodies targeted against neuro-filament subunits for the detection of
diffuse axonal injury in humans. Journal of Neuropathology and Experimental
Graham, D. I. (1985) The pathology of brain ischaemia and possibilities for
therapeutic intervention. British Journal of Anaesthesia, 57, 325.
Graham, T. W., Williams, F. C., Harrington, T. and Spetzler, R. F. (1990) Civilian
gunshot wounds to the head: a prospective study. Neurosurgery, 27, 696700.
Gurdjian, E. S., Lissner, H. R. and Hodgson, V. R. (1966) Mechanisms of head
injury. Clinical Neurosurgery, 12, 112128.
Hall, E. D. and Braughler, J. M. (1989) Central nervous system trauma and
stroke pathophysiological and pharmacological evidence for involvement
of oxygen radicals and lipid peroxidation. Free Radicals and BioMedicine, 6,
303313.
Harders, A., Kakarieka, A., Braakman, R. et al. (1996) Traumatic subarachnoid
hemorrhage and its treatment with nimodipine. Journal of Neurosurgery, 85,
8289.
Hayashi, K., Handa, H., Nagasawa, S. et al. (1980) Stiffness and elastic
behavior of human intracranial and extracranial arteries. Journal of
Biomechanics, 13, 175184.
Holbourne, A. H. S. (1943) Mechanics of head injury, Lancet, ii, 438441.
Inao, S., Marmarou, A., Clark, G. E. et al. (1988) Production and clearance of
lactate from brain tissue, cerebrospinal fluid, and serum following
experimental brain injury. Journal of Neurosurgery, 69, 736744.
Islam, N., Aftabuddin, M., Moriwaki, A. and Horri, Y. (1995) Detection of
DNA damage induced by apoptosis in the rat brain following incomplete
ischemia. Neuroscience Letters, 188, 159162.
Ito, J., Marmarou, A., Barzo, P. et al. (1996) Characterization of edema by
diffusion weighted imaging in experimental traumatic brain injury. Journal
Jenkins, L. W., Lyeth, B. G., Lewelt, W. et al. (1988) Combined pretrauma
scopolamine and phencyclidine attenuate post traumatic increase sensitivity to delayed secondary ischemia. Journal of Neurotrauma, 5, 275282.
Jenkins, L. W., Monzynski, K., Lyeth, B. G. et al. (1989) Increased vulnerability
of the mildly traumatized brain to cerebral ischemia: the use of controlled
secondary ischemia as a research tool to identify common or different
mechanisms contributing to mechanical and ischemic brain injury. Brain
Research, 477, 212224.
Jones, T. H., Morawetz, R. B., CrowelI, R. M. et al. (1981) Threshold of focal
cerebral ischemia in awake monkeys. Journal of Neurosurgery, 54, 773782
Katayama, Y., Becker, D. P., Tamura, T. et al. (1990) Massive increases in
extracellular potassium and the indiscriminate release of glutamate
following concussive brain injury. Journal of Neurosurgery, 73, 889900.
Kawamata, T., Katayama, Y., Hovda, D. A. et al. (1995) Lactate accumulation
following concussive brain injury: the role of ionic fluxes induced by
excitatory amino acid. Brain Research, 674, 196204
Kimelberg, H. K. and Norenberg, M. D. (1989) Astrocytes. Scientific American,
260, 6676.
King, L. R., McLaurin, R. L. and Knowles, H. C. (1974) Acid base balance and
arteriole and CSF lactate levels following human head injury. Journal of
REFERENCES
Kirino, T. (1982) Delayed neuronal death in the gerbil hippocampus following
ischemia. Brain Research, 39, 5764.
Kirino, T., Yamure, A. and Sano, K. (1984) Delayed neuronal death in the rat
hippocampus following transient forebrain ischemia. Acta Neuropathologica
(Berlin), 64, 139147.
Kontos, H. A. (1985) Oxygen radicals in cerebral vascular injury. Circulation
Research, 57, 508516.
Kuijpers, A. H. W. M., Claessens, M. H. and Sauren, A. A. (1995), The
influence of different boundary conditions on the response of the head to
impact: a two dimensional finite element study, in Traumatic Brain Injury
Bioscience and Mechanics, (eds F. Bandak, R. Eppinger, A. Ommaya and M. A.
Liebert), New York, pp. 197206.
Kuroda, Y. and Bullock, R. (1992) Failure of cerebral blood flow-metabolism
coupling after acute subdural hematoma in the rat. Neurosurgery, 31,
10621071.
Kuroda, Y., Dewar, D. and Bullock, R. (1993) Early changes in second
messenger but not receptor binding sites after acute subdural hematoma
an in vitro autoradiographic study. Journal of Neurotrauma, 10, 4755.
Kuroda, Y., Ingles, F., Miller, J. et al. (1992) Transient glucose hypermetabolism
after acute subdural hematoma in the rat. Journal of Neurosurgery, 76,
471477.
Lang, D. A., Hadley, D. M., Teasdale, G. T. et al. (1991) Gadolinium DTPA
enhanced magnetic resonance imaging in acute head injury. Acta Neurochirurgica (Vienna) 109, 511.
Long, D. M. (1984) Microvascular changes in cold injury edema, in Recent
Progress in the Study and Therapy of Brain Edema 4, (eds K. G. Go and A.
Baethmann), Plenum Press, New York, pp. 4554.
Le Coma Traumatique, (eds I. Papo, F. Cohadon and M. Massarotti), Liviana
Editrice, Padua, pp. 165185.
Marmarou, A. (1994) Traumatic brain edema: an overview. Acta Neurochirurgica, 60(Suppl.), 421424.
Marmarou, A., Fatouros, P., Bandoh, K. et al. (1993) The contribution of brain
edema to brain swelling, ICP and craniospinal dynamic, in Intracranial
Pressure VIII, (eds C. J. J. Avezaat, J. H. M. van Eindhoven, A. I. R. Maas and
J. T. J. Tans), Springer-Verlag, Berlin, pp. 525528.
Marmarou, A., Foda, M., Vandenbrink, W. et al. (1994) A new model of diffuse
brain injury in rat. part 1: pathophysiology and biomechanic. Journal of
Massa, S. M., Swanson, R. A. and Sharp, F. R. (1996) The stress gene response
in brain. Cerebrovascular Brain Metabolism Review, 8, 95158.
Mesenge, C., Verrecchia, C., Allix, M. et al. (1996) Reduction of the
neurological deficit in mice with traumatic brain injury by nitrate oxide
synthase inhibitors. Journal of Neurotrauma, 13, 209214.
Miller, J. D., Butterworth, J. F., Gudeman, S. K. et al. (1981) Further experience
289299.
Miyazaki, S., Katayama, Y., Lyeth, B. G. et al. (1992) Enduring suppression of
hippocampal long-term potentiation following traumatic brain injury in the
rat, Brain Research, 585, 335339.
Muizelaar, J. P., Wei, E. P., Kontos, H. A. et al. (1983) Mannitol causes
compensatory cerebral vaso-constriction and vasodilatation to blood
viscosity changes. Journal of Neurosurgery, 59, 822828.
Myseros, J. S., Bullock, M. R., Zauner, R. A. et al. (1995) Brain extracellular
fluid glucose content after severe human head injury: microdialysis
evaluation of contributing factors. Journal of Neurotrauma, 12, 410.
Narayan, R., Willberger, J. and Povlishock, J. (eds) (1966) Neurotrauma,
McGraw-Hill, New York.
Nath, F. and Galbraith, S. (1986) The effect of mannitol on cerebral white
matter water content. Journal of Neurosurgery, 65, 4143.
Newman, E. A. (1986) High potassium conductance in astrocyte end feet.
Science, 223, 453454.
Nilsson, P., Hillered, L., Olsson, Y. et al (1993) Regional changes in interstitial
potassium and calcium levels following cortical compression, contusion,
trauma in rat. Journal of Cerebral Blood Flow and Metabolism, 13, 183192.
Obrist, W. D., Langfitt, T. W., Jaggi, J. L. (1984) Cerebral blood flow and
metabolism in comatose patients with acute head injury. Journal of
Ommaya, A. K., Hirsch, A. E., Flamm, E. S. et al. (1966) Cerebral concussion in
the monkey: an experimental model. Science, 153, 211212.
141
Orkand, R. K., Nicholls, J. G. and Kuffler, S. W. (1966) Effect of nerve impulses

on the membrane potential of glial cells in the central nervous system of
amphibia. Journal of Neurophysiology, 29, 778806.
Peerless, S. J. and Rewcastle, N. M. (1967) Shear injuries of the brain. Canadian
Medical Association Journal, 96, 557582.
Pellerin, L. and Magistretti, P. J. (1994) Glutamate uptake into astrocytes
stimulates aerobic glycolysis: a mechanism coupling neuronal activity to
glucoutilization. Proceedings of the National Academy of Sciences of the USA,
91, 1062510629.
Pettus, E. H., Christman, C. W., Gieble, M L and Povlishock, J. T. (1994)
Traumatically induced altered membrane permeability: its relationship to
traumatically induced reactive axonal change. Journal of Neurotrauma, 11,
507522.
Prasad, R. M., Dhillon, H. S., Carbary, S. et al. (1994) Enhanced phosphodiestetic breakdown of phosphatidylinositol bisphosphate after experimental
brain injury. Journal of Neurochemistry, 63, 773776.
Pulsinelli, W. A., Brierley, J. B. and Plum, F. (1982) Temporal profiles of
neuronal damage in a model of transient forebrain ischemia. Annals of
Rehcrona, S., Hauge, H. N. and Siesjo,
B. K. (1989) Enhancement of iron
catalysed free radical formation by acidosis in brain homogenates:
differences in effect by lactate acid and CO2. Journal of Cerebral Blood Flow
and Metabolism, 9, 6570.
Reilly, P. L., Adams, R. H., Graham, D. I. and Jennett, B. (1975) Patients with
head injury who talk and die. Lancet, ii, 375377.
Rink, A. D., Fung, K. M., Fan, L. et al. (1994) Cell death by apoptosis after lateral
fluid percussion brain injury. Society of Neuroscience Abstract, 20, 250.
Sachs, F. (1991) Mechanical Transduction by Membrane Ion Channels: A Mini
Review. Molecular and Cellular Biochemistry, 104, 5760.
Schroeder, M. L., Muizelaar, J. P. and Kuta, A. J. (1994) Documented reversal
of global ischemia immediately after removal of an acute subdural
hematoma. Journal of Neurosurgery, 80, 324327.
Schroeder, M. L., Muizelaar, J. P., Bullock, R. et al. (1994) Focal ischemia due to
traumatic contusion documented by stable, xenon CT, and ultrastructural
studies. Journal of Neurosurgery, 82, 966971.
Schwartz, J. H. and Kandell, E. R. (1991) Synaptic transmission mediated by
second messengers, in Principles of Neuroscience, (eds E. R. Kandell, J. H.
Schwartz and T. M. Jessell), Elsevier Science, New York, pp. 173193.
Siegelbaum, S. A. and Koester, J. (1991) Ion channels, in Principles of
Neuroscience, (eds E. R. Kandel, J. H. Schwartz and T. M. Jessell), Elsevier
Science, New York, pp. 6679.
Siesjo,
B. K. (1992a) Pathophysiology and treatment of focal cerebral ischemia.
Part II: Mechanisms of damage and treatment. Journal of Neurosurgery, 77,
337354.
Siesjo,
B. K. (1992b) Pathophysiology and treatment of focal cerebral ischemia.
Part I: Pathophysiology. Journal of Neurosurgery, 77, 169184.
Strich, S. J. (1956) Diffuse degeneration of the cerebral white matter in severe
dementia following head injury. Journal of Neurology, Neurosurgery and
Sutton, R. L., Hovda, D. A., Adelson, P. D. et al. (1994) Metabolic changes
following cortical contusion: relationships to edema and morphological
changes. Acta Neurochirurgica (Supplement), 60, 446448.
Symon, L., Pasztor, E. and Branston, N. M. (1974) The distribution and density
of reduced cerebral blood flow following acute middle cerebral artery
occlusion: an experimental study by the technique of hydrogen clearance in
baboons. Stroke, 5, 355364.
Tavalin, S. J., Ellis, E. and Satin, S. J. (1995) Mechanical perturbation of
cultured cortical neurons reveals a stretch-induced delayed depolarization.
Journal of Neurophysiology, 74, 27672773.
Todd, N. V. and Graham, D. I. (1990) Blood brain barrier damage in traumatic
brain contusion. Acta Neurochirurgica (Supplement) (Vienna), 51, 296299.
Yoshino, A., Hovda, D. A., Kawamata, T. et al. (1991) Dynamic changes in local
cerebral glucose utilization following cerebral concussion in rats: evidence of
a hyper- and subsequent hypo-metabolic state. Brain Research, 561, 106119.
Zauner, A., Bullock, R., Kuta, J. et al. (1996a) Glutamate release and cerebral
blood flow after severe head injury. Acta Neurochirurgica.
Zauner, A., Doppenberg, E., Woodward, J. J. et al. (1996b) Multi-paremetric
continuous monitoring of brain metabolism and substrate delivery in
neurosurgical patients. Neurological Research.
Part Two
MEASURING AND
MONITORING INJURY
Introduction
CT scanning has provided a fast, non-invasive and
detailed method of assessing the morphology of brain
injury and is now an essential part of any trauma
management system. CT scanners are increasingly
available at the hospital of first contact and connected
to a central trauma center by teleradiology. This has
led to an increase in the speed of diagnosis and
treatment of surgical hematomas, better planning and
retrieval, and more discriminating admission policies
for patients with severe head injury or skull fracture.
Our limited understanding of the pathophysiological events that occur after brain injury, and the
high frequency of secondary delayed neurological
deterioration, have stimulated the search for accurate,
continuous monitoring techniques, particularly for the
first days after injury. Some of these techniques are
listed in Table II.1. None of the present techniques are
ideal, and academic head injury centers continue to
search for better methods.
The reasons for monitoring the injured brain are
simple:

to detect harmful pathophysiological events before

they cause irreversible damage to the brain;
to allow the type of harmful pathophysiological
process to be diagnosed and effectively treated;
to provide on-line feedback to guide therapy
directed at these processes.
In addition they should be non-invasive and preferably cheap.

In general, monitoring techniques may be divided
into two types those that assess substrate delivery to
the injured brain and those that assess brain function.

Brain function is particularly difficult to assess, since
most severely head-injured patients are in coma.
Monitoring techniques that aim at providing continuous or near continuous information must be
differentiated from intermittent measurement techniques (e.g. CT scanning) that provide a static image
of the injured brain at intervals. Modern management
clearly uses both techniques.
Duration of monitoring
The optimal duration of monitoring will vary from
patient to patient, depending upon the pathophysiological process in question. In a recent large series
from the Medical College of Virginia, the mean
duration of ICP monitoring ranged between 5 and 7
days. Often it is not practical to continue to monitor a
patient until consciousness returns. On the other
hand, it is seldom necessary to continue monitoring
once a patient is able to obey commands; the patient
can then be followed adequately with the neurological
exam. It is desirable to continue monitoring until ICP
is beginning to decline and until cerebrovascular
autoregulation has been reestablished. Studies have
shown that autoregulation is generally recovering by
the seventh day after injury and is almost always back
to normal for all three physiological stimuli blood
pressure, PaCO2 and PO2 by 2 weeks after injury.
In the following chapters, monitoring of ICP, CBF,
transcranial Doppler, brain oxygen uptake and electrophysiological function are highlighted. Many of
these techniques are complementary.
144
INTRODUCTION
Table II.1
Advantages and disadvantages of monitoring techniques used for acute brain injury
Method
Advantages
Disadvantages
Approximate cost
(US$; 1995)
Invasiveness/
risk
Techniques for assessing substrate delivery

ICP Monitoring
Ventriculostomy
May drain CSF to lower

ICP; may calculate CPP;
may use waveform
analysis and measure PVI.
Highest infection risk

(510%).
Hemorrhage risk ~ 0.5%.
May be difficult to insert.
450
++++
Easy to insert and use.

Lower infection/
hemorrhage risk; may use
waveform analysis.
Cannot drain CSF.

Difficult to recalibrate.
Camino 6500
Each sensor 485
++++
May select the brain region

to monitor. May guide
hyperventilation therapy.
Measures relative
change, not absolute
flow. Thermal diffusion
requires a craniotomy for
optimal insertion.
Instrument
470016 500
Sensor 300
+++
May guide
hyperventilation and
pressor therapy.
Intermittent. Indirect
measure of brain O2
extraction and flow.
Contaminated by
extracranial venous flow.
Catheter 100
Sample analysis
500/day
+
+
Fiberoptic sensor e.g.

Oxymetrix
Continuous.
Inaccurate in up to 40%
of readings.
28 000 (instrument)
765 (each sensor)
Transcutaneous
near-infrared
spectroscopy (NIRS)
Correlates well with brain

O2 (Hammamatsu system)
Accuracy/specificity not
yet proven in trauma.
Not reliable when
intracranial bleeding is
present.
50 000
(Hammamatsu
system)
Brain oxygen
measurement
Reflects true substrate

delivery: accurate. (May
measure, CO2 , pH and
temperature).
Fragile; microregional.
20 000 (system)
300 (each sensor)
+++
Transcranial
Doppler
Wave-form analysis may

indicate high ICP. Detects
vasospasm in ~ 28%.
Qualitative; difficult to
fix the head; operatordependent. Significance
unclear.
15 00030 000
(system)
Microdialysis
Many analytes available

for analysis; highly
sensitive.
Not on-line; requires

HPLC for analysis; laborintensive.
20 000 (system)
26 000 (HPLC)
150 (probes)
++
Parenchymal
electronic sensors
e.g. Camino,
Codman
CBF flow probes
Thermal diffusion or
laser Doppler
AVDO2
Jugular catheter
Techniques for assessing brain function

Neurological
observation (coma
scale)
Most sensitive and specific

indicator of brain function.
Lost in coma; may be

preserved while focal
damage is occurring, in
silent areas?
Nursing time
EEG/EEG
Spectral array
Sensitive, even in coma;

indicates seizures, even
when patient is on muscle
relaxants.
Useful in prognosis.
Difficult to interpret;
non-specific; significance
uncertain.
system 20 00080 000
Evoked potentials
Loss of evoked potentials

correlates well with death/
vegetative outcome. Useful
for prognosis. May detect
focal deficit or spinal
injuries.
Not a useful guide to

therapy. Operatordependent.
system
40 000100 000
CLINICAL EXAMINATION
AND GRADING
Donald A. Simpson
8.1
Introduction
8.1.1 ROLES AND LIMITATIONS OF CLINICAL

EXAMINATION
Severe closed head injuries are now routinely investigated by early computed tomography (CT), which
visualizes most pathological lesions of immediate
surgical importance. It is also routine practice to
monitor the intracranial pressure and other parameters of cerebral physiology, providing objective
data to control the use of artificial ventilation and
other forms of conservative therapy. It is therefore
legitimate to ask what now are the roles of clinical
neurology in the management of head injuries in
general, and severe head injuries in particular.
The initial clinical evaluation is still crucially important, in triage and as a baseline in assessing progress.
The prognosis depends to a large extent on the
findings of the initial examination, and the neurological status at a specified time after injury is widely
used as a measure of head injury severity. Moreover,
valuable as they are, the neuroradiological findings
must be interpreted in the light of the clinical findings.
Thus clinical examination of the head-injured patient
continues to be indispensable. But modern strategies
of severe closed head injury management have
brought one very important change in the nature of
this examination: because it is often wise to perform
endotracheal intubation as soon as possible, the first
neurological examination is now usually performed at
the accident site or in the emergency room, and often
by someone with no special training in neurology.
This means that the paramedic, the intensivist and the
emergency physician must be competent in performing
an appropriate neurological examination before intubation and respiratory paralysis are instituted.
This does not mean that all the rites of medical
neurology should be taught to everyone who may
intubate an unconscious patient; it means that such
persons must be skilled in making a few basic
neurological observations and recording them accu-
rately. These observations are usually done in two

phases. In the primary survey, the conscious level
and the pupillary reactions are tested (Table 8.1). If
resuscitation includes endotracheal intubation, limb
movements should be quickly tested before a muscle
relaxant is given. In the secondary survey, done after
resuscitation, these findings are checked, and in
addition the examiner assesses limb motor function
and, if possible, vision and cutaneous sensation
(Table 8.2).
A definitive or tertiary examination by a surgical
neurologist retains its value in the late evaluation of a
severe head injury. It is often done in collaboration
with a neuropsychologist. This definitive examination
considers especially the neurological functions that
cannot be tested in the unconscious patient, notably
speech, mentality, cognitive functions, smell, vision,
hearing and sensorimotor function. The definitive
assessment also includes a retrospective judgment of
the duration of unconsciousness and amnesia (see
below). The long-term effects of severe head injury are
outside the scope of this book, but the care of a headinjured person should not be compartmentalized:
evaluation like counseling is a continuous process
and the early findings bear important relationships
with what is found as the patient emerges from coma.
8.1.2
THE HISTORY
The history retains its importance in the evaluation of

severe head injury. It is usually obtained from eyewitnesses of the accident and from the family or friends
of the injured person. The site and cause of the impact
may give clues to the pathology; blunt weapons, falls
and road crashes show more or less characteristic
patterns of intracranial damage. The sequence of
events after impact may distinguish between primary
brain damage and secondary cerebral compression.
The health before injury may be relevant. Medicalert
bracelets identify serious illness or medication, and are
often helpful in unconscious patients. However people
146
CLINICAL EXAMINATION AND GRADING
Table 8.1 The primary survey. The A-B-C-D-E summary is a mnemonic for the early evaluation of severe trauma. It
gives basic data on the neurological status after head and/or spinal injury (after American College of Surgeons
Committee on Trauma, 1993)
Survey
Check
Note, record and correct
A Airway
Patent?
Noisy?
Obstruction
B Breathing
Effective?
Rate and depth

Chest movements
Air entry
Cyanosis
C Circulation
Adequate?
Pulse rate and volume

Skin color
Capillary return
Hemorrhage
Blood pressure
D Disability
(= neurological status)
Normal?
Conscious level AVPU or preferably GCS

Pupillary light reactions
E Exposure
(= undress)
Other injuries?
Limb movements on command or on painful stimulus
with serious illnesses may refuse to wear these,

especially if the disease is seen as a stigma.
8.2
The initial examination
8.2.1
CONSCIOUS LEVEL
For the clinician, this is the best empirical measure of

impaired cerebral function after closed head injury.
Table 8.2 The secondary survey. This is done after

resuscitation. It includes a minimal neurological
examination, which should be within the competence of
any practitioner undertaking the early management of
severe trauma. It commonly leads at once to special
investigations e.g. X-ray of the cervical spine, CT scan,
blood screen for ethanol and other drugs
History
From patient and from observers
Reassessment of
vital signs
As in primary survey
External signs of
injury
Inspection and palpation of scalp,

face, eyes and neck
Conscious level
GCS is repeated; this entails

evaluation of speech and record of
airway
Vision
Pupils: evaluation of size, symmetry

and light reaction
Acuities and fields: done by covering
each eye and checking visual
perception of examiners face
Limb weakness
lateralized or
localized?
Detected from spontaneous

movement, movement on commands,
movement on finger- and toenail
pressure
Impairment of consciousness is stratified in terms of

the responses to external stimuli, and serial records of
these responses are important in head injury management. The early postinjury conscious level may be
preserved prior to secondary deterioration; this means
a much better prognosis and also implies that a
hematoma may be present. Though the role of nursing
records in the early detection of cerebral compression
has been diminished by the advent of CT scanning,
serial progress records of the conscious level remain
standard practice in head-injury observation, especially when the head injury initially appears to be less
serious (Figure 8.1).
The conscious level is also a valuable index of injury
severity. In early evaluation, the depth of impairment
of consciousness can be used as a measure of cerebral
impairment, provided that the dimension of time after
impact is taken into account and provided that
confounding causes of impaired consciousness, such
as ethanol, drugs or hypoxia, can be excluded. In late
evaluation, a retrospective estimate of the duration of
loss of consciousness can be used as a definitive
measure of cerebral injury, though with certain reservations that are discussed below. For both purposes, it
is necessary to define and quantify impairments of
consciousness.
8.2.2
COMA SCALES
In 1941, a wartime committee of British clinical

neuroscientists recognized the need for a standardized
terminology for states of impaired consciousness after
head injury, and published a glossary of descriptive
terms. This committee included three brilliant neurosurgeons Hugh Cairns, Geoffrey Jefferson and
THE INITIAL EXAMINATION
147
Figure 8.1 Neurological observation sheet. The record shows rapid deterioration of the consciousness level with
appearance of ipsilateral pupillary dilatation and a contralateral hemiparesis. The GCS score falls from 15 to 5. Since the
indications for action were evident by 15.30 h at latest, the record would imply unacceptable management! (Hypothetical
case recorded on the 15-point version of the Glasgow Coma Scale, by courtesy of Mr M. Fearnside.)
Norman Dott. The committee had great authority and

the terminology was widely accepted. Although the
original publication (Medical Research Council Brain
Injuries Committee, 1941) did not set out a graded
hierarchy of levels of impaired consciousness, the
recommended terms were easily used to formulate

such a hierarchy, and appeared in graphs of clinical
progress in Rowbothams (1945) very influential textbook (Table 8.3). Those who used this system soon
found that it was often misleading in communication
148
Table 8.3 Depth of unconsciousness, simplified from

Medical Research Council Brain Injuries Committee, 1941.
This terminology for post-traumatic impairments of
consciousness is now only of historical interest: it has
however influenced later thinking (Starmark, 1988)
Coma
Absolute unconsciousness, judged by

absence of any psychologically
understandable response to external
stimuli or inner need
Primitive reflexes may or may not be
present
Semicoma
Psychologically understandable responses

are elicited only by painful or other
disagreeable stimuli
Primitive reflexes present
Confusion
Impaired capacity to think clearly and

rapidly, to perceive and remember current
stimuli; also disorientation
Severe
The patient occasionally responds to

simple commands, if necessary reinforced
by gestures, e.g. put out your tongue,
take my hand
Moderate
The patient, though out of touch with his

surroundings, can be got to give relevant
answers to simple questions, e.g. what
work do you do?, where do you live?
Mild
Confused but capable of coherent

conversation and appropriate behaviour
between staff members without neurological training,

and many neurosurgeons tried to formulate scales that
used explicit grades of response to specified stimuli,
given in simple descriptive terms. It was from such
endeavors that Teasdale and Jennett (1974) devised the
Glasgow Coma Scale (GCS).
This scale has been in use for more than 20 years,

and its value is so widely accepted that a description
may seem superfluous. However, the scale has its
critics, and its application has changed somewhat
since it was first reported. The authors described
hierarchies in the levels of response for movement of
the upper limb, verbal or vocal utterance and eye
opening (Table 8.4). In their first paper, they listed only
five levels of motor response; they noted that it is
possible to distinguish normal and abnormal limb
flexion, but concluded that the distinction was not
appropriate for general clinical use. As a painful
stimulus, nailbed pressure with a pencil was advised.
For verbal responses and for eye opening, the scale
provided five and four levels respectively. In describing the painful stimulus for eliciting eye opening, the
authors specifically warned against using supraorbital
pressure or retromandibular pressure, as these stimuli
may evoke eye closure. This warning has not always
been heeded. Although the value of the scale as a
means of communication between different hospitals
was emphasized, no attempt was made in this paper
to give the findings in an aggregated score. Tests of
inter-rater accord were later reported, and showed
reasonably close agreement in evaluations by nurses,
neurosurgeons and other medical practitioners (Teasdale, Knill-Jones and Jennett, 1974; Teasdale, KnillJones and Van der Sande, 1978). Similar tests have
shown good agreement between ratings by emergency
physicians and paramedics(Menegazzi et al., 1993).
In later papers, the authors formulated six levels of
motor responses by including the distinction between
abnormal or spastic flexion and flexor withdrawal
(Teasdale and Jennett, 1976; Jennett et al., 1977) though
Table 8.4 Variant forms of the Glasgow Coma Scale. In their first publication, Teasdale and Jennett (1974) did not
discuss the use of a summated score; in later publications, the 15-point version of the GCS has been used in giving
Coma Score
14-point scale
(Teasdale and Jennett, 1974)
15-point scale
(Teasdale and Jennett, 1976)
Pediatric scale
(Simpson and Reilly, 1982)
Eye opening
Spontaneous
To sound
To pain
None
4
3
2
1
The same
The same
Best verbal response
Orientated
Confused
Inappropriate
Incomprehensible
None
5
4
3
2
1
The same
Orientated
Words
Vocal sounds
Cries
None
5
4
3
2
1
Best motor response
Obeying
Localizing
Flexing
Extending
None
5
4
3
2
1
Obeys commands
Localizes pain
Flexionwithdrawal
Flexionabnormal
Extension
None
Obeys commands
Localizes pain
Flexion
Extension
None
5
4
3
2
1
Maximum sum
14
15
6
5
4
3
2
1
14
they recognized that the original simpler scale might be

preferable for clinical purposes (Jennett and Teasdale,
1981). Abnormal flexion was recorded if there were any
two of the following: stereotyped flexion posture,
extreme wrist flexion, abduction of the upper arm and
fisting of the fingers over the thumb. This additional
distinction made the GCS somewhat more demanding
for the less skilled observer, but increased its analytic
power, especially in severe closed head injuries; so
much so that Jagger et al. (1984) argued that, as a
prognostic guide, the Glasgow motor score alone was
more informative, at least in comatose patients.
Teasdale and Jennett (1976) also recommended the
use of the total score (summed scores for eye opening,
verbal and motor responses) for comparison of head
injury severity between patients and between series,
and as a rough definition of coma. A patient giving no
verbal response, not obeying commands and not
opening the eyes was judged to be in coma; by this
definition, all patients showing a GCS score of 7 or less
were comatose, and so were the majority (53%) of those
with GCS score 8, when the maximum score was 15.
149
The GCS found many supporters, but also a few

critics and skeptics. Some critics were clinicians who
already used a measure of consciousness of some
kind, and were reluctant to discard it; others devised
supposedly more convenient variants of the GCS to
meet local needs. Other criticisms related to the
admitted limitations of the GCS in cases with periorbital swelling, which may eliminate the eye-opening
response, or endotracheal intubation, which eliminates the verbal response. Starmark, Holmgren and
Stlhammer (1988) reviewed 96 head injury studies
published in the period 19831985; in these, GCS data
were interpreted or aggregated in many ways, and
other methods of grading consciousness were used in
23 papers (24%). Starmark et al. (1988) compared the
Swedish form of the GCS (using retromandibular
pressure and nailbed pressure as painful stimuli) with
their own Reaction Level Scale (RLS 85; Table 8.5) and
found better inter-rater agreement with the RLS 85;
however, Johnstone et al. (1993) could find no significant differences between these scales in discriminating between grades of head injury severity, though
Table 8.5 The Swedish Reaction Scale (RLS85); in the manual for this scale, the responses are further explained by
diagrams (Source: after Starmark et al., 1988, with simplification of the explanatory column. The RLS85 Manual is
published by Acta Neurochirurgica (Wien).)
Mentally responsive
1. Alert. No delay in response
Alert: not drowsy, orientated (intubated patient: no signs of delay in

reaction)
2. Drowsy or confused. Responsive to

light stimulation (verbal or touch)
Drowsy: the patients seems drowsy and shows delay in reaction

Confused: the patient gives the wrong answer to at least one of three
questions: What is your name? Where are you? What is the year and the
month?
3. Very drowsy or confused.

Responsive to strong stimulation
(loud verbal, shaking, pain)
Arousable: performs at least one of the following functions: oral response

with words; orientating eye movements; obeying commands; warding off
pain
Mentally unresponsive
4. Unconscious. Localizes but does
not ward off pain
Unconscious. No mental activity. Cannot perform any of the activities listed

above for mentally responsive
Localizes pain. Examination is done in supine position: retromandibular
pressure elicits movement of arm above chin level; nailbed pressure elicits
movement of other hand across the midline
5. Unconscious. Withdrawing
movements to pain
Withdrawing movements. On retromandibular pressure patient turns face

away; on nailbed pressure, patient does not localize the pain but makes
clear withdrawing movements
6. Unconscious. Stereotyped flexion

movements to pain
Stereotyped flexion movements. On retromandibular pressure or on

nailbed pressure, patient makes slow and mechanical flexion movements of
elbows and wrists but no localizing or withdrawing movements
7. Unconscious. Stereotyped
extension movements to pain
Stereotyped extension movements. On retromandibular pressure or on

nailbed pressure, patient makes extension movements, straightening arms or
legs. No flexion is seen; if both flexion and extension are seen, the better
response (i.e. flexion) is recorded
8. Unconscious. No response to pain
No response to pain. Repeated strong pain from retromandibular or nailbed

pressure gives no movement in arms, legs or face
150
the RLS 85 was regarded as a simpler procedure. The

GCS has been in service for more than 20 years, and it
should now be possible to give an appraisal of the
present status of the GCS in the three roles for which
it has been used: neurological observation, prognosis
and severity grading.
For the first purpose, it seems that most nurses,
ambulance officers and surgeons accept the scale as a
means of detecting changes in the conscious level and
as a convenient currency in communication. In some
centers, other in-house coma scales are used for
supposedly greater simplicity, but in general the GCS
has been found to be very serviceable. Many standard
textbooks recommend the GCS. The American College
of Surgeons Committee on Trauma (1993), in its
manual Advanced Trauma Life Support Program for
Physicians, sanctions the simpler non-quantitative
AVPU variant, which has four levels alert (A),
responsive to vocal stimuli (V), responsive to pain (P),
and unresponsive (U). The AVPU scale is used in the
initial assessment, but the GCS is recommended in the
secondary assessment.
In modern surgical practice, the chief clinical uses
of the GCS are in the early evaluation of the primary
effects of a head impact, in routine monitoring of less
severe head injuries to detect changes due to complications and in monitoring the progress of recovery.
The relative merits of the 14- and 15-point scales have
received little discussion in these contexts. In 1990, a
questionnaire was sent to senior Australasian neurosurgeons and, of the minority who responded, seven
out of 11 preferred the 14-point scale for clinical use.
However, the 15-point scale is internationally accepted
for research studies and this is a powerful argument
for using it routinely.
It is less easy to determine the status of the GCS
in prognosis. The depth of coma is certainly one
factor in clinical decisions on head injury treatment.
Jennett (1992) has reviewed the circumstances in
which continued treatment of a head-injured person
is futile or disproportionately burdensome and in
this review age and depth of coma stand out as the
chief predictors of death. In an earlier study of 1000
severe head injuries, Jennett et al. (1979) had shown
the prognostic value of GCS summated scores: death
or vegetative survival was the fate of 87% of those
whose best score in the first 24 hours was 3/4,
whereas these bad outcomes were recorded in only
53% of those with scores of 5/6/7. To exclude the
effects of alcohol, hypoxia and other confounding
factors, this study included only cases remaining in
coma for more than 6 hours. If, as is now often the
case, initial resuscitation includes immediate endotracheal intubation and paralysis, the effects of these
confounding factors on conscious level cannot
always be excluded. The ethanol level can and
should be measured; Jagger et al. (1984) have shown

that over a level of 0.20%, ethanol may significantly
depress the GCS score, though there is much individual variation in this effect. But even when allowance
is made for intoxication, the initial GCS level has to
be viewed with reserve as a prognostic factor and, in
determining whether to cease supportive treatment,
the clinician will usually rely on repeated evaluations
of the coma level after temporary cessation of respiratory paralysis and sedation. In establishing the
coma scale, and in subjecting it to rigorous statistical
analysis, Teasdale, Jennett and colleagues made a
very great contribution to clinical neuroscience, and
their work has enduring value. But, as these authors
presciently forecasted, newer methods of therapy
have reduced the availability of GCS scores for
prognostic purposes.
The value of the GCS as a measure of head injury
severity is considered below.
8.2.3
PEDIATRIC COMA SCALES
It is not easy to assess the consciousness level in

infants and young children, and mistakes are often
made. Sometimes the severity of a head impact is
overestimated, but the converse error is much commoner: because an injured infant cries or whimpers, it
is thought to be fully conscious and serious brain
damage may be overlooked.
The verbal and motor responses that indicate full
consciousness in the GCS are obviously not achievable
by preverbal infants. Even after speech is attained, a
frightened but fully conscious child may withhold
speech or cooperation. There have therefore been
many attempts to devise a scale of consciousness
appropriate to the first 5 years of life. These have been
reviewed by Yager, Johnston and Seshia (1990) and
Simpson et al. (1991).
Pediatricians and neurological nurses are well
aware of the subtlety and scope of preverbal responses, and some of the reported scales try to quantify
these. Thus, Seshia, Seshia and Sachdevan (1977)
devised an elaborate grading that tested social, adaptive, vocal and motor responses, and also suck/cough
responses, both spontaneous and stimulus-evoked;
each was given a 04 value. Hahn et al. (1988) devised
a scale using the Glasgow scale for eye opening and
motor responses but with a more complex range of
verbal responses, including subscores for smiling, eye
orientation, consolability and interaction. Simpson
and Reilly (1982) preferred a much simpler system,
directly based on the original GCS but with agerelated norms for the verbal and motor responses. This
scale expresses the concept that the range of responses
in head-injured infants and young children is narrower than is the case over the age of 4 years. Norms
151
Figure 8.2 Norms in the Paediatric Glasgow Coma Scale (PGCS). The expected norms for successive age ranges are set out
on a standard ward chart, which shows the modifications of the best verbal responses used in the PGCS. In teaching the use
of this scale, it is emphasized that actual performance is often better than the expected norms: many children in the 35 year
range will demonstrate awareness of place or personal relations.If necessary, a standard adult scale can be used, but it must
then be emphasized that adult performance is not to be expected and the record should show what responses are actually
elicited.
(Figure 8.2) for best anticipated responses at birth, 6

months, 1 year, 2 years and 35 years are derived from
accepted developmental milestones (Reilly et al.,
1988); actual responses are recorded in ward charts
identical with the 14-point GCS except in the verbal
scale, in which minor changes have been made to
allow a simple grading of preverbal responses (Table
8.4). The Paediatric Glasgow Coma Scale (PGCS),
was independently compared by Yager, Johnston and
Seshia (1990) with five other systems of quantifying
the consciousness level in early life, and found to be
one of the two best from the viewpoint of observer
disagreement (< 0.10). We believe that the PGCS, in its
simplicity and its close resemblance to the GCS, is
well adapted to hospital use, though some instruction
is needed for nurses used to the adult scale. A video
film has been made for this purpose. In Melbourne, G.
Klug (personal communication) has successfully
developed a 15-point version of the PGCS, in which
abnormal limb flexion is recognized as in the adult
scale.
As a prognostic tool, the PGCS has not been fully
tested. In a series of 23 infants and young children
with impaired consciousness, bad outcomes were
recorded in six of seven cases with summated PGCS
scores of 3/4, in three of five in the range 5/6 and in
none of 11 cases in the range 7/8 (Simpson et al., 1991).
At the lower levels, the PGCS is based on observations
identical with those recorded in the adult scale, and it
seems likely that the predictive value is similar. Hofer

(1993) compared outcomes with scores derived from
the standard 15-point adult scale in a sample of 41
children (age range 217 years, mean age 8.8 years)
and found that the lowest GCS scores were strong
predictors of death. However, this study excluded
infants and included only two children less than 4
years old; moreover, it related to observations made 24
hours after operation. There is need for further
research on the prognostic significance of post-traumatic coma in infants.
8.2.4
THE PUPILS
Pupillary size, shape and reactions are routinely

recorded at the initial examination, and routinely
checked at specified intervals thereafter. If the pupillary light reflex is impaired on one side, the consensual light reflex is tested to exclude an optic nerve
lesion. Even iconoclastic clinicians approve these timehallowed practices, but one must nevertheless ask
what may be learned from them.
Pupillary abnormalities may be bilateral or unilateral; they may be present from the time of injury
(Figure 8.3(a)), or may appear after an interval of time.
If the initial examination shows that both pupils are
widely dilated, and if there is no reaction to a strong
light not always available in an emergency room!
then the pathological basis may be an irreparable
152
(a)
(b)
(c)
Figure 8.3 Oculomotor paralysis. A young girl was admitted in coma after a road accident. (a) The left pupil was
dilated (5 mm) and fixed to light. The right pupil was
smaller (3 mm) but varied in diameter; initially there was no
light reaction, but later this pupil reacted sluggishly to light.
(b) 4 weeks later, rotation of the head to the left and right
elicited a small change in the deviation of the right eye; the
left eye did not move (positive horizontal oculocephalic
response in association with left third-nerve paralysis).
(c) Flexion and extension of the head elicited no change in
the position of the eyes (absent vertical oculocephalic
response). Photographs reproduced with permission.
primary midbrain lesion or advanced bilateral transtentorial herniation. There are, however, other causes.
The pupils may be fixed and dilated in the aftermath
of an epileptic fit, or from inadequate cerebral perfusion (Narayan, 1989), or from local trauma to the iris
or its innervation on both sides, or from the use of a
mydriatic to view the fundi, a practice to be prohibited
in the early period after trauma. (Homatropine was
the cause of dilated fixed pupils in an injured
motorcyclist admitted in coma after visiting an
ophthalmologist, whose use of this mydriatic possibly
caused the accident as well as confusing the diagnosis.) In general, the finding of bilateral fixed dilated
pupils soon after injury is a very adverse sign, and the
appearance of this sign after initial normality often
indicates irreversible cerebral compression. However,
the finding must be interpreted in its context and with
regard to other findings.
Bilateral fixed pupils of normal shape and size
may indicate a midbrain lesion; bilateral sluggish
pupils associated with ptosis and impaired upward
gaze are an almost pathognomonic sign of central or
posterior transtentorial herniation. Bilateral optic
nerve injury may give bilateral fixed or sluggish
pupils, sometimes with pupillary escape, and this
should be remembered especially in head injury from
a frontal impact; in such cases, the pupils typically
show spontaneous fluctuation (hippus) in diameter.
Bilateral small pupils, often appearing fixed, are a
classical sign of a pontine lesion. This is a relatively
rare finding in closed head injuries. Large doses of an
opiate give similar appearances; in many intensive
care units, morphia is infused to control the reflex
responses of intubated patients, but in the doses now
used the pupillary reactions are usually well preserved, though the diameters may be small. In deep
barbiturate coma, the pupils become fixed and nonreactive.
Previous neurological disease may be associated
with bilateral or unilateral pupillary abnormalities.
Neurosyphilis was the confusing cause of small fixed
pupils in a workman who fell from a scaffold and
sustained compound skull fractures. The sluggish
tonic pupils of HolmesAdie syndrome could be
misleading, and it should be kept in mind that tendon
areflexia is not an invariable finding in this condition
(Bacon and Smith, 1993).
Unilateral dilatation and loss of light reflex in one
pupil commonly means a third-nerve paralysis, often
accompanied by ptosis and a divergent squint (Figure
8.3(a)). This may be a primary effect of the initial head
impact (Heinze, 1960), as a traction injury of the nerve
or from damage in the skull base or orbit. Delayed
onset of a third-nerve paralysis is of course the
classical sign of lateral transtentorial herniation. In
modern practice this is most often due to an acute
subdural hematoma or massive hemispheric swelling;

extradural hematomas are now commonly diagnosed
before this dangerous complication has developed.
Jones et al. (1993) studied a series of 366 cases of
extradural hemorrhage treated by a single neurosurgical service in a 35-year period and found that before
the advent of CT scanning, pupillary abnormalities
were recorded in 70.4%, but were recorded in only
34.3% of those treated in recent years.
A unilateral fixed pupil of normal or fluctuating
size may be due to an optic nerve lesion; consensual
testing will usually establish this diagnosis. A fixed
pupil of large or normal size may result from trauma
to the iris and/or ciliary body; other signs of a local
impact may or may not be detectable. A more
obvious cause of a unilateral fixed pupil is a prosthetic eye.
Irregularities of pupil shape are not uncommon in
terminal stages of cerebral compression; the mechanism is uncertain. Marshall et al. (1983a) have drawn
attention to the finding of an oval pupil, often
eccentric, as an early sign of transtentorial herniation
(Figure 8.4). An ectopic pupil (corectopia) may also be
the result of ocular injury, sometimes long-standing; in
a recent case of severe head injury, relatives had
known of the abnormal pupil for many years, but had
not been questioned about it until its diagnostic
significance was discussed.
In view of the clinical importance attached to the
pupillary light reflex, it might be expected that its
prognostic significance would be substantial. This is
indeed so in certain categories of head injury, notably
extradural hematomas; in the study by Jones et al.
(1993) quoted above, death or persistent vegetative
states were recorded in 8.3% of cases presenting with
normal pupils, but in 26.9% of those with a pupillary
abnormality (p = 0.0002). In using the pupils as a
153
general prognostic factor for outcome after severe

head impact, difficulties arise from the great diversity
of pupillary abnormalities and their variable significance. Nevertheless, Choi et al. (1988) concluded
that the pupillary light response was one of the three
most accurate predictors of final outcome, the others
being age and GCS motor score. In this analysis, these
authors recognized three grades of pupillary response
on admission: bilaterally absent, unilaterally absent,
and normal. Braakman et al. (1980) used a similar
grading, but selected the best state of the pupils in the
first 24 hours as the prognostic factor: where both
reacted, the mortality rate was 29%, rising to 54%
when only one reacted and to 90% when both were
fixed.
Pupillary size has received less attention. In the
hope of earlier diagnosis of intracranial mass lesions,
Chesnut et al. (1994) considered pupillary inequality,
irrespective of reaction to light, in a series of 608
comatose head injuries. They found that inequality of
more than 1 mm was present in 35% of these
patients; when present, this inequality indicated the
presence of an intracranial mass lesion in only 30%
not always ipsilateral and not always an extracerebral clot. Greater asymmetry (more than 3 mm) was
more often associated with a mass lesion; nevertheless, in more than half the cases with this degree
of asymmetry no such association was found. Commenting on this study, Narayan (1994) drew the
conclusion that pupillary inequality is an unreliable
sign and no substitute for routine early CT scanning
in patients with severe head injury. Nevertheless,
pupillary size and reactivity to light are valuable
signs if taken in context, both in diagnosis and in
prognosis.
Pupillary testing is often made impossible by
orbital swelling, and this emphasizes the importance
(a)
(b)
Figure 8.4 Oval pupil. A young man was admitted in coma after a road accident.On admission, the left pupil was dilated
(7 mm) and fixed to light. The right pupil was smaller (23 mm) and also fixed. (a) 10 weeks after injury, the left pupil was
fixed to light as a result of a third-nerve paralysis. The right eye was deviated down and was abducted. (b) The right pupil
was oval in shape, and reacted sluggishly to light. Photographs reproduced with permission.
154
of accurate early examination and recording before

swelling is established. A slovenly initial examination may miss an optic nerve injury, which may then
be detected days later when the swelling subsides.
Desmarress eyelid retractors sometimes allow exposure of an otherwise inaccessible pupil, but should
be used with great caution to avoid corneal
abrasion.
8.2.5
EYE MOVEMENTS
In the routine neurological examination of the unconscious patient, spontaneous eye movements should be
noted. If there are none, the oculocephalic reflexes are
tested by rotating the head fully in the horizontal and
vertical planes the oculocephalic or dolls eye test
(Figure 8.3(b), (c)). This is done only when a cervical
spinal injury has been excluded by adequate radiographs demonstrating all vertebrae including C7. The
findings relate to the functional integrity of the
midbrain, the pons and the third, fourth and sixth
cranial nerves. Thus, spontaneous roving eyes with
parallel visual axes suggest normal central and
peripheral innervation of the extraocular muscles.
Lesions of the third and sixth cranial nerves show up
as limitation of eye movements effected by the
paralyzed muscles. The fourth nerve is untestable in
the unconscious patient. Forced downward ocular
deviation suggests a midbrain lesion. Absence of
upward movement (vertical oculocephalic reflex) has
the same significance, but may be hard to elicit in a
convincing manner in unconscious patients. Forced
lateral gaze suggests an irritative lesion which may be
in the brain stem or in the supratentorial brain;
absence of lateral gaze may indicate a paralytic lesion
in the same sites. Vertical divergence of the visual axes
(skew deviation) is usually taken to mean a pontine
lesion.
In addition to their localizing value, the eye movements have been considered as indices of head-injury
severity. Visual fixation and tracking are preserved in
relatively mild injuries; the capacity to fix on a target
and to follow it is a favorable finding, and is especially
useful in examining a preverbal infant or an aphasic at
any age. Spontaneous roving eye movements usually
indicate a milder impairment of consciousness, in the
GCS range 7/8 or better. At the other end of the
severity spectrum, absence of eye movements is an
ominous finding. Absence of eye movements on
irrigating the external auditory canal with up to
100 ml ice-cold water (oculovestibular reflex) is indicative of profound brain-stem failure and is one of the
accepted criteria of brain death (Walker, 1985). The test
should not be done if there has been anything to
suggest a cranioaural fistula, such as cerebrospinal
fluid otorrhea, intracranial air or a middle fossa skull-
base fracture. However, between these extremes of

severity, there are many ill-defined disturbances of eye
movement, and it is questionable whether these have
much if any diagnostic or prognostic meaning.
Jennett et al. (1977) constructed a composite score for
eye movements, but this has not been widely used.
From Liège, Born et al. (1982) have reported on a
composite coma scale combining the 15-point GCS
with a five-point reflex scale. In the reflex scale,
cephalic reflexes are ordered in a hierarchy corresponding to their supposed clinical significance as a
measure of rostrocaudal brain-stem function (Table
8.6). The fronto-orbital blink reflex is elicited by a light
tap on the glabella. The pupillary light reflex and the
oculocephalic reflexes are elicited as described above.
If a cervical spinal injury makes head rotation dangerous, the oculovestibular reflexes are obtained by aural
irrigation with iced water bilaterally for vertical eye
movement, unilaterally for horizontal movement. The
oculocardiac reflex is obtained by pressure on the
globe, failure to demonstrate slowing of heart rate
being the lowest level in the 05 reflex scale. The sum
of the GCS score and the reflex score is termed the
GlasgowLiège (GLS) score. Born et al. (1987) have
confirmed good inter-rater agreement for the reflex
scale. In a comparison of the predictive value of the
reflex score with the motor component of the GCS in
severe head injuries, Born (1988) found that, in the
first 24 hours, the reflex score is superior as a
prognostic tool. However, in surviving cases, the
reflex score usually returns to normal within 2 weeks,
even in patients destined to be severely disabled,
whereas the GCS motor score tends to remain low for
longer periods, even in cases showing a favorable
outcome. This interesting composite scale has not
received as much attention as it deserves, perhaps
because two of the reflexes the fronto-orbital and the
vertical oculocephalic reflex are not in general use in
most neurosurgical centers.
8.2.6
FUNDI
Fundal abnormalities are not usually of great importance in the early management of severe head injuries,
and the examination is often difficult: the pupils may
Table 8.6 Liège Reflex Scale. (Source: After Born et al. 1982)
Brain-stem reflexes
Fronto-orbicular
Vertical oculocephalic
Pupillary light
Horizontal oculocephalic
Oculocardiac
No response
Score
5
4
3
2
1
0
be small, there may be orbital swelling, and the use of

corneal lubricants is often a further impediment. There
is however some diagnostic yield from early examination of the fundi. The finding of retinal hemorrhages
may indicate a period of sudden increase in intracranial pressure and massive intraocular bleeding
may be a threat to vision. In infants, retinal hemorrhages are often due to child abuse, though this is by no
means a specific association (Duhaime et al., 1992).
The diagnosis of child abuse is of such importance
that it may be justifiable to use a mydriatic for
fundoscopy in head-injured infants though only
after a CT scan has been done. Early fundoscopy also
gives a baseline against which later possibly abnormal
findings can be evaluated. Papilledema is not a
common finding in severe head injury: Selhorst et al.
(1985) found swollen discs in only 15/426 (3.5%) cases.
When papilledema does occur, it is usually not gross,
and knowledge that the discs were previously normal
can be helpful in deciding on the significance of
indistinct disc margins. This is true also of the late
appearance of optic disc pallor from nerve injury.
8.2.7
LIMB MOVEMENTS AND REFLEXES
Spontaneous and evoked limb movements are studied

as part of the GCS examination. This records the best
motor response; less good responses are also noteworthy, since asymmetrical or localized impairment of
movement may reveal a hemiparesis, monoparesis,
paraparesis or limb fracture.
Muscle tone is assessed after inspection, by putting
the limbs through a full range of passive movement
keeping the possibility of a long bone fracture in mind.
The tendon reflexes have very little value in the
diagnosis of acute cerebral injuries, but localized
absence of tendon jerks may disclose a nerve injury.
The plantar reflexes are usually extensor in severe
head injuries. With a coexisting acute spinal cord
transection, a slow flexor plantar response is often
seen, and this can be a useful corroborative finding in
an unconscious patient. Absence of sweating may then
clinch the diagnosis of cord damage: this is best
detected by running the dorsal surface of the examiners fingers up the body the change from dry to
moist skin is unmistakable.
The motor findings have considerable significance.
The prognostic importance of the GCS motor response
has been generally confirmed, though in young
children the finding of generalized extensor patterns is
not so adverse as in adults (Robertson and Pollard,
1955). A lateralized limb weakness may mean a
contralateral (rarely ipsilateral) intracranial clot, especially if serial records have shown that movements
were previously symmetrical. Other important causes
include cerebral infarction from internal carotid occlu-
155
sion. Bilateral flaccid leg weakness very strongly

suggests a spinal lesion; bilateral leg weakness is
sometimes seen after cranial injury in the vertex
region, but is likely to be spastic, and should not show
absence of sweating. A flaccid arm weakness means a
brachial plexus paralysis until proven otherwise,
though in the acute phase of cerebral injury, a
hemiplegic arm sometimes shows reduced tone and
even depression of tendon reflexes (Russell, 1947).
Abnormalities of the motor examination can only be
reliably inferred in the absence of neuromuscular
paralyzing drugs or sedatives such as barbiturates and
benzodiazepines. Use of a nerve stimulator may be
helpful, if paralytic drugs have been used. Nurses and
residents should avoid repeated painful stimuli for
motor-system testing in patients who are pharmacologically paralyzed or heavily sedated for ICP
control.
8.2.8
CARDIORESPIRATORY PARAMETERS
Pulse rate, blood pressure and respiratory rate have

long been recognized as valuable indicators of raised
intracranial pressure (ICP). In 1901, Theodor Kocher of
Bern clearly associated higher grades of raised ICP.
with slowed pulse and slow, shallow breathing interrupted by deep breaths; thanks to the experimental
work of his brilliant young American protege Harvey
Cushing he was also aware that when raised ICP
compromised the medullary vasomotor centers, the
blood pressure would rise (Kocher, 1901).
These signs of brain-stem failure now have less
diagnostic value in cases of severe head injury, being
seen chiefly as late events after failure of treatment.
Nevertheless, the pulse, blood pressure and respiration should be monitored routinely and in less severe
head injuries may give valuable advance warning of
raised ICP. In children especially, a slowing pulse
sometimes appears before an obvious fall in the
consciousness level. Conversely, tachycardia and a
falling blood pressure may be of great importance in
detecting an extracranial lesion such as a ruptured
spleen or other abdominal or thoracic organ.
The prognostic significance of the cardiocirculatory
parameters received surprisingly little attention in
earlier studies of severe closed head injury (Teasdale
and Jennett, 1976; Jennett and Teasdale, 1977). However, with increasing awareness of the critical importance of cerebral perfusion pressure, the significance of
a low blood pressure has received more attention.
Stening et al. (1986) found records of an arterial blood
pressure below 90 mmHg persisting for more than 60
minutes in 90/290 (31.0%) cases of acute subdural
hematoma, and analysis showed that this was a strong
predictor of bad outcome. These authors saw arterial
hypotension as a preventable cause of bad outcome,
156
and this is indisputable when hypotension is due to

inadequate resuscitation or failure to deal with an
extracranial lesion. But arterial hypotension may also
be seen as a manifestation of terminal brain-stem
failure, and then a bad outcome may be inevitable. Lyle
et al. (1986), in a study of severe head injury, found that
a systolic arterial blood pressure below 90 mmHg was
significantly associated with death on univariant
analysis; however, in this study hypotension correlated
closely with a low GCS score and on multivariant
analysis the only significant variables were the GCS
score and the pupillary light reflexes. Choi et al. (1988)
also found that blood pressure on admission did not
correlate significantly with outcome.
The converse finding of hypertension has received
less attention, though Robertson et al. (1983) have
given prominence to arterial hypertension as an
adverse event in severe head injury. Fearnside et al.
(1993) considered the prognostic significance of paroxysmal arterial hypertension as a clinical variable in a
series of 315 severe head injuries; it was seen in
association with profuse sweating and tachycardia
and did not adversely affect mortality.
Respiratory abnormalities were studied by North
and Jennett (1974) in a series of acute neurological
disorders. They found that bad outcomes were associated with abnormal breathing patterns, and in
particular with tachypnea combined with hyperventilation. With the advent of routine early intubation and control of blood oxygen saturation by pulse
oximetry, spontaneous respiratory patterns have lost
their diagnostic and prognostic significance; the initial
ABC examination always includes an assessment of
the adequacy of breathing, but a detailed description
of the rhythm does not add much to the overall
evaluation of injury severity.
8.2.9
EXTERNAL FINDINGS
In the unconscious acute head injury, concern with the

state of neurological function may overshadow the
rest of the clinical examination. It is easy to omit an
examination for external signs of injury. The scalp and
face should be carefully inspected and palpated.
Wounds, abrasions and swellings should be recorded
on a diagram and also photographed if there are
medicolegal implications. If for any reason CT scanning is not possible, the site of the primary impact is
an important guide to the siting of an emergency
craniotomy or burrhole exploration. The impact site(s)
may also be important in future litigation, for example
in determining whether a helmet might have given
effective protection, or in a victim of child abuse or
other form of criminal assault.
Orbital swelling has obvious diagnostic importance
as a sign of anterior fossa fracture; auscultation may
detect a carotidcavernous fistula. Bruising behind the

ear (Battles sign) is a well-known sign of a fractured
petrous bone. However, if bruising is seen immediately after injury, it may point to a local impact. If
the bruising appears after a time interval, it is likely to
be due to a fracture of the skull base.
Blood or cerebrospinal fluid (CSF) leakage from the
nose or ear should be noted and a sample of fluid
(> 0.1 ml) should be sent for immunochemical analysis
for 2-transferrin as a marker of the presence of CSF
(Ryall, Peacock and Simpson, 1992).
8.3
The definitive examination
8.3.1
TIMING
This depends on the speed and degree of recovery.

Ideally, a full neurological examination is done when
the patient is conscious, cooperative and fully oriented. Since recovery from a severe head injury is
usually slow and often incomplete, more or less
selective serial neurological examinations are usually
done at intervals before full cooperation is complete,
and when the patient is still in the phase of posttraumatic amnesia (see below). These progress examinations can be very informative. It is very important
in planning and assessing rehabilitation, and also for
medicolegal purposes, to ensure that all residual
neurological disabilities are documented before the
patient goes home or is transferred to another hospital. In particular, residual amnesia should be assessed
and the senses of sight, smell and hearing should be
tested. Therefore, a definitive neurological examination is mandatory before discharge or transfer.
8.3.2
ORGANIZATION
Who does the definitive examination? In many centers,

the neurosurgeon is a surgical neurologist and examines the patient in person or delegates to a properly
trained proxy. In some centers, a medical neurologist
may be consulted. Increasingly and beneficially, parts
of the examination are now subcontracted to a
neuropsychologist, neuro-ophthalmologist, neurootologist or specialist in neurorehabilitation; pediatric
neurosurgeons may consult a specialist in developmental neurology. Whatever the division of labor,
there needs to be a final common synthesis and
evaluation, preferably made by a single person with
final clinical responsibility.
8.3.3
ORIENTATION AND AMNESIA
Orientation is always confirmed as part of the

definitive examination. As a minimum, the patient
should be asked to name the day of the week, date and
THE DEFINITIVE EXAMINATION
place. The patient should also be asked to describe the

last event recalled before the injury and the first event
recalled after the injury. These questions should define
and quantify the periods of retrograde amnesia (RA)
and post-traumatic amnesia (PTA). In patients emerging from a prolonged amnesic state, one of the amnesia
questionnaires described below can be used.
8.3.4
SPEECH, MENTAL STATE AND COGNITION
Speech is assessed in conversation, both for impaired

articulation and for fluency and thought content.
Minor degrees of dysarthria can be brought out by
tongue-twisting words, such as hopping hippopotamus. A tape recording of speech may be made for
future comparison.
The preferred hand is recorded, though it must be
remembered that many left-handed persons have full
or partial left-hemisphere dominance. If there is reason
to suspect injury to the dominant hemisphere, the
patient is asked to name a series of test objects of
increasing complexity: for adults, the final challenge
can be the parts of a watch and for children a toy. This
test for nominal dysphasia can also be used to test
recent visual memory, by asking the patient to recall as
many as possible of the objects immediately after the
naming test; 8/10 is a good score. Other standard tests
of memory include recall of a name, an address and a
flower after 5 minutes, digit retention forwards and
backwards, and timed serial subtraction of 7 from 100.
Tests of cognition are best done by a neuropsychologist; Walsh (1985) and Wood and Woodroffe (1995)
have reported on tests found useful in head injury
practice. If the services of a neuropsychologist are not
available, there are many simple tests that require little
special expertise; literacy can be checked with an agegraded word list and non-verbal intelligence can be
assessed by the Raven colored matrix test (Raven,
1986), which also probes function in the non-dominant
parietal lobe.
The emotional state and the degree of insight are
noted in conversation and in discussions of future
plans for rehabilitation.
For head-injured children, cognitive capacities and
mental attitudes can be assessed in play and games
appropriate to the age; a childs attempts to draw a
man are very informative, and so are games based on
family relationships. At some stage, a formal developmental assessment (Griffiths, 1970) should be done;
this requires considerable pediatric experience.
8.3.5
VISION
This is always assessed in severe head injuries, though

the depth and scope of the assessment vary with the
nature of the injury. Covering one eye may bring out
157
subjective visual loss or blurring. The Snellen and/or

reading test types should always be used in the
definitive examination of a severe head injury. For
illiterates, the E tests can be used; alternatively, the
STYCAR toys provide a simple and quick way of
estimating visual acuity (Sheridan, 1976). Peripheral
visual fields are tested by confrontation with the
examiners fingers as stimulus. Central fields can be
tested with a small white or red object, such as the
head of a mapping pin, or a bead on a black stick. It is
also possible to assess the central fields very effectively by asking the patient to fix on the examiners
nose and to say if any feature is missing or blurred. If
there is a field defect, formal perimetry is done
usually by a neuro-ophthalmologist. The optic fundi
are always examined; as noted above, a mydriatic
should not be used in the early period after a head
injury, but may be used later when definitive fundoscopy is done.
8.3.6
SMELL AND TASTE
Olfaction must always be tested. Tar or phenol is a

good strong test odor, but should be complemented by
a milder odor such as banana or raspberry, or cloves.
Each nostril is tested separately and, to exclude
guessing, the patient is warned that the test bottle may
be empty. A more objective system of smell testing is
provided by the University of Pennsylvania Identification Test (Doty et al., 1984), a quantitative smell test
that permits the examiner to determine whether there
is normal olfaction, microsmia, anosmia or
malingering.
Taste is rarely of importance, but may be tested
when there is a facial paralysis, by dropping strong
syrup or salt solution on each side of the tongue;
electrical tests of taste are not always reliable, as the
patient may report tingling as a taste.
8.3.7
HEARING
This is tested with particular care when there is

evidence of a skull-base fracture. A simple check is
done by whispering words or numbers into each ear,
hearing by the opposite ear being masked by gentle
circular rubbing with a finger tip pressed on the
opposite tragus to occlude the external auditory
meatus. If deafness is found, a 1024 or 512 Hz tuning
fork is used to distinguish inner- and middle-ear
deafness by the Weber and Rinne tests. The external
auditory canals are examined with an otoscope and
the color of the ear drum is noted. An otologist should
be consulted if hearing is impaired, or if there is a
hemotympanum. An audiogram should also be performed, both to aid prognosis for recovery and for
medicolegal reasons.
158
8.3.8
CRANIAL NERVES
The eye movements are tested in lateral, vertical and

oblique planes, and note is made of ptosis, diplopia,
squint or nystagmus.
The trigeminal nerve is tested in head injuries
associated with facial or skull-base fractures. The
corneal reflex is tested with a wisp of cotton wool.
Mild trigeminal hypesthesia may be brought out by
stroking parts of the face and asking the patient if
there is any qualitative difference. Differences in
pinprick perception may be checked, especially if
there has been a facial injury; a blunted disposable
needle is used. Two-point testing on the lips is
sometimes useful.
The facial nerve is tested by asking the patient to
screw up the eyes, whistle or smile; an emotional
facial weakness of upper motor neuron type may be
brought out by watching the spontaneous smile. If
there is any facial weakness, it is most important to
know whether it is of late onset. In unconscious
patients, unless in deep coma, facial movements can
be elicited by strong pressure on the supraorbital
nerve or mandibular ramus. Schirmers test of lacrimation is very useful in cases of peripheral facial
paralysis: thin strips of filter paper are placed in the
conjunctival sac for 30 seconds or more, and the extent
of saturation is measured (Trott and Cooter, 1995).
The lower cranial nerves (nerves IXXII) are tested
by examining the movements of the palate, pharynx,
tongue, trapezius and sternomastoid muscles.
Peripheral paralyses of these nerves are occasionally
seen after closed head injuries. Much commoner are
dysphagias and dysarthrias from brain-stem damage.
For these an assessment by a speech pathologist is
necessary and a radiological swallow study is
advisable.
8.3.9
SENSORIMOTOR LIMB FUNCTIONS
Limb function is tested with respect to muscle tone,

power against graded resistance, and coordination;
ataxia is a very common sequel in severe closed head
injuries, presumably from injuries of the superior
cerebellar peduncle. Quantitative measures of limb
function are desirable if cooperation is good. In most
units, these are done by physiotherapists and/or
occupational therapists, but their value and limitations should be understood by all clinicians concerned
in head injury evaluation. Dynamometry should be
used to give an objective measure of hand grip
(Mathiowetz, 1990); hand-held dynamometers can be
used to test other muscle groups. Manual facility and
coordination can be quantified by finger tapping or by
pegboard tests, and for these consultation with an
occupational therapist is advisable. Wood and Hammerton (1995) have reported on the Purdue pegboard
test in evaluating head injury, both in adults and in

children over the age of 7 years. This test was devised
for selecting industrial workers (Tiffin, 1968); Gardner
and Broman (1982) found it to be an excellent test of
minimal brain damage. In the simplest form of the
test, the subject inserts metal pegs in a row of holes in
a standard board as rapidly as possible. A score is
obtained for each hand over a period of 30 seconds.
The reproducibility of pegboard testing is impressive,
and the test is of value as a guide in determining when
a stable level of recovery has been attained. Other
quantitative tests of handarm function are described
by Gloss and Wardle (1982). If the patient can walk,
the gait is described, and the ability to hop on either
leg is noted; this is a good quick test of lower-limb
motor competence, though limb or spinal injury may
falsify the interpretation. In patients unable to walk,
the degree of mobility in bed or in a wheelchair is
recorded. The progress of motor recovery can be
documented by video.
The tendon and plantar reflexes are again tested,
and with more attention: persisting reflex abnormalities have much more significance than the evanescent reflex changes seen in the acute phase.
Sensation, except in the trigeminal area, is rarely
affected in closed head injuries, but occasionally one
sees what appears to be a spinothalamic sensory loss
in cases of brain-stem damage. A full sensory examination is needed if there is an associated spinal injury
or suspicion of a lesion in the parietal lobe or basal
ganglia.
8.4
Evaluation of injury severity
8.4.1
PROSPECTIVE GRADING
Estimates of head-injury severity may be made prospectively, as aids in triage, prognosis and family
counseling. For these purposes, the estimate can take
into account many factors and nuances; most clinicians will admit that intuition enters assessments done
for prognosis. But when the estimate is done for
statistical purposes, as in therapeutic trials, the criteria
should be as few as possible, and they should be based
on observations that have good inter-rater reliability.
The consciousness level, assessed at a specified
period after injury, has been widely used in definitions
of a severe head injury, both for prognosis and for
research purposes. In most reports, the chief criterion
of severity is a GCS score of 8 or less. For the US
National Coma Data Bank the definition for inclusion as a severe head injury is:

GCS score of 8 or less following resuscitation,

which may include endotracheal intubation; or
GCS score deteriorating to 8 or less within 48 hours
of injury (Marshall et al., 1983b).
EVALUATION OF INJURY SEVERITY
In this definition the use of endotracheal intubation

could reduce the best verbal score to 1. In theory this
might result in the inclusion of less deeply unconscious patients, and it is reasonable to use the best
motor score to control the reliability of the GCS
summated score. If this is done, the six-level Glasgow
motor scale should be used. Data from the six-level
motor scale should be routinely recorded for audit and
research studies.
This definition of head-injury severity is widely
accepted, and is being used as a basis for inclusion in
therapeutic trials. Thus, in a well designed phase 2
trial of an oxygen radical scavenger, Muizelaar et al.
(1993) accepted for entry into the trial cases with a
GCS score of 8 or less who were unable to follow
commands after resuscitation; the time at which the
GCS score was estimated was not specified. The
criteria for exclusion included the likelihood of brain
death after resuscitation; presumably this exclusion
would remove cases with GCS score of 3 and other
adverse signs.
The GCS has also been used to stratify cases within
a more broadly inclusive trial. In a randomized trial of
nimodipine therapy for head injury, Bailey et al. (1991)
accepted all patients who were unable to obey
commands, thus including GCS scores as high as 13.
For statistical analysis, the best motor response was
used, the scale being collapsed into three classes nil
or extending; flexing; localizing. (These classes were
further subdivided by the presence or absence of an
intracranial lesion requiring operation.) This trial
protocol shows the flexibility of the GCS in stratifying
injury severity and it seems likely that, in further trials
of neuroprotective agents, components of the GCS will
be used in various ways, depending on the level of
severity at which the agent is expected to be beneficial
and the size of the series to be analyzed.
8.4.2
RETROSPECTIVE GRADING: COMA DURATION
Head injury severity may also be assessed retrospectively, for epidemiological and other research
studies, especially in correlation with measures of
outcome (see below). The duration of impaired consciousness has been much used as a retrospective
measure of injury severity. In contemporary neurosurgical practice, this is commonly done in two
different ways.
The duration of coma can be measured on the basis
of serial clinical observations of responsiveness. Thus,
Bricolo, Turazzi and Feriotti (1980), in a very thoughtful study, reported on 135 cases who were in coma 14
days after head injury, coma being defined as unresponsive . . . or incapable of obeying simple commands or showing any rapport with their environment. Outcomes were assessed at 1, 3, 6 and 12
159
months by the Glasgow Outcome Scale (Jennett and

Bond, 1975): by the categorization given in that scale,
13.3% made good recoveries, 17.7% were left with
moderate disabilities, 31.1% were left with severe
disability, 8.1% remained in persistent vegetative
states (Jennett and Plum, 1972) and 29.6% were dead.
This study illustrates both the usefulness and the
complexity of the duration of coma as a measure of
unconsciousness. Recovery from coma was identified
in terms of the three components of the GCS.
Spontaneous or evoked eye opening appeared after 1
month in 76.3%. Response to commands came later,
and was achieved after 3 months in only 52%. Speech
restoration was achieved by 3 months in only a third
of cases, rising to 51% by the end of the study. There
was a correlation between duration of coma and
quality of final outcome. In this study, each measure of
responsiveness was separately correlated with quality
of recovery. In theory, it should always be possible to
do this if accurate GCS records are kept, but many
reports on the prognosis of traumatic coma have made
the simpler distinction between comatose and not
comatose on the basis of the summated GCS score.
Lyle et al. (1986), who did this, noted that the GCS
score is a relatively insensitive measure of recovery,
since the early return of spontaneous eye opening has
little prognostic value. The GCS descriptors can be
combined with terms such as akinetic mutism, apallic
state, and persistent vegetative state, but the significance of these terms is not always clear and they
can be dangerous labels if given a prognostic importance during the first few months after injury.
8.4.3
RETROSPECTIVE GRADING: AMNESIA
The other widely used measure of duration of

impaired consciousness is the period of PTA. Ritchie
Russell, in his pioneering study of the neurology of
head injury (Russell, 1932), argued that the duration of
unconsciousness is best estimated as the period before
return of memory. He found that the return of
memory could be timed by the patients recollection of
when he woke up. Russell believed that this wake-up
time could be estimated with fair accuracy long after
the accident. On this basis, he graded surviving cases
into three groups those unconscious for less than 1
hour, those unconscious for 124 hours and those
unconscious for longer periods. With later experience,
Russell became aware that the first clear recollection
could be followed by a further period of amnesia, and
the PTA was therefore measured by the return of
continuous memory (Russell and Nathan, 1946). The
PTA was correlated with return to full wartime duties
after head injury and appeared to be a robust
prognosticator. Russell and his colleagues also studied
the period of amnesia before the injury, retrograde
160
amnesia (RA); this was found to have less significance

as a measure of injury severity. The RA is still
generally recorded, as it has some diagnostic value;
RA is often important in medicolegal issues, both as
confirmation of a cerebral insult and because it
obviously affects the victims capacity as an accident
witness. But the duration of RA often shrinks with the
passage of time; Richardson (1990), reviewing the
abundant literature, concluded that the RA has no
practical value as an indicator of injury severity or in
prognosis.
There has been general agreement that the PTA is a
very valuable measure, especially for less severe
injuries, but there has been doubt as to the reliability
of amnesia endpoints ascertained by simple retrospective questioning. In the first place, the period of
PTA may be interrupted by islands of recollection;
Gronwall and Wrightson (1980) found such islands in
26 (39%) of 67 minor head injuries. Awareness of this
phenomenon led to the definition of the PTA as the
period of continuous memory loss after injury (see
above), but this also has proved to be hard to define by
retrospective interrogation. Much of Russells very
productive work was done on British soldiers transferred to Oxford for assessment and rehabilitation. In
their passages from accident site or battlefield to
evaluation, they had usually experienced a series of
well-defined and well-documented events that made
retrospective estimation of return of continuous memory easier than in cases where all treatment has been
undertaken in a single institution. Furthermore, in
Russells earlier reports the proportion of cases with
very prolonged periods of amnesia was not high
(Russell and Nathan, 1946; Russell, 1954). In later
studies on a larger database, Russell accepted that
factors other than the severity of injury influenced the
PTA, notably the presence of focal brain lesions, severe
associated extracranial injuries, and the age of the
injured person (Russell and Smith, 1961; Richardson,
1990). Present-day practice is especially concerned
with cases slowly emerging from prolonged coma or
confusional states, and retrospective interrogation is
often done after the patient has been (quite rightly)
briefed by family members on the course of events: the
patient may then confuse what is remembered with
what has been told, giving a falsely short PTA. This is
especially likely to happen in children, and retrospective PTA measurements in children under the age
of 810 years are very unreliable. On the other hand, a
retrospective PTA evaluation done weeks or months
later may give a falsely long measure, since the patient
may have forgotten some landmark event or may have
become confused between true recollections and
second-hand information. For these and other reasons,
efforts have been made to determine the PTA by more
reliable means.
Richardson (1990) and Forrester and Geffen (1996)

have reviewed the development of prospective measurements of the PTA prospective in the sense that the
aim is to detect by ongoing assessments the time at
which the return of continuous memory can be
demonstrated objectively. These tests embody standardized questionnaires, which are presented to the
patient at regular (usually daily) intervals until the
answers are considered to indicate that the patient has
emerged from PTA. The best known is the Galveston
Orientation and Amnesia Test (GOAT) designed by
Levin, ODonnell and Grossman (1979). This tests
orientation in considerable detail (Table 8.7), allotting
error points for disorientation. The test also gives error
points for PTA and RA; for each, five error points are
deducted for inability to recall a verifiable, or at least
plausible event before or after injury, and an additional five error points when the patient cannot give
details of this event. A final score is made by
subtracting the sum of the error points from 100; a
score of 75 or more is said to be within normal limits.
The GOAT is open to the obvious objection that this
normal score is in theory obtainable when a patient is
Table 8.7 Galveston Orientation and Amnesia test

(Source: after Levin, ODonnell and Grossman, 1982)
Questions
What is your name?
Where do you live?
Where were you born?
Where are you now?
City
Hospital (need not be correctly named)
On what date were you admitted to this
hospital?
How did you get here?
What is the first event you remember after
the injury?
Describe event in detail, e.g. date, time,
companions
What is the last event you remember before
the injury?
Describe event in detail, e.g. date, time,
companions
What time is it now? (1 error point per half
hour removed from correct time)
What day of week is it now? (1 error point
per day removed from correct day)
What day of month is it now? (1 error point
per day removed from correct day to
maximum of 5)
What is the month? (5 error points per
month removed from correct month to
maximum of 15)
What is the year? (10 error points per year
removed from correct year to maximum of
30)
Maximum no.
error points
2
4
4
5
5
5
5
5
5
5
5
5
3
5
15
30
161
fully oriented but still in an amnesic state; Gronwall

and Wrightson (1980) found that a head-injured
person may be oriented but amnesic, or vice versa.
Nevertheless, Levin, ODonnell and Grossman (1979)
found that GOAT scores correlated well with the
duration of GCS impairment and with the final
outcome.
Ewing-Cobbs et al. (1990) have devised a pediatric
version of the GOAT, the Childrens Orientation and
Amnesia Test (COAT). This does not attempt to test
the PTA in any way, but does include a quantitative
evaluation of temporal orientation, using the classic
test of forward digit retention and the more modish
capacity to recall television programs. The COAT is
applicable in full to children in the age range 815,
and without the memory tests in children as young as
3 years. Baryza and Haley (1994) have used the COAT
as a screening test for otherwise undetected impairments in memory and orientation in children who
appear to have recovered from head injury, and this
may be its most valuable application.
Two leading Australian neurorehabilitation units
have designed simple questionnaires for evaluation of
the PTA. From Sydney, Shores et al. (1986) have
reported on the Westmead PTA scale. This embodies
two biographical questions, five questions related to
time date and place and three questions assessing
recollection of pictures of objects (Table 8.8). The
ability to remember the examiners face and name are
also tested. The patient is deemed to be out of the PTA
when able to give correct answers in all components of
the scale on three consecutive days. Shores (1989)
compared this scale with the duration of coma
measured with the GCS and concluded that the
Westmead scale was a better predictor of outcome.
Haslam et al. (1994) have further explored the relation
between long-term cognitive impairments and PTA
established with the Westmead scale. These authors
have reported on a new variable, the post-coma
disturbance (PCD). This is the period of confusion
after emergence from coma and is derived by subtracting the duration of coma from the duration of PTA. In
this study, it appeared that the PCD was a significant
predictor of impairment in recent memory 12 months
after injury, whereas the PTA better predicted poor
performance in information processing. With both the
PCD and the PTA, the relations between duration and
cognitive impairment were non-linear.
Forrester and Geffen (1996) have criticized the
Westmead scale on practical grounds, and advocate
the scale used in the Julia Farr Centre, Adelaide. In this
scale (Table 8.9), the questionnaire has six orientation
items and five memory items. In the memory items,
the patient is asked to memorize the name attached to
a photograph, a gesture, and the names of three
objects shown in photographs. Memory is not tested
Table 8.8 Westmead PTA Scale questionnaire this is

presented daily until the patient achieves a perfect score of
12 on three successive days; the PTA is deemed to have
ended on the first of the three days (after Shores et al., 1986)
Questions
Maximum
No. points
1. How old are you?
2. What is your date of birth?
3. What month are we in?
4. What time of day is it?
5. What day of the week is it?
6. What year are we in?
7. What is the name of this place? If the

patient does not know, a multiple
choice is given home, name of
hospital, name of another hospital
8. The patient is asked to remember the

examiners face. On the following day,
he/she is shown three photographs, one
of the examiner, and asked to identify
the examiner
9. The patient is asked to remember the

examiners first name. On the following
day, he/she is asked to recall this
name; if unable to do so, he/she is
asked to select the name from a series
including this name and two
phonologically similar names or names
with an equal number of syllables
10. Pictures I, II and III: the patient is

shown three colored pictures of
common objects and asked to name
them. On the following day, the patient
is asked to name the pictures. If unable
to do so, he/she is asked to identify
these pictures in a series of 12 pictures,
being a random assortment of the three
original pictures and nine distractor
pictures
Total
12
until full orientation is confirmed. In this test also,

PTA is deemed to be ended when the patient scores
correctly in all orientation and memory tests on three
consecutive days.
G. Geffen (personal communication) finds that the
GOAT and Julia Farr Centre PTA scale correlate
closely, and the choice of scale may be a matter of unit
preference rather than theoretical advantage.
8.4.4
APPLICATIONS OF CLINICAL EVALUATION
It seems that for most clinical trials, severe head

injuries are best defined prospectively by the GCS
score, either summated or as the best motor score,
162
Table 8.9 Julia Farr Centre Post-traumatic Amnesia Scale. Orientation tests: These include four autobiographical
questions, one question on time of day (cf. Question 4 in Westmead Scale) and one on place (cf. Question 7 in Westmead
Scale). Orientation is tested daily until a score of 6 or more is achieved on three successive days. Memory tests: When
oriented in person, time and place, the patients is taught to memorize a gesture, the name of a person seen in a
black-and-white photograph and three objects seen in black-and-white photographs a cup, a comb and an umbrella.
On the following and successive days, the patient is asked to recall these, first freely and if unable to do so, after cued
prompting by showing the test gesture or photograph together with a distractor. When the patient is oriented and
achieves a minimum memory score of 5 (gesture 1, name 1, picture 1) on three successive days, the PTA endpoint
is recorded for the first of the three days. (Source: from Forrester and Geffen, 1996)
Orientation tests
Questions
Answers
Score
a. Personal orientation:
1. What is your name?
No answer or wrong answer

Correct answer
0
1
2. Are you married/Do you live with a partner?

Correct answer
0
1
3. Do you have any children?

Correct answer
0
1
4. What is your job?

Correct answer
0
1

Prompted correct answer
Unprompted correct answer
0
1
2

Prompted correct answer
Unprompted correct answer
0
1
2
Maximum score
b. Orientation in time:
5. What time of day is it?
(Is it morning, afternoon or night?)
c. Orientation in place:
6. Where are we now?
(Are we at home, in a hospital or a hotel?)
Orientation daily total

Memory tests
Maximum
score
Test
Response
1. Gesture
Free recall
Cued recall
Recognition
3
2
1
2. Name of person in photograph
Free recall
Cued recall
Recognition
3
2
1
3. Pictures (i)
Free recall
Recall after cue
Recall when shown and reject the distractor
3
2
1
(ii)
Free recall
Recall after cue
3
2
1
(iii)
Free recall
Recall after cue
3
2
1
Picture total
Memory total
(i) + (ii) + (iii)
9
15
determined after resuscitation. For prognostic purposes, the GCS score can be strengthened by taking
into account other predictors, notably age, arterial
blood pressure and the absence of pupillary light
reflexes on one or both sides. Choi et al. (1988) have
tried to refine the prognosis after severe head injury
by preparing three graphs relating outcome to motor
score, pupils, and age.
The prognostic importance of increasing age is
now well documented. Jennett and Teasdale (1981)
studied outcomes in severely head-injured persons
who remained in coma for at least six hours. There
was a linear relationship between age and bad
outcome (death or vegetative survival) and over 70
years there were no good recoveries. Luerssen,
Klauber and Marshall (1988) also found an increase
in mortality rates with advancing age; for comatose
(GCS < 8) patients, the death rate rose steeply in the
4549-year age group, and remained between 60%
and 80% thereafter. Conversely, comatose children
over the age of 5 years tend to have better outcomes.
Infants and young children do not show this favorable tendency, but the difficulties in grading coma
below the age of 5 years make it necessary to be
cautious in using the consciousness level for prognoses in this age group (Simpson et al., 1991), and
especially in extrapolating adult experience to the
prognosis in injured infants.
For retrospective classification of head-injury severity, it is still debatable whether duration of coma or
length of PTA is the better yardstick. Much effort has
gone into this debate, and some of this represents
quests for precision in a field where in reality precision
is impossible. For patients who emerge from coma to
become responsive and cooperative, the time of
recovery from confusion and amnesia has pragmatic
importance in rehabilitation and prognosis and a
simple standardized endpoint test is desirable; to
establish this, a questionnaire is certainly useful.
Wilson et al. (1993) have emphasized that the PTA is of
value as a measure of injury severity, even when
ascertained by the traditional method of retrospective
questioning; when compared with coma duration, the
PTA correlated better with lesion severity measured in
MRI scans. Nevertheless, duration of coma is also
important. There appears to be agreement that coma
persisting after 14 days is a very adverse finding and
usually predicts a severe disability. Clinical records
should be maintained to ensure that this period is
well documented for future reference. It is also
desirable to record separately the return of eye
opening, responsiveness to commands and capacity to
communicate.
This book is concerned only with severe head
injuries, but it should be noted that Rimel et al. (1981,
1982) have used the GCS within 1 hour of admission
163
to identify minor (GCS score 1315) and moderate

(GCS score 912) head injuries. This has an attractive
simplicity, but Johnstone et al. (1993) found poor interrater consistency in defining a moderate head injury
on the basis of a coma scale, especially when used by
a relatively inexperienced observer. For less severe
head injuries, PTA estimated at discharge seems to be
a preferable criterion. When an exact definition of a
relatively short period of PTA is needed for research
purposes, it may be advisable to administer a simple
orientation questionnaire at short intervals, taking
accurate answers to a standard set of questions as the
endpoint (Gronwall and Wrightson, 1980).
8.4.5
CLINICAL EXAMINATION AND OUTCOME
Outcome grading is discussed in Chapter 22. The

assessment of outcome requires a synthesis of medical
and social data, and the whole spectrum of posttraumatic neurological disability, unique in each individual, has to be taken into account. An appropriate
clinical assessment is therefore an essential part of
outcome evaluation. The purpose of the assessment
determines its depth and scope.
Epidemiological outcome studies require broad
categorizations based on social functional evaluations;
these categorizations are discussed in Chapter 22.
Rehabilitation requires an atomistic analysis of
function, repeated over time. A comprehensive neurological examination is an essential prelude to any
rehabilitation program. A preliminary and selective
neuropsychological evaluation is also desirable, and
when a planned program has been concluded, a full
reassessment is essential. It is often desirable to repeat
such assessments at scheduled times. Victims of head
injury often express resentment that rehabilitation has
been discontinued too soon, and this may in part stem
from failure to promise a future review of progress
and perhaps a further cycle of rehabilitation.
In the USA, Australia and many other countries, the
legal system requires definitive or interim medical
assessments of disability. These have to be both
atomistic and holistic. In assessing the outcome of a
severe head injury for medicolegal purposes, the
impact of the cerebral injury must be evaluated in
terms of physical, cognitive and behavioral effects,
but also in its effects on the victims social status and
quality of life.
Table 8.10 sets out the chief complaints voiced after
head injuries; each requires detailed analysis and
objective verification where possible. It should be
noted that the causes of some of these complaints
include non-neurological injuries of the facial skeleton,
the facial viscera or even extracranial structures.
Efforts are being made to formulate the effects of
injury in percentages of total impairment (American
164
Table 8.10 Medicolegal checklist the first column lists

the chief complaints made after severe head injury (the list
is not exhaustive), the second sets out the chief
impairments or other conditions usually associated with
each complaint
Complaint
Common causes
Personality change
Cerebral damage, especially

frontal lobe damage
Loss in lifestyle, depression
Poor memory, poor

concentration, reduced
intelligence
Cerebral damage
Depression
Speech defect
Cerebral damage
Cerebellar damage
Cranial nerve injury
Blackouts, giddiness
and funny turns
Epilepsy
Vasovagal attacks
Postural vertigo
Headache, facial pain,

scalp pain
Tension states
Nerve injury
Migraine (rare)
Loss of smell/taste
Olfactory nerve injury
Deafness, noise in ears
Injury to ear and/or auditory

nerve
Cranial bruit
Visual loss
Injury to eye and/or visual

pathways
Double vision
Injury to cranial nerves III, IV, VI

Eye injury (rare)
Impaired swallowing
and/or chewing
Injury of lower cranial nerves

Maxillofacial injury
Injury to dentition
Limb weakness,
tremor, unsteadiness,
gait change
Cerebral injury affecting motor

or sensorimotor pathways
Cerebellar injury
Spinal or limb injury
Incontinence and/or
impotence
Spinal injury; cerebral injury

especially frontal lobe injury
Disfigurement
Scar, loss of hair, cranial or facial

deformity, eye injury
Medical Association, 1993). These formulations are

difficult to apply in a realistic way when the cognitive
and behavioral outcomes of head injury are under
consideration, but can be meaningful if given as broad
assessments of social incapacity.
The purpose of outcome evaluation also determines
the timing of the final examination. For research
purposes, a relatively early evaluation may be acceptable. Choi et al. (1994) have argued for an assessment
of outcome at 6 months after injury. For medicolegal
evaluations, a later period is necessary, both because
much functional improvement may become evident
after 6 and even 12 months, and also because all
parties should be satisfied that a plateau in recovery

has been reached. As a rule, 2 years or more should
elapse. In children, the final examination is usually
deferred until adolescence, so that the impact of the
head injury on the childs educational experience can
be assessed and quantified by neuropsychological
tests. Such a prolonged deferral may have adverse
financial consequences for the childs parents, and
there should be legal provision for interim evaluation
at an earlier date, if the childs upbringing is in any
way dependent on a monetary settlement.
8.5
References
American College of Surgeons Committee on Trauma (1993) Advanced Trauma

Life Support Program for Physicians. Instructor manual, 5th edn, American
College of Surgeons, Chicago, IL.
American Medical Association (1993) Guides to the Evaluation of Permanent
Impairment, 4th edn, American Medical Association, Chicago, IL.
Bacon, P. J. and Smith, S. E. (1993) Cardiovascular and sweating dysfunction
in patients with HolmesAdie syndrome. Journal of Neurology, Neurosurgery
Bailey, I., Bell, A., Gray, J. et al. (1991) A trial of the effect of nimodipine on
outcome after head injury. Acta Neurochirurgica (Vienna), 110, 97105.
Baryza, M. J. and Haley, S. M. (1994) Use of the Childrens Orientation and
Amnesia Test at hospital discharge for children with neurological and nonneurological traumatic injuries. Brain Injury, 8, 167173.
Born, J. D. (1988) The GlasgowLiège scale. Prognostic value and evolution of
motor response and brain stem reflexes after severe head injury. Acta
Neurochirurgica (Vienna), 91, 111.
Born, J. D., Hans, P., Dexters, G. et al. (1982) Evaluation pratique du
dysfonctionnement encephalique chez le traumatise cranien. Neurochirurgie,
28, 17.
Born, J. D., Hans, P., Albert, A. and Bonnal, J. (1987) Inter-observer agreement
in assessment of motor response and brainstem reflexes. Neurosurgery, 20,
513517.
Braakman, R., Gelpke, G. J., Habbema, J. D. F. et al. (1980) Systematic selection
of prognostic features in patients with severe head injury. Neurosurgery, 6,
362370.
Bricolo, A., Turazzi, S. and Feriotti, G. (1980) Prolonged posttraumatic
unconsciousness. Therapeutic assets and liabilities. Journal of Neurosurgery,
52, 625634.
Chesnut, R. M., Gautille, T., Blunt, B. A. et al. (1994) The localizing value of
asymmetry in pupillary size in severe head injury: relation to lesion type
and location. Neurosurgery, 34, 840845.
Choi, S. C., Narayan, R. K., Anderson, R. L. and Ward, J. D. (1988) Enhanced
specificity of prognosis in severe head injury. Journal of Neurosurgery, 69,
381385.
Choi, S. C., Barnes, T. Y., Bullock, R. et al. (1994) Temporal profile of outcomes
in severe head injury. Journal of Neurosurgery, 81, 169173.
Doty, R. L., Shaman, P., Kimmelman, C. P. and Dann, M. S. (1984) University
of Pennsylvania Smell Identification Test: a rapid quantitative olfactory
function test for the clinic. Laryngoscope, 94, 176178.
Duhaime, A.-C., Alario, A. J., Lewander, W. J. et al. (1992) Head injury in very
young children: mechanisms, injury types, and ophthalmologic findings in
100 hospitalized patients younger than 2 years of age. Pediatrics, 90,
179185.
Ewing-Cobbs, L., Levin, H. S., Fletcher, J. M. et al. (1990) The Childrens
Orientation and Amnesia Test: relationship to severity of acute head injury
and to recovery of memory. Neurosurgery, 27, 683691.
Fearnside, M. R., Cook, R. J., McDougall, P. and McNeil, R. J. (1993) The
Westmead Head Injury Project outcome in severe head injury. A comparative analysis of pre-hospital, clinical and CT variables. British Journal of
Forrester, G. and Geffen, G. (1996). The Julia Farr Services Post Traumatic
Amnesia Scales (PTAS), in ACER Psychological Tests and Materials Catalogue
19967, Australian Council for Educational Research, Sydney, NSW.
Gardner, R. A. and Broman, M. (1982) The Purdue Pegboard: normative data
on 1334 school children. Journal of Clinical Child Psychology, 8, 156162.
Gennarelli, T. A. (1984) Emergency department management of head injuries.
Emergency Medicine Clinics of North America, 2, 749760.
Gloss, D. S. and Wardle, M. G. (1982) Reliability and validity of American
Medical Associations guide to ratings of permanent impairment. Journal of
the American Medical Association, 248, 22922296.
Griffiths, R. (1970) The Abilities of Young Children: A Comprehensive System of
Mental Measurement for the First Eight Years of Life, Young, Chard.
REFERENCES
Gronwall, D. and Wrightson, P. (1980) Duration of post-traumatic amnesia
after mild head injury. Journal of Clinical Neuropsychology, 2, 5160.
Hahn, Y. S., Chyung, C., Barthel, M. J. et al. (1988) Head injuries in children
under 36 months of age: demography and outcome. Childs Nervous System,
4, 3440.
Haslam, C., Batchelor, J., Fearnside, M. R. et al. (1994) Post-coma disturbance
and post-traumatic amnesia as nonlinear predictors of cognitive outcome
following severe closed head injury: findings from the Westmead Head
Injury Project. Brain Injury, 8, 519528.
Heinze, J. (1969) Cranial nerve avulsion and other neural injuries in road
accidents. Medical Journal of Australia, 2, 12461249.
Hofer, T. (1993) Glasgow Scale relationships in pediatric and adult patients.
Journal of Neuroscience Nursing, 25, 218227.
Jagger, J., Fife, D., Vernberg, K. and Jane, J. A. (1984) Effect of alcohol
intoxication on the diagnosis and apparent severity of brain injury.
Jennett, B. (1992) Severe head injuries: ethical aspects of management. British
Journal of Hospital Medicine, 47, 354357.
Jennett, B. and Bond, M. (1975) Assessment of outcome after severe brain
damage. A practical scale. Lancet, i, 480484.
Jennett, B. and Plum, F. (1972) Persistent vegetative state after brain damage.
A syndrome in search of a name. Lancet, i, 734737.
Jennett, B. and Teasdale, G. (1977) Aspects of coma after severe head injury.
Lancet, i, 878881.
Jennett, B. and Teasdale, G. (1981) Management of head injuries, F. A. Davis,
Philadelphia, PA.
Jennett, B., Teasdale, G., Galbraith, J. et al. (1977) Severe head injuries in three
countries. Journal of Neurology, Neurosurgery and Psychiatry, 40, 291298.
Jennett, B., Teasdale, G., Braakman, R. et al. (1979) Prognosis of patients with
severe head injury. Neurosurgery, 4, 283289.
Johnstone, A. J., Lohlun, J. C., Miller, J. D. et al. (1993) A comparison of the
Glasgow Coma Scale and the Swedish Reaction Level Scale. Brain Injury, 7,
501506.
Jones, N. R., Molloy, C. J., Kloeden, C. N. et al. (1993) Extradural haematoma:
trends in outcome over 35 years. British Journal of Neurosurgery, 7,
465471.
Kocher, T. (1901) Hirnerschutterung,
Hirndruck und chirurgische Eingriffe bei

Krankheiten, Holder,
Vienna.
Levin, H. S., ODonnell, V. M. and Grossman, R. G. (1979) The Galveston
Orientation and Amnesia Test: a practical scale to assess cognition after
head injury. Journal of Nervous and Mental Disease, 167, 675684.
Luersson, T. G., Klauber, M. R. and Marshall, L. F. (1988) Outcome from head
injury related to patients age. A longitudinal prospective study of adult
and paediatric head injury. Journal of Neurosurgery, 68, 409416.
Lyle, D. M., Pierce, J. P., Freeman, E. A. et al. (1986) Clinical course and
outcome of severe head injury in Australia. Journal of Neurosurgery, 65,
1518.
Marshall, L. F., Barba, D., Toole, B. M. and Bowers, S. A. (1983a) The oval
pupil: clinical significance and relationship to intracranial hypertension.
Marshall, L. F., Becker, D. P., Bowers, S. A et al. (1983b) The National Traumatic
Coma Data Bank. Part 1: design, purpose, goals, and results. Journal of
Mathiowetz, V. (1990) Grip and pinch strength measurements, in Muscle
Strength Testing, (ed. L. F. Amundsen), Churchill Livingstone, New York,
ch. 7.
Medical Research Council Brain Injuries Committee (1941) A glossary of
psychological terms commonly used in cases of head injury. MRC War
Memorandum No. 4. HMSO, London.
Menegazzi, J. J., Davis, E. A., Sucov, A. N. and Paris, P. M. (1993) Reliability of
the Glasgow Coma Scale when used by emergency physicians and
paramedics. Journal of Trauma, 34, 4648.
Muizelaar, J. P., Marmarou, A., Young, H. F. et al. (1993) Improving the
outcome of severe head injury with the oxygen radical scavenger
polyethylene glycol-conjugated superoxide dismutase: a Phase II trial.
Narayan, R. K. (1989). Emergency room management of the head -injured
patient, in Textbook of Head Injury, (eds D. P. Becker and S. K. Gudeman), W.
B. Saunders, Philadelphia, PA, ch. 2.
Narayan, R. K. (1994) Comments on Chesnut et al., 1994. Neurosurgery, 34,
845846.
North, J. B. and Jennett, S. (1974) Abnormal breathing patterns associated with
acute brain damage. Archives of Neurology, 31, 338344.
Raven, J. C. (1986) Coloured Progressive Matrices, H. K. Lewis, London.
Reilly, P. L., Simpson, D. A., Sprod, R. and Thomas L. (1988) Assessing the
conscious level in infants and young children: a paediatric version of the
Glasgow Coma Scale. Childs Nervous System, 4, 3033.
Richardson, J. T. E. (1990) Clinical and Neuropsychological Aspects of Closed Head
Injury, Taylor & Francis, London.
Rimel, R. W., Giordani, B., Barth, J. T. et al. (1981) Disability caused by minor
head injury. Neurosurgery, 9, 221228.
165
Rimel, R. W., Giordani, B., Barth, J. T. and Jane, J. A. (1982) Moderate head
injury. Completing the clinical spectrum of brain trauma. Neurosurgery, 11,
344351.
Robertson, R. C. L. and Pollard, C. (1955) Decerebrate state in children and
adolescents. Journal of Neurosurgery, 12, 1317.
Robertson, C. S., Clifton, G. L., Taylor, A. A. and Grossman, R. G. (1983)
Treatment of hypertension associated with head injury. Journal of Neurosurgery, 59, 455460.
Rowbotham, G. F. (1945) Acute Injuries of the Head, 2nd edn, E. & S.
Livingstone, Edinburgh, pp. 104, 144.
Russell, W. R. (1932) Cerebral involvement in head injury. A study based on
the examination of two hundred cases. Brain, 55, 549603.
Russell, W. R. (1947) The neurology of brain wounds. British Journal of Surgery
War Supplement, 1, 250252.
Russell, W. R. (1954) Studies on head injury. British Medical Bulletin, 10,
6568.
Russell, W. R. and Nathan, P. W. (1946) Traumatic amnesia. Brain, 69,
280300.
Russell, W. R. and Smith, A. (1961) Post-traumatic amnesia in closed head
injury. Archives of Neurology, 5, 417.
Ryall, R. G., Peacock, M. K. and Simpson, D. A. (1992) Usefulness of
2-transferrin assay in the detection of cerebrospinal leaks following head
Selhorst, J. B., Gudeman, S. K., Butterworth, J. F. et al. (1985) Papilledema after
acute head injury. Neurosurgery, 16, 357363.
Seshia, S. S., Seshia, M. M. K. and Sachdeva, R. K. (1977) Coma in childhood.
Developmental Medicine and Child Neurology, 19, 614628.
Sheridan, M. (1976) STYCAR Vision Tests, NFER, Windsor.
Shores, E. A. (1989) Comparison of the Westmead PTA Scale and the Glasgow
Coma Scale as predictors of neuropsychological outcome following
extremely severe blunt head injury (letter). Journal of Neurology, Neurosurgery and Psychiatry, 52, 126127.
Shores, E. A., Marosszeky, J. E., Sandanam, J. and Batchelor, J. (1986)
Preliminary validation of a clinical scale for measuring the duration of posttraumatic amnesia. Medical Journal of Australia, 144, 569572.
Simpson, D. A. and Reilly, P. L. (1982) Paediatric coma scale (letter). Lancet, ii,
450.
Simpson, D. A., Cockington, R. A., Hanieh, A. et al. (1991) Head injuries in
young children: the value of the Paediatric Coma Scale. Childs Nervous
System, 7, 183190.
Starmark, J.-E. (1988) Analysing Coma Scales. The Introduction of a New Scale
The Reaction Level Scale, University of Goteborg,.
Goteborg.
Starmark, J.-E., Holmgren, E. and Stlhammer, D. (1988) Current reporting of

responsiveness in acute cerebral disorders. A review of the neurosurgical
literature. Journal of Neurosurgery, 69, 692698.
Starmark, J.-E., Stlhammer, D., Holmgren, E. and Rosander, B. (1988) A
comparison of the Glasgow Coma Scale and the Reaction Level Scale
(RLS85). Journal of Neurosurgery, 69, 699706.
Stening, W. A., Berry, G., Dan, N. G. et al. (1986) Experience with acute
subdural haematomas in New South Wales. Australian and New Zealand
Teasdale, G. and Jennett B. (1974) Assessment of coma and impaired
consciousness. A practical scale. Lancet, ii, 8184.
Teasdale, G. and Jennett, B. (1976) Assessment and prognosis of coma after
head injury. Acta Neurochirurgica (Vienna), 34, 4555.
Teasdale, G. M., Knill-Jones, R. and Jennett, W. B. (1974) Assessing and
recording conscious level. Journal of Neurology, Neurosurgery and Psychiatry,
37, 1286.
Teasdale, G., Knill-Jones, R. and Van der Sande, J. (1978) Observer variability
in assessing impaired consciousness and coma. Journal of Neurology,
Tiffin, J. (1968) Purdue Pegboard Examiner Manual, Science Research Associates,
Chicago, IL.
Trott, J. and Cooter, R. (1995) Systematic clinical assessment, in Craniomaxillofacial Trauma, (eds D. J. David and D. A. Simpson), Churchill Livingstone,
Edinburgh, ch. 6.
Walker, A. E. (1985) Cerebral Death, 3rd edn, Urban & Schwarzenberg,
Baltimore, MD, ch. 3.
Walsh, K. W. (1985) Understanding Brain Damage. A Primer of Neuropsychological
Evaluation, Churchill Livingstone, Edinburgh.
Wilson, J. T. L., Teasdale, G. M., Hadley, D. M. et al. (1993) Post-traumatic
amnesia: still a valuable yardstick. Journal of Neurology, Neurosurgery and
Wood, M. and Hammerton, M. (1995) Impairments and disabilities, in
Craniomaxillofacial Trauma, (eds D. J. David and D. A. Simpson), Churchill
Livingstone, Edinburgh, ch. 22.
Wood, M. and Woodroffe, P. (1995) Neuropsychological testing, in Craniomaxillofacial Trauma, (eds D. J. David and D. A. Simpson), Churchill Livingstone,
Edinburgh, appendix III.
Yager, J. V., Johnston, B. and Seshia, S. S. (1990) Coma scales in pediatric
practice. American Journal of Diseases in Children, 144, 10881091.
IMAGING THE INJURY

Evelyn Teasdale and Donald M. Hadley
9.1
Introduction
Since its development in the early 1970s computed

tomography (CT) has remained the radiological
examination of choice in the acute assessment of a
patient with a head injury. CT illustrated for the first
time the multiplicity and complexity of the structure
of living pathology in head injury (see Figures 9.42,
9.48). This improved our understanding of the mechanisms involved, allowed more rational treatment
and thus improved the outcome for the patient. The
increasing availability of high-resolution rapid acquisition scanners will hopefully further improve the
early diagnosis, management and hence outcome in
these patients.
While CT is adequate for making acute surgical
management decisions it has become increasingly
apparent that other techniques are more sensitive
and give a more thorough pathophysiological assessment of the extent and nature of the whole injury
especially any associated potentially reversible cerebral damage. Magnetic resonance imaging (MRI),
with its high sensitivity to minute alterations in
cerebral water content and blood products, can show
injuries invisible on CT (Figure 9.6). It is particularly
useful in the subacute and chronic phases. Its major
and uniquely important role, however, in the acute
neurotrauma patient is in the diagnosis of suspected
or actual spinal cord injury, especially when the
patient is paralyzed and ventilated. Positron emission tomography (PET) and single photon emission
tomography (SPECT) are sensitive functional imaging modalities useful in the assessment of the disrupted metabolism and blood flow that are prevalent
in many types of brain injury. These potentially
reversible changes shown by PET and SPECT pose a
new challenge for cerebral protection therapies and
confirm an increasing complementary role for functional and structural imaging. It gives us the basic
pathophysiological knowledge to develop rational
imaging and treatment strategies for the management of head injury.
9.2
The role of plain film radiography
9.2.1
SKULL FILMS
It is neither advisable nor possible to perform a CT

scan on all patients with a head injury. Plain skull
films remain important in the triage of minor injuries.
However if neurosurgical referral is already deemed
necessary, or if CT is indicated on clinical grounds,
then performing a skull X-ray will only delay the
patients transfer or the definitive CT examination
and will give no additional management information.
A skull X-ray should now be performed only to
identify those patients with a minor head injury in
whom a fracture is present so that they may be
scanned electively within the following 24 hours
(Teasdale et al., 1990). Skull films also remain an
important screening modality in centers where no CT
scan is available, and may thus guide the referral of
patients with mild head injury for CT.
If skull X-ray is required at least two views of the
skull at right angles should be taken. The lateral
should be taken with the patient supine and the X-ray
beam parallel to the floor. This ensures that a nonrotated film can be achieved, that the upper cervical
spine is not moved but is included in the radiograph,
that a fluid level may be seen in the paranasal sinuses
and that any intracranial air can be detected. The
frontal view is taken to examine the frontal or occipital
bone. Fractures are shown as sharply demarcated
black lines with no corticated margins. They do not
branch, tend to run in straight lines and stop at
sutures, which may themselves be diastased or abnormally separated. These features are readily distinguished from vascular markings, which involve only
the inner or the outer table of the skull or lie within the
diploe and therefore appear gray rather than black.
Vascular markings also have corticated branching
paths which cross sutures lines. A depressed fracture
may be suspected if there is an area of reduced density
with an associated area of increased density where the
displaced bone superimposes upon adjacent normal
168
IMAGING THE INJURY
skull. An X-ray taken with the film tangential to the

suspected fracture will confirm the degree of displacement. Patterns of reduced density can also be caused
by air within the subarachnoid space, the ventricles or
the brain itself. A fluid level caused by blood within
the sphenoid sinus is a common indirect sign of a
skull-base fracture. Skull-base fracture lines themselves are rarely identified on plain skull films but are
readily shown by axial CT.
The assessment of complex facial fractures should
be delayed until the patient has been stabilized and is
fit enough to cooperate. Usually this CT assessment is
done when reparative surgery is imminent (Trott and
David, 1995). A lateral facial view and three frontal
facial views with the X-ray beam angled at 10, 20 and
30 to the orbitomeatal baseline are traditional, but
thin-section low-dose CT in the direct coronal plane
gives optimal visualization of the number and position of the fractures. Any associated soft-tissue displacements, such as may be present in medial or
inferior blow-out orbital fractures (Figure 9.1), can be
clearly demonstrated. If the patient is unable to
achieve a coronal position or if there are severe
artifacts from a dental prosthesis, axial CT with
coronal reconstruction is an alternative, but the detail
is inferior to that obtained from a direct coronal
acquisition. Three-dimensional (3D) reconstruction of
a stack of contiguous or overlapping sections can give
the surgeon an interactive dynamic interpretation of
the image, highlighting areas of flattening or distortion and showing major bony displacements (Figure
9.2). Although attractive images can be produced, the
algorithms used in the formation of the 3D image can

obscure or falsify fracture lines and their value to
patient management lies in planning treatment strategies and the construction of complex prosthetic skull
plates.
9.2.2
In the head-injured patient with a severely impaired

consciousness level it may be impossible to elicit the
clinical signs or symptoms necessary to exclude
associated injury to the cervical spine.
Lateral and frontal cervical films, which must
include the craniocervical junction and the first
thoracic vertebra, are therefore essential. Some 70% of
fractures occur below the level of the third cervical
vertebra and 10% involve the cervicothoracic junction. Adequate X-rays of C7/T1 are available in less
than 50% of patients. In other words, plain films will
miss 50% of fractures at C7/T1 (Annis et al., 1987;
Figure 9.3) and therefore axial thin-section CT
(2.5 mm) from C6 to T2 should be carried out with
midline and oblique sagittal reformations to highlight
any subluxation present. If satisfactory plain films of
the cervical spine are not available the scout headplanning CT view may be sufficient to identify a
suspicious area, which may then be scanned with
contiguous thin sections taken from a level above to a
level below the possible fracture (Figure 9.4).
This coverage ensures that adequate multiplanar
reformations can be made if necessary (Figure 9.5).
When the cervicothoracic junction is not visualized on
plain films then it may be scanned immediately after
the head has been examined, minimizing patient
movement. MRI may be required to demonstrate or
exclude soft-tissue injury involving the cord, ligaments, disk or paraspinal tissues.
9.2.3
Figure 9.1 Direct coronal CT of the orbit and facial bones

showing fractures of the inferior and medial orbital walls
with herniation of orbital contents. They are clearly seen
despite some streak artifact from dental fillings.
THE CERVICAL SPINE
THE CHEST
A frontal supine chest X-ray is required in the full

assessment of the head-injured patient as it may reveal
a treatable cause of hypoxia such as lung collapse,
consolidation, contusion, pneumothorax or pleural
effusion. Injury to the great vessels should be suspected if there is widening of the upper mediastinal
shadow, if a small apical pleural effusion is present or
if there is a fracture of the posterior aspect of the upper
ribs. Aortography is the definitive examination but
dynamic contrast-enhanced CT can show the false
lumen of a dissection and may be useful if the chest
X-ray is normal but the history and clinical situation is
very suggestive (White and Mirvis, 1995). The position
of endotracheal tubes and subclavian/jugular lines
should be scrutinized to ensure that they are correctly
sited.
CT IN HEAD INJURY
(a)
169
(b)
Figure 9.2 3D CT reconstructions. (a, b) A patient with a

self-inflicted gunshot wound. Amazingly, vision was preserved. The CT data was used to form a full-size model used
in operative planning. (c) Following reconstruction of a
cranio-orbital fracture. (Reproduced by courtesy of the
Australian Cranio-Maxillo Facial Unit, Adelaide.)
(c)
9.3
CT in head injury
The selection criteria for patients to have a CT

examination have already been covered but CT is
now so widely available that it should be carried out
on all patients in whom admission and observation
were previously recommended. In addition to all
those minor head injuries with a skull fracture, this

includes all those with a GCS less than 15, those with
clinical signs of a skull-base fracture, those with a
seizure or focal neurological sign and those with
persistent headache (Miller, 1990; Jeret et al., 1993), as
well as those in whom a penetrating injury is
suspected (Table 9.1).
170
IMAGING THE INJURY
Figure 9.3 Lateral cervical spine in the neutral position.

Note the anterior displacement of C4 on C5, involving both
the vertebral body and spinous process, indicating a true
rather than a pseudosubluxation.
Figure 9.4 (a) Lateral scout film for the head CT shows an
increase in the soft tissue in the pre-vertebral space at C1/C2
and a fracture/dislocation of the odontoid peg (arrowhead
on the anterior margin of the body of C2). (b) Midline
sagittal reformation of the axial CT confirms the abnormality
on the scout film.
(a)
(b)
MRI IN HEAD INJURY
Figure 9.5 Axial CT of the C7/T1 region shows a fracture

of the C7 transverse process that was not visible on plain
films.
CT maps the way in which different tissues absorb

or attenuate the beam of X-rays and so can display
normal structures and pathology. Blood clot, because
it contains large compacted protein molecules, has a
high electron density and relatively little water, and
absorbs the X-rays more than normal brain. It is
displayed as a hyperdense or white area of increased
beam attenuation. Conversely, edema or ischemia are
displayed as dark areas of reduced beam attenuation
because there is a relative increase in water content
and hence a reduction in electron density. Normal
gray and white matter have attenuations between
these two extremes and as gray matter is more cellular
and vascular it has a higher attenuation than white
matter.
9.4
It is necessary to ensure that the patient is scanned
in such a way that all the required information is
obtained and to ensure that further harm to the patient
is minimized while the scan is carried out (Miller,
1990). Resuscitative measures must continue during
scanning and attending staff should be protected from
radiation hazards. Ideally the patient should be
scanned in the supine position but if the patient is
restless a lateral position may be adopted and still
provide diagnostic images. This maneuver may avoid
the need for extra sedation or anesthesia. The scan
should extend from the foramen magnum to the
vertex and be angled parallel to the orbitomeatal line
to avoid irradiating the lens of the eye (MacLennan
and Hadley, 1994). In the posterior fossa the slice
thickness should not exceed 5 mm in order to minimize partial volume artifact, to optimize the signal-tonoise ratio and improve contrast resolution. If a slice
thickness of 10 mm is used, small hematomas will be
missed. Initially it is important to perform a multiple
thin section examination as a baseline for accurate
definition of the radiological pattern of injury. Subsequent scans can then be performed with thicker
(10 mm) slices and reduced cumulative radiation dose.
Both soft-tissue and bone window images can be
obtained from the same data set and 3D reconstructions are possible later. Acquisition time is usually 15
minutes.
Table 9.1

Indications for CT
GCS < 15
Minor head injury with skull fracture
Clinical signs of skull-base fracture
Seizure
Focal neurological signs
Persistent headache
171
MRI in head injury
Now that MRI-compatible cardiorespiratory monitoring and support equipment including intracranial
pressure monitors are available, critically ill patients
can be monitored in an MRI scanner at any stage after
a head injury (Hadley et al., 1988). However, if the
patient is not unconscious or paralyzed and ventilated, a far higher degree of cooperation is required
than for CT. These patients may have sustained
multiple injuries and screening is required to exclude
ferrometallic foreign bodies or electronic implants.
Patients who have cardiorespiratory instability in
particular should not undergo MRI, as infusion
pumps (e.g. for vasopressors) are not yet MRIcompatible. There is therefore a reticence to undertake
acute MRI in the routine clinical setting away from
research centers. Acquisition time for a brain sequence
is about 3060 minutes.
The contrast between normal and pathological
tissues on MRI can be many times that of CT, but is
dependent on the sequence parameters chosen. Routinely, a long TR spin echo dual echo set of axial
sections provide T2-weighted and proton-densityweighted contrast, while a further set of short TR spin
echo or inversion recovery sections demonstrate
T1-weighted contrast differences. Gradient echo
T2-weighted sequences highlight changes in magnetic
susceptibility, making them very sensitive to acute
and chronic hemorrhage (Figure 9.6), and flow within
vessels and also air. By using reduced flip angles (less
than 90) diagnostic quality can be improved by
reducing data acquisition time (with a slight loss of
signal to noise), limiting motion artifacts and consequently the need for sedation. Unfortunately the
magnetic susceptibility effects are not all useful with
regions close to the paranasal sinuses or other brain
air interfaces, giving an artifactual loss of signal due to
diamagnetic susceptibility gradient effects in the very
brain regions frequently injured by head trauma the
172
IMAGING THE INJURY
(a)
Figure 9.7 Subacute right temporal hematoma hyperintense on Tl -weighted sections, associated with a thin but
extensive SDH around the temporal and occipital lobes.
(b)
Figure 9.6 Hemorrhagic contusions in both frontal lobes 26
hours after injury. (a) T2-weighted MRI. (b) T2*-weighted
MRI. The hypointensity of the acute hemorrhage is more
pronounced on the T2*-weighted images.
gyral crests, the orbitofrontal, inferior and medial

temporal cortex (Figure 9.7).
Recently, sequences have been developed that are
three to four times faster than conventional spin
echoes. The increased speed of acquisition can be used
to carry out a quicker study in children or severely
injured patients. Alternatively, more acquisitions can
be obtained in the same time as a conventional scan to
markedly improve anatomical detail, as in the posterior fossa and petrous bone. Because of its multiple
1800 pulses, this sequence is less sensitive to magnetic
susceptibility effects and therefore acute and chronic
hemorrhage (Bradley, 1993; Jolesz and Jones, 1993).
This gives it the advantage of better imaging close to
airbonebrain interfaces and if necessary the time
gain can be used to carry out a quick gradient echo
sequence to show occult hemorrhage.
Ultra-fast echo planar imaging is becoming available on conventional imagers (Slavin et al., 1995) and
cerebral diffusion, perfusion and functional studies
will be available at high anatomical resolution, competing with the lower resolution data from PET and
SPECT. Magnetic resonance spectroscopy is also possible now (Felber et al., 1993; Vink, 1993) but so far it is
largely used as a research tool (Chapter 14).
9.5
SPECT in head injury
Although PET scanning has been used for metabolic

functional studies of the traumatized brain its complexity and cost, and the need for an on-site cyclotron,
have limited its use to a few research centers (Rao et
al., 1984). The development of radionuclide tracers for
SPECT scanning such as 99mTc hexamethyl propylene
amine oxime (99mTc-HMPAO) over the last decade has
allowed the investigation of regional cerebral blood
flow (RCBF) using either rotating gamma cameras or
dedicated head imagers in the routine clinical setting
INTRACEREBRAL LESIONS
Figure 9.8 Acute left frontal contusion. SPECT shows an

area of no perfusion surrounded by reduced perfusion.
There is also a small focus of hyperperfusion noted in the
cortex posterolaterally. Is this where autoregulation has been
compromised?
(Roper et al., 1991; Gray et al., 1992). After intravenous

injection, the delivery of 99mTc-HMPAO, to the brain is
proportional to RCBF (Figure 9.8). As this substance is
lipophilic it crosses the bloodbrain barrier, where it is
rapidly converted into a hydrophilic compound and is
trapped for several hours. It has been shown that this
compound reflects RCBF under both normal and
pathological conditions (Lear, 1988; Bullock, Patterson
and Park, 1991), although there is a systematic
underestimation in regions of very high perfusion
(Lassen et al., 1988).
9.6
173
develop as a complication of a contusion. This is more

likely if the patient has even a minor coagulopathy
such as is common in alcohol-related trauma. In many
cases there is only a semantic distinction between a
hemorrhagic contusion and a contusional hematoma
(Figures 9.99.13).
On CT acute hemorrhage is seen as an area of
increased attenuation or hyperdensity (white) with
surrounding hypodense (dark) edema. As the clot
ages, edema increases over approximately 4 days and
the clot becomes isodense with the brain over several
weeks. A fluid level within a hematoma indicates
coagulopathy and clot liquefaction, or extensive associated cerebral liquefaction and a worse prognosis
(Katayama et al., 1992).
In the absence of associated contusions it may be
difficult to distinguish a spontaneous (e.g. hypertensive) from a traumatic clot on purely radiological
features. A traumatic clot forms as a result of rupture
of an intracerebral or occasionally a subarachnoid
vessel, but angiography may be required to exclude
aneurysmal rupture as the cause of an unexplained
hematoma.
The presence of a clot at an unexpected site or in a
peripheral-to-deep linear orientation should always
raise the possibility of a penetrating injury causing
direct or indirect vascular damage. In these patients
angiography is required urgently to show the vascular
component of the lesion and allow appropriate ther-
Classification
There are many ways to classify head injury (Teasdale,

Teasdale and Hadley, 1992), but most require a
combination of clinical, imaging and pathological data
and have already been discussed. To avoid confusion
and repetition the radiological description given here
classifies lesions simply as intracerebral or
extracerebral.
9.7
Intracerebral lesions
9.7.1
INTRACEREBRAL HEMATOMA
Traumatic intracerebral hematoma can result from a

variety of causes. Most intracerebral hematomas (ICH)
Figure 9.9 Blood in the sphenoid sinus indicates a skullbase fracture even if the fracture lines cannot be seen. Left
subtemporal clot is noted but the exact compartment cannot
be defined because of the axial plane of the CT.
174
IMAGING THE INJURY
(a)
(b)
Figure 9.10 (a) Admission CT shows hemorrhagic bifrontal and right occipital contusions. a surface collection over the
right frontal lobe and a left occipital ICH. (b) The patient deteriorated on the following day, with the development of a new
left extracerebral collection and further hemorrhage with fluid levels in the left frontal and occipital lesions. These caused
midline shift and third ventricular compression.
Figure 9.11 Large cerebellar hematoma with bleeding into

the displaced fourth ventricle (arrows). Hydrocephalus is
shown by the dilated third ventricle and the dilated
temporal horns of the lateral ventricles. There is a large left
temporal ICH with an old left temporal fracture and
evidence of old cerebromalacia in the right temporal lobe
from a previous injury. Bifrontal contusions and diffuse SAH
are also noted.
Figure 9.12 Complex injuries in an alcoholic with minor

coagulation problems: a large right frontal ICH with surface
and Sylvian SAH in addition to a clot in the quadrageminal
cistern. There is a small clot in the right thalamus and an
IVH in the lateral and third ventricles.
175
(a)
(b)
Figure 9.13 (a) Admission CT showing a right ICH with a thin SDH and left-sided contusions. Midline shift is less than
5 mm but the cisterns and third ventricle are obliterated so ICP > 30 mmHg. (b) Follow up at 3 months shows significant
hydrocephalus with periventricular low attenuation.
apy to avoid further hemorrhage or emboli from a

dissection or false aneurysm (du Trevou and van
Dellen, 1992; Stein et al., 1993; Bula and Loes, 1994;
Figure 9.14). False aneurysms have been described
also after blunt trauma.
An ICH need not be present on the admission CT
but may develop later. This is termed a delayed
hematoma and was originally thought to be a rare
phenomenon. As more and more patients are scanned
within hours of their injury delayed hematomas have
been recognized more commonly and are found in up
to 1520% of patients with severe head injury (Gentleman, Nath and Macpherson, 1989). This presumably
reflects the evolution of the injury related to the timing
of the CT. Delayed clots rarely occur at a site
previously completely normal and a contusion is the
commonest precursor but they may develop in the
extradural or, more rarely, the subdural space. They
are also seen more frequently once another lesion has
been resected or after resuscitation (reperfusion
effect). The relief of vasospasm and coagulopathies
have also been suggested as causes. Unfortunately a
patient with a delayed clot has a worse prognosis
overall (Figures 9.10, 9.15).
Small hematomas occur in association with diffuse
white matter injury where the pattern of distribution
indicates the pathogenesis (see below). Occasionally,
hemorrhage will develop after some days in an area of

established arterial or venous infarction.
Hematomas in the posterior fossa may be difficult to
identify but significant clots will always displace the
fourth ventricle. In children an occipital fracture is a
usual association while in adults it is commoner in the
atrophic cerebellum of the aged and the alcoholic
(Figures 9.12, 9.16).
There are three specific types of brain-stem hematoma, which can usually be defined on imaging. Most
common is the hematoma associated with DAI, which
lies in the rostral brain stem lateral or posterolateral to
the fourth ventricle (see below; Figure 9.17(d)). When
there is a large supratentorial mass with tentorial
herniation, secondary brain-stem hemorrhage occurs.
This is thought to be due to distortion and consequent
rupture of the pontine perforating vessels. This has
been termed a Duret hematoma and is seen centrally
within the pons and mesencephalon (Figure 9.17(a),
(b)). Occasionally in severe primary trauma a hematoma is seen in a similar central position but without
evidence of transtentorial herniation. This is thought
to be due to an acute transient descent of the brain
stem at the time of impact, resulting in rupture of the
perforating vessels (Figure 9.17(c)).
MRI is more sensitive to hematomas than CT but
their appearance is more complex and depends on
176
IMAGING THE INJURY
(c)
(a)
Figure 9.14 (a) CT shows a bifrontal fresh hematoma with

some edema 6 days after the patient fell downstairs.
(b) Left carotid angiogram demonstrates a traumatic aneurysm. (c) Bone window image of the CT shows the tiny
fracture through which the point of a Ninja star, a throwing
weapon, had penetrated. This history became available only
after repeated questioning.
(b)
multiple factors such as the paramagnetic form of

hemoglobin present, clot matrix formation, changes in
erythrocyte hydration and changes in the degree of
red blood cell packing. Hence there is a characteristic
time sequence of intensity patterns as the hematoma
forms, hemoglobin matures, denatures and is metabolized, eventually leaving a hemosiderin-lined gliotic
cleft (Gomori and Grossman, 1987; Gomori et al., 1987;
Hadley et al., 1988). This evolving pattern is summarized in Table 9.2.
At the hyperacute stage, which lasts from minutes
to a few hours, coagulation mechanisms are initiated
and the few patients who have been imaged show

hyperintensity on T2-weighted sections with hypointensity on T1-weighted and hyper- or isointensity on
proton-density-weighted images. At this stage the
liquid hematoma behaves like a protein solution.
Following clot formation and absorption of serum
over the next 112 hours the acute phase is entered.
Oxyhemoglobin is reduced to deoxyhemoglobin at a
rate dependent on the local oxygen tension and pH.
This lasts for up to a week, although changes
representing the subacute stage can be demonstrated
as early as 3 days. On MRI in the acute stage,
177
(a)
(b)
Figure 9.15 (a) Admission CT demonstrates hemorrhagic contusion in the left frontal region and low attenuation
contusions in the right. (b) The patients consciousness level deteriorated and repeat CT shows a large delayed left frontal
ICH with blood in the compressed third ventricle. Note that the basal cisterns are now obliterated due to the increase in ICP.
Compare with the normal cisterns in (a).
Figure 9.16 A small superficial cerebellar hematoma with

compression and displacement of the fourth ventricle and
early hydrocephalus shown by the dilated temporal horns in
this 10-year-old child.
hematomas appear as well-circumscribed regions of

hypointensity on T2-weighted images (Figures 9.6,
9.18). On T1-weighted sections the hematomas appear
isointense with gray matter and usually hypointense
compared to white matter. The short T2 (hypointense
signal) of acute hematomas is more pronounced on
more heavily T2-weighted images, by higher field
strengths or by using T2* gradient echo sequences
(Figures 9.7, 9.19). The low intensity on T2-weighted
images corresponds exactly to the high attenuation of
hematomas on CT. All hematomas are surrounded by
varying amounts of absorbed serum and edema and
this is shown by a broad irregular hyperintense rim on
T2-weighted images (Figures 9.7, 9.19).
The subacute stage can start as early as 3 days and
is firmly established from a week to around a month
postinjury. It is marked by the change from hypo- or
isointensity to hyperintensity on T1-weighted images
(Figure 9.7). This starts as a peripheral ring, which
extends inwards over the next few weeks, depending
on the size of the lesion, to fill the entire hematoma.
The hyperintense signal is due to the conversion of
deoxyhemoglobin to methemoglobin. This has been
further divided into early and late stages on the
appearance of the T2-weighted images. Initially there
178
IMAGING THE INJURY
(b)
(a)
(c)
(d)
Figure 9.17 Brain stem hemorrhage. (a, b) Duret-type brain-stem hemorrhage secondary to a massive supratentorial SDH.
(c) Anterior midline hemorrhage and clot posterolateral to the fourth ventricle (arrow). No mass lesion is present as the
temporal horns are small and symmetrical. This is a primary impact brain-stem hemorrhage associated with DAI. (d) Large
clot in the rostral brain stem typical of DAI.
179
Table 9.2 Summary of the signal changes on T1- and T2-weighted images at each stage in the maturation of
hemorrhage (Source: after Gean, 1994)
Stage
Biochemistry
T1
T2
Comments
1. Hyperacute
Oxyhemoglobin: no unpaired
electrons; diamagnetic, Fe2+
Dark
Bright
Behaves as a simple fluid collection

(long Tl, long T2). Low-spin state.
2. Acute
Deoxyhemoglobin: four
unpaired electrons;
paramagnetic, Fe2+
Isointense
Dark
Susceptibility effect. High-spin state.

More apparent on high-field-strength
magnets.
3. Subacute: early
Methemoglobin,
hemichromes: five unpaired
electrons, paramagnetic, Fe3+
Bright
Dark
Intact RBCS. Appearance due to

changes in protein concentration,
exchange processes, and a decrease in
paramagnetism.
Bright
Bright
RBC lysis.
Isointense
Dark
Crystalline storage form of iron.

Marked susceptibility effect.
4. Subacute: late
5. Chronic
Ferritin/hemosiderin:
paramagnetic, Fe3+
is hypointensity due to intracellular methemoglobin.

As red cell lysis occurs the methemoglobin is released
and the maturing hematoma becomes hyperintense on
T2-weighted images.
In the chronic stage from 1 month onwards the
initial hyperintensity on T1-weighted sections can
persist somewhere in the hematoma for some time but
reactive macrophages accumulate at the periphery,
engulfing the hemoglobin breakdown products. These
(a)
hemosiderin-laden macrophages persist indefinitely

in the hematoma periphery and appear as a focus or
ring of marked hypointensity on T2-weighted images
in a region that is isointense or hypointense on
T1-weighted images (Figure 9.20). The surrounding
edema gradually becomes less sharply demarcated
and the space-occupying effect of the whole lesion
decreases. The signal changes due to edema gradually
resolve, returning to normal over 13 months.
(b)
Figure 9.18 Multiple small contusions in both frontal lobes shown on the T2-weighted MRI (a) but more clearly
demonstrated on the T2*-weighted MRI (b), which also shows petechial hemorrhages in the subcortical white matter and
hemorrhagic foci on the floor of both lateral ventricles. Note that the subgaleal extracranial hematoma of the same age is
hyperintense because of the different oxygen tension and pH compared to the intracerebral compartment.
180
IMAGING THE INJURY
Figure 9.19 T2-weighted MRI follow-up of a 20-year-old

with a focus of residual hemosiderin in the right side of the
splenium of the corpus callosum at the site of an ICH noted
on the acute CT scan 9 months earlier.
9.7.2
CONTUSIONS
Contusions are bruises of the brain formed by coalescing petechial hemorrhages caused by acute cerebral deformation, usually due to impaction against
the inner table of the skull. They are the most
common complication of head injury and have a
great variety of appearances, which change with
time. These areas of brain disruption commonly
occur symmetrically on the undersurfaces of the
frontal and temporal lobes where there is direct
contact with bony irregularities of the frontal and
middle fossa. Cytotoxic edema begins soon after the
injury, evolving to a maximum at about 5 days.
Surrounding the contusion is a variable amount of
ischemic brain, which further enlarges the area of
swelling (Figure 9.20). In this surrounding ischemic
area bloodbrain barrier breakdown can be demonstrated by radionuclide uptake on SPECT or by
contrast enhancement on CT or MRI after approximately day 3 or 4 (Lang et al., 1990; Kushi et al.,
1994). Contrast-enhanced CT should only be performed using non-iodinated contrast media, and
with a good clinical reason, as leakage of older
iodinated contrast agents may be toxic to the brain
and so may compound the brain insult.
The CT and MRI appearances of contusions are
varied but specific and their progression over time is
predictable. Figures 9.219.27 illustrate the various

appearances commonly seen. Hemorrhage is always
present pathologically but may be microscopic and
undetectable on CT or MRI. Minor contusions may
be seen only as an area of dark low attenuation. The
amount of blood present varies from a small quantity
on the surface to substantial collections extending
deeply into the subcortical white matter. As contusions heal by progressive degradation and maturation of hemorrhage with eventual gliosis they shrink
to leave a low attenuation area with some loss of
white matter and the hallmark of an old contusion
loss of cortex. This lesion is termed cerebromalacia. It
is typical in appearance and is commonly found in a
subfrontal and/or temporal location (Figure 9.24).
On MRI the border zone adjacent to the macrocystic cerebromalacia of a healed region of hemorrhagic contusion where there was persistent vasogenic edema may show an irregular, relatively
narrow border of hyperintensity on T2-weighted
images and isointensity to gray matter on T1- and
proton-density weighted images (Figure 9.27). This
represents a region of microcystic cerebromalacia or
gliosis, which often contains foci or a rim of hypointensity on T2-weighted images due to residual hemosiderin. When extensive tissue loss occurs due to
hemorrhagic, ischemic or mechanical lesions the area
of cerebromalacia is larger. The space-occupying
effect resolves and an ex vacuo effect supervenes with
enlarged adjacent sulci and ventricles.
9.7.3
PRIMARY WHITE MATTER INJURY
Primary white matter injury is commonly known as

diffuse axonal injury (DAI) or shearing injury because
the neuropathological marker of this entity is thought
to result from shearing stresses set up by angular
acceleration. Until recently this lesion was thought to
be complete immediately after injury. It has now been
shown that this is not the case and so a therapeutic
window may exist when treatment may be able to
limit the extent of the injury. This makes it even more
important to be able to recognize this DAI pattern of
abnormalities radiologically as soon as possible after
injury. Unfortunately the full extent of this cerebral
damage may only be illustrated later when diffuse
atrophy may be found (Figure 9.28).
In patients with a mild concussive injury, small
lesions with increased water content are seen at the
junction of the cortical gray/white matter on MRI.
On CT these lesions are rarely visible (Jenkins et al.,
1986; Mittl et al., 1994; Figure 9.29). They are thought
to represent the gliding contusions associated with
DAI, although as yet pathological correlation with
181
(b)
(a)
(c)
(d)
Figure 9.20 (a, b) Admission CT showing a right temporal contusional hematoma with an overlying SDH and mixed
hemorrhagic contusions on the left with blood in the Sylvian fissure. The temporal horn of the left lateral ventricle is not
visible. (c, d) CT 2 days later shows an increase in the size of the hematomas with increasing shift and now the temporal
horn on the left is dilated. Reduced attenuation edema has collected around both the right- and left-sided contusions,
increasing their overall mass effect.
182
IMAGING THE INJURY
Figure 9.21 Typical bifrontal and bitemporal contusions

with differing amounts of hemorrhage.
Figure 9.22 Mixed hemorrhagic contusions in the right

temporal lobe with a right parafalcine SDH and contralateral
temporal horn dilatation indicating incipient herniation.
(a)
(b)
Figure 9.23 (a) Admission CT shows minor mixed attenuation contusions in the right posterior temporal region with a
diffuse hemispheric mass effect. (b) Follow-up at 6 months shows extensive residual cerebromalacia and occipital
ischemia.
183
Figure 9.24 Typical bifrontal and bitemporal hemorrhagic contusions on the undersurface of the brain with low
attenuation edema and some hemorrhage extending into the subcortical white matter.
(a)
(b)
Figure 9.25 (a, b) Severe hemorrhagic contusions with bihemispheric swelling and obliteration of the CSF pathways. The
abnormally low position of the calcified pineal gland indicates transtentorial herniation.
184
IMAGING THE INJURY
(b)
(a)
Figure 9.26 (a) Minor hemorrhagic contusions on the left, but midline shift is from right to left. (b) Same slice viewed at
a wider window width of 200 Hu (Hounsfield units) rather than 100 Hu, which confirms the thin right SDH.
(a)
(b)
Figure 9.27 Follow-up MRI scan of a patient with bifrontal hemorrhagic contusions. T2-weighted MRI (a) and protondensity-weighted MRI (b) showing bilateral frontal macrocystic cerebromalacia, microcystic cerebromalacia, a thin
hemosiderin rim most prominent on the left, with a generalized ex vacuo effect causing slight dilatation of the frontal horns
of the lateral ventricles.
EXTRACEREBRAL COLLECTIONS
185
(b)
(a)
Figure 9.28 (a) CT on admission shows a small hematoma of the type associated with DAI in the white matter. ICP was
normal. (b) 1 month later there is diffuse atrophy, indicating the true extent of the injury.
MRI appearances is not available. As the severity of

the primary injury increases so other lesions become
apparent in the corpus callosum and in the rostral
brain stem. MRI is very sensitive to these lesions
even in the absence of associated hemorrhage, which
is necessary before the injury can be visualized on
CT. The sensitivity to petechial hemorrhage can be
maximized by using gradient echo T2-weighted MRI
sequences, where the lesions will show as hypointense foci occasionally without any edema (Figure
9.29, see p. 186). The classical triad on CT or MRI
of hemorrhage in the brain stem, the corpus callosum
and the cortical gray/white matter junction is rarely
found. More commonly, small hemorrhages in the
basal ganglia region develop secondary to differential rotational forces between the long white matter
tracts and the gray matter of the basal ganglia,
causing rupture of small blood vessels (Figures
9.309.34, see pp. 187189). Intraventricular hemorrhage is also associated with white matter injury
either because there is direct extension of a callosal
hemorrhage into the ventricular system or because
the torsional forces tear the subependymal veins or
the choroid plexus of the ventricles. As the injury
ages the blood becomes isodense on CT and if
scanning is delayed beyond a few days from the
injury the scan may be falsely interpreted as normal.
If delayed still further low attenuation mass lesions

may be seen at the site of the resolving clots (Figure
9.35, see p. 189). In this subacute phase with the
formation of methemoglobin, MRI is far superior in
the diagnosis of the extent of primary white matter
injury (Figure 9.29, see p. 186).
9.8
Extracerebral collections
After trauma, hemorrhage is commonly found over

the surface of the brain where it can arise in the extraor subdural space, within the subarachnoid space or
within the ventricles.
9.8.1
EXTRADURAL HEMATOMA
Hemorrhage occurs in the epidural potential space,

usually in direct association with a fracture. Impact
strips the dura from the inwardly and outwardly
deforming skull to form an epidural pocket. Blood,
coming from torn meningeal arteries or veins or from
a laceration of the dural sinuses generates a clot which
further strips the dura from the inner table of the skull
but is usually limited in its extent by the adjacent skull
sutures where the dura is tethered. The classical
description of the CT appearance of an established
186
IMAGING THE INJURY
(a)
(b)
Figure 9.29 (a) Normal acute CT in an unconscious fiveyear-old child. T2-weighted (b) and T2*-weighted MRI
(c) show hypointense foci of acute petechial hemorrhage in
the parasagittal frontal white matter. This represents a
gliding contusion associated with DAI. Note its increased
conspicuousness on the T2*-weighted section.
(c)
extradural clot is of a uniformly high attenuation

biconvex lesion based against the skull vault (Figure
9.36). As more patients are being scanned acutely,
many extradural clots are now seen as lesions of
mixed or even mainly low CT attenuation (Figures
9.37, 9.38). This is because the blood is still liquid and
clot is still forming and does not yet resemble the
dense, dehydrated mature hematoma. An EDH is
usually very obvious unless it lies in the axial plane of
the scan (Figure 9.39). Such clots in the floor of the
temporal fossa can be mistaken for an intratemporal
lesion and those on the vertex of the skull can be

missed altogether if the scan is not continued to the
top of the head.
On MRI extracerebral hematomas age in a similar
way to intracerebral hemorrhage, but their appearance is modified in the acute and subacute stages by
decreased reabsorption of serum and varying degrees
of liquefaction. They therefore mature more quickly
than their intracerebral counterparts. If air enters the
collection through a compound fracture, attempted
aspiration or surgery, the process of oxidative
(b)
(a)
Figure 9.30 (a, b)
187
Multiple hematomas in the capsular/basal ganglia region with associated IVH typical of DAI.
Figure 9.31 Hematoma in the genu of the corpus callosum

and IVH, typically associated with DAI.
Figure 9.32 Basal ganglia clot typically associated with

DAI, plus a small cerebellar gliding contusion at the gray/
white matter interface with the dentate nucleus. Elevated
ICP with compressed cisterns. Note the trace of blood in the
third ventricle (arrow) and slight hydrocephalus shown by
the dilated temporal horns of the lateral ventricles.
188
IMAGING THE INJURY
(a)
(b)
Figure 9.33 (ac) Callosal, basal ganglia and gray/white

interface hemorrhage indicative of DAI. There is a thin right
SDH and raised ICP indicated by third ventricular and basal
cistern obliteration.
(c)
reduction of deoxyhemoglobin to methemoglobin is

accelerated.
Unlike hemorrhage into the subdural space, a
chronic EDH is rare. It shows a similar low or mixed
attenuation pattern with dural marginal enhancement
after contrast on either CT or MRI. Rarely an EDH will
develop as a delayed clot at a site where injury was
not previously suspected.
9.8.2
SUBDURAL HEMATOMA
These collections lie between the arachnoid and the

inner meningeal layer of the dura. They are most
commonly found over the convexity of the brain but
can also arise along the falx and the tentorium. They
are usually due to tearing of the relatively fragile veins
that cross the subdural space but they can arise
189
(b)
(a)
Figure 9.34 (a, b) Multiple gray/white interface hemorrhage with a tiny clot anterior to the fourth ventricle. Basal ganglia
and IVH complete the pattern of DAI.
Figure 9.35 This patient clinically has a primary diffuse

injury. CT shows low attenuation in the basal ganglia/
capsule region and tiny gliding contusions (arrow). This is
thought to represent non-hemorrhagic axonal-injuryinduced edema. Note the normal CSF spaces indicating no
increase in ICP.
Figure 9.36 Bifrontal asymmetric lentiform dense mature

EDH due to rupture of the anterior sagittal sinus.
190
IMAGING THE INJURY
Figure 9.37 Sudden deterioration in a previously stable

patient, 6 hours after head injury. The acute EDH is mostly
of lower attenuation than usual in an acute clot since it has
formed rapidly. Note the hyperdense inner rim. Is this the
original EDH formed earlier?
directly from adjacent severe contusions and subarachnoid lacerations. When contusions are associated
with an adjacent SDH the lesion is termed a burst
lobe. In order to measure the exact thickness of the
crescentic collection the CT image should be looked at
with a wide window to distinguish hyperdense clot
from bone (Figure 9.40).
The low signal generated by bone on MRI has
allowed us to see that a small SDH is almost always
present with moderate to severe contusions (Figure
9.41). However, a clinically significant SDH will not be
missed on CT. A thin SDH may be suspected only from
the compressive pattern it causes. In young people an
associated swollen hypointense ipsilateral hemisphere
(presumed to be ischemic) is a rare but classical finding
indicative of a very poor, usually fatal outcome (Figure
9.42). SPECT has been able to demonstrate diffuse
hypoperfusion superficially in the brain deep to SDH.
This hypoperfused tissue may well become hyperperfused in the subacute phase due to loss of autoregulation. (Figure 9.43). Severe continued hypoperfusion predicts eventual tissue loss. Similar findings
have been seen by xenon-enhanced CT CBF methods
(Chapter 11). The diffuse effects of the surface collection
can also be inferred from the hemispheric atrophy seen
(a)
(b)
Figure 9.38 (a) and (b) illustrate mixed attenuation biconvex EDHs. This appearance is now common as CT is carried out
soon after the injury when the clot is still forming. The low attenuation areas are the liquid blood, and the white represents
mature contracted clot.
Figure 9.39 Established mature hematoma which is a

subtemporal EDH, not an ICH.
(a)
191
Figure 9.40 Acute mixed attenuation SDH with continuing

hemorrhage associated with a small frontal ICH and
intraventricular blood. Tentorial herniation is indicated by
the dilated right lateral ventricle. A small EDH with a gas
bubble is present in the left occipital region at the fracture
site. The gas bubble implies a compound fracture, unless the
patient has had a lumbar puncture unlikely in this setting.
Gas-forming organisms can occur but should only be a late
complication.
(b)
Figure 9.41 Thin bilateral subdural hematomas causing generalized hemisphere compression without midline shift shown
on T2-weighted (a) and proton-density-weighted MRI (b). Axial sections both show hyperintense hemorrhage not obscured
by the bone-induced artefact found on CT.
192
IMAGING THE INJURY
(b)
(a)
Figure 9.42 (a, b) Extensive SAH with dilated right temporal horn secondary to a large SDH with 15 mm midline shift.
Note the loss of gray/white matter differentiation in the frontal regions, indicating severe ischemia.
(a)
(b)
Figure 9.43 (a) SPECT in the early subacute stage shows hyperperfusion in cortex adjacent to a thin SDH causing moderate
compression. (b) 6 months later, the SPECT shows hypoperfusion in a similar distribution to the previous
hyperperfusion.
Figure 9.44 Acute mixed attenuation SDH with tentorial

herniation and a contralateral ICH.
193
on follow-up CT. More frequently, SDH is now seen as

iso- or hypodense on CT in the acute phase (Figures
9.449.46). This is thought to be caused by acute
anemia and hemodilution sometimes associated with
resuscitation from multiple injuries and continued
subdural hemorrhage.
If a SDH is not removed or if the patient presents a
week or more after the bleed the clot will show
evidence of resolution: the hematoma becomes isodense with brain at approximately 710 days and
hypointense by 2130 days. Fresh hemorrhage can
also occur within a subacute SDH, which will give a
mixed or layered pattern of high and low attenuation
clot (Figure 9.47). On occasion bilateral SDH will be
present and the absence of midline shift may make
diagnosis more difficult (Figure 9.48). Fortunately, the
abnormal apparent increase in cortical thickness associated with symmetrical posterior displacement of the
anterior horns of the lateral ventricles and the compression or absence of the third ventricle ensure the
correct diagnosis even if neither SDH can be identified
as an alteration in attenuation.
MRI will always show a SDH separate from the
underlying brain and if at least two sequences are
(b)
(a)
Figure 9.45 (a, b) Typical acute non-uniform density SDH with extension along the ipsilateral side of the falx, posterior
displacement of the anterior horn, 1 cm of midline shift and contralateral temporal horn dilatation indicating tentorial
herniation.
194
IMAGING THE INJURY
(a)
(b)
Figure 9.46 (a) A complex acute SDH with a uniform high density outer layer and an inner mixed density layer with a fluid
level (arrow). (b) At a higher level an acute EDH is also present.
carried out one will always show that the collection has
a different signal to brain (Gomori et al., 1987; Hadley et
al., 1988). An EDH can easily be distinguished from a
SDH by the sharply delineated, low-intensity dura
lying between the hematoma and the displaced brain in
the former, while the latter is often associated with
displaced cortical pial veins demarcated by flow voids.
If there is doubt about the compartment involved or if
the full extent of an extra-axial collection has to be
assessed, the direct coronal imaging available with MRI
is ideal. Collections lying along the falx, the peritentorial space and along the floor of the middle fossa are
clearly demarcated while the volumetric perception of
those lying over the convexity is more accurately
assessed. Subdural collections thicker than a few
millimeters are ovoid rather than crescentic when
viewed in the coronal plane by MRI or CT.
9.8.3
SUBARACHNOID HEMORRHAGE
CT evidence of subarachnoid hemorrhage is found in

2533% of severely head-injured patients irrespective
of the age of the patient (LeRoux et al., 1992; European
Study Group on Nimodipine in Severe Head Injury,
1994; Figures 9.12, 9.20, 9.49, 9.50) and extensive SAH
Figure 9.47 Extensive acute on chronic SDH with gross

midline shift in an alcoholic patient with cerebellar
atrophy.
(a)
195
(b)
Figure 9.48 (a) The scan looks fairly normal but there is midline shift from left to right with posterior displacement of the
frontal horn of the left lateral ventricle. (b) The section above confirms the suspicion of bilateral isodense subdural
collections. The brain/subdural interface is arrowed.
Figure 9.49 Compound occipital fracture with underlying

hemorrhagic contusions with air and blood in the subarachnoid space.
Figure 9.50 Diffuse basal cistern SAH with blood in the

fourth ventricle and early hydrocephalus.
196
IMAGING THE INJURY
correlates with a poorer admission GCS (Shigemori,

Tokutomi and Hirohata, 1990; Demircivi et al., 1993). It
is usually found isolated in the basal cisterns after
skull-base fractures while SAH over the hemispheres
may be related to a vault fracture, local hemorrhage or
cerebral contusion. SAH may coexist with any type of
traumatic cerebral lesion and in patients with preexisting atrophy thick subarachnoid collections overlying a hemisphere may mimic a subdural hemorrhage. SAH localized around the tentorium is
increasingly recognized but its significance remains in
doubt. It is now clear that a worse outcome than that
expected from a particular cerebral injury occurs in
trauma complicated by SAH even if this is the only CT
abnormality (Eisenberg et al., 1990; Kakarieka, Braakman and Schakel, 1994). SAH occurs without overt
evidence of an increase in ischemic events and so
seems not to be related to vasospasm (Shigemori,
Tokutomi and Hirohata, 1990; Demircivi et al., 1993).
Occasionally, angiography may be required to exclude
aneurysm rupture or traumatic vessel damage.
MRI does not routinely show SAH in the acute stage
because of the high oxygen tension and the pulsating
flow effects in the CSF, but if sequences are specially
selected it can be demonstrated (Jenkins et al., 1988;
Noguchi et al., 1994; Figure 9.51). If there is sufficient
Figure 9.52 Superficial siderosis of the vermis and brain

stem due to recurrent trauma showing as a thin hypointense
outline of the pia on this heavily T2-weighted section.
clot left in the subacute stage, SAH will be seen as

hyperintense on T1-weighted images. As the hemorrhage associated with trauma is not usually extensive
or recurrent, the hemosiderin-laden macrophages
from maturing clot are washed away and the build-up
of superficial or ependymal siderosis giving a hypointense penciled outline on T2-weighted images is rare.
SAH is sometimes seen with the recurrent trauma
associated with alcoholics (Figure 9.52).
9.8.4
Figure 9.51 Acute basal cistern subarachnoid hemorrhage

following trauma seen only on the FLAIR (fluid attenuated
inversion recovery) sequence, where the normally hypointense CSF (e.g. fourth ventricle) has become hyperintense in
the prepontine cistern because of the increased protein
content secondary to the subarachnoid hemorrhage.
INTRAVENTRICULAR HEMORRHAGE (IVH)
This is relatively uncommon, occurring in only about

3% of cases overall but in about 10% of severe injuries
(Figure 9.53). It can occur as an isolated finding,
especially in old patients, where it is likely to be due to
rupture of a subependymal vein. In this situation the
outcome is usually favorable. IVH is more usually due
to breakthrough bleeding from adjacent hemorrhagic
contusions or an intracerebral clot, and the greater the
amount of IVH the lower the admission GCS (LeRoux
et al., 1992). It has been shown that IVH is commonly
associated with a callosal tear (Gentry, Thompson and
Godersky, 1988) and so in the correct clinical setting
IVH can act as one of the markers for DAI. Many of
these patients with IVH have a poor outcome, especially if a large volume of blood is present throughout
the ventricles or there is a cerebral lesion ipsilateral to
RAISED INTRACRANIAL PRESSURE AND HERNIATION
197
adjacent to a bone defect the examination should be

repeated following contrast opacification of the cisterns (CSF). This is achieved by the instillation of 5 ml
of 300 mgI/ml non-ionic contrast medium via a
lumbar puncture. The contrast is then run up to the
basal cisterns and positioned over the appropriate
portion of the skull base and the CT is repeated.
However, this is unlikely to be diagnostic unless the
leak is continuing.
9.10 Raised intracranial pressure and

herniation
Figure 9.53
sive IVH.
A thin SDH with developing ICH and exten-
the lateral ventricle filled with blood. This is not

related to complicating hydrocephalus, which is a rare
sequel, but the ICP is raised in almost half of these
patients (LeRoux et al., 1992). The poor outcome
associated with IVH was previously thought to be due
to the blood itself but it is much more likely that it
relates to the cause of the IVH.
9.9
One of the major factors affecting the severity of a

head injury is the presence or absence of raised
intracranial pressure (ICP); therefore some detailed
consideration of the radiological features of this entity
is appropriate (Chapter 10). The brain is enclosed
within the rigid confines of the skull between the
dural folds of the falx and tentorium. Brain shifts
secondary to focal mass lesions or generalized swelling tend to occur in a predictable way. These major
displacements can then cause additional ischemic or
even hemorrhagic damage. If the mass lesion is small
the shift will be localized: anterior contusions causing
displacement of the anterior horns of the lateral
ventricles posteriorly, temporal or middle fossa
masses causing displacement of the temporal horns
medially and general swelling or multiple unilateral
diffuse contusions causing effacement of the cerebral
sulci of the involved lobes even if the third ventricle is
not compressed or displaced. This is readily identified
when both hemispheres are compared (Figure 9.54).
Pneumocephalus
Air can enter any space or potential space around or

within the brain when there is a communication
between it and the surrounding atmosphere.
Although air can occasionally enter the head after a
compound fracture, it usually does so via a fracture
involving the paranasal sinuses, most often the frontal, the mastoid air cells or the middle ear. A CSF leak
is commonly present and the intracranial pressure is
usually low. If no CSF escapes, air will accumulate and
cause a mass lesion, often in the subdural space or
rarely as a pneumatocele in the brain itself.
An overt CSF leak is present in about 25% of
patients with CT-diagnosed pneumocephalus but in
the majority of these there is spontaneous cessation
within 710 days. If the leak persists or if there is
evidence of late post-traumatic meningitis then a
search for the fistula site should be made.
Direct coronal high-resolution CT of the anterior
fossa is used to investigate CSF rhinorrhea and similar
sections of the petrous bones in patients with CSF
otorrhea. If this fails to demonstrate an opaque sinus
Figure 9.54 Asymmetry of the sulci is obvious due in this

case to a thin SDH.
198
IMAGING THE INJURY
The fourth ventricle should always be identified. It

should be central and symmetrical in a reasonably
well-positioned axial scan. If it is displaced or compressed then a local cause for this should be sought
and a thin extracerebral collection should be excluded
by interrogating the scan at different window levels or

by carrying out MRI.
Large supratentorial masses give a shift pattern
which is a combination of lateral and downward
herniation. There is compression of the ipsilateral
(a)
(b)
(c)
(d)
Figure 9.55 SDH causing raised ICP and a transtentorial herniation cone. T2-weighted (a) and proton-density-weighted
MRI (b) axial sections show the compressed upper brain stem and edematous posteromedial temporal lobes. T1-weighted
coronal sections (c, d) show the mixed-signal SDH and bilateral uncal herniation through the tentorial notch with
compromised posterior cerebral arteries (arrows).
RAISED INTRACRANIAL PRESSURE AND HERNIATION
ventricle, shift of the midline structures below the falx

and dilatation of the contralateral trigone and temporal horn once the foramen of Monro is occluded. This
dilatation is often associated with some periventricular edema. These appearances have been correlated with the development of brain-stem compression signs (Stovring, 1977) and are therefore a
radiological sign of incipient transtentorial herniation. The third ventricle will be occluded by the time
this stage is reached. With CT, in the absence of
underlying atrophy, it is usually not possible to
identify the herniated uncus because the basal cisterns, which normally provide the CT contrast
between structures, are already obliterated. This is
noted particularly in young patients where obliteration of the basal cisterns is a consequence of downward herniation.
With modern CT scanning, no matter how young
the patient it should always be possible to identify the
third ventricle and the basal cisterns in a normal
situation. The ambient cistern is the smallest and most
easily compressed so it will be obliterated first,
followed by the quadrigeminal cistern. If there is
doubt about the state of the basal cisterns or if they
contain blood, then the state of the third ventricle will
give a reliable indication of the ICP. An increase in
edema in the cortex of the medial aspect of the
199
temporal lobe shown by hyperintensity on axial

T2-weighted MRI indicates that herniation is occurring while on coronal T1-weighted sections the anatomical distortion is easily demonstrated (Figure 9.55).
The degree of downward shift is less pronounced in a
patient with an atrophic brain such as is found in the
elderly or in the head-injury-prone alcoholic. It is this
downward shift that correlates with the harmful
effects of increased ICP, to the extent that even severe
lateral shift can be tolerated if the third ventricle and
the basal cisterns remain patent. If the brain is
atrophic, transtentorial uncal herniation can occur,
associated with contralateral temporal horn dilatation
but without obliteration of the cisterns, and the lowattenuation CSF contrast will outline the herniated
uncus (Figure 9.56). Even the ipsilateral temporal horn
may be seen herniating in extreme cases. With the
multiplanar sections available with MRI, subcallosal
and uncal herniation can be more easily appreciated
(Figure 9.55).
Technically it is not possible to show directly
tonsillar herniation by CT although this may be clearly
shown by sagittal sections on MRI. Isolated masses
within the posterior fossa can cause an acute obstructive hydrocephalus of the third and lateral ventricles
by compressing the fourth ventricle. The earliest sign
of developing hydrocephalus is dilatation of the
(a)
(b)
Figure 9.56 (a) The uncus (arrow) is seen pushed through the tentorial hiatus by the ipsilateral chronic SDH. Note how the
brain stem rotates as it is displaced. The dilated left temporal horn is usually the only indication of this as, in the absence
of cerebellar atrophy, there is insufficient CSF around the brain stem to outline the uncus. (b) A similar, though less severe,
example.
200
IMAGING THE INJURY
(a)
(b)
Figure 9.57 (a) Admission CT shows that the third ventricle and cisterns are both patent. Tiny right frontal and posterior
temporal hematoma are noted. (b) CT 6 hours later shows that there is now complete obliteration of the third ventricle and
compression of the basal cisterns indicating an ICP above 30 mmHg.
temporal horns which in normal young people are

seen only as narrow curved slits (Figure 9.16). Upward
tentorial herniation is unusual because pure infratentorial trauma is uncommon. On CT it can be
inferred from a pattern comprising obliteration of the
basal and supracerebellar cisterns with associated
hydrocephalus. Sagittal and coronal MRI again demonstrates this shift pattern directly.
When there are bilateral, similar-sized mass lesions
or if there is general swelling of the hemispheres then
the midline may remain undisplaced but a significant
downward herniation can still be present with an
associated raised ICP. In this case the third ventricle
and/or the basal cisterns will be obliterated (Teasdale
et al., 1984; Figure 9.57). With modern CT scanning, no
matter how young the patient, it should always be
possible to identify the third ventricle and basal
cisterns as low-attenuation CSF-containing spaces in
the normal situation. The sensitivity of this sign has
been used to classify head-injured patients (Marshall
et al., 1991). Controversy still surrounds the cause of
generalized hemisphere swelling, which is commonest in children and young adults. There is some
evidence to suggest that it is due to hyperemia (Bruce
et al., 1981) and some to ischemia, but measurements
of the Hounsfield numbers of the brain are not a
reliable or accurate method for such an assessment

and should not be used clinically.
9.11
Patterns of ischemia
Infarction and ischemia on CT are indicated by a

reduction in the X-ray beam attenuation giving a
darker than usual appearance of the brain. This change
in attenuation may be reversible and so with CT it is not
always possible to differentiate ischemia from infarction. Cerebral ischemia is a universal postmortem
finding in fatal head injuries but it is rarely demonstrated on CT in life. The commonest pattern of
ischemia seen is that associated with contusions, which
cause low attenuation due to cytotoxic edema. The
diffuse multifocal neuronal ischemic cell changes seen
commonly by the pathologist (Graham, Adams and
Doyle, 1987) cannot be specifically shown by routine
CT, although it has been demonstrated by xenon CT
CBF (Bouma et al., 1992) and on SPECT (Sakas et al.,
1991; Figure 9.58; Chapter 5).
Arterial vascular territory ischemia is the most
commonly recognized ischemic complication. It frequently involves the posterior cerebral artery ipsilateral to a mass lesion causing severe midline shift
due to compression of the artery against the tentorial
PATTERNS OF ISCHEMIA
Figure 9.58 SPECT shows multifocal regions of reduced

perfusion in the frontal and occipital lobes.
(a)
201
edge (Figures 9.59, 9.60). Pericallosal artery ischemia

can be produced in a similar way, although it is less
common and is reported to occur only if there is preexisting vascular disease. If middle cerebral territory
ischemia is present on CT or MRI damage to the
carotid artery in the neck or skull base should be
suspected. Local dissection or embolic thrombosis
from more proximal dissection may also be the cause
(Figure 9.61). Early angiography will identify lesions
that can be treated surgically, although treatment for
small-vessel dissection or thrombosis remains controversial after trauma (Figure 9.62). Ischemia in the
basilar territory is rarely seen in life on CT (Figure
9.63) but it may be seen as a diffuse low attenuation in
the brain stem and midbrain, usually contrasted
against the preserved normal density pattern of the
cerebellum. Vertebral dissection may show only as an
infarction in the ipsilateral PICA territory. Ischemia of
all types is much better visualized with MRI, especially that in the posterior fossa and medial aspects of
the temporal lobes. Well defined focal regions of
(b)
Figure 9.59 (a) Acute right SHD with blood extending over the falx anteriorly. Low attenuation is present throughout the
right occipital lobe, indicative of posterior cerebral artery compression with distal ischemia. (b) The ischemia is more
obvious. The ventricular distortion indicates a large isodense SDH as the cause.
202
IMAGING THE INJURY
(a)
(b)
Figure 9.60 (a, b) The low attenuation in the right temporal region may be partly due to operation and contusion, but there
is evidence of complete posterior cerebral artery ischemia. The brain stem also shows reduced attenuation. In addition,
pericallosal and middle cerebral artery ischemia is present, indicating carotid and basilar ischemia.
Figure 9.61 Diffuse, ill-defined low attenuation in the right

frontal lobe with loss of the gray/white differentiation. This
is most probably caused by damage to the carotid artery by
the basal fracture.
hyperintensity on T2- and proton-density-weighted

sequences with corresponding less prominent areas of
hypointensity on T1-weighted sections may be seen
(Figure 9.64).
Fracture involving a major venous sinus or secondary infection can produce dural venous sinus thrombosis, which can result in cortical venous infarction
(Satoh et al., 1993; Taha et al., 1993). On CT this shows
as peripheral mixed hemorrhagic high-attenuation
areas surrounded by edematous low attenuation.
These lesions are minimally space-occupying regions
and lie in a non-arterial territorial distribution. They
involve the white matter more than gray matter and
can often mimic gliding contusions with their parasagittal predilection. Both the venous occlusion and
the consequent infarction are better defined by MRI.
This will show an abnormal intermediate or hyperintense signal in the occluded sinus on routine
sequences, which normally show a clear flow void. In
the acute stage when hemorrhage is hypointense there
can be confusion if only a T2-weighted sequence is
performed in one plane; however the thrombosis can
be most convincingly demonstrated on a 3D time-offlight MR venogram (Figure 9.65). The area of infarc-
(a)
203
(b)
Figure 9.62 (a, b) 30-year-old car passenger after a side impact. No evidence of a head injury. A dense left hemiplegia
developed with a normal CT. Angiography showed an internal carotid dissection with distal branch middle cerebral
occlusion.
tion gives similar signals to an arterial infarct but is

less well defined with a pseudovasogenic pattern of
edema involving the white matter. There are usually
signs of hemorrhage, with signals that depend on the
timing of the scan. Soon after venous occlusion
intravenous contrast may show intense enhancement
of the adjacent dura and collateral cortical and central
veins.
If there has been global hypotension then secondary watershed ischemia can develop. This is seen on
CT as areas of reduced attenuation while on MRI
there is hyperintensity on T2- and proton-densityweighted images with hypointensity on T1-weighted
images. These changes are seen in the frontal region
at the watershed between the anterior and middle
cerebral artery territories, in the parafalcine region
and posterosuperiorly in the parietal region between
the middle and the posterior cerebral artery territories. Occasionally in a young patient with an acute
subdural hematoma there is diffuse low attenuation
on CT throughout the ipsilateral hemisphere. CT and
MRI show a general loss of contrast between gray
and white matter with loss of the CSF in the sulci
giving the hemisphere a bland appearance. This is
Figure 9.63 This temporal ICH is associated with some low

attenuation. The strikingly symmetrical bithalamic and
occipital low density is diagnostic of terminal basilar
ischemia in this patient with skull-base fracture.
204
IMAGING THE INJURY
(a)
(b)
Figure 9.64 Posterior cerebral artery territory infarction

due to coning caused by a left-sided SDH after failing
downstairs. (a, b) T2-weighted axial sections showing
hyperintense signal. (c) T1-weighted coronal section showing hypointense infarction in the posteromedial temporal
lobes.
(c)
invariably associated with gross hemisphere swelling

at operation and is thought to be due to panhemispheric ischemia, although the reason for its development is not known (Figure 9.66). If the patient
survives, follow-up imaging shows panhemispheric
atrophy. Profound persistent hypotension, ischemia
or a generally elevated ICP results in loss of the
normal gray/white matter differentiation. This
usually involves both hemispheres and is associated
with an apparent increase in the density or intensity
of the normal cerebellum and tentorium (Figure

9.67). These appearances are virtually limited to
children and young adults and are pathognomonic of
a non-perfused cerebrum. This appearance is often
associated with child abuse (the shaken baby syndrome).If severe hypoxia occurs in the absence of significant hypotension the sensitive basal ganglia are
the areas often rendered ischemic. This type of injury
is seen after near drowning and toxic inhalation
(Figure 9.68).
205
Figure 9.65 3D time-of-flight MR venogram showing a partially thrombosed superior sagittal sinus due to a depressed
fracture involving the sinus.
Regions of abnormally high and abnormally low

blood flow may be shown by SPECT acutely in the
same patient (Reid et al., 1990; Choksey et al., 1991;
Roper et al., 1991; Wyper et al., 1991) and do not
necessarily relate to lesions seen on CT or MRI. All
focal traumatic mass lesions such as contusions and
ICH, so long as they are within the resolution of the

SPECT system show zones of severely reduced RCBF
(Figure 9.8). These abnormally perfused areas involve
the surrounding brain but do not extend outside the
abnormal area of signal intensity seen on MRI or
abnormal density seen on CT. This hypoperfusion can
Figure 9.66 A thin surface collection: (?subdural, ?subarachnoid) is present, but the most striking feature is the
diffuse low attenuation throughout the cerebrum, with loss
of gray/white differentiation, indicating pancerebral
ischemia.
Figure 9.67 A thin SDH is responsible for the midline shift.

The apparent hyperdensity of the cerebellum is due to
bihemisphere ischemia despite preservation of the gray/
white matter differentiation.
206
IMAGING THE INJURY
agents (Orrison et al., 1994); however, planar or SPECT

scanning with 99mTc-HMPAO may be a better test. If
there is irreversible brain damage there will be no
uptake or retention of radionuclide, indicating a lack
of viable neuronal activity. The methodology is simple
and imaging can be repeated to demonstrate continuing lack of metabolism. So far false positives have not
been reported (de la Riva et al., 1992; Schlake et al.,
1992; Wieler et al., 1993).
9.13
Figure 9.68 Low density throughout the basal ganglia and

hippocampus indicates severe hypoxic damage in this
3-year-old. Note ventricular and cisternal obliteration secondary to local swelling.
persist from days to months in the edematous area

and is likely to be due to capillary compression caused
by the massive astrocyte swelling found on electron
microscopy at the edges of contusions and hematomas
(Bullock et al., 1992). In brain compressed by extraaxial lesions there is often a paradoxical increase in
perfusion, presumably due to a loss of autoregulation
(Figures 9.8, 9.43). Focal hyperemia has been found in
generally normal cerebral tissue in up to 42% of
patients. It occurs across a wide spectrum of head
injury severity and is usually an apparently benign
and transient phenomenon persisting for up to 2
weeks after the injury. Occasionally in individual
regions this hyperperfusion resolves to abnormal
hypoperfusion over several months, although the
brain retains its normal structural appearance on CT
or MRI. This may correlate with some of the late
neuropsychological sequelae found after head injury
(Wiedmann et al., 1988, 1990).
9.12 Radiology in the diagnosis of brain

death
Brain death is a clinical diagnosis, in which radiological tests need play no part. Cerebral angiography
will demonstrate failure of the contrast to progress
beyond the skull base. This can also be shown by
dynamic bolus contrast-enhanced CT or similar flowsensitive MRI techniques with or without contrast
Conclusion
Although MRI can be used as an alternative modality

in acutely head-injured patients when appropriate
monitoring and anesthetic support facilities are available, CT remains the first choice for detecting surgically
significant lesions in the first few days after injury. If the
patients clinical status does not correlate with the CT
findings MRI will often show the cause of the deficit.
This is especially so in patients who have suspected
diffuse axonal injury but have a normal CT. The higher
sensitivity of MRI may well show underlying hemorrhagic or non-hemorrhagic lesions, although it cannot
exclude the presence of microscopic axonal tears.
Patients may remain in a vegetative state, due to DAI,
with MRI appearances that are normal, or near normal,
in the first weeks. In the subacute and chronic stages
MRI is superior to CT, especially in the posterior and
middle fossae where its multiplanar visualization of
hemorrhagic lesions is unequaled. SPECT and xenon
CT CBF examinations are practical in patients with
severe head injury and can give valuable information
on cerebral perfusion. Each of these modalities has
helped to refine the treatment and prognostic assessment of the head-injured patient.
9.14
References
Annis, J. A. D., Finlay, D. B. L., Allen M. H. et al. (1987) A review of cervical

spine radiographs in casualty patients. British Journal of Radiology, 60,
10591061.
xenon-enhanced computerized tomography. Journal of Neurosurgery, 77,
360368.
Bradley, W. G. Jr (1993) MR appearance of hemorrhage in the brain. Radiology,
189, 1526.
Bruce, D. A., Alari, A., Bilannik, L. D. C. et al. (1981) Diffuse brain swelling
following head injuries in children. The syndrome of malignant brain
Bula, W. I. and Loes, D. J. (1994) Trauma to the cerebrovascular system.
Neuroimaging Clinics of North America, 4, 753772.
Bullock, R., Patterson, J. and Park, C. (1991) Evaluation of 99mTc-hexamethylpropyleneamine oxime cerebral blood flow mapping after acute focal
ischemia in rats. Stroke, 22, 12841290.
Bullock, R., Sakas, D., Patterson, J. et al. (1992) Early post-traumatic cerebral
blood flow mapping: correlation with structural damage after focal injury.
Acta Neurochirurgica, 55, 1417.
Choksey, M. S., Costa, D. C., Iannotti, F. et al. (1991) 99Tcm-HMPAO SPECT
de la Riva, A., Gonzales, F. M., Llamas-Elvira, J. M. et al. (1992) A diagnosis of
brain death: superiority of perfusion studies with 99mTc-HMPAO over
conventional radionuclide cerebral angiography. British Journal of Radiology,
65, 289294.
REFERENCES
Demircivi, F., Ozkan, N., Buyukkececi, S. et al. (1993) Traumatic subarachnoid
haemorrhage: Analysis of 89 cases. Acta Neurochirurgica, 122, 4548.
du Trevou, M. D. and van Dellen, J. R. (1992) Penetrating stab wounds to the
brain the timing of angiography in patients presenting with the weapon
already removed. Neurosurgery, 31, 905912.
Eisenberg, H. M., Gary, H. E., Aldrich, E. et al. (1990) Initial CT findings in 753
patients with severe head injury. A report from the NIH Traumatic Coma
European Study Group on Nimodipine in Severe Head Injury (1994) A
multicentre trial of the efficacy of nimodipine on outcome after severe head
Felber, S. R., Ettl, A. R., Birbamer, G. G. et al. (1993) MR imaging and proton
spectroscopy of the brain in post traumatic cortical blindness. Journal of
Magnetic Resonance Imaging, 3, 921924.
Gean, A. D. (1994) Imaging of Head Trauma, Raven Press, New York, p. 178.
Gentleman, D., Nath, F. and Macpherson, P. (1989) Diagnosis and management of delayed traumatic haematomas. British Journal of Neurosurgery, 3,
367372.
Gentry, L. R., Thompson, B. and Godersky, J. C. (1988) Trauma to the corpus
callosum: MR features. American Journal of Neuroradiology, 9, 11291138.
Gomori, J. M. and Grossman, R. I. (1987) Head and neck hemorrhage, in
Magnetic Resonance Annual, (ed. H. Y. Kressel), Raven Press, New York, pp.
71112.
Gomori, J. M., Grossman, R. I., Hackney, D. B. et al. (1987) Variable
appearances of subacute intracranial hematomas on high-field spin-echo
MR. American Journal of Neuroradiology, 8, 10191026.
Graham, D. I., Adams, J. H. and Doyle, D. (1987) lschaemic brain damage
in fatal non-missile head injury. Journal of Neurological Sciences, 39,
213234.
Gray, B. G., Ichise, M., Chung, D. G. et al. (1992) Technetium-99m-HMPAO
SPECT in the evaluation of patients with remote history of traumatic brain
injury: a comparison with X-ray computed tomography. Journal of Nuclear
Medicine, 33, 5258.
Hadley, D. M., Teasdale, G. M., Jenkins, A. et al. (1988) Magnetic resonance
imaging in acute head injury. Clinical Radiology, 39, 131139.
Jenkins, A., Teasdale, G., Hadley, M. D. M. et al. (1986) Brain lesions detected
by magnetic resonance imaging in mild and severe head injuries. Lancet, ii,
445446.
Jenkins, A., Hadley, D. M., Teasdale, G. M. et al. (1988) Magnetic resonance
imaging of acute subarachnoid haemorrhage. Journal of Neurosurgery, 68,
731736.
Jeret, J. S., Mandell, M., Anziska, B. et al. (1993) Clinical predictors of
abnormality disclosed by computed tomography after mild head trauma.
Jolesz, F. A. and Jones, K. M. (1993) Fast spin-echo imaging of the brain. Topics
in Magnetic Resonance Imaging, 5, 113.
Kakarieka, A., Braakman, R. and Schakel, E. H. (1994) Clinical significance of
the finding of subarachnoid blood on CT scan after head injury. Acta
Neurochirurgica, 129, 15.
Katayama, Y., Tsubokawa, T., Kinoshita, K. et al. (1992) lntraparenchymal
blood-fluid levels in traumatic intracerebral haematomas. Neuroradiology,
34, 381383.
Kushi, H., Katayama, Y., Shibuya, T. et al. (1994) Gadolinium DTPA-enhanced
magnetic resonance imaging of cerebral contusions. Acta Neurochirurgica
(Supplement), 60, 472474.
Lang, D. A., Hadley, D. M., Teasdale, G. M. et al. (1990) Gadolinium DTPA
enhanced magnetic resonance imaging in acute head injury. Acta Neurochirurgica, 51, 293295.
Lassen, N. A., Andersen, A. R., Friberg, L. et al. (1988) The retention of
[99mTc]-d,I-HMPAO in the human brain after intracarotid bolus injection
a kinetic analysis. Journal of Cerebral Blood Flow and Metabolism, 8,
S13S22.
Lear, J. L. (1988) Quantitative local cerebral blood flow measurements with
technetium-99m HMPAO: evaluation using multiple radionuclide digital
quantitative autoradiography. Journal of Nuclear Medicine, 29, 13871392.
LeRoux, P. D., Haglund, M. M., Newell, D. W. et al. (1992) Intraventricular
hemorrhage in blunt head trauma: an analysis of 43 cases. Neurosurgery, 31,
678685.
MacLennan, A. C. and Hadley, D. M. (1994) Radiation dose to the lens from
CT scanning in a neuroradiology department. British Journal of Radiology, 68,
1922.
75, S14S20.
207
Miller, J. D. (1990) Assessing patients with head injury. British Journal of

Surgery, 77, 241242.
Mittl, R. L., Grossman, R. I., Hiehle, J. F. et al. (1994) Prevalence of MR
evidence of diffuse axonal injury in patients with mild head injury and
normal head CT findings. American Journal of Neuroradiology, 15,
15831589.
Noguchi, K., Ogawa, T., lnugami, A. et al. (1994) MR of acute subarachnoid
hemorrhage: a preliminary report of fluid-attenuated inversion-recovery
pulse sequences. American Journal of Neuroradiology, 15, 19401943.
Orrison, W. W., Champlin, A. M., Kesterson, O. L. et al. (1994) MR hot nose
sign and intravascular enhancement sign in brain death. American Journal
of Neuroradiology, 15, 913916.
Rao, N., Turski, P. A., Poicyn, R. E. et al. (1984) 18F Positron emission
computed tomography in closed head injury. Archives of Physical Medicine
and Rehabilitation, 65, 780785.
Reid, R. H., Gulenchyn, K. Y., Ballinger, J. R. et al. (1990) Cerebral perfusion
imaging with technetium-99m HMPAO following cerebral trauma: initial
experience. Clinical Nuclear Medicine, 15, 38388.
Roper, S. N., Mena, I., King, W. A. et al. (1991) An analysis of cerebral blood
flow in acute closed-head injury using technetium-99m-HMPAO SPECT
and computed tomography. Journal of Nuclear Medicine, 32, 16841687.
Sakas, D., Patterson, J., Bullock, R. et al. (1991) Focal post-traumatic
hyperaemia a benign process? Journal of Cerebral Blood Flow and
Metabolism, 2, S834.
Satoh, H., Uozumi, T., Kiya, K. et al. (1993) Venous thrombosis after closed
head injury: a report of two cases presenting as intracranial hypertension.
Neurological Surgery, 21, 953957.
Schlake, H. P., Bottger I. G., Grotmeyer, K. H. et al. (1992) Determination of
cerebral perfusion by means of planar scintigraphy and 99mTc-HMPAO in
brain death, persistent vegetative state and severe coma. Intensive Care
Medicine, 18, 7681.
Shigemori, M., Tokutomi, T. and Hirohata, M. (1990) Clinical significance of
traumatic subarachnoid haemorrhage. Neurologia Medico-Chirurgica (Tokyo),
30, 396400.
Slavin, G. S., Butts, K., Rydberg, J. N. et al. (1995) Dual-echo interleaved echoplanar imaging of the brain. Magnetic Resonance in Medicine, 33, 264270.
Stein, S. C., Spettell, C., Young, G. et al. (1993) Delayed and progressive brain
injury in closed-head trauma: radiological demonstration. Neurosurgery, 32,
2531.
Stovring, J. (1977) Contralateral temporal horn widening in unilateral
supratentorial mass lesions: A diagnostic sign indicating tentorial herniation. Journal of Computer Assisted Tomography, 1, 319323.
Taha, J. M., Crone, K. R., Berger, T. S. et al. (1993) Sigmoid sinus thrombosis
after closed head injury in children. Neurosurgery, 32, 541545.
Teasdale, E., Cardosa, E., Galbraith, S. et al. (1984) CT scan in severe diffuse
head injury: physiological and clinical correlations. Journal of Neurology,
Teasdale, G., Murray, G., Anderson, E. et al. (1990) The risks of intracranial
haematoma after head injury in adults and children. British Medical Journal,
300, 363367.
Teasdale, G., Teasdale, E. and Hadley, D. (1992) Computed tomographic and
magnetic resonance imaging classification of head injury. Journal of
Trott, J. and David, D. J. (1995) Definitive management principals, priorities
and basic technique, in Cranio-maxillofacial Trauma, (eds D. J. David and, D.
A. Simpson), Churchill Livingstone, Edinburgh, pp. 233250.
Vink, R. (1993) Nuclear magnetic resonance characterization of secondary
mechanisms following traumatic brain injury. Molecular Chemistry and
Neuropathology, 18, 279297.
White, C. S. and Mirvis, S. E. (1995) Pictorial Review: Imaging of traumatic
aortic injury. Clinical Radiology, 50, 281287.
Wiedmann, K. D., Patterson, J., Hadley, D. M. et al. (1988) Correlates of focal
lesions in closed head injury: distribution of cerebral blood flow and
neuropsychological performance. Journal of Clinical and Experimental Neuropsychology, 11, 158.
Wiedmann, K. D., Wilson, J. T. L., Wyper, D. et al. (1990) SPECT cerebral blood
flow, MR imaging and neuropsychological findings in traumatic head
injury. Neuropsychology, 3, 267281.
Wieler, H., Marohl, K., Kaiser, K. P. et al. (1993) Tc-99m HMPAO cerebral
scintigraphy. A reliable, noninvasive method for determination of brain
death. Clinical Nuclear Medicine, 18, 104109.
Wyper, D. J., Sakas, D., Bullock, R. et al. (1991) Traumatic cerebral contusions
cause severely reduced perifocal CBF and ischaemic damage. Journal of
10
MONITORING
Brian North
10.1
Introduction
Intracranial pressure (ICP) measurement is an

extremely important part of the neurosurgical armamentarium. Not only is raised ICP the commonest
cause of death in neurosurgical patients, it is
extremely common in patients suffering from head
injury. In this latter group, 40% of patients who are
admitted in an unconscious state have raised ICP, and
in 50% of those who die, raised ICP is the main cause
(Miller et al., 1977). Numerous investigations have
shown that sustained intracranial hypertension is
associated with a poor prognosis (Becker et al., 1977;
Johnston and Jennett, 1973; Langfitt, 1969; Narayan et
al., 1982; Marshall, Smith and Shapiro, 1979a; Miller et
al., 1981). The effective treatment of high ICP has been
shown to decrease mortality (Marshall, Smith and
Shapiro, 1979a, b). Even so, ICP must be measured
before a diagnosis of raised ICP can be made.
Obviously, an understanding of the principles of
ICP measurement is an important prerequisite to
considering the disturbances of brain function that
follow head injury.
The quest for developing the ideal method of
recording ICP has been a difficult exercise. The first
requirement for any method is that it is accurate; and
it should also be safe and simple (North and Reilly,
1990). The search is still incomplete, since all current
methods are invasive. The necessity to breach the skull
to record ICP has resulted in a significant number of
neurosurgeons being reluctant to embrace this technique. It took at least 15 years before ICP monitoring
became fully accepted into clinical neurosurgical
practice in more than a few centers. Even now,
opinions vary as to the value of the technique, from
those who claim that it makes no difference to the
outcome of any neurosurgical disease to those who
assert that it is an indispensable part of neurosurgical
practice, without which many patients would surely
die. The truth lies somewhere between these two
extremes and it depends on the facilities and personnel available in any given neurosurgical unit (Miller,
1987).
10.2
Historical aspects
10.2.1
LUMBAR PUNCTURE
Lumbar puncture was introduced into clinical medicine in 1897 (Quincke, 1897) and following this, the
spinal CSF pressure was used as an indirect measure
of ICP. CSF pressure is defined as the pressure just
necessary to prevent escape of fluid into a needle
introduced into the lumbar subarachnoid space.
Sharpe published a monograph on head injury in
1920 and stated that his principal indication for the
operation of subtemporal decompression was a spinal
fluid pressure above 15 mmHg (Sharpe, 1920). Jackson
also advocated the use of lumbar puncture and
pressure measurement in head injury in 1922 (Jackson,
1922), but there was much disagreement on the place
and dangers of lumbar puncture, and the reliability of
the procedure in accurately measuring ICP. Most
authors did agree that a pressure in excess of
200 mmH2O was definitely abnormal.
The two principal objections to lumbar puncture in
the diagnosis of intracranial hypertension have been
the danger of inducing brain-stem compression
through tentorial or tonsillar herniation and the
contention that spinal fluid pressure is not always an
accurate reflection of ICP. Langfitts work was particularly important in demonstrating this lack of correlation between ICP and spinal CSF pressure under
conditions of high ICP (Langfitt et al., 1964).
10.2.2
VENTRICULAR PUNCTURE
Ventricular puncture for the relief of increased ICP is

one of the oldest practices in neurosurgery. Pressure
measurements during this procedure were often done
210
INTRACRANIAL PRESSURE MONITORING
The development of strain gauges allowed ICP measurement to be performed directly using a ventricular
catheter and an external transducer. The pioneering
neurosurgeons in its development were Janny (Guillaume and Janny, 1951) and Lundberg (1960). Since
then, the technique has been widely adopted, with
some variations which will be discussed later.
In a commonly used transducer such as the Statham

P23 series, four strain-sensitive wires are connected to
two frames, one of which fits inside the other. The
wire frame is attached to the diaphragm of the
transducer (upon which the pressure acts) and the
outer frame is fixed (Figure 10.1). The set of wires form
a Wheatstone bridge network (Figure 10.2), which is
energized by a direct current (e.g. 10 V DC). A stable
DC amplifier is used to detect the imbalance of the DC
bridge, and this signal can then be used to drive a pen
recorder or be displayed on an oscilloscope. It goes
without saying that the staff in an intensive care unit
must be familiar with calibrating and maintaining pen
recorders.
10.3
10.4
but prolonged pressure measurements were performed infrequently because water and mercury
manometers were cumbersome and also because of
the risk of intracranial infection.
10.2.3
Transducers
Transducers for measuring pressure are based on

strain gauges, which were originally developed by
engineers who needed to be able to measure the effects
of tension and compression in beams. The applied
force (per unit area) is called the stress and the
resulting increase in length (per unit length) is called
the strain. The operation of a wire strain gauge
depends on the fact that if a length of wire is stretched,
its electrical resistance will increase, and vice versa.
Methods of measuring ICP
Intracranial pressure varies widely from minute to

minute, particularly when it is raised, so that single
observations can be misleading. Therefore it is
important to examine a continual record of ICP, so
that waves of raised ICP are not missed. Nevertheless, achieving this ideal poses problems not
least in storing large quantities of chart paper. Marmarou found that the ICP level recorded by the ICU
nurses at the end of each hour (end hour recording)
Figure 10.1
A Statham pressure transducer.
Figure 10.2
An unbonded strain gauge; four active elements.
METHODS OF MEASURING ICP
of ICP was a good estimate of ICP for the entire

hour, finding that 83% of the paired observations of
computer- and nurse-recorded ICP values differed by
less than 6 mmHg (Marmarou et al., 1991). We use
the end hour recording for data entry in research
projects and visual scanning of the chart output in
routine cases.
10.4.1
INTRAVENTRICULAR PRESSURE RECORDING
The methodology for measuring ICP has evolved

progressively, with many workers preferring a fluid
coupled system using a ventricular catheter and an
external transducer, considering it to be the gold
standard of ICP measurement (Miller, 1989). Ventricular ICP recording is the most reliable method in
current use and it has the advantage of minimal
expense and maximal accuracy, since the external
transducer can be calibrated against an external
reference at any time. The equipment required is
commonplace in any intensive care unit.
The reference point for an external transducer
should be the foramen of Monro, because it is close to
the center of the head 2 cm above the pterion is a
rough guide to its surface marking. The midpoint of a
line joining the two external auditory meati is another
suitable reference point, although somewhat posterior
to the interventricular foramen. Other workers use the
external auditory meatus (Kosteljanetz, 1987). Whatever reference point is used, the level of an external
transducer needs to be altered with each change in
head position.
The ventricular method obviously requires placement of a catheter into a lateral ventricle, and this may
be a technically difficult procedure because of a
narrow or displaced ventricle. Injury of the basal
ganglia can occur with ill directed or over enthusiastic
attempts at ventricular cannulation. The infection rate
in our hands is 3.6%, reaching a potentially serious
level after three days (North and Reilly, 1986). Other
quoted infection rates range from less than one per
cent to more than 5% (Rosner and Becker, 1976;
Sundbarg et al., 1972). A big advantage of the
ventricular method is that CSF can be withdrawn to
lower ICP.
All joints in the recording system must be watertight. If they are not, micro-leaks will invalidate the
pressure recordings. Each portion of the system must
be tested periodically by isolating the external system
from the patient temporarily and subjecting it to a
pressure head of about 50 mmHg. After being isolated,
the external system should maintain its intraluminal
pressure and if not, the connections must be tightened
or the system discarded and replaced with a water
tight system (Shields, McGraw and Garretson, 1984).
Sometimes, ventricular catheters block and this can
211
be overcome by flushing a small amount of sterile

saline through the system. However, repeated flushing should be avoided because of the real risk of
infection.
10.4.2
OTHER METHODS
The hollow skull bolt (Richmond screw) has been

widely used in many centers (Vries, Becker and
Young, 1973; Figure 10.3). There have been many
modifications to achieve a lower profile, CT compatibility, more side holes (Leeds screw), and a pediatric
version (Coroneos et al., 1973; James, Bruno and Schut,
1975; Landy and Villanueva, 1984; Mann and Yue,
1988). These devices are simple to insert but they have
a tendency to block and so produce a damped,
inaccurate trace. At high pressures, the subdural bolt
tends to read lower than a ventricular catheter
(Mendelow et al., 1983; Miller, Bobo and Knapp, 1986).
This question of accuracy presents a major problem
and is the main reason why the hollow bolt method
has declined in popularity (Miller, 1987). Subdural
catheters can be useful where the ventricle cannot be
cannulated, but they are also likely to underestimate
the true ICP.
The extradural site for monitoring has been used
and has the advantage avoiding penetration of the
dura. However, there are technical problems associated with the inelasticity of the dura and the need for
the transducer to lie flat (co-planar) on the dura.
Unfortunately, irregularities of the dura and inner
table of the skull are common. If co-planarity is not
achieved, stresses and strains in the dura can distort
the measurements and falsely record high pressure
(Dorsch and Symon, 1975; Coroneos et al., 1973). For
these reasons concerning accuracy, the extradural
method is now used very infrequently.
Figure 10.3
A subarachnoid screw.
212
10.4.3
CATHETER-TIP TRANSDUCER
We have used catheter-tip transducers for several

years and this is currently our preferred method of
recording ICP. Miniature implantable transducers
have been developed from intravascular transducers,
of which the Camino transducer is an example (Figure
10.4). Pressure is measured at the tip of a narrow
fiberoptic catheter where there is a flexible diaphragm.
Light is reflected off the diaphragm and these changes
in light intensity are interpreted in terms of pressure.
The outside diameter of the device is only 4 FG
(1.3 mm). The system is not dependent on a fluid
column, or on an external transducer where the height
needs constant readjustment depending on the level of
the patients head. Ostrup et al. (1987) and Crutchfield
et al. (1990) have reported excellent results but cost is
still a problem. There is a close correlation between
ICP measured by the Camino catheter-tip transducer
and the intraventricular method (Gambardella,
dAvella and Tomasello, 1992). The Innerspace transducer is a similar type of a fiberoptic catheter-tip
transducer, but the physical principle uses spectral
frequency. Marmarou has reported both experimental
(Yoshihara et al., 1993) and clinical (Marmarou et al.,
1994) tests on this transducer.
The main limitation of a catheter-tip transducer is
that it is not possible to calibrate it in situ and it should
be replaced if monitoring is to be continued for longer
than 5 days, because of possible drift. They are simple
to insert and we place the tip in the brain at a depth of
Figure 10.4
A catheter-tip transducer.
12 cm. The fiberoptic cable can be damaged by

restless patients or if it is bent acutely, and this
fragility is a practical problem and is one that limits
the usefulness of the method.
10.4.4
IMPLANTED MICROCHIP TRANSDUCER
Implanted microchip sensors have now been developed and an example is the Codman MicroSensor
transducer. It consists of a miniature solid state
pressure sensor (Figure 10.5) mounted in a very small
titanium case (diameter 1.2 mm = 3.6 FG) at the tip of
a 100 cm long flexible nylon tube (diameter 0.7 mm =
2.1 FG). The transducer tip contains a silicon microchip with diffused piezoresistive strain gauges which
are connected by wires in the nylon tube to complete
a Wheatstone bridge type circuit. When the transducer
is energized and pressure is applied, the silicon
diaphragm deflects a small amount (less than
0.001 mm3 for 100 mmHg applied pressure), inducing
strain in the embedded piezoresistors. This resistance
change is reflected in the form of a differential voltage
which is then converted into units of pressure, i.e.
millimeters of mercury. The bottom layer of the silicon
diaphragm is vented to the atmosphere along the
nylon tube, while the top layer is exposed to the
applied CSF or brain tissue pressure.
The microsensor transducer can be inserted directly
into the brain parenchyma but is also fine enough to
be passed through a catheter into the lateral ventricle.
Narayan and his colleagues found that this device had
WHICH SYSTEM?
Figure 10.5
A microchip transducer.
an average drift of less than 1 mmHg over a 9-day

period (Gopinath et al., 1993, 1994). This group
(Gopinath et al., 1995) also tested the Codman transducer in 25 patients, comparing it against a ventricular
catheter and an external transducer. Encouragingly
low levels of baseline drift were found and it showed
no tendency to under-read or over-read. Piper and
Miller (1995) evaluated the wave-form analysis capability of this transducer against a fluid coupled
transducer. They found that there were no significant
differences between the two transducers. Actually the
microchip transducer has a superior frequency
response, although this may not be clinically important for wave form analysis.
A variation of a ventricular catheter with a pressure
measuring transducer located at the tip of the catheter
is the Ventcontrol MTC (Piek and Raes, 1996). This
technology represents the direction in which ICP
monitoring will evolve.
10.5
213
Which system?
It is not easy to decide which system of ICP measurement is the best, because of the large number of
variables, including cost. If access to the ventricle is
required, then a ventricular catheter and external
transducer is both cost-effective and reliable; it is the
gold standard. However, most patients now being
monitored for ICP are likely to be suffering from head
injury and they will usually have narrow ventricles,
making cannulation potentially difficult for a young
neurosurgeon. In the head-injury situation, the preferred method is either a fiberoptic catheter-tip transducer (e.g. Camino or InnerSpace) or an implantable
transducer (e.g. Codman) inserted into brain parenchyma, and this can be done at the bedside very
simply. The choice between these two types of
transducer largely comes down to a question of cost,
which varies from country to country and is a
individual decision. The Camino and InnerSpace
transducers require a more expensive control monitor
(US$5000) whereas the Codman control unit is less
expensive (US$500). At the present time, the disposable transducer kit costs about the same for each of
these three products.
Pickard and his colleagues have examined the
Camino, Codman and InnerSpace transducers in a
pressure-flow test rig designed for the assessment of
hydrocephalus shunts (Czosnyka, Czosnyka and Pickard, 1996). They measured long-term and temperature
zero drifts, frequency response characteristics, the
accuracy of measurement of static and pulsatile
pressures and the slew rate. All three transducers
scored satisfactorily during bench testing, and performed according to the manufactures specifications,
giving high quality readings in testing conditions. The
Codman transducer scored best overall. Pickards
ranking is shown in Table 10.1.
I believe that the time has come to recommend that
implantable and catheter-tip transducers should
replace fluid coupled systems. The main disadvantage
of catheter-tip transducers is that they cannot be
214
Table 10.1 Results of testing ICP measurement transducers (1 = better; 2 = average or as good as another having the
same score; 3 = worse; these values were measured relative to the Camino transducer)
24-hour zero drift at 0 mmHg

24-hour zero drift at 20 mmHg
Temperature drift (2740C)
Accuracy in static measurement
Accuracy in measurement of pulse pressure
Bandwidth
Slew rate
calibrated in vivo, but this deficiency appears to be of

little practical importance.
10.6
Interpretation of ICP monitoring
It is conventional to calibrate ICP monitoring apparatus in units of mmHg to permit a direct comparison of
ICP with blood pressure and to enable the difference
between the two pressures (CPP) to be calculated.
ICP records offer two main kinds of information, the
baseline level and variations of the pressure, i.e.
waves. In other words, raised ICP may be steady or
periodic.
10.6.1
BASELINE PRESSURE
Normal ICP is pulsatile due to intracranial arterial

pulsations reflecting the cardiac and respiratory cycles.
Based on largely intuitive considerations, the normal
level of mean ICP is 010 mmHg and it is abnormal
over 15 mmHg. Lundberg suggested that mean levels
above 20 mmHg are moderately elevated and that
sustained levels above 40 mmHg are severely increased
(Lundberg, 1960). In head injury, it is more common to
observe a rise in baseline pressure, rather than waves of
raised ICP. If the bone flap has been removed surgically,
pressure readings can be unreliable.
10.6.2
PRESSURE WAVES
Lundberg identified three different types of ICP

variations, A, B and C waves (Lundberg, 1960).
Plateau waves (A waves) are clinically very important
because they indicate dangerously reduced intracranial compliance. They rise steeply from near normal
or slightly raised ICP to 50 mmHg or more and persist
for 520 minutes before falling precipitously, even to
below the original level. Although named for their
rather flat tops, there may be irregularities and peaks.
The most frequent type of pressure wave, although
of less adverse clinical significance than the plateau
wave, is the B wave. These are rhythmic oscillations,
sharply peaked and occurring once every 12 minutes, in which mean ICP rises in a crescendo manner
Camino
Codman
InnerSpace
2
2
3
1
2
2
2
2
2
2
2
1
2
1
3
3
3
3
from a variable baseline to a level 2030 mmHg

higher, and then falls abruptly with no intervening
period of sustained intracranial hypertension. C
waves seem to be of little clinical significance.
10.6.3
PULSE AMPLITUDE
As ICP increases above the resting level, the amplitude of the cardiac pulse component increases while
the relative magnitude of the respiratory component
may decrease. Thus ICP pulse amplitude increases
linearly with increases in ICP, an observation made by
Cushing over 90 years ago (Cushing, 1902). Pulse
pressure may also increase before mean ICP rises. This
has clinical importance as it may allow prediction of
deterioration before ICP rises. In other words, a
widening pulse amplitude in the absence of an
increased ICP indicates an impairment of intracranial
compliance or reserve.
10.6.4
INTRACRANIAL PRESSURE WAVE FORM
The ICP wave has a pulsatile quality at two different

frequencies one synchronous with the arterial pulse
while the other is slower, in time with breathing
(Figure 10.6).
The vascular waves are caused by arterial pulsations in the large vessels within the brain, producing
an oscillation in the volume of the ventricular system
(Bering, 1955). The shape of the CSF pressure wave is
similar to that of systemic blood pressure and it has
three fairly consistent components, the percussion
wave (P1), tidal wave (P2) and dicrotic wave (P3;
Figure 10.7). The dicrotic notch between P2 and P3
corresponds to the dicrotic notch of the arterial
pulsation. The respiratory wave is synchronous with
alterations in central venous pressure, reflecting intrathoracic pressure. They are seen prominently in
patients on ventilators. Normally, the amplitude of the
cardiac pulse is about 1.1 mmHg, and the combined
cardiac and respiratory variation is approximately
3.3 mmHg (Bradley, 1970).
Analyzing cerebrovascular pressure transmission by
applying Fourier analysis to both the arterial and
CONCLUSION
Figure 10.6
ICP waveform.
intracranial pressure waveforms has been examined as

a method of assessing cerebrovascular pathophysiology (Chopp and Portnoy, 1980). Experimental studies
have determined changes in the low-frequency component of the cerebrovascular pressure transmission
spectrum to be associated with vascular mechanisms
underlying raised intracranial pressure while nonvascular mechanisms are associated with increased
transmission of both low and high frequency components (Portnoy and Chopp, 1981; Takizawa, GabraSanders and Miller, 1986; Kasuga, Nagai and Hasegawa, 1989). Clinical studies have also demonstrated
specific patterns of cerebrovascular pressure transmission occurring in patients with severe head injury (Bray
et al., 1986; Piper et al., 1990). Although initial studies of
the high frequency band (415 Hz) demonstrated a
correlation with mortality (Robertson et al., 1989),
recent work has shown this frequency spectrum to be
affected significantly by heart rate (Contant et al., 1993).
Analysis of low-frequency pressure transmission may
provide a clinically useful indicator of the presence or
absence of autoregulation, based on experimental work
examining phase shift characteristics between the
fundamental of the BP to ICP waveform (Piper et al.,
Figure 10.7
215
CSF pressure wave.
1993). It may also provide a means of identifying

optimal CPP thresholds by identifying autoregulatory
breakpoints (Takizawa, Gabra-Sanders and Miller,
1987). Although this work is promising further development is required before cerebrovascular pressure
transmission analysis becomes a clinically useful
analytical tool.
10.7
Conclusion
ICP monitoring has developed into a very useful tool,

particularly in patients suffering from head injury. If a
decision is made to monitor ICP, then certain standards must be achieved so that reliance can be placed
on the data which are obtained. We have much
confidence in both the catheter-tip transducer and the
implanted microchip transducer, with the output
being directed to a slow-moving chart recorder so that
the paper trace of ICP is available to all members of
the team caring for the patient. Other physiological
variables such as arterial BP are also recorded whenever possible. We regard the catheter-tip and implanted microchip transducers as having replaced the
ventricular catheter and external transducer as the
gold standard in ICP measurement.
ICP monitoring provides the only sure way of
confirming or excluding intracranial hypertension. If
this is present, ICP monitoring provides the only
reliable method of assessing whether therapy works
and provides an early opportunity of switching to an
alternative therapy should treatment fail. If increased
ICP is not present, potentially dangerous treatment can
be avoided. If the patient is paralyzed or heavily
sedated, conventional neurological observation is useless and ICP monitoring provides a means of determining the patients cerebral perfusion pressure (CPP) and
an index of cerebral function (Sullivan et al., 1977).
216
10.8
References
Becker, D. P., Miller, J. D., Ward, J. D. et al. (1977) The outcome from severe head
injury with early diagnosis and intensive management. Journal of Neurosurgery 47, 491502.
Bering, E. A. (1955) Choroid plexus and arterial pulsation of cerebrospinal
fluid. AMA Archives of Neurology and Psychiatry, 73, 165172.
Bradley, K. C. (1970) Cerebrospinal fluid pressure. Journal of Neurology,
Bray, R. S. Jr, Sherwood, A. M., Halter, J. A. et al. (1986) Development of a clinical
monitoring system by means of ICP waveform analysis, in Intracranial
Pressure VI, (eds J. D. Miller, G. M. Teasdale and J. O. Rowan), SpringerVerlag, Berlin, pp. 260264.
Chopp, M. and Portnoy, H. D. (1980) Systems analysis of intracranial pressure.
Comparison with volumepressure test and CSF-pulse amplitude analysis.
Contant, C. F., Robertson, C. S., Narayan, R. K. et al. (1993) Effects of heart rate
on the shape of the intracranial pressure wave and related parameters, in
Intracranial Pressure VIII, (eds C. J. J. Avezaat, J. H. M. van Eindhoven, A. I. R.
Maas and J. T. J. Tans), Springer-Verlag, Berlin, pp. 356362.
Coroneos, N. J., McDowall, D. G., Gibson, R. M. et al. (1973) Measurement of
extradural pressure and its relationship to other intracranial pressures. An
experimental and clinical study. Journal of Neurology, Neurosurgery and
Crutchfield, J. S., Narayan, R. K., Robertson, C. S. and Michael, L. H. (1990)
Evaluation of a fiberoptic intracranial pressure monitor. Journal of Neurosurgery, 72, 482487.
Cushing, H. (1902) Some experimental and clinical observations concerning
states of increased intracranial tension. American Journal of Medical Science,
124, 375400.
Czosnyka, M., Czosnyka, Z. and Pickard, J. D. (1996) Laboratory testing of three
intracranial pressure microtransducers: technical report. Neurosurgery, 38,
219224.
Dorsch, N. W. C. and Symon, L. (1975) The validity of extradural measurement
of the intracranial pressure, in Intracranial Pressure II, (eds N. Lundberg, U.
Ponten and M. Brock), Springer-Verlag, Berlin, pp. 403408.
Gambardella, G., dAvella, D. and Tomasello, F. (1992) Monitoring of brain
tissue pressure with a fiberoptic device. Neurosurgery, 31, 918922.
Gopinath, S. P., Cherian, L., Robertson, C. S. et al. (1993) Evaluation of a
microsensor intracranial pressure transducer. Journal of Neuroscience Methods,
49, 1115.
Gopinath, S. P., Robertson, C. S., Narayan, R. K. and Grossman, R. G. (1994)
Evaluation of a Microsensor Intracranial Pressure Transducer, in Intracranial
Pressure IX, (eds H. Nagai, K. Kamiya and S. Ishii), Springer-Verlag, Berlin,
pp. 24.
Gopinath, S. P., Robertson, C. S., Contant, C. F. et al. (1995) Clinical evaluation of
a miniature strain-gauge transducer for monitoring intracranial pressure.
Guillaume, J. and Janny, P. (1951) Manometrie intracranienne continue. Interet
de la methode et premiers resultats. Revue de Neurologie, 84, 131142.
Jackson, H. (1922) The management of acute cranial injuries by the early, exact
determination of intracranial pressure, and its relief by lumbar drainage.
Surgery, Gynecology and Obstetrics, 34, 494508.
James, H. E., Bruno, L. and Schut, L. (1975) Intracranial subarachnoid pressure
monitoring in children. Surgical Neurology, 3, 313315.
Johnston, I. H. and Jennett, B. (1973) The place of continuous intracranial
pressure monitoring in neurosurgical practice. Acta Neurochirurgica (Vienna),
29, 5363.
Kasuga, Y., Nagai, H. and Hasegawa, Y. (1989) Transmission characteristics of
pulse waves in the intracranial cavity of dogs during normal intracranial
condition, intracranial hypertension, hypercapnia and hydrocephalus, in
Berlin, pp. 196200.
Kosteljanetz, M. (1987) Intracranial pressure: cerebrospinal fluid dynamics and
pressurevolume relations. Acta Neurologica Scandinavica (Supplement), 111,
123.
Landy, H. J. and Villanueva, P. A. (1984) An improved subarachnoid screw for
intracranial pressure monitoring. Technical note. Journal of Neurosurgery, 61,
606608.
Langfitt, T. W. (1969) Increased intracranial pressure. Clinical Neurosurgery, 16,
436471.
Langfitt, T. W., Weinstein, J. D., Kassell, N. et al. (1964) Transmission of
increased intracranial pressure: I. Within the craniospinal axis. Journal of
pressure in neurosurgical practice. Acta Psychiatrica et Neurologica Scandinavica, 36(Suppl. 149), 1193.
Mann, K. S. and Yue, C. P. (1988) A device for long-term monitoring of
intracranial pressure. British Journal of Neurosurgery, 2, 273275.
Marmarou, A., Anderson, R. L., Ward, J. D. et al. (1991) NINDS Traumatic Coma
Data Bank: intracranial pressure monitoring methodology. Journal of

Marmarou, A., Tsuji, O., Myseros, J. et al. (1994) Multicenter clinical evaluation
of a new ICP device, in Intracranial Pressure IX, (eds H. Nagai, K. Kamiya and
S. Ishii), Springer-Verlag, Berlin, pp. 1819.
Marshall, L. F., Smith, R. W. and Shapiro, H. M. (1979a) The outcome with
aggressive treatment in severe head injuries. Part I: The significance of
intracranial pressure monitoring. Journal of Neurosurgery, 50, 2025.
Marshall, L. F., Smith, R. W. and Shapiro, H. M. (1979b) The outcome with
aggressive treatment in severe head injuries. Part II: Acute and chronic
barbiturate administration in the management of head injury. Journal of
Mendelow, A. D., Rowan, J. O., Murray, L. and Kerr, A. E. (1983) A clinical
comparison of subdural screw pressure measurements with ventricular
pressure. Journal of Neurosurgery, 58, 4550.
Miller, J. D. (1987) ICP monitoring current status and future directions. Acta
Miller, J. D. (1989) Measuring ICP in patients: its value now and in the future?,
in Intracranial Pressure VII, (eds J. T. Hoff and A. L. Betz), Springer-Verlag,
Berlin, pp. 515.
Miller, J. D., Bobo, H. and Kapp, J. P. (1986) Inaccurate pressure readings for
subarachnoid bolts. Neurosurgery, 19, 253255.
Miller, J. D., Becker, D. P., Ward, J. D. et al. (1977) Significance of intracranial
hypertension in severe head injury. Journal of Neurosurgery, 47, 503516.
Miller, J. D., Butterworth, J. F., Gudeman, S. K. et al. (1981) Further experience in
the management of severe head injury. Journal of Neurosurgery, 54, 289299.
Narayan, R. K., Kishore, P. R. S., Becker, D. P. et al. (1982) Intracranial pressure:
to monitor or not to monitor? Journal of Neurosurgery, 56, 650659.
North, B. and Reilly, P. (1986) Comparison among three methods of intracranial
pressure recording. Neurosurgery, 18, 730732.
North, B. and Reilly, P. (1990) Raised Intracranial Pressure: A Clinical Guide,
Heinemann Medical Books, Oxford.
Ostrup, R. C., Luerssen, T. G., Marshall, L. F. and Zornow, M. H. (1987)
Continuous monitoring of intracranial pressure with a miniaturized
fiberoptic device. Journal of Neurosurgery, 67, 206209.
Piek, J. and Raes, P. (1996) Pressure-controlled drainage of cerebrospinal fluid:
clinical experience with a new type of ventricular catheter (Ventcontrol MTC)
and an integrated piezo-resistive sensor at its tip: technical note. Neurosurgery, 38, 216218.
Piper, I. R. and Miller, J. D. (1995) The evaluation of the wave-form analysis
capability of a new strain-gauge intracranial pressure microsensor. Neurosurgery, 36, 11421145.
Piper, I. R., Miller, J. D., Dearden, N. M. et al. (1990) Systems analysis of
cerebrovascular pressure transmission: an observational study in headinjured patients. Journal of Neurosurgery, 73, 871880.
Piper, I. R., Chan, K. H., Whittle, I. R. and Miller, J. D. (1993) An experimental
study of cerebrovascular resistance, pressure transmission, and craniospinal
compliance. Neurosurgery, 32, 805815.
Portnoy, H. D. and Chopp, M. (1981) Cerebrospinal fluid pulse wave form
analysis during hypercapnia and hypoxia. Neurosurgery, 9, 1427.
Quincke, H. (1897) Ueber Meningitis serosa und verwandte Zustande. Deutsche
Zeitschrift fur
Nervenheilkunde, 9, 149168.
Robertson, C. S., Narayan, R. K., Contant, C. F. et al. (1989) Clinical experience
with a continuous monitor of intracranial compliance. Journal of Neurosurgery, 71, 673680.
Rosner, M. J. and Becker, D. P. (1976) ICP monitoring: complications and
associated factors. Clinical Neurosurgery, 23, 494519.
Sharpe, W. (1920) Diagnosis and Treatment of Brain Injuries With and Without a
Fracture of the Skull, J. B. Lippincott, Philadelphia, PA.
Shields, C. B., McGraw, C. P. and Garretson, H. D. (1984) Accurate intracranial
pressure monitoring: technical note. Neurosurgery, 14, 592593.
Sullivan, H. G., Miller, J. D., Becker, D. P. et al. (1977) The physiological basis of
intracranial pressure change with progressive epidural brain compression.
An experimental evaluation in cats. Journal of Neurosurgery, 47, 532550.
Sundbarg, G., Kjallquist, A., Lundberg, N. and Ponten, U. (1972) Complications
due to prolonged ventricular fluid pressure recording in clinical practice, in
Intracranial Pressure, (eds M. Brock and H. Dietz), Springer-Verlag, Berlin, pp.
348352.
Takizawa, H., Gabra Sanders, T. and Miller, J. D. (1986) Analysis of changes in
intracranial pressure and pressurevolume index at different locations in the
craniospinal axis during supratentorial epidural balloon inflation. Neurosurgery, 19, 18.
Takizawa, H., Gabra Sanders, T. and Miller, J. D. (1987) Changes in the
cerebrospinal fluid pulse wave spectrum associated with raised intracranial
pressure. Neurosurgery, 20, 355361.
Vries, J. K., Becker, D. P. and Young, H. F. (1973) A subarachnoid screw for
monitoring intracranial pressure. Technical note. Journal of Neurosurgery, 39,
416419.
Yoshihara, M., Marmarou, A., Dunbar, J. et al. (1993) A fiberoptic device suitable
for subdural pressure measurementICP82024.
11
MEASURING CEREBRAL BLOOD

FLOW AND METABOLISM
11.1
Overview of CBF measurements
11.1.1
INTRODUCTION
The optimal method for measuring cerebral blood

flow (CBF) has yet to be discovered. Because of the
anatomical difficulties inherent in accessing the central nervous system, quantitative measurements of
cerebral blood flow are difficult. An ideal method for
measuring CBF should provide all the following.

Quantitative results
High spatial resolution
Continuous measurements, if clinically required
No influence on the normal brain function
No or minimal risk to the patient
Cost-effectiveness
Application in the clinical setting
The first well-recognized CBF studies on humans were

performed by Kety and Schmidt in the 1940s using
nitrous oxide (N2O) gas. The measurement of CBF was
based on the principle that the rate of uptake and
clearance of an inert diffusible gas is proportional to
blood flow. The tissue concentration of the tracer at
equilibrium must therefore be dependent on blood
flow rather than diffusibility. The Kety and Schmidt
equation can be used for calculating CBF using N2O,
krypton-85 and xenon-133. In fact, complete equilibrium of N2O between arterial blood, brain tissue and
venous blood occurs slowly or not at all. Thus CBF
calculated from a short saturation period will overestimate CBF (Hoyer, 1985).
The Kety and Schmidt equation (1948) is:
C(t) = f t Ca (u)ek(t u)du
k = f/ = relative perfusion index ( is the partition
coefficient of the tracer); C(t) = total amount of tracer
delivered to tissue between time zero and t; f = flow;
u = tissue concentration of the tracer at time t;
Ca (t) = input function, i.e. the amount of tracer
delivered to the brain, an integral of the arterial

concentration from 0 to t.
In the 1960s extracranial recordings of the uptake
and clearance of radioisotopes for measuring local
CBF became available using krypton-85 and xenon133 (133Xe; Lassen and Ingvar, 1961). Krypton-85 is a
weak beta emitter and has been abandoned, whereas
xenon-133, a strong gamma emitter, has been widely
used in humans. Xenon-133 is a freely diffusible inert
gas that does not interfere with cerebral metabolism.
Elimination occurs very rapidly and almost completely through the lungs.
11.1.2 CLINICAL METHODS FOR MEASURING CBF
AND BRAIN METABOLISM
Even 50 years after the first quantitative clinical

measurements, CBF is not routinely measured in
neurosurgery or neurology. This is surprising considering its potential importance in the management
of patients with severe head injury, stroke and
subarachnoid hemorrhage.
Clinical measurements of CBF and metabolism can
be divided into:

quantitative measurements;
qualitative measurements;
indirect measurements of CBF and metabolism;
measurements of cerebral metabolism.
The most important techniques for obtaining CBF are

explained in detail below.
(a)
Quantitative measurements
Quantitative CBF measurements mainly use xenon133, either injected into a carotid artery or intravenously, or given by inhalation (Bruce et al., 1973;
Obrist et al., 1975; Agnoli et al., 1969). The stable
(non-radioactive) xenon technique uses repeated
computer tomography scanning (Winkler et al., 1977;
218
MEASURING CEREBRAL BLOOD FLOW AND METABOLISM
Got et al., 1982). Positron emission tomography (PET)

for measuring CBF and cerebral metabolism is well
established for other purposes and is being used
more frequently in the study of patients with severe
head injury (Langfitt et al., 1986). The use of magnetic resonance imaging (MRI) for measuring CBF is
currently under investigation in several centers.
(b)
Qualitative measurements
Qualitative CBF measurement includes single photon

emission computed tomography (SPECT), laser Doppler flowmetry and transit time techniques using
radioactive contrast agents with CT or radioisotopes
with SPECT.
Attempts have been made to obtain quantitative
CBF values with SPECT, i.e. in ml/100 g/min, using
the tracer 99mTC-HMPAO, but at present only the
visual interpretation of SPECT images is of clinical
use. This may change as newer tracers are developed.
SPECT may also be used to study the bloodbrain
barrier after trauma (Bullock et al., 1990; Schroder
et
al., 1994).
Laser Doppler measures the local, relative (not
absolute) CBF in a small brain volume (1 mm3 ) and
can be used for tests of autoregulation and CO2
reactivity (Bolognese et al., 1993).
The recently developed thermal diffusion technique
is based on the thermal conductivity of cortical tissue,
allowing continuous recordings of CBF in a small
region of the cortex (Carter, 1991). Indirect methods
for obtaining information on CBF and metabolism
include jugular venous oximetry, transcranial Doppler
sonography and near-infrared spectroscopy.
Continuous monitoring of jugular venous oxygen
saturation (SjvO2) assesses global cerebral oxygenation.
Episodes of cerebral desaturation can be detected
which would not been seen otherwise. Intermittent
calculations of arteriovenous oxygen difference
(AVDO2) and the metabolic rate of oxygen consumption (CMRO2) help to determine the status of cerebral
oxygen metabolism (Chapter 5; Cruz et al., 1991;
Sheinberg et al., 1992).
Transcranial Doppler sonography (TCD), a noninvasive bedside test, is now more frequently used in
intensive care settings (Chan, Dearden and Miller,
1993; Steiger et al., 1994). Doppler sonography is
described in detail in Chapter 13.
Near-infrared spectroscopy (NIRS) is another noninvasive monitoring now being evaluated for its
clinical value in detecting changes in brain oxygenation, CBF and cerebral blood volume. NIRS
depends upon the relative transparency of biological
tissue to light in the near-infrared spectrum. The
absorption of oxyhemoglobin, deoxyhemoglobin and
oxidized cytochrome can be detected transcranially.
NIRS is most widely used to assess neonatal cerebral

hemodynamics and appears reliable and useful in this
setting. In adults only certain brain areas can be
monitored because bone thickness produces a scattering of the light spectrum (Elwell et al., 1993; Gopinath
et al., 1993). Intracerebral microdialysis is undergoing
clinical evaluation for measuring certain products of
cerebral metabolism. With this method extracellular
brain tissue electrolytes, excitatory amino acids, glucose and lactate can be measured (Persson and
Hillered, 1992; Bullock et al., 1994; Zauner and Bullock,
1995). Continuous on line pH measurements in the
dialysate are also possible and need further evaluation. The measurement of cerebral PO2, PCO2, pH and
temperature with one fiberoptic sensor is another
application still in the experimental stage that may
find a place in managing head-injured patients (Zauner, Bullock and Young, 1995; Zauner et al., 1995).
11.2
Xenon CBF measurements
11.2.1
THE XENON-133 METHOD
Whatever the route of administration, this method

depends on direct monitoring of the radioactive tracer
tissue concentration by external recording with multiple scintillation detectors placed against the scalp.
Assessment of regional CBF in a number of brain areas
is possible depending on the number and location of
the detectors, but such information is not truly
tomographic and it is degraded by various methodological artifacts. The reader is referred to Wood et al.
(Jaggi and Obrist, 1987; Ewing et al., 1987) for a
detailed review.
(a)
The intracarotid xenon-133 method
A bolus of xenon-133 is injected into one internal

carotid artery and the peak count rate measured by
external detectors reflects the isotope concentration in
brain tissue. CBF can be calculated from the clearance
curve as the height/area under the curve, providing
the mean CBF (non-compartmental index; Meier and
Zierler, 1954). By calculating the initial slope index a
fast and a slow compartment can be differentiated in
normal brain. However this may not be valid in
injured brain (Risberg et al., 1975).
Over the last decade the intracarotid xenon method
has been increasingly displaced by less invasive
techniques using inhalational xenon-133 or stable
xenon-enhanced computer tomography.
(b)
The non-invasive xenon-133 methods
The modified intravenous xenon-133 injection method

developed by Austin et al. (1972) and the inhalation
XENON CEREBRAL BLOOD FLOW MEASUREMENTS
methodology developed by Obrist et al. (1975) are two

less invasive radioactive xenon techniques. In contrast
to the intracarotid method, they result in greater scalp
and extracerebral contamination, which needs correction. Assessment of the arterial isotope input can be
achieved by monitoring the xenon-133 concentration
in end-tidal air. There is a close relationship between
end-tidal air xenon-133 and arterial concentration if
pulmonary diffusion is normal (Hausen et al., 1990).
All these factors depend on a reliable mathematical
model for calculating CBF.
Because of the instability of compartmental analyses
in patients after severe head injury, other blood flow
indices, such as CBF15 and CBFinf, equivalent to the
height over area method, have been developed (Obrist
and Wilkinson, 1985).
In our studies, severely head-injured patients studied with the inhalation method inhaled a gas mixture
containing 58 mCi 133Xe/l for 1 minute. Those studied
with the intravenous injection technique were injected
with xenon-133 dissolved in saline (0.3 mCi/kg body
weight). Immediately following the injection, ventilation was interrupted for 20 seconds to prevent large
amounts of xenon from being expired during its first
passage through the pulmonary circulation.
After xenon-133 administration, radioactivity from
the brain was monitored for 15 minutes (CBF15). A
system with 16 detectors concentrically arranged in a
Plexiglass helmet or a portable instrument with ten
probes may be used (Bouma et al., 1991).
For better comparison, CBF15 values are often
recalculated to a PaCO2 of 34 mmHg, assuming the CO2
response to be 3% change in CBF per 1 mmHg change
in PaCO2. Normal CBF values in young volunteers
with a PaCO2 of 34 mmHg are 44.1 5.6 ml/100 g/min
and increase to about 50 ml/100 g/min at a PaCO2 of
40 mmHg (Obrist et al., 1984).
Advantages of this technique are quantitative
results, repeatable measurements, a portable bedside
system and cost-effectiveness. The limitations of noninvasive CBF measurements are the inability to detect
small regions of low CBF and poor spatial resolution,
especially for deep brain structures.
11.2.2 STABLE-XENON-ENHANCED COMPUTED
TOMOGRAPHY
(a)
219
al. proposed in 1977 that the radiodensity of xenon,

which is close to that of iodine, would provide CT
enhancement in proportion to blood flow. Thus for
CBF measurements stable non-radioactive xenon can
be inhaled during CT scanning.
Commercially available software can be easily
added to a standard CT scanner. In studying patients
with severe head injury, good teamwork is required
between
neurosurgery,
neuroradiology
and
intensivists.
CBF is computed by measuring the brain tissue
uptake during arterial build-up of xenon. The end-tidal
xenon concentration is assumed to be in equilibrium
with the concentration in arterial blood. Subsequently,
the time-dependent change in arterial xenon concentration is obtained and can be converted to equivalent
CT units (Obrist et al., 1975). Additionally, a correction
for the hematocrit is required. The relationship for the
hematocrit is shown in the following equation:
Ca (t) = kXe(t) (1 + 1.1 hct)
Ca(t) = arterial xenon concentration; k is a conversion
constant; Xe(t) = end-tidal xenon concentration; hct =
hematocrit.
The baseline images are averaged and subtracted
from the enhanced images obtained during xenon
inhalation, producing a sequence of raw enhanced
images. CBF is calculated by using the Kety and
Schmidt equation on a pixel by pixel basis (Kety and
Schmidt, 1948). CBF images are generated by computed flow values which are converted into Hounsfield units. Software post-processing techniques and
smoothing procedures reduce image artifacts. A confidence image for each CBF image yields information
about the quality and reliability of the CBF assessment. Most centers use 612 minute xenon washin
protocols for calculating CBF (Bouma et al., 1991; Got,
Yonas and Good, 1989; Good, Got and Yonas, 1992;
Nakano et al., 1992). Xenon washin and washout
(clearance) protocols have the advantage of capturing
more data points, which increases the accuracy of
blood flow measurements and a shorter inhalation
time of xenon. Fatouros et al. showed that a 3 minutes
washin, i.e. 3 minutes of xenon inhalation only, and 3
minutes washout procedure is superior to 6 minutes
washin alone, especially if low blood flow is present
(Fatouros et al., 1995a).
Methodology
The technical limitations and the poor spatial resolution of the non-invasive xenon-133 method were
reasons for searching for better ways of measuring
CBF with xenon as the diffusible tracer.
Sequential CT scanning allows visualization and
quantification of time-dependent changes following
the administration of a contrast medium. Winkler et
(b)
Clinical studies
At the Medical College of Virginia, CT-CBF measurement for severely head-injured patients (GCS 8 or less)
is a routine diagnostic procedure. The first CBF scan is
usually performed on admission following a regular
diagnostic head scan, unless the patient is clinically
unstable or needs the immediate evacuation of a
220
hematoma or any other emergency surgical treatment.

Prior to CT-CBF scanning, arterial blood gas analysis
and hematocrit are required. To avoid artifacts, the
patient must not move during the study and sedation
and/or muscle paralysis may be necessary . At present
an 8 minute xenon washin study is performed. As
noted above, a 3 minute washin and 3 minute washout
protocol is currently being investigated. Two baseline
non-enhanced scans are obtained at three levels of the
brain determined from the diagnostic CT scan. A gas
mixture containing 30% xenon and 3060% oxygen is
than delivered for 8 minutes via a specially designed
enhancer connected between the respirator and
patient. Eight sequential CT scans at each level are
obtained during the 8 minute period using a rapid
sequence CT technique. Xenon end-tidal values are
obtained and recorded during the study. End-tidal
PCO2 is measured with a capnograph and O2 saturation is monitored using a pulse oximeter. Blood
pressure and intracranial pressure may be continuously monitored throughout the study and in some
patients transcranial Doppler measurements are also
performed prior to or during the study. CBF studies
add about 1530 minutes to a regular CT scan. Followup CBF studies are performed between days 3 and 5 or
as needed, to distinguish brain ischemia from hyperemia. Depending on the clinical situation cerebral
blood volume (CBV) may also be measured by
injecting 50 ml iodinated non-ionic contrast material
intravenously after completing the xenon study. CBV
is calculated by multiplying CBF by the mean transit
time (MTT) for contrast (see below). CBF images are
analyzed according to regions of interest (ROI) immediately after the measurements. Hemispherical or
regional values are calculated by outlining these
structures using drawing tools on the CT screen.
(c) Advantages and limitations of the stable
xenon technique
Quantitative, non-invasive blood flow measurements
in the superficial cortex as well as in deeper structures,
such as brain stem and basal ganglia, can be studied.
The direct correlation with anatomy on the baseline
CT images enables easy recognition of specific regions
of interest. Problems associated with the xenon-133
technique, such as extracerebral contamination and
the look-through phenomenon do not occur with this
technique. Added to a regular CT scan, it can be
undertaken in severely head-injured patients without
putting them at risk, although careful monitoring of
blood pressure, O2 saturation, ICP and end-tidal CO2
is necessary.
There are several reports of the direct effects of
xenon inhalation on CBF and ICP. At greater than
50% concentration, xenon is an anesthetic agent and
may cause respiratory depression and cerebral vasodilation thereby increasing ICP (Bruce et al., 1973;
Giller, Purdy and Lindstrom, 19990; Takasago et al.,
1992; Obrist, Jaggi and Harel, 1985; Winkler and
Turski, 1985). By using shorter xenon inhalation
times (washin and washout procedures) and a maximal xenon concentration of 32%, a compromise can
be achieved between xenon side effects and accuracy
which depends on sufficient enhancement for a
good signal-to-noise ratio. The radiation exposure of
approximately 828 cGy per studied level must be
also considered (Good, Got and Yonas, 1992). In an
awake or agitated patient head motion is another
limitation, whereas in the comatose head-injured
patient, with proper sedation and head fixation, this
is usually not a problem.
(d) Cerebral blood volume (dynamic CT
technique)
In patients with severe head injury, cerebral blood
volume (CBV) is of great interest, especially if ICP is
raised as a result of brain edema and/or vascular
engorgement. Increased CBV may accompany high,
normal or low CBF, suggesting that CBF and CBV are
not simply related (Grubb et al., 1978; Muizelaar et
al., 1989; Phelps et al., 1979). The ratio of CBF/CBV
may help to differentiate between irreversible and
reversible ischemia in some patients (Leblanc et al.,
1987; Powers and Raichle, 1985).
With CT scan times of 2 seconds and very short
interscan delays it is possible to assess the kinetics of
the first passage of an i.v. bolus of iodine contrast
through brain tissue. The mean transit time (MTT) of
the bolus through the cerebral vasculature can then be
estimated and CBV be calculated as CBV =
CBF MTT.
The MCV method is manually to inject a bolus of
50 ml of iodinated non-ionic contrast medium through
an i.v. line in less than 5 seconds. The hemispheres or
any other regions, excluding major vessels and the
ventricular system, may be chosen as ROIs and the
mean Hounsfield number in each region is calculated.
The CT enhancement versus time curves are fitted to a
gamma variate function:
C(t) = k(t ta ) e (t - ta )/
where t = time after injection; ta = indicator appearance time; k, , are fit parameters; C(t) = indicator
concentration that is proportional to the CT number.
The arrival time, as well as a cut-off point on the
downslope to avoid recirculation, must be specified by
the operator or computed automatically. The use of a
gamma variate function greatly simplifies the computation of the MTT. The interval between the first and
second inflection points, which is the inflection width
FURTHER DIRECT CLINICAL METHODS FOR OBTAINING CBF
(IW) of the time density curve can be used as an

estimate of MTT. The IW represents the transit time of
the densest part of the bolus and this is close to the
MTT (Axel, 1980; Oldendorf, 1964; Bouma et al., 1992;
Fatouros et al., 1995b).
In summary, a CT-based method for measuring CBV
is practicable, by combining CBF measurements with
the stable xenon CT technique and using the inflection
width following an intravenous iodine bolus as the
cerebral transit time. However, additional work is
required to further validate this method.
11.3 Further direct clinical methods for

obtaining CBF
11.3.1
THERMAL DIFFUSION TECHNIQUE
The thermal diffusion measurement of cortical CBF

(CoBF) is based on the thermal conductivity of cortical
tissue (Carter et al., 1982; Dickman et al., 1991). A
sensor consisting of two small flat gold plates, one of
which is neutral and the other heated, is placed
through a burr hole or craniotomy on a region of
interest of the cortex. The CoBF in ml/100 g/min is
estimated from the temperature difference between
these two plates. If CoBF increases, the temperature
difference between the two plates becomes smaller,
the heat loss being conducted through the capillary
bed. A microprocessor continuously converts the
temperature difference to CBF in ml/100 g/min.
Placement of the sensor over large surface vessels
should be avoided.
Cortical blood flow measurements have been used
in patients following subarachnoid hemorrhage and
head injury, as well as after surgery for tumors,
arteriovenous malformations or epilepsy (Carter, 1991;
Carter et al., 1982; Dickman et al., 1982). In headinjured patients, continuous monitoring of CoBF may
assist in differentiating cortical ischemia from hyperemia, and permit autoregulation and CO2 reactivity
tests to be done. Hyperventilation during increased
ICP may be monitored with less risk of inducing
ischemia. A correlation between CoBF and whole
brain CBF measured by the Kety and Schmidt N2O
method has been shown (Carter, Weinand and Oommen, 1993; Robertson, 1993). Although CoBF is not
representative of blood flow in the whole hemisphere
or in deeper regions of the brain, continuous measurements of relative changes in the days after severe head
injury may be of great value in guiding clinical
management.
11.3.2
LASER DOPPLER FLOWMETRY
Laser Doppler flowmetry (LDF) is a new method for

continuous measurement of local microcirculatory
221
blood flow. The method is based on the principle of

Doppler shift (Chapter 13). Monochromatic laser light
reflected from stationary tissue remains unchanged in
frequency. When it is reflected from moving blood
cells it undergoes a frequency shift and therefore light
reflected from the microcirculation measures the
movements of red blood cells (Frerichs and Feuerstein,
1990; Meyerson et al., 1991). Monochromatic laser
light, wavelength 600780 nm, which is above the
maximal absorption of hemoglobin and below the
absorption of water, is used. The magnitude and
frequency distribution of the wavelength changes are
directly related to the number and velocity of red
blood cells but unrelated to the direction of their
movement. The light is carried by a fiberoptic probe
and the reflected light is converted into an electrical
signal from which several values are derived. The flux
(an arbitrary flow unit) represents the movement and
concentration of blood cells through the microvasculature. It is derived from the concentration of moving
blood cells (CMBC) in the measured volume multiplied by the mean velocity. All three values, flux,
CMBC and velocity, are continuously recorded and
can be displayed by a computer (Bolognese et al.,
1993). Absolute CBF in ml/100 g/min can be theoretically estimated from LDF measurements, but must be
interpreted with great caution. LDF measures the
local flow within a very small sample volume of
approximately 1 mm3, depending on the type of laser
being used, and varies markedly with the location of
the probe and its relationship to the microvasculature
(Meyerson et al., 1991; Haberl, Villringer and Dirnagl,
1993). At present the measurement of relative values
and their changes over time seems to be more accurate
than relying on absolute values.
LDF has been developed for bedside monitoring
(Meyerson et al., 1991; Bolognese et al., 1993) and
requires surgical implanting. The device may be used
alone or with other intracranial monitoring devices
(Meyerson et al., 1991; Bolognese et al., 1993; Haberl,
Villringer and Dirnagl, 1993; Steinmeier, Bondar and
Bauhuf, 1993). In patients with severe head injury,
LDF can be used to assess CO2 reactivity and the state
of autoregulation (Le Bihan and Turner, 1992).
11.3.3 MAGNETIC RESONANCE IMAGING FOR CBF
MEASUREMENTS (CHAPTER 14)
The washin and washout kinetics of exogenously

administered MRI-detectable tracers, such as
2
H-water, 19F-trifluoromethane and gadolinium permit measurements of cerebral perfusion (Zhang, Williams and Koretsky, 1993). Most of these tracers are
still under investigation and MRI studies for calculating CBF in acute clinical settings such as head injury
are not yet possible.
222
Recently a technique for proton magnetic resonance

imaging (MRI) has been developed using water as the
freely diffusible tracer. This technique involves labeling the water proton nuclear spins in the arterial blood
by continuously inverting them in the neck region
before they enter the brain. This permits the calculation of CBF (Williams et al., 1992). Le Behian et al. have
taken another approach for calculating CBF (Le Bihan
and Turner, 1992). They have focused on the capability of magnetic resonance to image and measure
molecular diffusion and capillary blood flow or
perfusion. By using diffusion imaging techniques,
perfusion can be measured non-invasively on the
basis of intravoxel incoherent motion (IVIM). This
MRI model considers that capillary segments are
randomly oriented in each voxel and the perfusion
seen by the IVIM technique is quantified in terms of
active capillary density or average blood flow velocity
rather than as ml/100 g/min.
At present for patients with severe head injury CBF
measurements using MRI are in an experimental
stage, but MRI has great potential and warrants
further exploration.
11.4 Indirect methods for obtaining CBF

and metabolism
11.4.1
JUGULAR VENOUS OXIMETRY
Jugular bulb oxygen saturation can be used to

calculate AVDO2 and to monitor cerebral oxygenation.
The cerebral metabolic rate for oxygen (CMRO2) can
be calculated as:
CMRO2 = CBF AVDO2.
Under physiological conditions CBF and CMRO2 are
coupled and the ratio of these two parameters, and
AVDO2, remains constant (Robertson et al., 1989). The
normal value for AVDO2 is approximately 6.5 ml
O2/100 ml blood and for CMRO2 3.2 ml O2/100 g/min
(Muizelaar and Schroder,
1994). In about 45% of

patients who are comatose following severe head
injury, CMRO2 and CBF remain normally coupled
(Obrist et al., 1984). Only in these patients does a high
AVDO2 indicate low CBF relative to CMRO2, as the
brain compensates for the decreased CBF by extracting a greater amount of oxygen, and conversely a low
AVDO2 indicates increased CBF (Robertson et al.,
1989). Cruz has suggested that measuring the cerebral
oxygen extraction, i.e. the difference between arterial
and jugular bulb oxyhemoglobin saturation, is more
accurate than AVDO2, especially in the presence of
anemia (Cruz, 1993a).
Fiberoptic catheters that measure jugular venous
oxygen saturation (SjvO2) continuously in the jugular
bulb are a major advance over intermittent jugular
monitoring (Cruz, 1993a, b; Sheinberg et al., 1992). A

no. 4 or 5G fiberoptic oxygen saturation catheter is
inserted percutaneously in a retrograde direction into
the internal jugular vein. The tip of the catheter must
be placed within the jugular bulb and its position
verified by X-ray or CT-scan. The catheter must be
calibrated before and after insertion and thereafter
every 8 hours by drawing a blood sample from the
catheter and measuring the oxygen saturation on a
co-oximeter. The reflected light intensity of the fiberoptic device must be checked regularly. It depends on
correct placement of the catheter tip within the
lumen and will be reduced if the catheter becomes
attached to the vein wall. SjvO2 in normal individuals
is 65%. However, in comatose patients with severe
head injury, readings over 50% are considered normal (Sheinberg et al., 1992). An SjvO2 less than 50%
and AVDO2 above normal indicate ischemia or a
mismatch between CBF and metabolism. Because
arterial oxygen saturation and arterial hemoglobin
affect cerebral metabolism and oxygenation, they
must also be taken into consideration (Cruz, 1993a,
b). Whenever SjvO2 drops below 50% for more than
15 minutes, the cause of the desaturation must be
sought. First, the light intensity of the catheter
should be checked and the catheter calibrated by cooximeter. If both are within the normal range then
cerebral desaturation has occurred. Episodes of cerebral desaturation occur most frequently within the
first 48 hours after injury, most often in relation to
high ICP (Muizelaar and Schroder,
1994; Sheinberg et
al., 1982).
There are several limitations with jugular bulb
oximetry.

It measures the global saturation of one hemisphere

only.
SjvO2 may not decrease during raised ICP until
tentorial herniation has occurred (Sheinberg et al.,
1982).
There is contamination with blood from the contralateral hemisphere and from extracranial venous
blood.
A previous study found that almost 50% of readings

that suggested desaturation were incorrect (Sheinberg
et al., 1982), usually as a result of low light intensity.
At present jugular venous oximetry is the only
method available to give continuous monitoring of
cerebral oxygenation and metabolism. Even though
the technique can be time-consuming, several authors
have found it helpful in management, even though
experience at the Medical College of Virginia has, in
comparison, been disappointing. In the near future
cerebral tissue O2 and CO2 measurements may prove
to be a better way of monitoring comatose patients
(Zauner and Bullock, 1995; Zauner et al., 1995).
DIRECT MEASUREMENT OF CEREBRAL METABOLISM
11.4.2
TRANSCRANIAL DOPPLER SONOGRAPHY
Transcranial Doppler sonography is discussed in

detail in Chapter 13.
The technique was introduced in 1982 by Aaslid,
Markwalder and Nornes (1982). Using a 2 MHz
pulsed Doppler it is possible to penetrate the skull at
certain areas and to record the blood flow velocity
from the major basal cerebral arteries. The most
frequent measurement site is transtemporal (via the
temporal window), where the middle cerebral artery,
internal carotid artery and anterior cerebral artery can
be identified (McCormick et al., 1991). Vessel identification depends on the cranial window used, the
depth of the sample volume, the flow direction and
the spatial relationships of vessels. The vertebral
arteries and the basilar artery may be insonated
through the foramen magnum.
A standard nomenclature has become accepted, in
which peak systolic velocity refers to the highest
velocity during systole and the end-diastolic velocity
represents the maximum velocity at the end of the
diastole. The mean velocity is the time averaged
maximum velocity during the cardiac cycle.
Blood flow velocity does not give direct information on CBF. In an artery with a lumen area A (cm2),
the blood flow Q (ml/s) and cross sectional average
blood velocity V (cm/s) are related by the following
equation:
Q = VA.
However cross-sectional blood velocity is difficult
to measure, since the anatomy of cerebral arteries is
very complex. Blood flow velocity is further influenced by several factors, including arterial blood
pressure, ICP, hematocrit, PaCO2 and the status of
autoregulation, thus making a direct comparison of
flow velocity and CBF very complex. All these factors
may be altered in severe brain injury and may make
Doppler measurements difficult to interpret (Steiger et
al., 1994).
11.4.3
NEAR-INFRARED SPECTROSCOPY
Near-infrared spectroscopy (NIRS) was recently

developed as a continuous bedside technique suitable
for clinical use particularly in neonates. Near-infrared
light with a wavelength between 6501100 nm can
penetrate the scalp, skull and brain (Jobsis,
1977;
McCormick et al., 1991). Penetration into brain tissue is
limited by the thickness of the skull. In adults it
extends for a few centimeters only, whereas in
neonates penetration is much deeper. Thus relative
changes in CBF and cerebral oxygenation can be
investigated transcranially by means of the transparency and absorption of oxyhemoglobin, deoxyhemoglobin and oxidized cytochrome in the near-infrared
223
spectrum. The absorption of hemoglobin occurs at a

wavelength between 670 and 760 nm, whereas oxidized cytochrome aa3 is detected near 830 nm. No
separation into arterial, venous or capillary compartments in the measured brain areas can be discerned.
Recent studies in adults have reported that NIRS is at
least as sensitive in detecting progressive human
cerebral hypoxia as EEG (McCormick et al., 1991).
Gopinath et al. (1993) have also found NIRS useful in
detecting and localizing intracranial hematomas.
Potential limitations of this application are that only
acute hematomas could be found and that bilateral
hematomas might not be detected.
With further refinement NIRS may prove to be
useful in managing patients with severe head injury in
the intensive care unit or the emergency room, both to
detect brain oxygen desaturation in selected areas and
localize intracranial hematomas. At present, NIRS is
more reliable in neonates than in adults.
11.5 Direct measurement of cerebral

metabolism
11.5.1
INTRACEREBRAL MICRODIALYSIS
Intracerebral microdialysis enables endogenous substances in the extracellular fluid (ECF) of the brain to
be retrieved. Many experimental studies have suggested that excitatory amino acids (EAAs), neurotransmitters, electrolytes and energy-related metabolites are released in abnormal quantities after severe
head injury. EAAs such as glutamate and aspartate
have toxic effects on cell membranes and cause cell
damage particularly in states of prolonged ischemia
(Ungerstedt, 1991; Beneviste and Hiittemeier, 1990).
The membranes of microdialysis probes in current use
have a length of 410 mm and a diameter of 0.5 mm.
They may be implanted during surgery or together
with an ICP monitoring device (Persson and Hillered,
1992; Bullock et al., 1994). The probe is perfused with
sterile normal saline or Ringers solution using a
microinjection pump (2 l/min). The dialysate is
collected in vials over a defined interval (for example:
30 min = 60 l) in a cooled fraction collector. In our
studies the probe is left in place for 4 days. The
samples are analyzed by high performance liquid
chromatography for EAAs, lactate and glucose, and
by flame photometry for electrolytes (Bullock et al.,
1994, 1995; Zauner and Bullock, 1995). In the first 50
patients with severe head injury, there was a six- to
eightfold increase in EAAs if there were no secondary
ischemic events. However, if secondary ischemia
occurred, EAA release was 2050 times greater than
normal and persisted over days (Figure 11.1; Persson
and Hillered, 1992; Bullock et al., 1994, 1995; Zauner
and Bullock, 1995). An increase in ECF K+ following
224
(a)
(b)
(d)
(c)
(e)
Figure 11.1 (ac) Regular CT scan and xenon cerebral blood flow images. The regular scan (a) shows a subdural hematoma
overlying the left hemisphere in a patient who later died. Regions of interest can be drawn and projected on the
corresponding xenon images to calculate local (b) or hemispherical (c) blood flow. In this particular patient, flow was
approximately 6 ml/100 g/min at the time of the study. (d, e) Regular CT scan (d) and xenon cerebral blood flow images
(e) in a patient with favorable outcome. Hemispherical blood flow reached approxiamtely 27 ml/100 g/min. A region of
interest is drawn in the area of placement of a microdialysis probe (for measuring glucose, lactate, amino acids and ions) and
a multiparameter sensor (for measuring brain PO2, PCO2, pH and temperature; see section 11.6 for further detail) to correlate
cerebral blood flow with these data. The blood flow in the region of interest is shown in the right upper corner of the flow
image (24.2 ml/100 g/min).
severe head injury was also seen and correlated with

the increase of glutamate. In ischemia, a rise in ECF K+
and an influx of Ca+ is thought to trigger glutamate
release and activation of cation and anion conductance. (Bullock et al., 1995). A marked reduction in ECF
glucose can be seen after head trauma. This is most
probably due to low CBF and increased anaerobic
metabolism (Figure 11.2).
11.5.2
POSITRON EMISSION TOPOGRAPHY
Positron emission topography (PET) scanning is currently the most versatile and widely used functional
imaging modality for the human brain, both in health
and disease. The theory and methodologies of the
technique are beyond the scope of this chapter and the
reader is directed to Alavi et al., 1982, 1996; Huang et
al., 1980; Phelps, Mazziotta and Huang, 1982; and
Heiss et al., 1984. However, in spite of the enormous

cost and complexity of PET scanning, it has made little
impact on either research or clinical practice in the
neurosciences. Since the isotopes must be obtained online from a cyclotron, PET scan studies need to be
planned and scheduled well in advance. Patients need
to be stable and in close physical proximity to the
scanner and this limits its utility in patients with acute
head injuries. The anatomical resolution of detail is
quite limited, being about 0.5 cm3, which is similar to
that of SPECT scans.
Despite these limitations, the technique has been
applied to studies in patients with head injury. CBF
and metabolism were studied in stable patients during
the first and second weeks after head injury(Langfitt et
al., 1986). There were uniform reductions in both CBF
and metabolism, the most profound reductions being
found in the parietal and occipital lobes. More
DIRECT MEASUREMENT OF CEREBRAL METABOLISM
Figure 11.2
225
Methods of placement of a microdialysis probe.
recently, elegant studies of glucose metabolism using

fluorodeoxyglucose found that in about one-third of
13 patients cerebral metabolism for glucose was
increased within 48 hours of injury. The greatest
increases were seen in zones surrounding focal cere-
bral contusions and adjacent to hematomas (Hovda et

al., 1995). These findings accord exactly with animal
studies. At later time points, however, glucose metabolism had decreased by 3040%, also in accord with
findings in animal studies. It appears, therefore, that
Figure 11.3 Instrumentarium used for placement of a triple lumen bolt (drill bit, taper), together with the multiparameter
sensor, microdialysis probe and triple lumen blot itself.
226
there is a initial phase of hypermetabolism for glucose

after trauma, which is followed by a phase of
hypometabolism. The potential of PET scanning to
demonstrate changes in neurotransmitter metabolism
and protein synthesis has yet to be fully realized.
11.6
Comprehensive neuromonitoring
There are now a number of monitoring and diagnostic

devices that have been developed for use in comatose
patients. They include devices for ICP monitoring,
jugular bulb oximetry, CBF measurement, transcranial
Doppler sonography and EEG. However none of these
techniques is ideal, and only indirect information on
cerebral oxygen delivery and metabolism can be
obtained from them.
At the Medical College of Virginia, a comprehensive
neuromonitoring system is being developed that
includes an ICP monitor with a microdialysis probe
and a combined sensor for measuring brain oxygenation, CO2 generation, and temperature (Zauner,
Bullock and Young, 1995; Zauner et al., 1995) With this
system it is possible to continuously measure ICP,
CPP, brain oxygen, CO2, pH and temperature, as well
as brain chemistry for glucose, lactate, ions and amino
acids, for several days after injury (Figure 11.3). The
three measuring devices are secured to a three lumen
bolt for skull and brain fixation. This system is being
assessed in patients with severe head injury but
further work is needed to establish its value in
intensive care management.
11.7
References
Aaslid, R., Markwalder, T. M. and Nornes, H. (1982) Noninvasive transcranial

Doppler ultrasound recording of flow velocity in basal cerebral arteries.
Agnoli, A., Prencipe, M., Priori, A. M. et al. (1969) Measurements of rCBF by
intravenous injection of 133Xe: a comparative study with the intra-arterial
injection method, in Cerebral Blood Flow, (eds M. Brock, C. Fiescbi, D. H.
Ingvar et al.), Springer-Verlag, Berlin, pp. 3134.
Alavi, A., Reivich, M. and Jones, S. C. (1982) Functional imaging of the brain
with positron emission tomography, in Nuclear Medicine Annual 1982, (eds
H. S. Weissman and L. M. Freeman), Raven Press, New York, pp.
319372.
Alavi, A., Dann, R., Chawluk, J. et al. (1996) Positron emission tomography
imaging of regional cerebral glucose metabolism. Seminars in Nuclear
Medicine, 16, 234.
Austin, G., Horn, N., Rouhe, S. et al. (1972) Description and early results of an
intravenous radioisotope technique for measuring regional cerebral blood
flow in man. European Neurology, 4351.
Axel, L. (1980) Cerebral blood flow determination by rapid sequence CT, a
theoretical analysis. Radiology, 137, 679686.
Benveniste, H. and Hiittemeier, P. C. (1990) Microdialysistheory and
application. Progress in Neurobiology, 35, 195215.
Bolognese, P., Miller, J. I., Heger, I. M. et al. (1993) Laser-Doppler flowmetry in
neurosurgery. Journal of Neurosurgical Anesthesiology, 5, 151158.
metabolism after severe traumatic brain injury: the elusive role of ischemia.
Bouma, G. J., Fatouros, P. P., Muizelaar, J. P. et al. (1992) Determination of
cerebral blood volume by simultaneous xenon enhanced computed
tomographic cerebral blood flow measurement and dynamic computed
tomography enhanced by intravenous contrast, in Cerebral Blood Flow
Measurements With Stable Xenon-Enhanced Computer Tomography, (ed. H.
Bruce, D. A., Langfitt, T. W., Miller, J. D. et al. (1973) Regional cerebral blood
flow, intracranial pressure, and brain metabolism in comatose patients.
Bullock, R., Statham, P., Patterson, J. et al. (1990) The time course of vasogenic
oedema after focal human head injury-evidence from SPECT mapping of
blood brain barrier defects. Acta Neurochirurgica (Supplement), 51,
286288.
Bullock, R., Zauner, A., Tsuji, O. et al. (1994) Excitatory amino acid release after
severe human head trauma: effect of intracranial pressure and cerebral
perfusion pressure changes, in Intracranial Pressure IX, (eds H. Nagai, K.
Kamiya and S. Ishii), Springer-Verlag, Berlin, pp. 264267.
Bullock, R., Zauner, A., Tsuji, O. et al. (1995) Patterns of excitatory amino acid
release and ionic flux after severe human head trauma, in Neurochemical
Monitoring in the Intensive Care Unit, (eds T. Tsubokawa, A. Marmarou, C.
Robertson and G. Teasdale), pp. 6471.
Carter, L. P. (1991) Surface monitoring of cerebral cortical blood flow.
Carter, L. P., Weinand, M. E. and Oommen, K. J. (1993) Cerebral blood flow
(CBF) monitoring in intensive care by thermal diffusion. Acta Neurochirurgica (Supplement), 59, 4346.
Carter, L. P., Erspaner, J. R., White, W. L. et al. (1982) Cortical blood flow
during craniotomy for aneurysms. Surgical Neurology, 17, 204208.
Chan, K. H., Dearden, N. M. and Miller, J. D. (1993) Transcranial Doppler
sonography in severe head injury. Acta Neurochirurgica (Supplement), 59,
8185.
Cruz, J. (1993a) Cerebral oxygenation monitoring and management. Acta
Cruz, J. (1993b) On-line monitoring of global cerebral hypoxia in acute brain
injury: relationship to intracranial hypertension. Journal of Neurosurgery, 79,
228233.
Cruz, J., Miner, M. E., Allen, S. J. et al. (1991) Continuous monitoring of
cerebral oxygenation in acute brain injury: assessment of cerebral hemodynamic reserve. Neurosurgery, 29, 743749.
Dickman, C. A., Carter, L. P., Baldwin, H. Z. et al. (1991) Technical report.
Continuous regional cerebral blood flow monitoring in acute craniocerebral
trauma. Neurosurgery, 28, 467472.
Elwell, C. E., Owen-Reece, H., Cope, M. et al. (1993) Measurement of adult
cerebral hemodynamics using near infrared spectroscopy. Acta Neurochirurgica (Supplement), 59, 7480,
Ewing, J. R., Robertson, W. M., Brown, G. G. and Welch, K. M. A. (1987)
133
Xenon inhalation: accuracy in detection of ischemic cerebral regions and
angiographic lesions, in Cerebral Blood Flow: Physiological and Clinical
Aspects, (ed. J. Wood), McGraw-Hill, New York, pp. 202219.
Fatouros, P., Schroder,
M., Kota, J. et al. (1995a) Accuracy of low CBF

measurements using xenon-CT: effect of xenon inhalation time.
Fatouros, P. P., Kura, A. J., Schroder,
M. L. et al. (1995b) Method for cerebral

blood volume measurements in acute head injury.
Frerichs, K. U. and Feuerstein, G. Z. (1990) Laser Doppler flowmetry: a review
of its application for measuring cerebral and spinal cord blood flow.
Molecular Chemical Neuropathology, 12, 5561.
Giller, C., Purdy, P. and Lindstrom, W. (1990) Effect of inhaled stable xenon on
cerebral blood flow velocity. American Journal of Neuroradiology, 11,
117182.
Good, W. F., Got, D. and Yonas, H. (1992) Technical aspects, in Cerebral Blood
Flow Measurements With Stable Xenon-Enhanced Computer Tomography, (ed. H.
Gopinath, S. P., Robertson, C. S., Grossman, R. G. et al. (1993) Near-infrared
spectroscopic localization of intracranial hematomas. Journal of Neurosurgery, 79, 4347.
Got, D., Yonas, H. and Good, W. F. (1989) Local cerebral blood flow by xenonenhanced CT: current status, potential improvements, and future directions.
Got, D., Wolfson, S. K., Yonas, H. et al. (1982) Progress in cerebrovascular
disease: local cerebral blood flow by xenon enhanced CT. Stroke, 13,
750758.
Grubb, R. L., Raichle, M. E., Higgins, C. S. et al. (1978) Measurement of
regional cerebral blood volume by emission tomography. Annals of
Haberl, R. L., Villringer, A. and Dirnagl, U. (1993) Applicability of laserDoppler flowmetry for cerebral blood flow monitoring in neurosurgical
intensive care. Acta Neurochirurgica (Supplement), 59, 6973.
Hansen, M., Jakobsen, M., Enevoldsen, E. M. et al. (1990) Problems in cerebral
blood flow calculations using xenon133 in patients with pulmonary
disease. Stroke, 21, 745750.
Heiss, W. D., Pawlik, O., Herholz, K. et al. (1984) Regional kinetic constants and
cerebral metabolic rate for glucose in normal human volunteers determined
by dynamic positron emission tomography of [18F]-2-fluoro-2-deoxyD-glucose. Journal of Cerebral Blood Flow and Metabolism, 3, 250253.
Hovda, D. A., Lee, S. M., Smith, M. L. et al. (1995) The neurochemical and
metabolic cascade following brain injury: moving from animal models to
man. Journal of Neurotrauma, 12, 903906.
Hoyer, S. (1985) Metabolism of the human brain: the principle and limitation
of global measurements, in Cerebral Blood Flow and Metabolism Measurement,
(eds A. Hartmann and S. Hoyer), pp. 382390.
REFERENCES
Huang, S. C., Phelps, M. E., Hoffman, E. J. et al. (1980) Noninvasive
determination of local cerebral metabolic rate of glucose in man. American
Journal of Physiology, 238, E69E82.
Jobsis,
F. F. (1977) Noninvasive, infrared monitoring of cerebral and

myocardial oxygen sufficiency and circulatory parameters. Science, 198,
12641267.
Jaggi, J. L. and Obrist, W. D. (1987) Regional cerebral blood flow determined
by 133xenon clearance, in Cerebral Blood Flow: Physiological and Clinical
Aspects, (ed. J. Wood), McGraw-Hill, New York, pp. 189201.
Kety, S. S. and Schmidt, C. F. (1948) The nitrous oxide method for the
quantitative determination of cerebral blood flow in man: theory, procedure, and normal values. Journal of Clinical Investigation, 27, 476483.
Langfitt, T. W., Obrist, W. D., Alavi, A. et al. (1986) Computerized tomography,
magnetic resonance imaging and positron emission tomography in the
study of brain trauma. Preliminary observations. Journal of Neurosurgery, 64,
760767.
Lassen, N. A. and Ingvar, D. H. (1961) The blood flow in the cerebral cortex
determined by radioactive 85krypton. Progress in Nuclear Medicine, 17, 4243.
Le Bihan, D. and Turner, R. (1992) The capillary network: a link between IVIM
and classical perfusion. Magnetic Resonance Medicine, 27, 171178.
Leblanc, R., Tyler, J. L., Mohr, G. et al. (1987) Hemodynamics and metabolic
effects of cerebral revascularization. Journal of Neurosurgery, 66, 529535.
McCormick, P. W., Stewart, M., Goetting, M. G. et al. (1991) Regional
cerebrovascular oxygen saturation measured by optical spectroscopy in
humans. Stroke, 22, 596602.
Meier, P. and Zierler, K. L. (1954) On the theory of the indicator dilution
method for measurement of blood flow and volume. Journal of Applied
Physiology, 6, 731744.
Meyerson, B. A., Gunasekera, L., Linderoth, B. et al. (1991) Bedside monitoring
of regional cortical blood flow in comatose patients using laser Doppler
flowmetry. Neurosurgery, 29, 750755.
Muizelaar, J. P. and Schroder,
M. L. (1994) Overview of monitoring of cerebral

blood flow and metabolism after severe head injury. Canadian Journal of the
Neurological Sciences, 21, S6S11.
Muizelaar, J. P., Ward, J. D., Marmarou, A. et al. (1989) Cerebral blood flow in
severely head injured children. Part 1: Relation with GCS, outcome, ICP
and PVI. Journal of Neurosurgery, 71, 6371.
Nakano, S., Yamashita, T., Kashiwagi, S. et al. (1992), in Cerebral Blood Flow
Measurements With Stable Xenon-Enhanced Computer Tomography, (ed. H.
Obrist, W. D., Jaggi, J. L. and Harel, D. (1985) Effect of stable xenon inhalation
on human CBF. Journal of Cerebral Blood Flow and Metabolism, 5, 557558.
Obrist, W. D. and Wilkinson, W. E. (1985) Stability and sensitivity of CBF
indices in the noninvasive 133Xe method, in Cerebral Blood Flow and
Metabolism Measurement, (eds A. Hartmann and S. Hoyer), pp. 3036.
Obrist, W. D., Thompson, H. K., Wang, H. S. et al. (1975) Regional cerebral
blood flow estimated by 133Xenon inhalation. Stroke, 6, 245256.
Oldendorf, W. H. (1964) Measurement of mean transmit time of cerebral
circulation by external detection of an intravenously injected radioisotope.
Journal of Nuclear Medicine, 3, 382398.
Persson, L. and Hillered, L. (1992) Chemical monitoring of neurosurgical
intensive care patients using intracerebral microdialysis. Journal of Neurosurgery, 76, 7280.
227
Phelps, M. E., Mazziotta, J. C. and Huang, S. C. (1982) Study of cerebral

function with positron computed tomography. Journal of Cerebral Blood Flow
Phleps, M. E., Huang, S. C., Hoffman, E. J. et al. (1979) Validation of
tomographic measurement of cerebral blood volume with C-11-labeled
carboxyhemoglobin. Journal of Nuclear Medicine, 20, 328334.
Powers, W. J. and Raichle, M. E. (1985) Positron emission tomography and its
application to the study of cerebrovascular disease in man. Stroke, 16,
361376.
Risberg, J., Ali, Z., Wilson, E. M. et al. (1975) Regional cerebral blood flow by
133
xenon inhalation: preliminary evaluation of an initial slope index in
patients with unstable flow compartments. Stroke, 6, 142148.
Robertson, C. (1993) Measurement of cerebral blood flow and metabolism in
severe head injury using the KetySchmidt technique (1993) Acta Neurochirurgica (Supplement), 59, 2527.
Robertson, C. S., Narayan, R. K., Gokastan, Z. L. et al. (1989) Cerebral
Schroder,
M., Muizelaar, J. P., Bullock, R. et al. (1994) Focal ischemia due to

traumatic contusions, documented by SPECT, stable Xenon CT, and
ultrastructural studies. Journal of Neurosurgery, .
Steiger, H. J., Aaslid, R., Stooss, R. et al. (1994) Transcranial Doppler
monitoring in head injury: relations between type of injury, flow velocities,
vasoreactivity, and outcome. Neurosurgery, 34, 7986.
Steinmeier, R., Bondar, I. and Bauhuf, C. (1993) Assessment of cerebral
haemodynamics in comatose patients by laser Doppler flowmetry
preliminary observations. Acta Neurochirurgica (Supplement), 59, 6973.
Takasago, T., Yamashita, T., Nakano, S. et al. (1992) Complication and
problems, in Cerebral Blood Flow Measurements With Stable Xenon-Enhanced
Computer Tomography, (ed. H. Yonas), Raven Press, New York.
Ungerstedt, U. (1991) Microdialysis principles and applications for studies
in animals and man. Journal of Internal Medicine, 230, 365373.
Williams, D. S., Detre, J. A., Leigh, J. S. et al. (I992) Magnetic resonance
imaging of perfusion using spin inversion of arterial water. Proceedings of
the National Academy of Sciences of the USA, 89, 212216.
Winkler, S. and Turski, P. (1985) Potential hazards of xenon inhalation.
American Journal of Neuroradiology, 6, 974975.
Winkler, S., Sacker, J., Holden, J. et al. (1977) Xenon inhalation as an adjunct to
computerized tomography of the brain: preliminary study. Investigative
Radiology, 12, 1518.
Zauner, A. and Bullock, R. (1995) The role of excitatory amino acids in severe
brain trauma: opportunities for therapy. Journal of Neurotrauma, 12,
547554.
Zauner, A., Bullock, R. and Young, H. F. (1995) Continuous brain oxygen, CO2,
pH and temperature monitoring in neurosurgical patients. Neurosurgery, 37,
570 (abstract).
Zauner, A., Bullock, R., Di, X. et al. (1995) Brain oxygen, CO2, pH, and
temperature monitoring: evaluation in the feline brain. Neurosurgery, 37,
66.
Zhang, W., Williams, D. S. and Koretsky, A. P. (1993) Measurement of rat brain
perfusion by NMR using spin labeling of arterial water: in vivo determination of the degree of spin labeling. Magnetic Resonance Medicine, 29,
416421.
12
ELECTRICAL FUNCTION
MONITORING
R. J. Moulton
12.1
Goals
The goals of electrophysiological monitoring are to

provide a continuous, objective and readily interpretable measure of brain function in the comatose
head-injured patient. As with any monitoring system,
one would like to detect potentially serious complications and intervene before they result in permanent
morbidity or mortality. Information about the localization of focal lesions would also be helpful, although
this has decreased in importance since the advent of
modern radiological imaging studies (i.e. computed
tomographic scanning and magnetic resonance imaging). Another potential use of electrical function
monitoring is the measurement of the therapeutic
efficacy of new drugs or other therapies for head
injury. Ideally one would be able to determine in the
acute phase whether the agent was exerting the
hoped-for beneficial effects, obviating the requirement
of waiting weeks or months for the standard clinical
outcome measures. Eventually one might be able to
titrate therapy based on real-time neurophysiological
response, in a similar way as is currently done in
respiratory and cardiovascular intensive care.
The monitored parameter or function should ideally
have low variability, so that a change over some
period of time indicates some real change in cerebral
function rather than normal random physiological
variability, yet should be sensitive enough to detect
change sufficiently early to allow treatment to be
rendered in a timely fashion. The monitored function
should also be relatively resistant to the effects of
typical sedative and analgesic drugs, used in customary dosages. As will be seen, the sensitivity of
monitored responses to sedative and analgesic medications varies widely among electrophysiological
measures. The interpretability and ease of use of the
measured parameter is a function of the volume and
complexity of information generated. A monitoring
technique that yields an overwhelming amount of
information will not be practical to use; nor will one
that produces information that is only meaningful to

experts trained in electrophysiology. To be clinically
useful in a monitoring application, the information
must be easily interpretable by nurses and non-expert
medical personnel. These issues have started to be
addressed within the last few years.
12.2
Problems and limitations
Problems with electrophysiological monitoring may

be broadly classified into three areas:
technical issues surrounding the recording of the

signal (most frequently corruption of signal by
noise);
problems with data display and interpretation
specific to the method of monitoring (alluded to
above);
most fundamentally, the fidelity with which the
monitored parameter reflects the pathophysiology
of the disease of interest.
Clearly, monitoring a parameter that changes significantly only in the end stages of a disease process is
likely to yield little or no benefit in terms of early
detection and prevention of morbidity. Measurements
that are overly sensitive to normal physiological
variation may disguise significant events or trends
within this normal variation until it is too late to act on
the evolving trend (the forest and trees problem).
Brain-stem auditory evoked potentials (BAEPs) offer
an example of the former, whereas the variability and
extreme sensitivity of the electroencephalogram (EEG)
and its processed derivatives to drug effects cause
problems of the latter type (see below).
Technical issues that have been problematic in the
past, such as the size of electrophysiological monitoring apparatus and contamination of signal with noise,
continue to shrink in relative importance with the
improvement in electronic components and the
advent of computerization of monitoring equipment.
Monitoring systems have been made smaller and
230
ELECTRICAL FUNCTION MONITORING
portable and now can be brought to the bedside with

relative ease. Improvements in analog and digital
signal filtering have made recording good-quality
signal possible in previously electrically hostile environments such as intensive care units. Recording in
special electrically shielded rooms is no longer essential. Computer control of recording systems makes the
unattended operation of these systems possible,
obviating the need for (and excessive cost associated
with) full-time attendance of a technician at the
bedside. Computerized display systems, digital storage of data, computerized feature extraction and data
reduction techniques are beginning to shrink the
enormous amount of data associated with continuous
or very frequent data recording, and to simplify its
interpretation by non-experts. Having surmounted
many of the technical obstacles to electrical function
monitoring, it now remains to answer the more
fundamental questions about how faithfully these
methods monitor the progress or deterioration of
patients with severe head injury and what are the
optimal electrophysiological parameters to monitor.
The balance of this chapter will be devoted to these
questions, as well as providing the reader with an
introduction to the technical principles concerning the
recording and physiological origins of these signals.
12.3
Methods and modalities
12.3.1
THE ELECTROENCEPHALOGRAM
The origins of the EEG date back to the end of the 19th
century. At this time investigators began to record
electrical signals from the brains of living animals. In
1924 Hans Berger (a psychiatrist) successfully recorded electrical potentials from the human brain. Berger
continued his work through the thirties and over the
succeeding decade electroencephalography saw
increasingly widespread application (Gibbs and
Gibbs, 1950). Standardization of the common scalp
electrode positions was accomplished in 1958 with the
publication of The ten twenty electrode system of the
International Federation (Jasper, 1958). Measurement
and amplification of cortical electrical potentials from
the scalp remain the common underlying methods of
all brain electrophysiological monitoring. Potentials
are conducted from the cerebral cortex to the scalp
through the intervening tissue by a process known as
volume conduction. The signal becomes progressively
attenuated as the distance from the signal generator
increases. The impedance and conductance of the
intervening tissue (CSF, dura, skull, scalp) vary, and
may be further modified by skull fractures, craniotomy flaps, scalp edema and hematomas, so that the
voltage detected at an electrode location is not simply
a function of the electrode placement, amplifier
settings, and underlying cortical electrical activity

(Stockard, Bickford and Aung, 1975).
In the case of the EEG the measured potentials arise
spontaneously. Gross reactivity to external stimulation
by the examiner (e.g. noise, light, pain) may be tested.
Other than routine bandpass filtering and amplification, there is no processing or averaging of signal. The
signal is typically on the order of 10100 V. Other
physiological signals such as the electrocardiogram
and electromyogram are typically two to three orders
of magnitude greater than the EEG so that great care
with respect to technical details must be taken in
recording the EEG. This is particularly true in the case
of ensuring a low impedance (less than 25 k at the
scalpelectrode interface), selection of appropriate
bandpass filters and selection of appropriate reference
and ground electrodes. The placement of reference
electrodes and the recording montage used will
depend to some extent on the purpose of the examination. Cephalic reference electrodes tend to minimize
non-cephalic signal, i.e. cardiac or EMG signal. Simultaneous monitoring of other physiological signals
(electrocardiogram, electro-oculogram) on separate
channels may help identify contamination of the EEG
with these signals. Ambient electrical noise (usually
60 Hz) in the environment may also be problematic,
and the recording problems are often magnified in the
hostile electric environment of the ICU, where there
are multiple sources of noise contamination (e.g. other
monitors, intravenous pumps, cooling blankets, respirators, etc.). With a combination of improved recording equipment, and some pragmatic bedside modification of potential sources of electrical interference,
it is usually possible to obtain technically acceptable
recordings. For practical purposes, the frequencies of
interest in the spontaneous EEG are below 30 Hz, so
that adjustment of the high-pass filter with this in
mind can help screen out a lot of electrical noise.
(a)
The EEG in traumatic coma
The EEG has been used in traumatic coma since the

1940s. The most common observation has been slowing of background frequencies, with the amount of
slowing being approximately proportional to the
depth of coma and prognosis (Rumpl et al., 1979;
Silverman, 1963; Stockard, Bickford and Aung, 1975;
Synek, 1988). Early serial EEG measurements following head injury showed a lag between the onset of
profound coma and slowing of the EEG early after
injury, and the authors felt that this detracted from the
diagnostic value of EEG early after head injury
(Dawson, Webster and Gurdjian, 1951). The authors
speculated that this might be due to increasing
cerebral edema or some metabolic derangement.
Given the paucity of clinical responses in profound
METHODS AND MODALITIES
coma, the clinical examination may well be insufficiently sensitive to detect evolving metabolic abnormalities. We routinely observe continued worsening
(slowing) of the background frequencies of the EEG
when monitoring the power spectrum, and this occurs
in tandem with loss of evoked potential activity, and
reduced arteriojugular oxygen extraction (Figure 12.1).
Other phenomena described in traumatic coma have
been the loss of EEG reactivity to external stimuli,
such as noise or eye opening, and the loss of normal
231
spontaneous variability in the EEG (Bricolo, 1976;

Hutchinson et al., 1991; Rumpl et al., 1979; Silverman,
1963; Synek, 1990a, b). Patients where such reactivity
has not been totally lost have tended to have better
outcomes than those without any evidence of reactivity in the EEG. A more recent study in children
with severe head injuries failed to show any difference
in prognosis between patients with and without EEG
reactivity to external stimulation (Dusser et al., 1989).
Patients who have evidence of periodic sleep patterns
(a)
(b)
Figure 12.1 (a) Deterioration of EEG activity within 48 hours of severe head injury. Monitoring of the EEG power spectrum
was begun immediately after evacuation of an acute subdural hematoma. The graph shows a progressive increase in the
relative amount of slow activity (delta) averaged from four leads over each hemisphere. This accompanied by a progressive
fall in cerebral oxygen extraction, measured as the difference in oxygen content between a radial artery catheter and a
catheter placed in the right jugular bulb. The ICP was not elevated during this period. (b) Median nerve somatosensory
evoked potentials recorded from the contralateral hemispheres in the same patient shown in (a) The number of hours postinjury at which the traces were recorded is shown alongside each SSEP.
232
within the EEG (spindle pattern coma) tend to have

a better prognosis (Bergamasco et al., 1968; Bricolo et
al., 1982; Chatrian, White and Daly, 1963). The most
profound pattern of EEG dysfunction in post-traumatic coma, short of complete electrical silence, is
burst suppression. Unless it is drug-induced (i.e.
barbiturates), and therefore potentially reversible, it
is usually a preterminal finding.
Alpha coma refers to coma in the presence widespread alpha (812 Hz) activity. In contrast to normal
alpha activity, which occurs over the occiput and can
be induced in the normal subject by asking them to lie
quietly with the eyes closed, the alpha activity seen in
alpha coma is distributed over the entire scalp, and
does not vary spontaneously or with external stimuli.
Most writers agree that it indicates a poor prognosis
(Hari, Sulkava and Haltia, 1982; Obeso et al., 1980;
Stockard, Bickford and Aung, 1975; Westmoreland et
al., 1975), although this opinion is not unanimous
(Sorenson, 1978). Various EEG patterns have been
graded according to increasing degrees of abnormality, and the resultant grades are correlated with
prognosis (Synek, 1988, 1990a, b). In one series
mortality varied from 13% of patients with diphasic
spindle activity to 86% in patients with monophasic
activity and 100% of patients with electrocerebral
silence (Bricolo, 1976). Care must be taken to avoid
misinterpretation of profoundly abnormal EEG patterns (burst suppression, electrical silence) where the
possibility of drug overdose exists, as the EEG (and
processed versions thereof) is exquisitely sensitive to
drug effects. Most patients in modern head-injury
practice will have been given drugs that affect the
EEG.
The paper chart EEG is probably most useful
following head injury in the diagnosis and management of seizure activity, particularly status epilepticus. Subclinical seizures may continue following the
cessation of clinically evident seizure activity or may
occur without antecedent tonicclonic seizure activity. Fortunately, the latter is an infrequent complication of head injury (Bricolo, 1976). Suspicion of
subclinical seizures may be raised by the failure of a
patients level of consciousness to improve following
the cessation of observable seizure activity. In this
circumstance, or if pharmacological paralysis and
treatment with large doses of barbiturates are necessary to control post-traumatic status epilepticus, bedside monitoring of the EEG is required to ensure that
cortical seizure discharge has actually ceased, in the
absence of observable tonicclonic motor movements.
The cortical seizure discharge itself is harmful to the
brain, independent of the additional systemic physiological derangements (hypotension, lactic acidosis)
produced by the sustained motor activity (DelgadoEscueta et al., 1982).
In the years prior to the introduction of modern

imaging studies, an effort was made to characterize
the accuracy of the EEG in localizing mass lesions. The
accuracy of localization was never more than mediocre, i.e. 5080% (Courjon, 1972; Ravagnati et al., 1982;
Rumpl et al., 1979; Silverman, 1963; Stockard, Bickford
and Aung, 1975). In cases of minor injury, slowing is
often in the posterior leads, no matter where the CT
lesion is located (Liguori et al., 1989). Brain mapping
techniques appear to improve the localization of
structural lesions (see below).
The major drawbacks to EEG monitoring have been
the tremendous amount of data generated by the EEG
and the requirement for expert interpretation of the
recorded data. These factors have driven the efforts to
compress and simplify the interpretation of the EEG.
The nature of the EEG recording mandates expert
interpretation, and this is not available on a continuous
basis at the bedside. Offline expert analysis of even a
few hours of EEG is difficult because of the volume of
data. Also, the necessity for offline expert interpretation blunts the effectiveness of EEG as a monitoring tool. The major advantages of unprocessed EEG are
that it remains the only technique for reliably showing
transient activity and wave morphology.
(b)
The cerebral function monitor
The difficulties with EEG collection and interpretation

alluded to above have spawned a number of technologies aimed at data reduction and simplification of
interpretation of the EEG. One of the earlier and more
popular of these was the cerebral function monitor.
The monitor was devised in the period antedating the
reduction in the cost and complexity of computer
technology and its consequent widespread dissemination. The cerebral function monitor is a simplified strip
chart recording of a single channel of EEG activity. The
recording is carried out across the vertex in the
parietal region. The signal is subjected to a bandpass
filter with the passband from 215 Hz. The signal is
further modified by logarithmic compression of the
peak to peak amplitudes and recorded on a very slow
strip chart recorder (Maynard, Prior and Scott, 1969).
The technique provides information on whether background EEG activity is increasing or decreasing
overall, but little information on the frequency content, or transient phenomena. It is most useful for
detecting catastrophic global changes in cerebral
function, as might occur with severe hypoxia, hypotension or cardiac arrest (Levy et al., 1980; Schwartz
et al., 1973). More recently, the cerebral function
monitor has been used as an indicator of the level of
cortical metabolism and a guide for the use of
anesthetic agents to control intracranial pressure
(Procaccio et al., 1988).
(c)
Power spectral analysis of the EEG
The development of the fast Fourier transform algorithm (FFT) provided a computationally efficient (and
therefore rapid) means of resolving complex wave
patterns such as EEG into their individual frequency
components. Combined with the steady reduction in
the cost, size, and complexity of computer equipment
in the 1970s, this permitted the widespread introduction of equipment that was capable of real-time power
spectral analysis (PSA) of the EEG at the bedside. This
is currently the most widespread technique of data
reduction and simplification of interpretation of the
EEG. Bickford pioneered the technique for use with
EEG and coined the term compressed spectral analysis, based on the technique of displaying the power
spectra of the EEG over time (Bickford et al., 1973;
Bickford, 1977).
In order to produce an EEG power spectrum one
selects an epoch of EEG, typically of 24 seconds
duration. The FFT resolves the complex EEG waveform
into its frequency components and displays the power
(amplitude squared) in each frequency bin. (Typically
the bins are chosen to be 1 Hz.) In compressed spectral
analysis the frequency histogram generated by the FFT
is smoothed and displayed as a smooth line drawn
through the top of the histogram. Successive lines from
each epoch are stacked on top of the previous epoch.
The display can be printed out continuously or
displayed on a monitor so that trends in the background frequency can be displayed over time. Other
display systems employ color or gray-scale coding of
the frequency histogram. The speed of the FFT is such
that the power spectra can be calculated on a typical
microprocessor and displayed in real time, while the
data collection system captures and digitizes the next
epoch of EEG. Power spectra are displayed for each
channel and systems vary in complexity from very
small, portable two-channel systems to more complex
(and larger) systems with 16-channel capability and
more flexible storage and display capabilities.
A further refinement of power spectral analysis of
the EEG is brain electrical mapping, which involves
the interpolation of the power spectra values between
electrode positions on the scalp and the display of
these values as contour maps (Duffy, Burchfiel and
Lombroso, 1979). This technique may improve the
localization accuracy of EEG for structural lesions. In
one recent study the accuracy of localization of
structural lesions (abscess, glioma, hematoma) was
100% . There was a tendency for the lesions to appear
larger with brain mapping, perhaps reflecting an area
of physiological abnormality that was larger than the
anatomic lesion. Ischemic lesions may be shown that
are not evident with CT/MRI imaging (Jerret and
Corsak, 1988).
233
From a practical perspective, brain mapping is not

likely to displace CT imaging in the diagnosis of acute
operable structural lesions, because of the difficulty in
distinguishing physiological abnormalities associated
with non-operative structural lesions, from electrical
abnormalities emanating from the brain surrounding
operable structural lesions. However, EEG topographic
mapping or other techniques of electrical function
monitoring may be useful in showing the progression
of existing lesions, or providing early evidence of
development of delayed structural lesions.
The above techniques considerably compress and
simplify the interpretation of the EEG and these
reasons alone are often sufficient to justify their use. In
addition, quantification of the EEG allows for statistical
analysis of EEG data and permits graphic displays of
frequency trends over prolonged periods of time. The
degree of quantitative frequency resolution achievable
by PSA is not something that can be duplicated by
visual interpretation of time domain analog EEG
signal. There are a number of trade-offs for these
advantages. Probably the most important is that the
FFT technique averages the signal over the duration of
the epoch. While this makes no difference for the
analysis of background frequencies, it does remove the
ability to detect transients such as seizure spikes. In
addition, it removes any ability to look at the
morphology of EEG waves. For example, a seizure
discharge may not appear any different from ordinary
background activity of the same frequency, unless the
seizure discharge involves some major frequency shift
or change in amplitude. Even if these occur, it will still
be necessary to examine the raw EEG tracings to
confirm the presence of seizure activity, based on the
morphology of the raw EEG. The inclusion of all
recorded signal in the Fourier transform makes the
incorporation of physiological or external electrical
noise in the transformed EEG potentially problematic.
A certain amount of experience looking at raw EEG is
necessary to make this distinction. The ability of the
monitoring equipment to display the raw EEG tracing
to detect ECG or other non-EEG signal is essential.
Selection of grounding and reference leads can often be
optimized to minimize inclusion of non-EEG signal.
(d)
Power spectral analysis in traumatic coma
A number of authors have looked at the prognostic

capabilities of PSA in traumatic coma. The most
frequently used criterion for predicting a poor prognosis was a pattern of unvarying activity with the
major frequency component in the delta (13 Hz.)
band. Variable spectral patterns were generally associated with a better prognosis (Bricolo et al., 1978;
Sironi et al., 1982). This variation probably represents
the PSA equivalent of spontaneous or induced varia-
234
bility seen in the raw EEG and previously associated

with a good prognosis. Other authors have found that
the persistence or return of a peak in the alpha or theta
frequency band indicated a good outcome from
traumatic coma (Cant and Shaw, 1984; Steudel and
Kruger, 1979). Karnaze, Marshall and Bickford (1982)
found that the compressed spectral array was equivalent in prognostic accuracy to the Glasgow Coma
Score. The background frequencies of the EEG measured with power spectral analysis are highly correlated with the Glasgow Coma Score and the correlation improves with increasing time postinjury
(Moulton et al., 1988). This improvement in correlation with the depth of coma is probably due to the lag
between injury and slowing of the background frequencies described previously with conventional EEG.
As noted above, the delayed slowing takes place in
association with progressive disturbance of other
electrical function parameters, and measurements of
oxygen extraction (Figure 12.1).
In our experience with monitoring the EEG power
spectrum we have found that the overall trend in
background frequency, viewed over several days,
usually mirrors patient improvement or deterioration.
We monitor the relative amplitude within the four
customary frequency bands (i.e. the fraction of total
power contributed by the delta, theta, alpha and beta
frequency bands) rather than the absolute power in
each band. This tends to reduce the amount of
variability over time and makes trends more easily
discernible (Figure 12.2). In general there is a steady
increase in the amount of slow (delta) activity from the
time of initiation of monitoring in patients who die, or
survive with severe disabilities. Patients who survive
with a functional outcome usually either show reduction in the amount of slow activity, or little or no net
change over time. Using discriminant analysis of
frequency content with the jack-knife technique of
error estimation we found that the prognostic accuracy for good or moderate outcome versus death,
vegetative survival, or severe disability was 75%.
There were a number of falsely pessimistic predictions
because of transient reversible complications that
were reflected in the EEG (Moulton et al., 1988). With
the luxury of viewing the PSA trends post hoc after
several hours (or days) post-injury, these trends are
usually easily seen. However, because of the variability inherent in the measure, particularly in patients
who do well, the meaning of background frequency
changes at any given moment in time is much less
clear (Figure 12.3). For this reason, although we do
monitor the EEG power spectrum continuously, we
find the information much less useful than that
obtained from evoked potentials, which have much
less variability over time. Occasionally, in cases of a
significant global insult, processed EEG may more
clearly reflect the impact on cerebral function than

other electrophysiological data (Figure 12.2(a)).
Other information available from analysis of the
EEG in the frequency domain includes phase and
coherence data. These reflect the lead or lag time
between shared activity in different regions of the
brain, and the amount of shared activity between two
regions in the brain, respectively (Thatcher et al., 1989).
Generally with increasing severity of head injury one
sees increased coherence and reduced phase. The
latter may be reflective of subcortical white matter
damage and tends to be more stable over time than
measures of relative or total power, possibly indicating a degree of resistance to acute injury dynamics
(brain edema, intracranial pressure etc.). Use of phase
and coherence data from the EEG significantly
improves the prognostic accuracy of EEG following
head injury, possibly reflecting better measurement of
the predominant pathology in blunt head injury, i.e.
axonal damage in the white matter of the hemispheres
(Thatcher et al., 1991).
12.3.2
(a)
EVOKED POTENTIALS
Technical principles
In contrast to EEG, which is a measurement of the

spontaneous electrical activity of the cerebral cortex,
evoked potentials measure cerebral hemispheric and/
or brain-stem responses to stimulation of a sensory
receptor (i.e. visual, auditory or somatosensory). The
typical response is of the order of 0.55 V as opposed
to 10100 V for spontaneous cortical activity (i.e.
EEG). A single evoked response is therefore lost in the
background random EEG activity. Signal averaging
techniques are required to demonstrate the response.
Repeated stimuli are used at fixed intervals that are
dependent on the particular evoked response to be
studied. The evoked responses, which occur at a fixed
time from the stimulus, are preserved by the averaging process, whereas the background EEG, which
occurs randomly with respect to the stimulus, is
canceled out by averaging. Evoked electrical potentials arrive at the scalp by two processes: conduction
along the fiber tracts of the sensory pathway in
question and volume conduction. For practical purposes, volume conduction is virtually instantaneous
with respect to conduction along axons, so that
potentials arising from synaptic relays along the
course of the sensory pathways arrive in an orderly
fashion from distal (brain stem) to proximal (cortex).
(b)
Visual evoked potentials
These are generally evoked by using a strobe light or

checkerboard stimulus. They yield information about
235
(a)
(b)
Figure 12.2 (a) EEG power spectrum monitoring in a 50-year-old man following severe closed head injury. This man
aspirated shortly after injury and developed severe bilateral aspiration pneumonia and ARDS. The monitored parameter is
the relative amount of slow activity averaged from four leads over each hemisphere. At 36 hours post-injury the patients
blood pressure fell to 60 mmHg and the PO2 fell to 60 mmHg on 100% oxygen. The large increase in slow activity occurred
at that time. Two days later a CT scan showed large bilateral frontal infarctions around two small contusions. The patient
died at 1 week post-injury. (b) EEG power spectrum monitoring in the same patient. The monitored parameter in this
instance is the absolute amplitude in the delta band rather than the relative amplitude. The large amount of variability from
hour to hour effectively conceals the increase in slow activity seen in (a).
236
Figure 12.3 The EEG power spectrum monitored over 3 days post-injury in a 40-year-old man following evacuation of an
acute subdural hematoma. The relative amount of slow activity (delta) is averaged from four leads per hemisphere. The
overall trend is one of decreasing slow activity and the patient went on to a functional recovery. The large amount of
variability over time, particularly in the mid-portion of the recording, makes the real-time interpretation of the data difficult.
These fluctuations did not occur in the presence of any detectable systemic physiological perturbation, nor any ICP
elevation. They may represent diurnal variation in the EEG in a patient with a good prognosis.
the functional integrity of the visual apparatus and

may be useful in indicating the presence of visual loss
in a comatose patient (Feinsod and Auerbach, 1973;
Mahapatra and Bhatia, 1989). In terms of predicting
the outcome from head injury, they are less useful than
somatosensory evoked potentials (see below) (Lindsay et al., 1981), although they do marginally improve
the prognostic accuracy of evoked potentials when
used in combination with other evoked potential
modalities (Narayan et al., 1981).
(c)
Brain-stem auditory evoked potentials
Brain-stem auditory evoked potentials (BAEPs) have

been used in a continuous monitoring application and
have been shown to progressively deteriorate in the
preterminal stages of coma following head injury
(Bertrand et al., 1987; Garcia-Larrea et al., 1987). This
deterioration is not necessarily related to uncontrollable intracranial pressure (Garcia-Larrea et al.,
1992). In our experience with hourly monitoring of
both somatosensory evoked potentials (SSEPs) and
BAEPs we have found that SSEP deterioration up to
and including complete loss of cortical activity
precedes loss of peaks in the BAEP (Figure 12.4). The
stage of brain dysfunction that precludes the possibility of a functional outcome may well occur before
there is significant change in the BAEP. This probably
explains the frequent occurrence of poor outcome from

head injury in the presence of normal or nearly normal
BAEPs (Cant et al., 1986; Cant, 1987; Lindsay et al.,
1981; Papanicolaou et al., 1986). While the absence of
BAEP activity certainly predicts a poor outcome, the
presence of brain-stem responses does not necessarily
predict survival or a functional outcome. The quality
of survival depends on cerebral hemispheric integrity,
and this is not something that is measured by the
BAEP. Near field auditory potentials have been found
to be of greater value in predicting functional outcome
(Lindsay et al., 1981; Ottaviani et al., 1986), but these
are not commonly done and cortical functional integrity is more commonly measured with somatosensory
evoked responses.
BAEPs are usually evoked by using a click stimulus
delivered to a single ear using an earphone or earplug
apparatus. Optimal recordings are obtained from C0
referenced to the ipsilateral ear. The duration of
recording of the response is typically 10 ms. Five peaks
are produced as the stimulus traverses the brain stem.
Wave I is thought to arise from the auditory nerve
itself. Waves II and III arise from the caudal pons and
waves IV and V from rostral pons and mesencephalon
(Legatt, Arezzo and Vaughan, 1988). The wave IV
interpeak latency therefore gives an indication of the
transit time of an electrical signal across the brain
stem. Abnormalities of BAEP conduction time do
237
Figure 12.4 Serial BAEPs and SSEPs (positive waves down for both) in a patient with a severe diffuse head injury.
Monitoring was begun at approximately 5 hours post-injury at 22.00h. There is progressive loss of cortical SSEP activity over
time from levels initially compatible with good quality survival. At 03.15h cortical activity has been absent bilaterally for at
least 90 minutes. At this time the BAEP shows only some prolongation of the lV interpeak latency. Waves IIIV disappear
some 2.5 hours later. ICP was maintained at less than 25 mmHg during the period in which cortical SSEPs disappeared. The
BAEPs were lost when the ICP subsequently became uncontrollable.
correlate with poor outcome, but the correlation is not

as good as that of the central conduction time of the
somatosensory evoked potential (Cant et al., 1986;
Lindsay et al., 1990). In head-injury patients the entire
BAEP, including wave I, may be absent because of
end-organ damage (e.g. petrous bone fracture).
Absence of wave I following closed head injury is a
fairly common circumstance and prevents satisfactory interpretation of BAEPs in a significant number
of patients. In the situation of an absent wave I one
cannot reliably draw any conclusions about brainstem function because of the possibility that its
absence is due to end-organ damage rather than
impaired brain-stem function (Cant, 1987). In contradistinction, if waves IIV are absent and wave I is
present, this indicates severe brain-stem disturbance
and indicates a very poor prognosis (Cant et al., 1986;
Garcia-Larrea et al., 1992; Hall, Huang-fu and Gennarelli, 1982). BAEP abnormalities correlate with pupillary abnormalities and other clinical evidence of brainstem dysfunction but do not correlate with motor
posturing (Barelli et al., 1991; Greenberg et al., 1977;
Lindsay et al., 1990).
(d)
Somatosensory evoked potentials
Somatosensory evoked potentials (SSEPs) are generated by stimulating a peripheral nerve containing
large myelinated sensory fibers, usually the median at
the wrist or the posterior tibial at the ankle. The
impulses are conducted centrally through the dorsal
column system, traverse the brain stem and produce a
series of peaks from the cerebral hemisphere. The
number and morphology of the latter depend in part
on the duration of the recording. The standard
recording montage for intermediate and long-latency
activity consists of recording just posterior to the C3
and C4 electrode positions (C3 and C4) referenced to
linked ears. Additional electrode positions at Erbs
point (located over the brachial plexus) and over the
spinous process of the second cervical vertebra are
used if one wishes to examine the somatosensory farfield potentials in greater detail. This is often done for
precise recording of the central conduction time, i.e.
the transit time of the somatosensory response from
the caudal medulla to the primary somatosensory
cortex in the postcentral gyrus.
238
The naming convention for SSEP peaks is based on

the polarity (positive P or negative N) and the poststimulus latency in milliseconds. The P15 is the first
deflection seen in the SSEP and arises from the caudal
medulla. The N20 is the first cortical peak and is felt to
arise from the postcentral gyrus. The initial cortical
peak may have a frontal component and parietal and
frontal components may be lost independently of each
other following injury (Gutling
et al., 1993). Later

peaks likely arise from association cortex or reverberating circuits between the cortex and subcortical
structures (Greenberg and Ducker, 1982; Yamada,
1988). While there is a reasonably established consensus about the origins of the P15 and N20, there is no
well-established consensus regarding the site of origin
of the later peaks. These later peaks also have
considerably more variability in their latencies, both
between and within patients (Greenberg and Ducker,
1982; Yamada, 1988).
Unlike the fairly constant morphology of the BAEP
and the relatively constant measure of the wave IV
latency, the variability of the SSEP leads to some
problems in classifying/grading the degree of abnormality of SSEP responses. The one exception to this is
the use of the central conduction time, as the peaks on
which calculation of this value depend are quite
constant (Hume and Cant, 1981; Lindsay et al., 1990;
Rumpl et al., 1983; Whittle, Johnston and Besser, 1987).
However, there may be some loss of sensitivity in
detecting SSEP change because the method ignores
cortical activity occurring after the N20 (Moulton,
Konaciewicz and OConnor, in press). Our work and
that of others, in cases of post-traumatic and postanoxic coma, suggests that, while the absence of early
cortical activity is an excellent predictor of survival
or death, it is activity occurring at or beyond 70 ms
that more accurately reflects the probability of a good
quality outcome (see below). A number of other
approaches to grading SSEPs have been described,
including fairly complex (and subjective) grading
schemes based on the presence or absence of peaks at
particular latencies and the number, amplitude and
morphology of these peaks (Greenberg et al., 1977;
Rappaport et al., 1981). Other approaches have
included simple peak counts (Lindsay et al., 1981), or
grading based on the longest latency peak present in
the SSEP response (Moulton et al., 1991; De La Torre et
al., 1978). There is a substantial degree of intercorrelation between these parameters, the more subjective grading schemes and the central conduction
time. Consequently the results are similar when these
grading paradigms are employed in the clinical
setting, i.e. correlation with patient outcome (Moulton, Konaciewicz and OConnor).
Over time the SSEP has come to be recognized as
the single best predictor of outcome among the
available evoked potential measurements (Cant, 1987;

Lindsay et al., 1981, 1990). This is very probably due
to the ability of the SSEP to sample the function of the
entire intracranial neuraxis from brain stem to cortex,
and the heavy dependence of survival and the quality
of survival on the integrity of both brain stem and
cerebral hemispheric function. The response is usually
relatively easy to record and, unlike the BAEP, is
seldom lost on account of associated injuries. The two
exceptions are injury to the spinal cord or brachial
plexus (the latter in the case of upper extremity
evoked responses).
The absence of any activity beyond the P15 is
highly predictive of death. The presence of increasingly long latency peaks correlates with increasing
quality of survival. Presence of SSEP activity beyond
5070 ms post-stimulus appears essential for functional survival (De La Torre et al., 1978; Moulton et
al., 1991). This appears to be true in cases of postcardiac-arrest anoxic coma (Madl et al., 1993), in
addition to post-traumatic coma . In our hands the
sensitivity, specificity, and positive predictive value
for the bilateral absence of activity at or beyond N70
in predicting death or vegetative survival were 0.61,
0.95 and 0.89 respectively. The sensitivity, specificity,
and positive predictive value of bilateral preservation
of activity at N70 or later in predicting good outcome
or moderate disability were 1.0, 0.75 and 0.55 respectively (Moulton et al., in press). The poor positive
predictive value for good outcome or moderate
disability seemed to be due to the effect of age on
outcome. Elderly patients tended to do poorly even
in the presence good evoked potentials, because of
the independent effect of age on outcome from head
injury (Shedden et al., 1990). On the other hand,
although bilateral intermediate and late cortical SSEP
activity is not sufficient grounds for good outcome or
moderate disability, it is a necessary one. In our
hands no patients had a good outcome or moderate
disability who did not have bilateral intermediate
and late cortical activity. Unilateral or bilateral
absence of activity has a high positive predictive
value for severe disability, vegetative survival or
death. The effect of bilateral absence of activity is
independent of age (Moulton et al.).
Early recordings of SSEPs tended to be done on a
sporadic basis, often 4872 hours postinjury or more.
These single or infrequent studies were adequate to
establish the prognosis from injury, but were not
adequate for patient monitoring. When serial testing
was initiated, it became apparent that changes in
SSEPs do occur. Newlon et al. felt that deterioration in
SSEPs was due to secondary insults such as raised
intracranial pressure (ICP), hypoxia or delayed hematomas (Newlon et al., 1982). Lindsay et al. also detected
changes in SSEPs recorded early after injury and 4872
239
Figure 12.5 Serial somatosensory evoked potentials in a man following evacuation of an acute subdural hematoma. On the
fourth post-injury day loss of SSEP activity above 25 ms latency on the left side triggered a repeat CT scan. A small left temporal
contusion present at the time of admission had increased in size and a left temporal lobectomy was subsequently undertaken.
There was immediate recovery of SSEP activity following operation. The ICP never exceeded 20 mmHg prior to evacuation of
the contusion. The labels on the tracings are the time post-injury (days, 24-hour clock beginning on fourth day).
hours later. They did not speculate on reasons for

these changes, but simply stated that early SSEP
recordings were not as reliable for determination of
prognosis (Lindsay et al., 1990). In our own recordings
of SSEPs, done on an hourly basis, we found that
changing SSEP activity was extremely common, with
deterioration occurring in approximately two-thirds
of patients whose monitoring was begun within 24
hours of injury. These changes occurred bilaterally
(Figures 12.1, 12.4). In the majority of patients no
clearcut secondary insult could be identified to
account for the deterioration. In a much smaller group
of patients we identified unilateral changes in SSEPs
occurring as the result of evolving focal pathology
(Figure 12.5). A small number of patients showed
improvement following injury (Moulton et al., 1991).
We have been unable to relate deterioration in SSEP
values to levels of ICP (Konasiewicz, Moulton and
Shedden, 1993; Moulton et al., in press). The deterioration does seem to correspond to reduced levels of
transcranial oxygen extraction (i.e. the arterio-jugular
oxygen difference) possibly implicating a role for
perturbation of oxidative metabolism in the genesis of
SSEP deterioration (Figure 12.1; Moulton et al., in
press). The correspondence between SSEP deterioration and AVDO2 decrease is not absolute, and the
observed deterioration in SSEP activity is probably
multifactorial.
(e) Multimodality evoked potential monitoring
The prognostic accuracy of the SSEP may be improved
to some extent by monitoring visual and brain-stem
auditory evoked responses as well (Narayan et al.,
1981). The additional improvement in accuracy is
marginal and may not justify the increased effort and
expense of monitoring other electrophysiological
parameters, including the EEG or its power spectrum.
Perhaps the one exception to this is the BAEP, as it
does have the desirable feature of being impervious to
the effects of the large doses of barbiturates that are
occasionally necessary to control refractory increases
in ICP (Newlon et al., 1983). While relatively resistant
to the usual analgesic or sedative doses of narcotics,
cortical SSEP responses can be abolished by large
doses of barbiturates. From time to time other
electrical function measurements may be useful in
diagnosing specific complications such as seizures
240
(EEG) or visual loss in an uncooperative patient

(VEPs), but the workhorse evoked potential for
prognosis and monitoring will probably remain the
SSEP. In patients who die, loss of cortical somatosensory evoked potentials (a highly reliable indicator of
poor outcome) precedes electrocortical silence by
hours or even 12 days (Ganes and Lundar, 1988). A
direct comparison of the prognostic capability of EEG
and somatosensory evoked potentials has shown the
latter to be clearly superior (Hutchinson, 1991). Motor
evoked responses have not been shown to be useful
in predicting outcome from coma (Zentner and
Rohde, 1992).
12.4
Conclusions
At the present time the value of evoked potential

examination is proven for prognostication in cases of
traumatic coma. In ordinary circumstances the accuracy of prediction is approximately that which is
achievable by clinical measures alone, so that measurement is probably not cost-effective in this circumstance. However, in cases where clinical monitoring is
difficult or impossible, e.g. pharmacological paralysis
for respiratory or ICP management, information from
evoked potentials may be extremely useful in deciding
on continuation or cessation of very aggressive
therapy.
The need for on-line, objective electrophysiological
monitoring of brain function seems self-evident. However, the combination of the expense, and difficulty of
use of such equipment on the one hand, and our
limited ability to usefully intervene in patients with
declining neurological status on the other, have
tended to restrict the application of such monitoring to
university-affiliated tertiary referral centers. Even in
this limited application, these techniques will remain
useful in the immediate future for defining important
pathophysiological sequelae of injury, and distinguishing these from epiphenomena. With the advent
of new therapies for closed head injury, further
improvement in monitoring equipment and techniques and a reduction of equipment cost, it is likely
that brain electrical function monitoring will play an
increasingly important role in evaluating and later
directing the application of these new therapies.
12.5
References
Barelli, A., Valente, M. R., Clemente, A. et al. (1991) Serial multimodalityevoked potentials in severely head-injured patients: Diagnostic and
prognostic implications. Critical Care Medicine, 19, 13741381.
Bergamasco, B., Bergamini, L., Doriguzzi, T. et al. (1968) EEG sleep patterns as
a prognostic criterion in post-traumatic coma. Electroencephalography and
Clinical Neurophysiology, 24, 374377.
Bertrand, O., Garcia-Larrea, L., Artru, F. et al. (1987) Brain-stem monitoring. I.
A system for high-rate sequential BAEP recording and feature extraction.
Electroencephalography and Clinical Neurophysiology, 68, 433445.
Bickford, R. G. (1977) Computer analysis of background activity, in EEG

Informatics. A Didactic Review of Methods and Applications of EEG Data
Processing, (ed. A. Remond), Elsevier/North-Holland, Amsterdam, pp.
215232.
Bickford, R. G., Brimm, J., Berger, L. et al. (1973) Application of compressed
spectral array in clinical EEG, in Automation of Clinical Electroencephalography, (eds P. Kellaway and I. Petersen), Raven Press, New York, pp.
5564.
Bricolo, A. (1976) Electroencephalography in neurotraumatology. Clinical
Electroencephalography, 7, 184197.
Bricolo, A., Turazzi, S., Faccioli, F. et al. (1978) Clinical application of
compressed spectral array in long-term EEG monitoring of comatose
patients. Electroencephalography and Clinical Neurophysiology, 45, 211225.
Bricolo, A., Faccioli, F., Turazzi, S. et al. (1982) EEG and evoked potentials in
brainstem traumatic lesions, in Advances in Neurotraumatology, (eds R.
Villani, I. Papo, M. Giovanelli et al.), Excerpta Medica, Amsterdam, pp.
7984.
Cant, B. R. and Shaw, N. A. (1984) Monitoring by compressed spectral array
in prolonged coma. Neurology, 34, 3539.
Cant, B. R. (1987) Evoked potential monitoring of post-traumatic coma and its
relation to outcome, in The London Symposia (Electroencephalography and
Clinical Neurophysiology Suppl. 39), (eds R. J. Ellingson, N. M. F. Murray and
A. M. Halliday), Elsevier, Amsterdam, pp. 250254.
Cant, B. R., Hume, A. L., Judson, J. A. et al. (1986) The assessment of severe
head injury by short-latency somatosensory and brain-stem auditory
evoked potentials. Electroencephalography and Clinical Neurophysiology, 65,
188195.
Chatrian, E. G., White, L. E. and Daly, D. (1963) Electroencephalographic
patterns resembling those of sleep in certain comatose states after injuries
to the head. Electroencephalography and Clinical Neurophysiology, 15,
272280.
Courjon, J. (1972) Traumatic Disorders, in Handbook of Electroencephalography
and Clinical Neurophysiology, vol. 14, Part B (ed. A. Remond), Elsevier,
Amsterdam, pp. 839.
Dawson, R. E., Webster, J. E. and Gurdjian, E. S. (1951) Serial electroencephalography in acute head injuries. Journal of Neurosurgery, 8,
613630.
Delgado-Escueta, A. V., Wasterlain, C., Treiman, D. M. et al. (1982) Management of status epilepticus. New England Journal of Medicine, 306,
13371340.
De La Torre, J. C., Trimble, J. L., Beard, R. T. et al. (1978) Somatosensory evoked
potentials for the prognosis of coma in humans. Experimental Neurology, 60,
304317.
Duffy, F. H., Burchfiel, J. L. and Lombroso, C. T. (1979) Brain electrical activity
mapping (BEAM): a method for extending the clinical utility of EEG and
evoked potential data. Annals of Neurology, 5, 309321.
Dusser, A., Navelet, Y., Devictor, D. et al. (1989) Short- and long-term value of
the electroencephalogram in children with severe head injury. Electroencephalography and Clinical Neurophysiology, 73, 8593.
Feinsod, M. and Auerbach, E. (1973) Electrophysiological examinations of the
visual system in the acute phase after head injury. European Neurology, 9,
5664.
Ganes, T. and Lundar T. (1988) EEG and evoked potentials in comatose
patients with severe brain damage. Electroencephalography and Clinical
Neurophysiology, 69, 613.
Garcia-Larrea, L., Bertrand, O., Artru, F. et al. (1987) Brain-stem monitoring. II.
Preterminal BAEP changes observed until brain death in deeply comatose
patients. Electroencephalography and Clinical Neurophysiology, 68, 446457.
Garcia-Larrea, L., Artru, F., Bertrand, O. et al. (1992) The combined monitoring
of brain stem auditory evoked potentials and intracranial pressure in coma.
A study of 57 patients. Journal of Neurology, Neurosurgery, and Psychiatry, 55,
792798.
Gibbs, F. A. and Gibbs, E. L. (1950). Atlas of Electroencephalography. Volume 1:
Methodology and Controls, Addison-Wesley, Cambridge, MA.
Greenberg, R. P. and Ducker, T. B. (1982) Evoked potentials in the clinical
neurosciences. Journal of Neurosurgery, 56, 118.
Greenberg, R. P., Mayer, D. J., Becker, D. P. et al. (1977) Evaluation of brain
function in severe human head trauma with multimodality evoked
potentials. Part 1: Evoked brain-injury potentials, methods, and analysis.
Greenberg, R. P., Becker, D. P., Miller, J. D. et al. (1977) Evaluation of brain
function in severe human head trauma with multimodality evoked
potentials. Part 2: Localization of brain dysfunction and correlation with
posttraumatic neurological conditions. Journal of Neurosurgery, 47,
163177.
Gutling,
E., Gonser, A., Regard, M. et al. (1993) Dissociation of frontal and

parietal components of somatosensory evoked potentials in severe head
injury. Electroencephalography and Clinical Neurophysiology, 88, 369376.
Hall, J. W., Huang-fu, M. and Gennarelli, T. A. (1982) Auditory function in
acute severe head injury. Laryngoscope, 92, 883890.
Hari, R., Sulkava, R. and Haltia, M. (1981) Brainstem auditory evoked
responses and alpha-pattern coma. Annals of Neurology, 11, 187189.
Hume, A. L. and Cant, B. R. (1981) Central somatosensory conduction after
head injury. Annals of Neurology, 10, 411419.
REFERENCES
Hutchinson, D. O., Frith, R. W., Shaw, N. A. et al. (1991) A comparison
between electroencephalography and somatosensory evoked potentials for
outcome prediction following severe head injury. Electroencephalography and
Clinical Neurophysiology, 78, 228233.
Jasper, H. H. (1958) The ten twenty electrode system of the international
federation. Electroencephalography and Clinical Neurophysiology, 10, 371375.
Jerrett, S. A. and Corsak, J. (1988) Clinical utility of topographic EEG brain
mapping. Clinical Electroencephalography, 19, 134143.
Karnaze, D. S., Marshall, L. F. and Bickford, R. G. (1982) EEG monitoring of
clinical coma: the compressed spectral array. Neurology (NY), 32, 289292.
Konasiewicz, S. J., Moulton, R. J. and Shedden, P. M. (1993) The relationship
of intracranial pressure to neurologic deterioration: a study using a
quantitative somatosensory evoked potential measurement. Journal of
Neurotrauma, 10(Suppl. 1), S58.
Legatt, A. D., Arezzo, J. C. and Vaughan H. G. (1988) The anatomic and
physiologic bases of brain stem auditory evoked potentials. Neurologic
Clinics, 6, 681704.
Levy, W. J., Shapiro, H. M., Maruchak, G. et al. (1980) Automated EEG
processing for intraoperative monitoring: a comparison of techniques.
Anesthesiology, 53, 223236.
Liguori, G., Foggia, L., Buonaguro, M. et al. (1989) EEG findings in minor head
trauma as a clue for indication to CT scan. Childs Nervous System, 5,
160162.
Lindsay, K. W., Carlin, J., Kennedy, I. et al. (1981) Evoked potentials in severe
head injury- analysis and relation to outcome. Journal of Neurology,
Neurosurgery, and Psychiatry, 44, 796802.
Lindsay, K., Pasaoglu A., Hirst, D. et al. (1990) Somatosensory and auditory
brain stem conduction after head injury: a comparison with clinical features
in prediction of outcome. Neurosurgery, 26, 278285.
Madl, C., Grimm, G., Kramer, L. et al. (1993) Early prediction of individual
outcome after cardiopulmonary resuscitation. Lancet, 341, 855858.
Mahapatra, A. K., and Bhatia R. (1989) Predictive value of visual evoked
potentials in unilateral optic nerve injury. Surgical Neurology, 31, 339342.
Maynard, D., Prior, P. F. and Scott, D. F. (1969) Device for continuous
monitoring of cerebral activity in resuscitated patients. British Medical
Journal, 4, 545546.
Moulton, R. J., Konasiewicz, S. K., and OConnor, P. (1994) Development and
validation of a quantitative measure for continuously monitored somatosensory evoked potentials. Canadian Journal of Neurological Sciences, 21(2),
517522.
Moulton, R. J., Marmarou A., Ronen, J. et al. (1988) Spectral analysis of the EEG
in craniocerebral trauma. Canadian Journal of Neurological Sciences, 15, 8286.
Moulton, R., Kresta, P., Ramirez M. et al. (1991) Continuous automated
monitoring of somatosensory evoked potentials in posttraumatic coma.
Moulton, R. J., Shedden, P. M., Tucker, W. S. et al. (1994) Somatosensory
evoked potential monitoring following severe closed head injury. Clinical
and Investigative Medicine, 17(3), 187195.
Narayan, R. K., Greenberg, R. P., Miller, J. D. et al. (1981) Improved confidence
of outcome prediction in severe head injury. A comparative analysis of the
clinical examination, multimodality evoked potentials, CT scanning, and
intracranial pressure. Journal of Neurosurgery, 54, 751762.
Newlon, P. G., Greenberg, R. P., Hyatt M. S. et al. (1982) The dynamics of
neuronal dysfunction and recovery following severe head injury assessed
with serial multimodality evoked potentials. Journal of Neurosurgery, 57,
168177.
Newlon, P. G., Greenberg, R. P., Enas, G. G. et al. (1983) Effects of therapeutic
pentobarbital coma on multimodality evoked potentials recorded from
severely head-injured patients. Neurosurgery, 12, 613619.
Obeso, J. A. Inagui, M. I., Marti-Masso, J. F. et al. (1980) Neurophysiological
assessment of alpha pattern coma. Journal of Neurology, Neurosurgery, and
Ottaviani, F., Almadori, G., Calderazzo, A. B. et al. (1986) Auditory brain-stem
(ABRs) and middle latency auditory responses (MLRs) in the prognosis of
severely head-injured patients. Electroencephalography and Clinical Neurophysiology, 65, 196202.
241
Papanicolaou, A. C., Loring, D. W., Eisenberg, H. M. et al. (1986) Auditory

brain stem evoked responses in comatose head-injured patients. Neurosurgery, 18, 173175.
Procaccio, F., Bingham, R. M., Hinds, C. J. et al. (1988) Continuous EEG and
ICP monitoring as a guide to the administration of althesin sedation in
severe head injury. Intensive Care Medicine, 14, 148155.
Rappaport, M., Hall, K., Hopkins, H. K. et al. (1981) Evoked potentials and
head injury. I. Rating of evoked potential abnormality. Clinical Electroencephalography, 12, 154166.
Ravagnati, L., Iacuone, M. M., Sironi, V. A. et al. (1982) Small cerebral lesions
in minor head injuries: CTEEG correlation, in Advances in Neurotraumatology, (eds R. Villani, I. Papo, M. Giovanelli et al.), Excerpta Medica,
Rumpl, E., Lorenzi, E., Hackl, J. M. et al. (1979) The EEG at different stages of
acute secondary traumatic midbrain and bulbar brain syndromes. Electroencephalography and Clinical Neurophysiology, 46, 487497.
Rumpl, E., Prugger, M., Gerstenbrand, J. M. et al. (1983) Central somatosensory conduction time and short latency somatosensory evoked potentials in
post-traumatic coma. Electroencephalography and Clinical Neurophysiology, 56,
583596.
Schwartz, M. S., Colvin, M. P., Prior, P. F. et al. (1973) The cerebral function
monitor: its value in predicting the neurological outcome in patients
undergoing cardiopulmonary by-pass. Anaesthesia, 28, 611618.
Shedden, P. M., Moulton, R. J., Sullivan, I. et al. (1990) Effect of population
characteristics on head injury mortality. Pediatric Neurosurgery, 16,
203207.
Silverman, D. (1963) Retrospective study of the EEG in coma. Electroencephalography and Clinical Neurophysiology, 15, 486503.
Sironi, V. A., Ravagnati, L., Signoroni, G. et al. (1982) Diagnostic and
prognostic value of EEG compressed spectral analysis in post-traumatic
coma, in Advances in Neurotraumatology, (eds R. Villani, I. Papo, M.
Giovanelli et al.), Excerpta Medica, Amsterdam, pp. 328330.
Sorenson, K., Thomassen, A. and Wernberg, M. (1978) Prognostic significance
of alpha frequency EEG rhythm in coma after cardiac arrest. Journal of
Neurology, Neurosurgery, and Psychiatry, 41, 84042.
Steudel, W. I. and Kruger, J. (1979) Using the spectral analysis of the EEG for
prognosis of severe brain injuries in the first post-traumatic week. Acta
Stockard, J. J., Bickford, R. G. and Aung, M. H. (1975) The electroencephalogram in traumatic brain injury, in Handbook of Clinical Neurology, vol. 23,
Injuries of the Brain and Skull, (eds R. J. Vinken and G. W. Bruyn), Elsevier,
Synek, V. M. (1988) EEG abnormality grades and subdivisions of prognostic
importance in traumatic and anoxic coma in adults. Clinical Electroencephalography, 19, 160166.
Synek, V. M. (1990a) Revised EEG coma scale in diffuse acute head injuries in
adults. Clinical and Experimental Neurology, 27, 99111.
Synek, V. M. (1990b) Revised EEG coma scale in diffuse acute head injuries in
adults. Clinical and Experimental Neurology, 27, 99111.
Thatcher, R. W., Walker, R. A., Gerson, I. et al. (1989) EEG discriminant
analyses of mild head trauma. Electroencephalography and Clinical Neurophysiology, 73, 94106.
Thatcher, R. W., Cantor, D. S., McAlaster, R. et al. (1991) Comprehensive
predictions of outcome in closed head-injured patients. The development of
prognostic equations. Annals of the New York Academy of Sciences, 620,
82101.
Westmoreland, B. F., Klass, D. W., Sharbrough, F. W. et al. (1975) Alpha-coma:
electroencephalographic, clinical, pathologic, and etiologic correlations.
Archives of Neurology, 32, 713718.
Whittle, I. R., Johnston I. H. and Besser, M. (1987) Short latency somatosensory-evoked potentials in children Part 3. Findings following head injury.
Yamada, T. (1988) The anatomic and physiologic bases of median nerve
somatosensory evoked potentials. Neurologic Clinics, 6, 705733.
Zentner, J. and Rohde, V. (1992) The prognostic value of somatosensory and
motor evoked potentials in comatose patients. Neurosurgery, 31, 429434.
13
TRANSCRANIAL DOPPLER
Peter J. Kirkpatrick and Kwan-Hon Chan
13.1
Introduction
The study of large numbers of head-injured patients

has identified clinical and radiological features which
indicate the severity of the initial cerebral trauma.
(Born et al., 1985; Jennett et al., 1979; Marshall et al.,
1991; Miller, 1985; Miller, 1992). In some individuals,
secondary neuronal injury follows the primary injury
(Chan et al., 1992a; Gopinath et al., 1994; Kirkpatrick et
al., 1995; Miller, 1986; Robertson et al., 1989). These
secondary insults may, in part, account for the poor
predictive value of early clinical findings. Until
recently, the nature of secondary neuronal injury
following severe head trauma has remained elusive.
However, modern methods for monitoring various
physical and biochemical parameters indicate that
potentially adverse episodes can be detected (Bouma
et al., 1992; Chan, Dearden and Miller, 1992; Chan,
Miller and Dearden, 1992; Chan, Miller and Piper,
1992; Chan et al., 1992a, b; 1993; Cruz et al., 1991; Cruz,
1993; Czosnyka et al., 1994a; Jones et al., 1993;
Kirkpatrick et al., 1994a, 1995, 1996; Kirkpatrick,
Czosnyka and Pickard, 1996) and that some are
important in terms of prognosis. Low cerebral blood
flow (CBF) values in the first few hours after injury,
and profound cerebral hypoxia are events which
predict a poor outcome (Bouma and Muizelaar, 1990;
Chan et al., 1992a; Gopinath et al., 1994). Episodes of
hypoxia, hypotension and reduced cerebral perfusion
due to high levels of intracranial pressure (ICP) are
important factors that reduce cerebral blood flow and
oxygen delivery (Siesjo,
1992). Cerebral ischemia
ensues if compensation by increased oxygen extraction is incomplete. Since histological ischemic changes
are found in the brains of up to 80% of patients dying
from head injuries, the provision for monitoring
cerebral ischemic episodes appears important, and has
received wide attention (Kirkpatrick, Czosnyka and
Pickard., 1996). The reliable detection and quantification of such episodes has, however, proven difficult.
The depth and duration of secondary cerebral
ischemic episode varies, and recent evidence indicates
that such episodes may last only a few minutes (Cruz,

1993; Kirkpatrick, Czosnyka and Pickard, 1996;
Robertson et al., 1989). The need for real-time measurements is therefore apparent. The imaging methods
for estimations of CBF and oxygenation using contrast
agents and isotopes (xenon-enhanced CT, SPECT and
PET) are invasive, and may require transfer of critical
patients to specialized facilities (Bouma and Muizelaar, 1990; Choksey et al., 1991; Gonalaves et al., 1994;
Marion, Darby and Yonas, 1991; Meixensberger, 1993;
Obrist and Wilkinson, 1990). Most importantly, they
only allow isolated estimations which will miss
transient secondary episodes. Their routine use in the
intensive care of head-injured patients is clearly
inappropriate. In response, monitors that allow the
real-time estimation of cerebral pathophysiological
parameters have evolved and include the continuous
measurement of ICP, arterial blood pressure (BP),
cerebral perfusion pressure (CPP = BP ICP), and the
indirect estimation of global cerebral oxygenation by
measuring jugular venous oxygen saturation (SjO2)
using jugular venous oximetry. Episodes of low CPP
and SjO2 are now recognized as being detrimental
(Chan et al., 1992a; Czosnyka et al., 1994a; Gopinath et
al., 1994) although the exact values at which cerebral
ischemia occurs probably varies with time and
between individuals (Chan, Miller and Dearden,
1992). The continuous measurement of CBF for clinical
application remains a goal shared by many.
Transcranial Doppler (TCD) provides a means of
measuring relative changes in CBF by observing blood
flow velocity (FV) in basal cerebral arteries (Aaslid,
Markwalder and Nornes, 1982). The method does
require a certain degree of technical expertise, but is
non-invasive, relatively inexpensive and provides realtime information with high temporal resolution. TCD
can be used for measuring flow velocities from several
vessels of the circle of Willis (Aaslid, 1986; Figure 13.1),
but most published data refer to the middle cerebral
artery (MCA). This vessel has a favorable orientation
(see below), is readily accessible to TCD insonation and
provides the most reliable flow velocity signal with a
244
Figure 13.1 Diagram showing the different anatomical locations of the basal cerebral arteries from which Doppler signals
are obtained. Flow detected from the middle cerebral artery (middle trace) is towards the probe and is shown as a positive
(above zero baseline) waveform. This is in contrast to the normal flow direction in the anterior cerebral artery, which is away
from the probe (negative flow upper trace). The posterior cerebral artery also produces a positive waveform.
high signal-to-noise ratio. Further, the MCA delivers

approximately 7080% of the ipsilateral carotid artery
blood flow and can therefore be considered to reflect
blood flow to the majority of the ipsilateral cerebral
hemisphere. Bilateral real-time estimations are now
available using a dual-channel facility. Most commercial machines provide analog signals with provision for
digital logging, making on-line analysis of the FV
signal possible and the collection of data on bedside
computers convenient.
Before discussing the application of TCD in the
head-injured patient, attention to the theoretical limitations of the technique and their relevance in
practical terms is necessary.
For a detailed description of TCD theory, the reader
is referred to the excellent descriptions provided by
respected workers in the field (Newell and Aaslid,
1992a, b).
13.2
The theory of TCD sonography
13.2.1
PHYSICAL PRINCIPLES
The shift in frequency of a wave when either the

transmitter or the receiver are moving with respect to
the wave propagating medium was described by
Doppler in 1843 and is accordingly known as the
Doppler effect. The difference in frequency is known
as the Doppler shift.
In a pulsed ultrasound Doppler instrument, the

same transducer is used for both transmitting and
receiving wave energy. The moving blood acts as a
reflector, first receiving the transmitted ultrasound
wave from the transducer and then reflecting the
ultrasound wave back toward the transducer.
The simplified formula for the Doppler shift (f)
from the moving blood with a velocity v is:
f = 2f0 v/c
where f0 and c are the frequency and velocity of the
emitted ultrasound wave respectively.
Insonation of the basal vessels of the circle of Willis
was made possible by the development of a high
frequency pulsed Doppler technique designed to
penetrate skull bone. (Aaslid, Markwalder and Nornes,
1982). The frequency necessary for transcranial Doppler (TCD) applications is in the order of 2 MHz.
It is seldom that blood within a vessel is moving
directly toward or away from the transducer. More
generally, it will be moving in a direction at an angle
of insonation to the ultrasound beam. The blood flow
velocity measured using TCD is thus dependent on
the angle of insonation, and can vary according to
technique. For the middle cerebral artery (MCA) the
error is small and of the order 3%, since the direction
of proximal segment of the MCA is such that, if
extrapolated, it would meet the pterional bone at a
near right angle (Figure 13.1).
TRANSCRANIAL DOPPLER MEASUREMENTS
The TCD recorded velocity varies with the real

blood velocity according to the cosine of the angle of
insonation. For angles of insonation between 0 and
30, the observed velocity varies between 0.87 and 1.0
of the true velocity. The anatomical limitations for
transtemporal insonation of the MCA are such that
signal capture is only possible at narrow angles. Thus
the observed velocity is a close approximation of true
velocity at a typical depth of insonation of 56 cm
from the scalp surface (1 cm less for children). Other
anterior circulation basal vessels have a less favorable
orientation, hence variation in FV derived from them
is much more dependent on the technique of insonation (Figure 13.1).
13.2.2 RELATIONSHIP BETWEEN INTRACRANIAL
MCA BLOOD FLOW VELOCITY AND CBF
TCD measures the velocity of red blood cell moving

within a vessel. It has a single dimension per unit of
time (usually quoted in cm/s). In contrast, CBF is a
three-dimensional measure of volume of blood
delivered per unit of cerebral tissue per unit of time
(usually quoted in ml/100 g/min). The two parameters are quantitatively different; hence the correlation between absolute values of FV and CBF is poor.
However, provided the cross-sectional area of the
insonated blood vessel remains constant, the CBF and
FV should vary directly with one another. The MCA
diameter appears to be relatively constant under
changing conditions of BP and carbon dioxide tension
in the normal brain, as demonstrated in healthy
volunteers (Aaslid et al., 1989; Dahl et al., 1989; Newell
et al., 1994) thus good correlations between relative
changes in FV and CBF have been reported experimentally (Barzo et al., 1991; Czosnyka et al., 1994c;
Richards et al., 1995) and in clinically stable patients
(Bishop et al., 1986; Romner et al., 1991). However in
unstable patients with acute brain injury, the MCA
diameter may be altered by changes in ICP and the
effects of vasospasm, reducing the reliability of the
relationship between relative CBF and FV (Kontos,
1989).
Some workers have indicated that the relationship
between FV and CBF varies according to the level of
CBF, with a stronger correlation occurring when CBF
was low (less than 20 ml/100 g/min; Halsey, McDowell
and Gelmon, 1986). These observations may reflect
different states of autoregulation and a changing MCA
diameter, although pressure autoregulation is believed
to primarily involve the microcirculation rather than
medium sized vessels. Alternatively, the flow characteristic of the formed elements may change at different
flow rates, and become non-laminar at extreme values
for CBF thus varying the relationship between flow and
velocity. The variable relationship between absolute
245
FV and absolute CBF is a clear disadvantage, reflected

by the normal variation in MCA FV of between 35 and
90 cm/s in the awake resting state. Thus TCD demands
caution in interpretation when attempting to estimate
CBF from FV data in isolation of other parameters in
head-injured patients.
13.3
TCD measurements
Three main pathways to access the intracranial arteries can be employed. They comprise the transtemporal
approach through the thin bone above the zygomatic
arch to the anterior, middle and posterior cerebral
arteries and circle of Willis, the transorbital approach
to the carotid siphon, and the suboccipital route to the
basilar and vertebral arteries. The transtemporal
approach is usually employed in critically ill patients.
In expert hands, about 5% of the individuals will yield
an unsatisfactory image employing the temporal
window (Aaslid, 1986; Harders and Gilsbach, 1985;
Harders, 1986).
Through the temporal window, the middle (MCA),
anterior (ACA) and posterior (PCA) cerebral arteries
can be readily examined (Figure 13.1). In each patient,
the same insonation window should be used throughout the entire study period. This could be accomplished
by putting a small marker at the patients temporal
region. The Doppler examination begins with the
identification of the bifurcation of the intracranial
portion of the internal carotid artery into the MCA and
ACA according to the method described by Aaslid
(1986). This bifurcation can usually be identified at a
depth of 6065 mm. The typical Doppler signal from
the carotid bifurcation is shown in Figure 13.2, which
consists of images above and below the xero line of
reference representing the flow directions towards and
away from the ultrasound probe of the MCA and ACA
respectively. The depth of insonation is then reduced to
follow the upward deflection image of the MCA flow
velocity as the vessel runs towards the skull. The MCA
can usually be traced up to a depth of 30 mm, which is
beyond the bifurcation of the MCA into the peripheral
branches. The proximal portion of the main trunk of the
MCA (the M1 segment) can be located at a depth of
around 4555 mm. The depth which gives the highest
velocity is usually chosen for measurement. In children, the depth which gives the highest MCA velocity
is usually 10 mm less than that of adults, but the same
principles apply. This method of obtaining the MCA
signal eliminates the possibility of mistaking the PCA
for the MCA because the PCA signal cannot be
obtained at a depth less than 55 mm for anatomical
reasons (Aaslid, 1982; Lindegaard, 1989). Mistaking the
PCA for the MCA would introduce a major error into
velocity measurement due to the lower velocity of the
former vessel.
246
Figure 13.2
Doppler tracing showing the carotid bifurcation.
After the MCA signal is obtained, the depth of

insonation is increased so that the image of the carotid
bifurcation can be seen again; the depth is increased
further with the probe directed slightly anteriorly so
that the ACA image can be found. The first part of the
ACA (the A1 segment) is recognized by a direction of
flow that is away from the probe. With the identification of the ACA, the depth of insonation is decreased
until the carotid bifurcation signal is obtained. The
probe is then angled slightly posteriorly until the
signal of the PCA is seen. The PCA can be distinguished from the MCA signal by a lower flow
velocity and failure to obtain the PCA signal when the
depth of insonation is decreased much below 55 mm.
The velocity parameters measured included the peak
systolic velocity (S), end-diastolic velocity (D), the
time-averaged mean velocity (M) and the pulsatility
index (PI). The latter is defined as S-D/M. The timed
mean velocity and PI are calculated by the Doppler
machine by averaging the signals from a number of
cardiac cycles by fast Fourier transformation, which
eliminates error as a result of beat-to-beat and respiratory variations. Beat-to-beat variation in adults suggests that about five heart beats are adequate for
measurement (Burns, 1987).
13.3.1 LONG-TERM TCD MEASUREMENT ON THE
INTENSIVE CARE UNIT PRACTICAL
CONSIDERATIONS
Fixation of the probe for continuous monitoring is

most conveniently achieved in the pterional region.
Both solid and elasticized devices are available for
probe fixation. None provide a universally accepted

means of keeping the probe fixed in an identical
orientation in the hostile environment of the intensive
care unit. With an elasticized band we have achieved
an approximately 60% signal reliability for MCA
insonation in patients monitored over several days
(Kirkpatrick et al., 1995). Nursing education is an
essential part of this process, and it is vital to inspect
the band and probe regularly to check for pressure
sore development. The latter complication usually
results from placement across injured scalp and bony
prominences. Elasticized bands should only be
applied across such areas with extreme care, since the
risk of scalp swelling and necrosis is increased. The
solid frame variety provides the advantage that
probes can still be applied to those patients with
significant scalp trauma and/or postoperative scalp
flaps. However, they are bulky and cumbersome, and
probe displacement is more likely during nursing
maneuvers. It is clear that there is room for considerable improvement in the design of attachment devices
for the purpose of long-term TCD monitoring.
The advantage of continuous monitoring is the
establishment of stable baseline conditions of variables that may affect FV over intermediate periods of
time. These variables include the insonation angle,
arterial oxygen and CO2 tension, blood pressure,
different anesthetic agents and hemoglobin levels
(Newell and Aaslid, 1992b). These concerns may be
overstated, since the basal cerebral arteries, do not
appear to dilate or constrict significantly with vascular
resistance and/or anesthetic changes (Matta and Lam,
1995; Schregel et al., 1992, 1994). Blood pressure and
SIGNAL PROCESSING AND DATA COLLECTION
247
carbon dioxide tension, important determinants of

cerebrovascular resistance, have little effect on the
diameter of the basal arteries (Huber and Handa,
1967). Some potent vasoactive agents (e.g. sodium
nitroprusside) have only a small effects on the
diameter of basal cerebral vessels (Giller et al., 1993),
whereas others (e.g. nitroglycerin) demonstrate significant vasodilatation (Dahl et al., 1989). While most
report the diameter of the proximal MCA to be
resistant to cerebrovascular variables, it would seem
prudent to maintain stability during acquisition of
TCD-related data. Under such stable conditions, FV
changes are likely to reflect changes in CBF rather
than changes in MCA diameter or technical
variation.
13.4 Signal processing and data

collection
13.4.1
COMPUTERS
A computer-based system is essential for logging,

collating and processing the large volume of data
acquired by TCD. Recognizing the changing cerebrovascular hemodynamics demands reliable monitoring techniques and sophisticated signal analysis
which can only be provided by dedicated computer
support.(Czosnyka et al., 1994f; Kirkpatrick, Czosnyka
and Pickard, 1996). Contemporary systems include
sophisticated multichannel digital recorders with
options for complex signal processing. The considerable flexibility of such systems permits signal analysis
and presentation of information that is comprehensible to medical and nursing staff. The importance of
monitoring multiple variables to increase the power of
data interpretation is increasingly recognized.
13.4.2
DERIVATIVES OF FV
Signal analysis can produce large volumes of information, leading to a state of data chaos. The recording of
user-defined targeted variables is therefore most
desirable. Various TCD signals recorded from flow of
formed elements within the MCA generate a spectrum
of flow velocities that are presented as a waveform
(Czosnyka et al., 1994d). The mean FV of the spectrum
theoretically varies with CBF and is therefore usually
presented. Since MCA flow is laminar, the maximum
FV (FVmax) varies in proportion with the mean FV. Thus
commercial machines take advantage of the superior
signal-to-noise ratio with FVmax and calculate FVmean
from the area under the curve. The FVmax signal is
frequently displayed as a FV envelope which can be
resolving into FVmax during diastole (FVd) and FVmax
during systole (FVs). It is these two components that
define the pulsatility of the waveform (Figure 13.3).
Figure 13.3 Relationship between the components of FV

(FVs and FVd ) with changing CPP during a plateau wave of
raised ICP and low CPP. As CPP falls, FVd is proportionally
affected to a greater extent than FVs resulting in the
divergence of these parameters, hence an increase in the pulse
amplitude (FVs minus FVd ) of the waveform is seen. ICP =
intracranial pressure (mmHg); CPP = cerebral perfusion
pressure (mmHg); FV = middle cerebral artery flow velocity
(cm/s); FVs = FV during systole; FVs = FV during diastole.
(Source: redrawn from Czosnyka et al., 1994b.)
13.4.3
PULSATILITY INDICES
The MCA FV waveform observed using TCD is

dependent on the arterial blood pressure waveform
and the viscoelastic properties of the cerebrovascular
bed provided the blood rheology remains constant.
Thus, if variables such as MCA diameter and BP
remain constant, the pulsatility of blood flow through
the conductance vessel reflects distal cerebrovascular
resistance (Czosnyka et al., 1994b; Gosling and King,
1974; Lindgaard, 1992). Common determinants of
changing MCA pulsatility include CO2 tension and
CCP. Several indices describing the pulsatility of blood
have been formulated, the most commonly adopted is
the pulsatility index (PI) of Gosling:
Gosling PI = (FVs FVd )/FVmean = FVamp /FVmean.
The key advantage of the Gosling PI is that it is
dimensionless and therefore independent of sampling
techniques, provided the signal-to-noise ratio is good
and the gain setting of the instrument is constant.
Most TCD software packages calculate the PI index as
averaged from several cardiac cycles. However, taken
alone the PI cannot distinguish the cause of the flow
pattern change (that is, an increase in PI can be due to
cerebral vasoconstriction or to high ICP), and it needs
to be interpreted in the light of other data.
248
(a)
(b)
Figure 13.4 (a) Continuous recordings of ICP, CPP and LDF signals during waves of raised ICP in a patient with a diffuse
head injury. Variations in relative CBF, estimated from FV, were closely coupled to changes in CPP indicating a state of nonautoregulation. (b) Regression analysis of the FV and LDF signal data from the period of recording shown in (a). The close
correlation indicates coupling between medium vessel flow and capillary perfusion. ICP = intracranial pressure (mmHg);
CPP = cerebral perfusion pressure (mmHg), FV = right middle cerebral artery flow velocity (cm/s); LDF = laser Doppler flux
from the right frontal region; AU = arbitrary units. (Source: reproduced from Kirkpatrick et al., 1994a, with permission.)
ANALYSIS USING TCD IN HEAD-INJURED PATIENTS
Thus TCD provides a basis for cerebrovascular

investigations in head-injured patients. These include
the non-quantified monitoring of changing CBF, cerebrovascular autoregulatory reserve, cerebrovascular
reactivity, cerebral perfusion pressure, cerebral hyperemia, post-traumatic spasm and the estimation of
cerebral tamponade.
13.5 Results of analysis using TCD in

head-injured patients
13.5.1
MEASUREMENT OF RAW BASELINE FV DATA
Collection of intermittent mean FV data from headinjured patients is of limited use (Chan, Miller and
Dearden, 1992; Weber, Grolimund and Seiler, 1990).
Chan, Miller and Dearden (1992) report that in
severely head-injured patients (GCS < 8), the mean FV
at admission was lower than in those with less severe
head injuries. Further, the mean FV remained
depressed in those same patients. However, the
dispersal of data points was so wide that raw
admission FV data was not a useful predictor of
outcome for individuals except where very low FV
(a)
249
was encountered (less than 28 cm/s = 80% death rate).

These findings are not surprising, bearing in mind the
variables that affect FV measurements in these unstable patients.
Continuous monitoring of mean FV is potentially of
value when used purely as a non-quantified trend
recorder, where dynamic changes from a variable
baseline are considered (Kirkpatrick et al., 1994a;
Kirkpatrick et al., 1995; Newell et al., 1992). Our
experience in Cambridge of continuous multimodality
monitoring systems, indicates that the TDC is able to
detect transient changes in relative CBF with high
resolution (Kirkpatrick et al., 1994a; Kirkpatrick,
Czosnyka and Pickard, 1996). Thus, episodes lasting
between 12 and 70 minutes were seen which corresponded to cortical cerebral perfusion changes as
assessed using laser Doppler flowmetry (Figure
13.4(a)). The correlation between relative changes was
high (Figure 13.4(b)) and could help to distinguish
between rises in ICP associated with low CPP and
those associated with hyperemia (Figure 13.5). Instances where there was significant uncoupling between
FV and cortical perfusion were rare (Figure 13.6). It is
the high dynamic resolution provided by TCD and the
(b)
Figure 13.5 (a) Recording of an event characterized by intracranial hypertension with an increase in ICP. A relative drop
in CBF is predicted from the fall in FV which is associated with cerebral desaturation (SjO2 and HbO2 signal changes).
(b) Recording suggesting cerebral hyperemia. In this event synchronous increases in all recorded parameters occurred
during the rise in ICP. ICP = intracranial pressure (mmHg), CPP = cerebral perfusion pressure (mmHg), Fv = right middle
cerebral artery flow velocity (cm/s) HbO2 = oxygenated hemoglobin (mol/l); tHb = total hemoglobin concentration
(mol/l), SjO2 = right jugular venous oxygen saturation (%). (Source: reproduced from Kirkpatrick et al., 1995, with
permission.)
250
Figure 13.6 An abrupt rise in ICP and fall in CPP recorded during attempted withdrawal of dopamine inotropic support.
Simultaneous falls in FV and LDF occurred. On recovery of CPP a transient hyperemia was registered by LDF, but not by
FV, indicating uncoupling of flow in different sized vessels. ICP = intracranial pressure (mmHg); CPP = cerebral perfusion
pressure (mmHg), FV = right middle cerebral artery flow velocity (cm/s); LDF = laser Doppler flux from the right frontal
region; AU = arbitrary units. (Source: reproduced from Kirkpatrick et al., 1994a, with permission.)
correlation with other hemodynamic modalities that

are proving encouraging to neurointensivists employing the technique as a real-time monitor.
13.5.2
AUTOREGULATION IN HEAD INJURY
Cerebral autoregulation describes the cerebrovascular

reflexes that maintain CBF during changing CPP
(Paulson, Olesen and Edvinsson, 1990). Failure of
autoregulation, which is frequent after severe head
injury, may aggravate cerebral ischemia during periods of relative poor perfusion.(Miller, 1985; Siesjo,
1992). Thus the autoregulatory status of the cerebral

vasculature may provide an index of injury severity,
help define the cerebrovascular reserve and indicate
the susceptibility of individuals to secondary ischemia
during low cerebral perfusion (Marmarou et al., 1991).
Assessment of autoregulation using TCD depends on
the assumption that relative changes in FV correlate
with relative changes in CBF. Thus, if autoregulation
is intact, FV should remain constant with a changing
CPP. Conversely, if autoregulation has failed, then the
relation between CPP and FV becomes linear (Figures
13.3, 13.7) The value of CPP at which autoregulation
Figure 13.7 Three characteristic patterns of linear regression between spontaneous waves of CPP and the components of the FV waveform captured
over a 15 minute epoch. FVm, FVs and FVd are resolved. A negative correlation between FVs, FVm, FVd and CPP (left graph) indicates intact
autoregulation for all components and is predictive of a favorable outcome. Disturbance of autoregulation initially affects the FVd component (middle
graph) resulting in a positive relationship between FVd, FVm and CPP, while FVs remains independent of CPP. In a fully depleted autoregulatory state
(right graph) all components of FV show a positive correlation with CPP, resulting in a pressure passive state. This is predictive of an unfavorable
outcome. ICP = intracranial pressure (mmHg); CPP = cerebral perfusion pressure (mmHg), FV = right middle cerebral artery flow velocity (cm/s); FVm
= mean flow velocity; FVs = FV during systole; FVd = FV during diastole. Source: redrawn from Czosnyka et al., 1995.
252
fails and FV begins to fall is called the autoregulatory

threshold or break point. Since the autoregulatory
threshold is dependent on PCO2 levels in states of
hypercapnia the threshold is exceeded at lower levels
of CPP (Markwalder et al., 1984; Paulson, Olesen and
Edvinsson, 1990) this variable has to be kept
constant for repeated evaluations.
The autoregulatory status can be assessed using
TCD in a variety of ways. Formal lowering of blood
pressure is no longer considered ethical in the headinjured. Observing the responses to spontaneous
changes in BP is an alternative (Czosnyka et al., 1994b),
but such transients are infrequent, often of low
magnitude, and assessments cannot be made at
designated times. This is especially so in the present
era of promoting a high CPP with fluid replacements
and inotropes (Rosner, Rosner and Johnson, 1991).
However, transient falls in CPP can be safely induced
by a three second carotid compression, or for longer
periods (2030s) by inflating and releasing large blood
pressure cuffs applied to the legs (Czosnyka, Pickard
and Whitehouse, 1992; Giller, 1991; Smielewski et al.,
1996). These indirect methods have been employed for
bedside assessments, but the results are techniquedependent. Although such methods have confirmed
that autoregulation is more frequently absent in the
most severe head-injured patient (Chan, Miller and
Dearden, 1992; Czosnyka et al., 1994e; Steiger et al.,
1994; Wong, Piper and Miller, 1994), further evaluation
is required.
The continuous assessment of the TCD wave profile
is potentially a sophisticated way of estimating
cerebral autoregulation. Thus with falling CPP, FVd
will reach the autoregulatory threshold before FVs,
resulting in divergence of these parameters and an
increase in pulse amplitude and PI (Figure 13.3). Thus a
decline in FVd, and an increase in PI gives an earlier
warning of impending autoregulatory failure than does
FVmean. If FVs also falls with a drop in CPP, then all
components of the FV waveform have reached the
autoregulatory threshold, indicating a severely
depleted cerebrovascular reserve. An evaluation of the
gradient of linear regression between the two main
components of the FV waveform (FVs and FVd) and the
CPP show that patients with failure of autoregulation
in both components are most likely to have an
unfavorable outcome when compared to those patients
who have intact autoregulation or failed autoregulation in FVd alone (Figure 13.7; Czosnyka et al., 1995).
13.5.3
NON-INVASIVE ASSESSMENT OF CPP
From the above considerations, it can be seen that the

different FV parameters may be used to indicate
failing autoregulation. If this occurs at a consistent
value of CPP between individuals, then changes in FV
may provide a non-invasive estimation of CPP. The

most sensitive index of falling CPP is the increase in
FV amplitude due to divergence of FVs and FVd
(Figures 13.3, 13.7). By calculating the PI from FV
amplitude and FVmean, an index of falling CPP is
provided which is independent of the angle of insonation (Figure 13.8). Clinical experience has demonstrated a close correlation between CPP and PI. Chan
et al. (1992a) demonstrated a better correlation
between CPP and PI than between FVmean and CPP,
and this relationship was independent of the mechanism of reduced CPP (hypotension or raised ICP).
Two breakpoints were identified on the CPP/PI curve
using sequential linear regressions: PI was independent of CPP when the latter was above 70 mmHg;
between 70 and 20 mmHg the PI increased (r = 0.767,
p < 0.0001, n = 41); and below 20 mmHg PI rose
rapidly. Thus at a CPP of between 20 and 70 mmHg
autoregulation is impaired but not completely
exhausted. Other groups have confirmed the existence
of a relationship between CPP and PI (Aaslid et al.,
1986; Czosnyka et al., 1994e; Newell et al., 1993),
although different breakpoints are quoted that probably reflect methodological variation (anesthetic
agents and differences in the preferred level of CO2
maintenance). Since the PI is dependent on BP, a
standardized pulsatility index (SPI) can be calculated
(SPI = PI/BP), which promises to improve the convergence of data describing the relationship between
Figure 13.8 Relationship between the pulsatility index of

the FV waveform (PI) and CPP. At high CPP values PI
remains constant. At low CPP values PI increases rapidly.
The exact threshold for CPP at which PI starts to increase
differs among patients, and can vary in the same patient at
different times. However, pooled data from different sources
indicates that the threshold lies between 55 and 75 mmHg.
(Source: redrawn from Czosnyka et al., 1994e.)
ANALYSIS USING TCD IN HEAD-INJURED PATIENTS
CPP and pulsatility (Kirkpatrick et al., 1993). This

work raises the concept of a therapeutic window
existing between the completely intact autoregulatory
state and the completely exhausted autoregulatory
state, the equivalent of the transient zone described
in experimental animals using TCD and laser Doppler
methods (Czosnyka et al., 1994c; Nelson et al., 1992;
Richards et al., 1995).
Mathematical models can be generated that attempt
to predict the CPP and cerebral autoregulatory status
of the cerebral circulation from TCD FV waveform
analysis. Aaslid and colleagues (1986) and Czosnyka et
al. (1994b) have suggested different formulae relating
CPP with flow velocity amplitude, flow velocity mean
and BP amplitude. The difficulty encountered by these
groups is that the convergence of data points from
pooled patient groups is relatively low, making
prediction of CPP for individuals imprecise. Further,
the various thresholds described for CPP, FV and PI
may vary for any patient with time (Chan, Miller and
Dearden, 1992).
13.5.4 RELATIONSHIP BETWEEN TCD CHANGES
AND OTHER CEREBRO-HEMODYNAMIC VARIABLES
Cerebral tissues can tolerate limited changes in CBF

before metabolic requirements fail and neuronal function becomes compromised. (Wong, Piper and Miller,
1994). Thus a fall in CPP below certain thresholds
results in loss of cerebral electrical activity and, on
further reduction, loss of membrane stability resulting
in neuronal death. Failure of pressure autoregulation
does not necessarily indicate impending cell death,
since the CPP threshold for pressure autoregulatory
failure and metabolic failure may be different. By
253
employing continuous multimodality monitoring

techniques, the response of both CBF and cerebral
metabolic variables to changing CPP can be observed
in individual cases (Chan et al., 1992a; Kirkpatrick et
al., 1995). Transient episodes of cerebral hypoperfusion due to hypotension or ICP plateau wave activity
usually result in relative falls in CBF, and cerebral
oxygenation (Figure 13.5(a)). A delay in the fall in
tissue oxygenation (read with near-infrared spectroscopy and jugular venous oximetry) of approximately
2 minutes is seen after the fall in CPP, suggesting that
parenchymal desaturation is secondary to the change
in tissue perfusion.
By pooling intermittent data from patients in whom
CPP and SjO2 measurements were recorded, a threshold of 71 mmHg was found below which SjO2 fell (r =
0.78, p < 0.0001, n = 22; Chan et al., 1992a). This is
almost identical to the CPP threshold defined by PI
(see above; Figure 13.8), suggesting that pressure
autoregulatory failure defined by TCD does imply the
onset of cerebral desaturation and the need for
increased oxygen extraction (Figure 13.9). The precise
threshold at which neuronal injury ensues has yet to
be defined and awaits the identification and monitoring of products of neuronal damage in the venous
effluent. Despite this, some reputable centers now
advocate the maintenance of CPP above 70 mmHg,
with very impressive results (Miller, 1992; Rosner,
Rosner and Johnson, 1991).
13.5.5 CEREBROVASCULAR REACTIVITY IN HEAD
INJURY
Cerebrovascular reactivity describes the near linear

relationship between cerebral artery CO2 levels and
Figure 13.9 Composite plot of cerebral perfusion pressure (CPP) versus jugular venous oxygen saturation (SjO2 ) and
doppler pulsatility index (PI) showing the CPP breakpoint of 70 mmHg.
254
CBF, and can be tested by observing the change in FV

in response to changes in CO2 (Dahl, 1992; Klingelhofer and Sander, 1992; Markwalder et al., 1984; Marmarou et al., 1991; Romner et al., 1991; Schalen,
Messeter and Nordstrom, 1991; Smielewski et al., 1994,
1995; Strebel et al., 1994). In normal individuals, a
1kPa increase in CO2 causes (on average) a 22%
increase in MCA FV, and thus TCD has been used to
assess CO2 reactivity in many clinical situations.
Reduced or absent CO2 reactivity indicates that the
ability of the cerebrovascular bed to vasodilate is
depleted hence indicating loss of cerebrovascular
reserve. Thus CO2 reactivity may provide predictive
information in patients with severe head injuries.
Klingelhofer and Sander reported on patients with
raised ICP due to cerebral hemorrhage, indicating that
the long term outcome was worse in those with
impaired CO2 reactivity (Klingelhofer and Sander,
1992). The same appears true of head-injured patients
(Grosset et al., 1993; Schalen, Messeter and Nordstrom,
1991). The use of acetazolamide, a potent cerebral
vasodilator, cannot assess vasoconstriction and has
not been employed in head-injured patients.
13.5.6
ABNORMALLY HIGH FV IN HEAD INJURY
Elevated levels of FV can either indicate a narrowed

MCA (vasospasm or stenosis) or high CBF (hyperemia). Both vasospasm and hyperemia are well
recognized following head injury (Chan, Dearden and
Miller, 1992; Chan et al., 1992b; Compton and Teddy,
1987; Muttaqin et al., 1993) and since they demand a
different therapeutic response their distinction is
potentially important (Pickard and Czosnyka, 1993).
If vasospasm is suspected, one can insonate the
extracranial internal carotid FV and calculate the
Lindegaard ratio (FVmca/FVica; Lindegaard et al.,
1988). Vasospasm is likely when this ratio exceeds 3
(Aaslid et al., 1986). Cerebral vasospasm after severe
head injury has long been recognized (Macpherson
and Graham, 1973) and is frequently associated with
traumatic subarachnoid blood. The flow velocities
recorded in such cases are usually between 100 and
150 cm/s, hence lower than those found after aneurysmal subarachnoid hemorrhage, and the time course
is also shorter, occurring within the first 25 days.
Nevertheless, CBF can be significantly impaired with
ischemic consequences (Chan et al., 1992b). Early
evidence from randomized trials suggests that such
patients may benefit from cerebral calcium antagonists (European Study Group, 1994).
Protracted cerebral hyperemia is also recognized
following head injury, especially in children (Bruce et
al., 1981; Muizelaar et al., 1989). Transient elevations
in FV due to hyperemia may also occur, and can be
captured with multimodality monitoring using near-
(a)
(b)
Figure 13.10 Doppler tracing showing waveforms of
(a) non-hyperemic and (b) hyperemic increase in Doppler
flow velocity.
infrared spectroscopy. Such episodes are accompanied

by highly correlated elevations in ICP, FV, SjO2 and
total blood volume (Figure 13.5(b)). Chan et al. (1993)
described a waveform notch in diastole in nonhyperemic patients with increased Fv (Figure 13.10).
The importance of recognizing hyperemia as a cause
of pathologically raised ICP is that conventional
methods for treating elevated ICP (such as mannitol,
and inotropes/colloid to elevate CPP) will increase
CBF and worsen the intracranial hypertension.
13.6
Role of TCD in monitoring therapy
TCD allows the opportunity to observe the response

of the MCA FV to different therapies (Chan et al., 1993;
Eng et al., 1992; Schregel et al., 1992; Thiel et al., 1992).
When therapy was instigated to improve CPP (n = 22),
a linear relationship with PI (r = 0.941, p < 0.0001) and
SjO2 (r = 0.837, p < 0.0001) was seen with a CPP
threshold at 70 and 68 mmHg respectively, above
which the correlations disappeared (Chan et al., 1993;
Figure 13.11). These thresholds are virtually identical
to those seen during spontaneous changes in CPP (see
above). In the same group of patients, treatment with
hypnotics frequently caused falls in CPP, SjO2 and PI
by virtue of arterial hypotension.
The real-time facility offered with a secured TCD
probe can also be used to monitor the dynamic
response to therapy (Eng et al., 1992; Kirkpatrick et al.,
1994b; Kirkpatrick, Czosnyka and Pickard, 1996; Mayberg et al., 1993; Schregel et al., 1992; Thiel et al., 1992).
The response to an infusion (n = 23) of a 200 ml bolus
of mannitol has been compared with that of a
preceding bolus of 200 ml saline, both given over 20
minutes in 14 severely head-injured patients with
raised ICP (Figure 13.12(a)). After mannitol only, a fall
SUMMARY
(a)
255
(b)
Figure 13.11 (a) Plot of cerebral perfusion pressure (CPP) versus jugular venous oxygen saturation (SjO2 ), showing CPP
breakpoint after intracranial pressure (ICP) reduction therapy. (b) Plot of CPP versus Doppler pulsatility index (PI), showing
CPP breakpoint after ICP reduction therapy.
in ICP (21%) was accompanied by an increase in FV

(+13%, p < 0.001) and fall in cerebrovascular resistance
(Figure 13.12(b)). The effect of mannitol on FV
decayed exponentially with a time constant of 34.0
minutes (Figure 13.12(c)) and was independent of the
pressure autoregulatory status of the patient (Kirkpatrick et al., 1994b; Kirkpatrick, Czosnyka and Pickard,
1996).
Although this data is of greater scientific interest
than clinical relevance, the short-term manipulation of
cerebrovascular hemodynamics is gaining in importance. Bearing in mind the short duration of many
secondary events, continuous TCD is of use in their
rapid detection and in monitoring the subsequent
response to chosen therapy. This is especially important if potentially dangerous maneuvers, such as
hyperventilation in states of cerebral oligemia, are to
be avoided (Gold, 1989).
13.7
TCD in the diagnosis of brain death
From Figure 13.3 it can be seen that, as CPP falls, FVd

approaches zero and forward movement of blood only
occurs during systole. When CPP falls further, a state
of reverberant flow can be seen, where reverse flow
occurs during diastole. It is at this point that a very
high PI is seen. Once backflow during diastole equals
forward flow during systole (oscillating flow pattern), no net flow of blood occurs and cerebral
tamponade in the territory of the insonated vessel is
complete (Feri et al., 1994; Hassler, Steinmetz and
Gawlowski, 1988; Hassler, Steinmetz and Pirschel,
1989; Newell, Grady and Sirotta, 1989; Perry et al.,
1990; Werner et al., 1990). Multimodality monitoring
using laser Doppler flowmetry and TCD during
cerebral tamponade has shown that positive flow

through the basal arteries can be seen after the patient
has coned with no net flow through the capillary bed
(Figure 13.13; Kirkpatrick et al., 1994a). Such real-time
observations concur with the reports using angiographic techniques, showing that the basal cerebral
arteries remain patent during early stages of circulatory arrest (Hassler, Steinmetz and Gawlowski,
1986). One explanation for this phenomenon would be
the presence of internal cerebrovascular shunts that
bypass the parenchymal small vessel networks, reinforcing the concept that initial blood flow obstruction
occurs at the level of the capillary.
Although high specificity and sensitivity have been
reported using various TCD criteria for brain death
(Perry et al., 1990), false positives have been reported,
more commonly in children. Consequently, TCD
measurements have not been included in the formal
testing for brain death. At present TCD can be
considered as a useful diagnostic aid in patients who
are progressing towards brain tamponade.
13.8
Summary
Although measurement of basal vessel FV using TCD

has restrictions in terms of quantification and spatial
resolution, monitoring of the MCA FV in head-injured
patients has produced data that have improved our
understanding of cerebral hemodynamic changes following cerebral injury. The most important contribution has been derived from observing the FV changes
with falling CPP and the evolution of the concept that
CPP should be maintained above a certain threshold
level. That threshold is different between patients, but
an average value of between 60 and 70 mmHg would
256
(a)
(b)
(c)
Figure 13.12 (a) A graphical display of multiple parameters measured during a 200 ml infusion of mannitol (20%), which was
preceded by an identical infusion of normal saline (arrowheads). Mannitol caused a rapid increase in FV and LDF signals,
which was associated with a fall in ICP. These changes were not seen after saline. BP remained unchanged throughout the
recording period. (b) Mean changes in FV ( s.e. of mean) recorded over approximately 60 minutes during and after a single
mannitol infusion from patients with diffuse head injuries (number of infusions = 23). The first point for each recording
represents the start of infusion, which was complete at t = 0 min. In all patients a rise in FV was seen, which decayed soon after
completion of infusion. Mean change in FV for all mannitol infusions shown. (t = 0: time of completion of infusion of
mannitol; t = 15: 15 minutes after completion of mannitol.) (c) The decay of the effect of mannitol on FV after completion of
mannitol infusion (time = 0 minutes). The FV has been normalized to a relative baseline of 1 and maximum effect of 2. The fall
of FV fits an exponential model with a decay time constant of 34 minutes. ICP = intracranial pressure (mmHg); BP = mean
arterial blood pressure (mmHg), FV = right middle cerebral artery flow velocity (cm/s); LDF laser Doppler flux from the right
frontal region; AU = arbitrary units. (Source: reproduced from Kirkpatrick et al., 1996, with permission.)
SUMMARY
257
Figure 13.13 Progressive fall in CPP during cerebral coning in a head-injured patient who did not respond to treatment.
The progressive failure of microcirculatory flow is seen in the decline in the LDF signal. FV remained positive and later
became reverberant. ICP = intracranial pressure (mmHg), CPP = cerebral perfusion pressure (mmHg), FV = right middle
cerebral artery flow velocity (cm/s); LDF laser Doppler flux from the right frontal region; AU = arbitrary units. (Source:
reproduced from Kirkpatrick et al., 1994a, with permission.)
seem appropriate for most individuals. Other contributions include the facility to discriminate between
different causes of raised ICP in real time, allowing an
earlier, more targeted hence more appropriate,
approach to therapy. Thus TCD can be used to
distinguish between raised ICP due to cerebral hypoperfusion, and that due to hyperemia.
Future development of TCD in head injury appears
to lie in its incorporation into multimodality monitoring systems, allowing identification of thresholds for
258
neuronal injury. Dynamic monitoring is also likely to

be of considerable use in the evaluation of compounds
that aim to assist the injured brain by mechanisms that
improve CPP or affect the cerebral vasculature.
13.9
References
Aaslid, R. (1982) Cerebral hemodynamics, in Transcranial Doppler, (eds D. W.

Newell and R. Aaslid), Raven Press, New York, p. 49.
Aaslid, R. (1986) Transcranial Doppler examination techniques, in Transcranial
Doppler Sonography, (ed. R. Aaslid), Springer-Verlag, Vienna, pp. 3960.
Aaslid, R., Markwalder, T. M. and Nornes, H. (1982) Non-invasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral
arteries. Journal of Neurosurgery, 57, 769774.
Aaslid, R., Lundar, T., Lindegaard, K. F. and Nornes, H. (1986) Estimation of
cerebral perfusion pressure and transcranial Doppler recordings, in
Intracranial Pressure VI, (eds J. D. Miller, G. M. Teasdale and J. O. Rowan),
Springer-Verlag, Berlin.
Aaslid, R., Lindegaard, K. F., Sorteberg, W. and Helge, N. (1989) Cerebral
autoregulation dynamics in human. Stroke, 20, 4552.
Barzo, P., Doczi, T., Csete, K. et al. (1991) Measurements of regional cerebral
blood flow and blood flow velocity in experimental intracranial hypertension
infusion via the cisterna magna in rabbits. Neurosurgery, 28, 821825.
Bishop, C. C. R., Powell, S., Rutt, D. and Brouse, N. I. (1986) Transcranial
Doppler measurement of the middle cerebral flow velocity: a validation
study. Stroke, 17, 913915.
Born, J. D., Albert, A., Hans, P. and Bonnal, J. (1985) The relative prognostic
value of best motor response and brain stem reflexes in patients with severe
head injury, Neurosurgery, 16, 595601.
Bouma, G. J. and Muizelaar, J. P. (1990) Relationship between cardiac output
and cerebral blood flow in patients with intact and with impaired
Bouma, G. J., Muizelaar, P. J., Stringer, W. A. et al. (1992) Ultra-early evaluation of
regional cerebral blood flow in severely head injured patients using xenonenhanced computerised tomography. Journal of Neurosurgery, 77, 360368.
edema. Journal of Neurosurgery, 54, 170178.
Burns, P. N. (1987) The physical principles of Doppler and spectral analysis.
Journal of Clinical Ultrasound, 15, 567590.
Chan, K. H., Dearden, M. and Miller, J. D. (1992) The significance of
posttraumatic increase in cerebral blood flow velocity: a transcranial
Doppler ultrasound study. Neurosurgery, 30, 697700.
Chan, K. H., Miller, J. D. and Dearden, N. M. (1992) Intracranial blood flow
velocity after head injury: relationship to severity of injury, time, neurological status and outcome. Journal of Neurology, Neurosurgery and Psychiatry,
55, 797791.
Chan, K. H., Miller, J. D. and Piper, I. R. (1992) Cerebral blood flow at constant
cerebral perfusion pressure but changing arterial and intracranial pressure:
relationship to autoregulation. Journal of Neurosurgery and Anesthesiology, 4,
189193.
Chan, K. H., Miller, D. J., Dearden, M. et al. (1992a) The effect of changes in
and jugular bulb venous oxygen saturation after severe brain trauma.
Chan, K. H., Dearden, M., Miller, J. D. et al. (1992b) Transcranial Doppler
waveform differences in hyperemic and non-hyperemic patients after
severe head injury. Surgical Neurology, 32, 443436.
Chan, K. H., Dearden, N. M., Miller, J. D. et al. (1993) Multimodality
monitoring as a guide to treatment of intracranial hypertension after severe
brain injury. Neurosurgery, 32, 547552.
Choksey, M. S., Costa, D. C., Iannotti, F. et al. (1991) 99TC-HMPAO SPECT
Compton, J. S. and Teddy, P. J. (1987) Cerebral arterial vasospasm following
severe head injury a transcranial Doppler study. British Journal of
Cruz, J. (1993) On-line monitoring of global cerebral hypoxia in acute brain
injury. Relationship to intracranial hypertension. Journal of Neurosurgery, 79,
228233.
Cruz, J., Miner, M. E., Allen, S. J. et al. (1991) Continuous monitoring of
cerebral oxygenation in acute brain injury: assessment of cerebral hemodynamic reserve. Neurosurgery, 29, 743749.
Czosnyka, M., Kirkpatrick, P., Guazzo, E. et al. (1994a) Assessment of the
autoregulatory reserve using continuous CPP and TCD blood flow velocity
measurement in head injury, in Intracranial Pressure IX, (eds H. Nagai, K.
Kamiya and S. Ishii), Springer-Verlag, Berlin, pp. 593594.
Czosnyka, M., Guazzo, E., Iyer, V. et al. (1994b) Testing of cerebral
autoregulation by waveform analysis of blood flow velocity and cerebral
perfusion pressure. Acta Neurochirurgica (Supplement), 60, 468471.
Czosnyka, M., Richards, H., Kirkpatrick, P. and Pickard, J. (1994c) Assessment

of cerebral autoregulation using ultrasound and laser Doppler waveforms
an experimental study in anaesthetized rabbits, Neurosurgery, 35, 287293.
Czosnyka, M., Richards, H., Pickard, J. D. and Iyer, V. (1994d) Frequencydependent properties of cerebral blood transport an experimental study
in rabbits. Ultrasound in Medicine and Biology, 20, 391399.
Czosnyka, M., Kirkpatrick, P., Guazzo, E. et al. (1994e) Can TCD pulsatility
indices be used for a non-invasive assessment of cerebral perfusion
pressure in head injured patients?, in Intracranial Pressure IX, (eds H. Nagai,
K. Kamiya and S. Ishii), Springer-Verlag, Berlin, pp. 146149.
Czosnyka, M., Whitehouse, R., Smielewski, P. et al. (1994f) Computer
supported multimodal monitoring in neuro intensive care. International
Journal of Clinical Monitoring and Computing, 11, 223232.
Czosnyka, M., Guazzo, E., Kirkpatrick, P. et al. (1995) Clinical significance of
simultaneous transcranial Doppler and ICP waveform analysis following
head injury. Journal of Neurotrauma, 12, 402.
Czosnyka, M., Smielewski, P., Kirkpatrick, P. J. and Pickard, J. D. (1996)
Monitoring of cerebral autoregulation in head injured patients. Stroke, 27,
18291834.
Dahl, A., Russell, D., Nyberg-Hansen, R. and Rootwelt, K. (1989) Effect of
nitroglycerin on cerebral circulation measured by transcranial Doppler and
SPECT. Stroke, 20, 17331736.
Dahl, A., Lindegaard, K. F., Russell, D. et al. (1992) A comparison of
transcranial Doppler and cerebral blood flow studies to assess cerebral
vasoreactivity. Stroke, 23, 1519.
Eng, C. C., Lam, A. M., Mayberg, T. S. and Mathisen, T. L. (1992) Influence of
propofol and propofolnitrous oxide anesthesia on cerebral blood flow
velocity and carbon dioxide reactivity in humans. Anesthesiology, 77,
872879.
European Study Group. (1994) A multicenter trial on the efficacy of
nimodipine on outcome after severe head injury. Journal of Neurosurgery, 80,
797804.
Feri, M., Ralli, L., Felici, M. et al. (1994) Transcranial Doppler and brain death
diagnosis. Critical Care Medicine, 22, 11201126.
Giller, C. A. (1991) A bedside test for cerebral autoregulation using
transcranial Doppler ultrasound. Acta Neurochirurgica, 108, 714.
Giller, C. A., Bowman, G., Dyer, H. et al. (1993) Cerebral arterial diameters
during changes in blood pressure and carbon dioxide during craniotomy.
Gold, G. E. (1989) Does acute hyperventilation provoke cerebral oligemia in
comatose patients after acute head injury? Acta Neurochirurgica, 96,
100106.
Goncalves, J. M., Vaz, R., Cereo, A. et al. (1994) HM-PAO SPECT in head
trauma. Acta Neurochirurgica (supplement), 55, 1113.
Gopinath, P. S., Robertson, C. S., Constant, C. F. et al. (1994) Jugular venous
Gosling, R. G. and King, D. H. (1974) Arterial assessment by Doppler shift
ultrasound. Proceedings of the Royal Society of Medicine, 67, 447449.
Grosset, D. G., Strebel, S., Straiton, J. et al. (1993) Impaired carbon dioxide
reactivity predicts poor outcome in severe head injury: a transcranial
Doppler study, in Intracranial Pressure VIII, (eds C. J. J. Avezaat, J. H. M. van
Eindhoven, A. I. R. Maas and J. T. J. Tans), Springer-Verlag, Berlin, pp.
322326.
Halsey, J. H., McDowell, H. A. and Gelmon, S. (1986) Transcranial Doppler
and regional cerebral blood flow compared in carotid endarterectomy.
Stroke, 17, 12061208.
Harders, A. (1986) Neurosurgical Applications of Transcranial Doppler Sonography,
Springer-Verlag, Vienna, pp. 2829.
Harders, A. and Gilsbach, J. (1985) Transcranial Doppler sonography and its
application in extracranial-intracranial bypass surgery. Neurology Research,
7, 129141.
Hassler, W., Steinmetz, H. and Gawlowski, J. (1988) Transcranial Doppler
ultrasonography in raised intracranial pressure and intracranial circulatory
arrest. Journal of Neurosurgery, 68, 745751.
Hassler, W., Steinmetz, H. and Pirschel, J. (1989) Transcranial Doppler study
of intracranial circulatory arrest. Journal of Neurosurgery, 71, 195201.
Huber, P. and Handa, J. (1967) Effect of contrast material, hypercapnia,
hyperventilation, hypertonic glucose and papaverine on the diameter of the
cerebral arteries angiographic determination in man. Investigational
Radiology, 2, 1732.
Jones, P. A., Andrews, P. J. D., Midgley, S. et al. (1993) Assessing the burden of
secondary insults in head injured patients during intensive care. Journal of
Neurology, Neurosurgery and Psychiatry , 56, 571572.
Kirkpatrick, P. J., Czosnyka, M., Iyer, I. et al. (1993) Transcranial Doppler
ultrasonography and use of a standardised pulsatility index in the
assessment of head injured patients. Journal of Neurology, Neurosurgery and
Kirkpatrick, P. J., Smielewski, P., Czosnyka, M. and Pickard, J. D. (1994a)
Continuous monitoring of cortical perfusion using Laser Doppler flowmetry in ventilated patients with head injury. Journal of Neurology,
REFERENCES
Kirkpatrick, P. J., Czosnyka, M., Smielewski, P. and Pickard, J. D. (1994b)
Continuous assessment of global cerebral blood flow using Laser Doppler
Flowmetry in head injured patients. Journal of Neurology, Neurosurgery and
Kirkpatrick, P. J., Smielewski, P., Czosnyka, M. et al. (1995) Provisional
observations with near infrared spectroscopy in head injured patients.
Kirkpatrick, P. J., Czosnyka, M. and Pickard, J. (1996) Multimodality
monitoring in neurointensive care (editorial). Journal of Neurology, Neurosurgery and Psychiatry, 60, 131139.
Kirkpatrick, P. J., Smielewski, P., Czosnyka, M. et al. (1996) Use of
multimodality monitoring to observe the mechanisms of mannitol in head
injured patients. Neurosurgery, 39, 716717.
Klingelhofer, J. and Sander, D. (1992) Doppler CO2 test as an indicator of
cerebral vasoreactivity and prognosis in severe intracranial hemorrhages.
Stroke, 23, 962966.
Kontos, H. A. (1989) Validity of cerebral arterial blood calculations from
velocity measurements. Stroke, 20, 13.
Lindegaard, K. F. (1992) Indices of pulsatility, in Transcranial Doppler, (eds D.
W. Newell and R. Aaslid), Raven Press, New York, pp. 6782.
Lindegaard, K. F., Nornes, H., Bakke, S. J. et al. (1988) Cerebral vasospasm
after subarachnoid hemorrhage investigated by means of transcranial
Doppler ultrasound. Acta Neurochirurgica (Vienna), 24, 8184.
Lindegaard, K. F., Nornes, H., Bakke, S. J. et al. (1989) Cerebral vasospasm
diagnosis by means of angiography and blood velocity measurements. Acta
Macpherson, P. and Graham, D. I. (1973) Arterial spasm and slowing of the
cerebral circulation in the ischaemia of head injury. Journal of Neurology,
Marion, D. W., Darby, J. and Yonas, H. (1991) Acute regional cerebral blood
flow changes caused by severe head injuries. Journal of Neurosurgery, 74,
407414.
Markwalder, T. M., Grolimund, P., Seiler, R. W. et al. (1984) Dependency of
blood flow velocity in the middle cerebral artery on end-tidal carbon
dioxide partial pressures a transcranial ultrasound Doppler study. Journal
of Cerebral Blood Flow and Metabolism, 4, 368372.
Marmarou, A., Anderson, R. L., Ward, J. D. et al. (1991) Impact of ICP
instability and hypotension in patients with severe head trauma. Journal of
75, S14S20.
Matta, B. F. and Lam, A. M. (1995) Isoflurane and desflurane do not dilate the
middle cerebral artery appreciably. British Journal of Anaesthesia, 74,
486P487P.
Mayberg, T. S., Lam, A. M., Eng, C. C. et al. (1993) The influence of alfentanil
on cerebral blood flow velocity and intracranial pressure during isofluranenitrous oxide anesthesia. Anesthesiology, 78, 288294.
Meixensberger, J. (1993) Xenon 133-CBF measurements in severe head injury
and subarachnoid haemorrhage. Acta Neurochirurgica (Supplement), 59, 2833.
Miller, J. D. (1985) Head injury and brain ischaemia. Implications for therapy.
British Journal of Anaesthesia, 57, 120129.
Miller, J. D. (1986) Minor, moderate and severe head injury. Neurosurgery
Review, 9, 135139.
Miller, J. D. (1992) Evaluation and treatment of head injury in adults.
Neurosurgery Quarterly, 2, 2843.
Muizelaar, J. P., Ward, J. D., Marmarou, A. et al. (1989) Cerebral blood flow and
metabolism in severely head-injured children. Part 2: Autoregulation.
Muttaqin, Z., Uozumi, T., Kuwabara, S. et al. (1993) Hyperaemia prior to acute
cerebral swelling in severe head injuries: the role of transcranial Doppler
monitoring. Acta Neurochirurgica, 123, 7681.
Nelson, R. J., Czosnyka, M., Pickard, J. D. et al. (1992) Experimental aspects of
cerebrospinal haemodynamics: the relationship between blood flow velocity waveform and cerebral autoregulation. Neurosurgery, 31, 705710.
Newell, D. W. and Aaslid, R. (1992a) Transcranial Doppler. Raven Press, New
York.
Newell, D. and Aaslid, R. (1992a) Transcranial Doppler: clinical and
experimental uses. Cerebrovascular Brain Metabolism Review, 4, 122143.
Newell, D. W., Grady, M. S. and Sirotta, P. (1989) Evaluation of brain death
using transcranial Doppler. Neurosurgery, 24, 509513.
Newell, D. W., Aaslid, R., Stooss, R., Reulen, H. J. (1992) The relationship of
blood flow velocity fluctuations to intracranial pressure B waves. Journal of
259
Newell, D. W., Aaslid, R., Stooss, R. and Reulen, H. J. (1993) Evaluation of

closed head injury patients using transcranial Doppler monitoring, in
Newell, W. D., Aaslid, R., Lam, A. M. et al. (1994) Comparison of flow and
velocity during dynamic autoregulation testing in humans. Stroke, 25,
793797.
Obrist, W. D. and Wilkinson, W. E. (1990) Regional cerebral blood flow
measurement in humans by xenon-133 clearance. Cerebrovascular Brain
Metabolism Review, 2, 283327.
Paulson, O., Olesen, J. and Edvinsson, L. (1990) Cerebral autoregulation.
Perry, G. W., Mohr, J. P., Pedley, T. A. et al. (1990) The role of transcranial
Doppler in confirming brain death: sensitivity, specificity, and suggestions
for performance and interpretation. Neurology, 40, 300303.
Pickard, J. D. and Czosnyka, M. (1993) Management of raised intracranial
pressure. Journal of Neurology, Neurosurgery and Psychiatry, 56, 845858.
Richards, U. K., Czosnyka, M., Kirkpatrick, P. and Pickard, J. D. (1995)
Estimation of laser Doppler flux biological zero using basilar artery flow
velocity in the rabbit. American Journal of Physiology, 268, H213H217.
Robertson, C. S., Narayan, R. K., Gokaslan, Z. L. et al. (1989) Cerebral
Romner, B., Brandt, L., Berntman, L. et al. (1991) Simultaneous transcranial
Doppler sonography and cerebral blood flow measurements of cerebrovascular CO2-reactivity in patients with aneurysmal subarachnoid
haemorrhage, British Journal of Neurosurgery, 5, 3137.
Rosner, M. J., Rosner, S. D. and Johnson, A. H. (1991) Cerebral perfusion
pressure: management protocol and clinical results. Journal of Neurosurgery,
83, 949962.
Schalen, W., Messeter, K. and Nordstrom, C. H. (1991) Cerebral vasoreactivity
and the prediction of outcome in severe traumatic brain lesions. Acta
Anaesthesiologica Scandinavica, 35, 113122.
Schregel, W., Schafermeyer, H., Muller, C. et al. (1992) The effect of halothane,
alfentanil and propofol on blood flow velocity, blood vessel cross section
and blood volume flow in the middle cerebral artery. Anaesthetist, 41,
2126.
Schregel, W., Schafermeyer, H., Sihle-Wissel, M. and Klein, R. (1994)
Transcranial Doppler sonography during isoflurane/N2O anesthesia and
surgery: flow velocity, vessel area and volume flow. Canadian Journal of
Anaesthesia, 41, 607612.
Siesjo,
B. K. (1992) Pathophysiology and treatment of focal cerebral ischaemia.
Parts I and II. Journal of Neurosurgery, 77, 169184, 337354.
Smielewski, P., Minhas, P., El Zayat, S. et al. (1994) Cerebral CO2 reactivity
testing using TCD ultrasound and near-infrared spectroscopy study in
normal volunteers. Cerebrovascular Diseases, 4(Suppl. 3), 23.
Smielewski, P., Kirkpatrick, P. J., Minhas, P. et al. (1995) Can cerebrovascular
reactivity be measured using near infrared spectroscopy? Stroke, 26,
22852292.
Smielewski, P., Czosnyka, M., Iyer, V. et al. (1995) Computerised Transient
Hyperaemic Response Test a method for the assessment of cerebral
autoregulation. Ultrasound in Medicine and Biology, 21, 599611.
Steiger, H. J., Aaslid, R., Stooss, R. and Seiler, R. W. (1994) Transcranial
Doppler monitoring in head injury: relations between type of injury, flow
velocities, vasoreactivity, and outcome. Neurosurgery, 34, 7985.
Strebel, S., Kaufmann, M., Guardiola, P.-M. and Schaefer, H.-G. (1994)
Cerebral vasomotor responsiveness to carbon dioxide is preserved during
propofol and midazolam anesthesia in humans. Anesthesia and Analgesia, 78,
884888.
Thiel, A., Zickmann, B., Zimmermann, R. and Hemplemann, G. (1992)
Transcranial Doppler sonography: effects of halothane, enflurane and
isoflurane on blood flow velocity in the middle cerebral artery. British
Journal of Anaesthesia, 68, 389393.
Weber, M., Grolimund, P. and Seiler, R. W. (1990) Evaluation of post-traumatic
cerebral blood flow velocity by transcranial Doppler ultrasonography.
Werner, C., Kochs, E., Rau, M. and Schulte am Esch, J. (1990) Transcranial
Doppler sonography as a supplement in the detection of cerebral
circulatory arrest. Journal of Neurosurgery and Anesthesiology, 2, 159165.
Wong, F. C., Piper, I. R. and Miller, J. D. (1994) Waveform analysis of blood
pressure, intracranial pressure and transcranial Doppler signals and their
relationship to cerebral perfusion pressure in head injured patients, in
Intracranial Pressure IX, (eds H. Nagai, K. Kamiya and S. Ishii), SpringerVerlag, Berlin, pp. 144145.
14
MAGNETIC RESONANCE
SPECTROSCOPY
Robert Vink
14.1
Introduction
Traumatic brain injury continues to be the major cause

of death in individuals under 44 years of age (Chapter
1). In addition to the high degree of mortality, the
morbidity associated with survivors of severe head
injury imposes significant social and financial costs
upon the community. Despite being such a major
public health issue, the mechanisms that result in the
irreversible tissue damage after brain trauma are still
unclear. It has become increasingly accepted that
irreversible tissue damage after brain trauma occurs
through both primary and secondary means. Primary
events are generally considered to be the mechanical
changes to nervous tissue that occur at the time of
insult, while secondary events are the physiological
and biochemical changes that occur over time after the
initiating insult (Chapter 4). The fact that the secondary events occur over hours to days after the trauma
encourages attempts to prevent, or at least attenuate,
the secondary injury cascade. Nonetheless, before
preventative therapies can be introduced, there is a
clear need to identify the factors involved in the
secondary injury process, and a need to be able to
identify whether a selected antifactor is indeed
having the desired effect on brain metabolism. While a
number of techniques can be used in experimental
studies to investigate metabolic events associated with
trauma and pharmacological intervention, these are
primarily invasive techniques that cannot be translated to a clinical setting. However, the recent application of non-invasive magnetic resonance spectroscopy
(MRS) to experimental trauma studies has now resulted in MRS techniques being applied to metabolic
studies of clinical head injury. The purpose of this
chapter is to provide an overview of magnetic
resonance spectroscopy and its utility in non-invasive
metabolic studies. Recent clinical applications of MRS
in head injury and how they correlate with previous
experimental studies will be summarized, and newer
magnetic resonance imaging (MRI) techniques for

studies of brain metabolism and function will be
briefly discussed.
14.2 Principles of magnetic resonance

spectroscopy
It is not the purpose of the current review to describe
in detail the basic principles of magnetic resonance,
but rather to illustrate the non-invasive nature of
magnetic resonance spectroscopy and its potential for
studies of brain metabolism. For readers interested in
a detailed description of magnetic resonance theory,
there are a number of excellent articles available
(Gadian, 1982; Iles, Stevens and Griffiths, 1982; Gordon, 1985).
Like all forms of spectroscopy, magnetic resonance
relies on the excitement of a particle from a ground
state to an excited state and the measurement of
energy as the particle relaxes back into the ground
state. In the case of MRS, the particle is the magnetic
dipoles within the nuclei of the particular molecules of
interest. When placed into a strong magnetic field, the
magnetic dipoles assume either a low-energy state
(where the dipoles align with the magnetic field) or a
high-energy state (where the dipoles align against the
magnetic field). Transitions between the two energy
states can be induced by the application of an
appropriate form of energy, which in this case is radiofrequency energy. Furthermore, different nuclei
require particular frequencies of radio-frequency
energy to induce the transition. Hence, proton nuclei,
for example, require a different frequency of excitation
from other nuclei such as phosphorus. Tuning in for a
particular nucleus is therefore similar to tuning in a
transistor radio to a particular radio frequency. Having tuned in to a nucleus of interest, applying a burst
of radio-frequency energy will induce transitions in
energy states in all molecules containing that nucleus.
The transition from a low-energy state to a high-
262
MAGNETIC RESONANCE SPECTROSCOPY
energy state and back again is known as resonance,

and the whole process as nuclear (for the nuclear
dipoles) magnetic resonance.
The relaxation of an excited nuclear dipole back into
the low-energy state will emit energy in the form of
radio frequency, which can be detected by a radiofrequency receiver coil. Again, the receiver coil must
be tuned into the correct radio frequency to detect this
release and it is not unusual that the one radiofrequency coil is used firstly as the transmitter and
then as the receiver. The detected signal will decay
with time as the dipole relaxes into the lower-energy
state. By converting the amplitude versus time decay
into an amplitude versus frequency plot (by Fourier
transformation), the resulting MR spectrum will show
a variable number of peaks whose amplitude is
directly related to the number of nuclei undergoing
resonance, and whose frequency identifies which
nuclei are being excited. Although it excites one type
of nucleus, nuclei in different chemical or ionic
environments have their resonant frequency affected
by that environment. The result is that slightly
different frequencies are observed for each different
environment experienced by the nuclei. For example,
in a phosphorus MR spectrum, inorganic phosphate
Figure 14.1 Typical phosphorus MRS spectrum of brain.

Identifiable peaks include the phosphomonoesters (PME),
inorganic phosphate (Pi ), phosphocreatine (PCr ) and the
three phosphate groups of adenosine triphosphate (ATP).
will have a slightly different frequency (chemical shift)

from a phosphate molecule attached to an adenosine
ring (adenosine monophosphate): the adenosine ring
confers a different chemical environment upon the
phosphate. Similarly, adenosine diphosphate will
have a different chemical shift from adenosine monophosphate because of the different chemical environments being experienced by the phosphate groups,
and so on. A typical phosphorus MRS spectrum of
brain is shown in Figure 14.1. A number of phosphate
metabolites can be recognized on the basis of previous
work (Vink, 1993), including phosphomonoesters,
inorganic phosphate, phosphocreatine, and the three
phosphate groups of adenosine triphosphate, labeled
as , and ATP.
Having considered how chemical environment
affects the resonant frequency of different nuclei, it
should be noted that ionic environment can also affect
resonant frequencies. This property has proven very
useful, particularly in phosphorus MRS where the
effect of protons (H+) on the inorganic phosphate
chemical shift has been exploited to gain information
on regional pH (Petroff et al., 1985). In addition, a
number of studies have used the chemical shift of the
ATP peak to calculate values for intracellular free
magnesium concentration (Gupta and Gupta, 1984).
14.3
MRS studies of neurotrauma
14.3.1
PHOSPHATES
The first applications of magnetic resonance spectroscopy to experimental traumatic brain injury were
published in 1987 and used phosphorus MRS to focus
on energy metabolism and pH after trauma (Ishige et
al., 1987; Vink et al., 1987b). It wasnt long before the
utility of MRS in clinical neurotrauma was recognized
and the first study on human head injury was
published in 1990 by Gennarelli and colleagues
(Rango et al., 1990). These authors found that in
patients with a mean Glasgow Coma Score of 6.1,
there were no significant changes in the phosphoruscontaining metabolites between days 2 and 21 after
injury when compared to normal controls. The lack of
any apparent energy failure in this study was consistent with the animal studies published earlier.
These results suggest that brain trauma, as opposed to
brain infarction (Williams, Crockard and Gadian,
1989), does not cause overt energy failure as determined from the level of high-energy phosphate compounds in the MRS spectra.
Nonetheless, subtle changes in phosphocreatine to
inorganic phosphate (PCr/Pi) ratio have been reported
in animal studies of trauma indicating an increased
energy demand after trauma (Vink et al., 1988a). These
changes in PCr/Pi ratios have been correlated to the
MRS STUDIES OF NEUROTRAUMA
appearance of neurological deficits in traumatized

animals (Vink et al., 1988a), and similar correlations
have been found clinically, at least in subdural
hemorrhage (Yoshida et al., 1994), birth asphyxia (Cady
et al., 1983; Younkin et al., 1984) and seizures (Younkin et
al., 1986). Moreover, the studies of subdural hemorrhage (Yoshida et al., 1994) have reported that evacuation of the hematoma results both in an improvement in
PCr/Pi ratio and the disappearance of any hemiparesis.
A closer inspection of PCr/Pi ratios in clinical trauma
may thus be warranted, at least with respect to
interventional strategies for subdural hematomas. A
final point to be made about phosphate MRS of trauma
is that animal studies have shown that imposition of a
second stressor, such as hypoxia or hypotension, on a
traumatized brain results in a significant decline in
brain phosphate energy stores (Ishige et al., 1987, 1988).
This finding suggests that the traumatized brain may
be sensitized to secondary insults and that the
subsequent appearance of a secondary event may
initiate a considerable energetic perturbation.
14.3.2
BRAIN pH
Phosphorus MRS spectra also contain information on

brain pH. The phosphorus data from the original 1990
clinical study of human head injury (Rango et al., 1990)
demonstrated that intracellular pH was never in the
263
acidotic range. In fact, there was a tendency toward

alkalosis in the first days to weeks after trauma, which
had resolved by 3 weeks. In a 1995 proton MRS study
of pediatric head injury, Sutton et al. (1995) confirmed
that diffuse axonal injury did not cause brain acidosis
by demonstrating an absence of any increase in lactate
concentration. Previous animal studies of trauma had
shown that any acidosis after trauma was correlated
with lactic acid accumulation (McIntosh et al., 1987).
Some cases of lactate accumulation were reported
after clinical head injury but this accumulation was
always associated with either localized contusion or
cerebral infarction (Sutton et al., 1995). Indeed, the
appearance of acidotic regions in human brain has
been proposed to be an indicator of focal ischemia
(Brooke et al., 1994). Moreover, their appearance was
correlated to impaired consciousness and the normalization of pH within these infarcted regions over time
coincided with an improvement in clinical condition.
It is significant that these clinical trauma studies
confirmed earlier MRS studies of experimental head
injury (McIntosh et al., 1987) which demonstrated
that any brain lactic acidosis after trauma was both
mild and transient irrespective of injury severity
(Figure 14.2). Indeed, significant acidosis only occurred under the most severe conditions, where mortality
was 100% and global ischemia was suspected to be a
contributing factor (McIntosh et al., 1987).
Figure 14.2 Temporal profile of lactate accumulation following low, moderate and high levels of experimental traumatic
brain injury in rats. Lactate accumulation is transient and levels remain below what is considered to be the injury threshold.
(Source: adapted from McIntosh et al., 1987.)
264
Previous studies had shown that CSF lactate

concentration may be a useful prognostic indicator
following brain injury (DeSalles et al., 1986). However, MRS studies of experimental trauma confirmed
that the injured brain may not be the source of the
CSF lactate and that much of the CSF lactate may in
fact originate from systemic sources (Inao et al.,
1988). Although CSF lactate may still be a useful
prognostic indicator after trauma, the conclusion
from both the experimental and clinical studies
described above is that therapies targeted at alkalinizing brain cells may only be of use in cases with
suspected brain contusion or infarction. Moreover,
while these conditions can be identified by MRI, it
would be beneficial to use MRS to determine efficacy
of the buffering agents.
14.3.3
FREE MAGNESIUM
The free magnesium ion is gaining increasing recognition as a critical factor in neuronal cell function.
The finding that magnesium ions gate the n-methylD-aspartate class of glutamate receptors (Mayer,
Westbrook and Guthrie, 1984) has rekindled earlier
enthusiasm for magnesium as an important regulator
of central nervous system function (Fishman, 1965).
Indeed, magnesium has now been shown to play an
active role in the regulation of calcium channels
(Agus et al., 1989), prostanoid synthesis (Nigam,
Averdunk and Gunther, 1986), membrane peroxidation and free radical formation (Gunther et al., 1994),
energy metabolism (Ebel and Gunther, 1980), glycolysis (Garfinkel and Garfinkel, 1985), neurotransmitter
release and binding (Rothman, 1983), cerebral vascular tone (Altura et al., 1984), protein synthesis and
DNA transcription (Terasaki and Rubin, 1985),
among others. It is therefore clear that potential
changes in magnesium ion concentration may have
far-reaching consequences in terms of metabolic
events and recovery after brain injury. The difficulty
in demonstrating a role for magnesium in secondary
injury after trauma has been that, unlike calcium,
measurement of free magnesium concentration has
been technically difficult and the available techniques largely inadequate. The recent development
of phosphorus MRS techniques for the determination
of intracellular free magnesium concentration has
since permitted the reliable analysis of free magnesium concentration prior to and following a traumatic event.
The first application of phosphorus MRS to measurement of free magnesium concentration after experimental trauma appeared in 1987 (Vink et al., 1987a)
and was quickly followed by a succession of papers
characterizing the potential role that magnesium
decline after trauma may have in the secondary injury
process (Vink et al., 1988b; Vink and McIntosh, 1990;

Vink, Faden and McIntosh, 1988; Vink, Golding and
Headrick, 1994; Vink, Portoghese and Faden, 1991). In
summary, these studies demonstrated that brain
intracellular free magnesium concentration declines
after trauma by as much as 60%. The significance of
this may not be readily apparent to those unfamiliar
with magnesiums role in biochemical and physiological processes, so Figure 14.3 shows the data
transformed to a logarithmic scale and plotted adjacent to a more familiar parameter, pH. Free magnesium concentration typically changes by 0.25 units. If
brain pH changed by 0.25 units (from 7.10 to 6.85), this
would undoubtedly cause considerable concern! As
opposed to the highly significant decline in free
magnesium concentration, brain pH does in fact not
change significantly after trauma.
Subsequent experimental trauma studies have demonstrated that pharmacological agents with neuroprotective properties in terms of outcome all
improved free magnesium homeostasis (Vink, 1993).
Notably, those agents that improved free magnesium
homeostasis the most also had the greatest improvement in neurological motor scores. Recently, Smith
and colleagues (1993) also demonstrated that
improvement in magnesium homeostasis after trauma
not only resulted in better neurological motor outcome after trauma but also significantly improved
cognitive performance. Although the mechanism of
action by which magnesium improves post-traumatic
outcome is unknown, the fact that magnesium is a
critical regulatory cation in such a large number of
cellular functions (see above) dictates that any alteration in free ion concentration may have wide-ranging
consequences.
Of course, the observation that magnesium declines
following experimental trauma may not be apparent
in clinical head injury. However, a recent study
presented at the Third International Neurotrauma
Symposium in Toronto (Lenkinski et al., 1995) suggests
that decline in free magnesium concentration is also
typical of clinical trauma. This group reported that
free magnesium concentration declines by approximately 50% and that this declines persists for at least
1 week. These observations have led to the recently
commenced clinical trial using magnesium as a posttraumatic therapy to attenuate secondary damage
(Lenkinski et al., 1995).
14.3.4
PROTON MRS
With the exception of Suttons recent publication on

pediatric head injury (Sutton et al., 1995), there have
been very few reports concerning applications of
proton MRS to clinical trauma. Nonetheless, the
results from proton MRS studies in clinical ischemia
MRI STUDIES OF BRAIN FUNCTION
265
Figure 14.3 Intracellular pMg and pH levels following moderate traumatic brain injury. Intracellular pH typically does not
change significantly after trauma whereas highly significant declines in free magnesium concentration occur immediately
after trauma and persist for up to 1 week after injury. * = p < 0.05 from preinjury by ANOVA.
and neonatal development demonstrate potential

applications to neurotrauma that are worthy of discussion. Of particular interest is the apparent correlation between n-acetyl aspartate (NAA) levels and
degree of neuronal injury. N-acetyl aspartate is an
intracellular
intermediate
metabolic
product,
required for neurotransmitter synthesis, which is
present in large amounts in normal functioning
mature neurons only. Its easily detected proton MR
spectral peak has been shown in numerous studies to
closely correlate with neuronal integrity and function
and its loss or attenuation is an important in vivo
indicator of acute or chronic neuronal depletion or
destruction. A number of studies at the experimental
and clinical level have now reported a decline in
NAA after ischemic injury to the brain. Moreover,
while phosphorus ratios derived from phosphorus
MRS spectra normalize over time following the
injury, NAA concentration as assessed from proton
MRS spectra remains depressed (Peden et al., 1990;
Houkin et al., 1993). Indeed, the NAA-to-creatine
and NAA-to-choline ratios in the proton MRS spectra
have been correlated to neurodevelopmental outcome in infants born with signs of hypoxicischemic
encephalopathy (Peden et al., 1993).
Declines in NAA concentration have also been
found in experimental trauma (Lenkinski, unpublished results; Vink, unpublished results) and it
remains to be determined whether such changes are
observed in clinical head injury and whether any

changes in NAA can be used for prognostic
purposes.
14.4
MRI studies of brain function
Most clinicians and basic scientists should be familiar

with the use of proton magnetic resonance as it is used
in MRI. Simply stated, MRI uses the concentration of
water to provide a spatial representation of water
distribution throughout the tissues (Zimmerman et al.,
1986). Because of the difference in water concentration
across various tissues, there is a 14% contrast available. Moreover, wherever water accumulates (as in
edema, tumors or hemorrhage) the high concentration
is easily detected by MRI. Recent technological developments in magnetic resonance now permit information about brain function, as opposed to brain anatomy, to be gained by such imaging techniques. By
exciting water molecules using strong magnetic field
gradients and obtaining images in a very short time,
the diffusion and perfusion of those excited molecules
in and out of a chosen plane can be monitored over
time. Hence changes in blood flow and volume,
increased blood oxygenation, intracellular accumulation of water and extracellular accumulation of water
can all be visualized using these functional-type MRI
techniques (for reviews, see Cohen and Bookheimer,
1994; Hossmann and Hoehnberlage, 1995). While few
266
applications of these techniques to trauma have been

reported to date, the potential is enormous.
14.4.1
APPLICATIONS TO TRAUMA
One of the first applications of MRI to gain functional

information after experimental trauma was in studies
of edema (Hanstock et al., 1994). The discrimination of
extracellular water accumulation (as in vasogenic
edema) from intracellular accumulation (as in cytotoxic edema) can be accomplished by a technique
known as diffusion-weighted imaging. In essence,
this technique results in an image whose intensity is
largely determined by the distance excited water
molecules can diffuse over time. The shorter the
distance a water molecule can diffuse the brighter the
image intensity (e.g. intracellular water is restricted to
within the boundaries of a cell membrane). In contrast,
the greater the diffusion distance of water (e.g.
interstitial water), the lower the image intensity. As
opposed to studies of brain ischemia (Moseley et al.,
1990), diffusion-weighted MRI studies of experimental
traumatic brain injury demonstrated that early water
accumulation following a traumatic event is largely
vasogenic in nature (Hanstock et al., 1994), at least for
the first 4 hours after the traumatic event. This
observation supported the notion that there was no
overt energy failure following trauma that might
precipitate the development of cytotoxic edema.
Rather, the brief opening of the bloodbrain barrier
after experimental trauma permits the accumulation
of proteins in the brain extracellular space and the
subsequent accumulation of water. In contrast, superimposition of a second insult such as hypotension or
hypoxia on brain trauma results in profound cytotoxic edema (Ito et al., 1996) consistent with the
previously proposed notion that combined insults
leads to ischemia with associated energy failure
(Ishige et al., 1987, 1988). Interestingly, Marmarou and
colleagues (Ito et al., 1995) also demonstrated that the
amount of water accumulation in the extracellular
space following trauma alone declines between 4
hours and 24 hours, implying that the extracellular
space is getting smaller. Various factors may account
for this reduction in extracellular volume, including
development of cytotoxic edema and/or astrocyte
swelling. Whatever the cause of the change with time,
these results re-emphasize that trauma has a clear
secondary component that develops over time.
Such secondary injury, at least with respect to
development of diffuse axonal injury, has also been
confirmed using magnetization transfer imaging
(Smith et al., 1995). These authors were able to
demonstrate that magnetization transfer imaging was
more sensitive than conventional imaging techniques
in detecting the presence of diffuse axonal injury. The
axonal injury was detected at 3 days after injury

despite no such indication on conventional MR
images. Because of the significance of diffuse axonal
injury in the development of severe injury, the early
identification of such injury may prove to be of
prognostic significance. Both diffusion-weighted
imaging and magnetization transfer imaging have
thus been reported to be more sensitive than any other
technique in the early detection of edema and diffuse
axonal injury, respectively.
14.4.2 POTENTIAL FUTURE APPLICATIONS TO
TRAUMA
Two further functional MRI techniques show particular promise for application to studies of traumatic
brain injury. The first is perfusion imaging, used to
measure blood flow. Cerebral blood flow has usually
been determined by measuring the washout of an
exogenously added tracer. Perfusion imaging works
on a similar principle except that the tracer is MR
excited water in the vasculature (arterial spin labeling) and washout is the signal decay in the region of
interest (Detre et al., 1994). Consequently, estimates of
cerebral blood flow and volume can be made noninvasively and repeatedly without the use of exogenous tracers. The technique has been successfully
developed for use in human brain where changes in
blood flow have been detected under conditions of
hyperventilation or breath-holding (Roberts et al.,
1994). More recently, perfusion imaging has been
nominated as a prognostic tool for predicting clinical
outcome following ischemic stroke (Warach, Dashe
and Edelman, 1996). Indeed, this early evidence
suggests that the technique will be a valuable tool in
distinguishing reversible from irreversible tissue
injury. In trauma studies, the technique will be
extremely useful in the determination of blood flow
values associated with the purported ischemic threshold that may or may not be similar to values obtained
in ischemia studies.
The second technique that shows promise for
application to trauma is what is known literally as
functional MRI. The technique takes advantage of the
effects of blood oxygenation on MR signal decay. The
lower the oxygen concentration, the higher the concentration of deoxyhemoglobin and the more rapidly MR
signal intensity decays. Thus, an increase in oxygen
concentration in venous blood (due to increased flow
relative to tissue oxygen consumption) results in an
increased intensity in the MR image (Cohen and
Bookheimer, 1994). Following the initial development
in animal studies, functional MRI has been used
extensively to demonstrate increased human functional brain activation in response to a specified task
(Shulman et al., 1993). In more recent clinical studies,
REFERENCES
the technique has been useful for mapping brain

functional deficits following stroke (Sorensen et al.,
1995) and in demonstrating cortical reorganization
following focal perinatal brain damage (Cao et al.,
1994). With respect to trauma, the technique has the
potential to be used to map functional deficits as well
as being a useful technique to predict functional
outcome and to chart the recovery of function in the
injured brain during recovery from severe head
injury.
14.5
Conclusion
There seems no doubt that magnetic resonance spectroscopy has made a significant contribution to the
understanding of pathophysiological processes following experimental traumatic injury to the brain.
Many of the early observations made in animals are
now being confirmed in a clinical setting and a
number of therapeutic strategies are being examined
on the basis of these MRS findings. Nonetheless,
despite the advances made in understanding the
mechanisms of injury, it remains unclear whether MRS
may develop into a useful prognostic indicator of
outcome. Recent reports of both proton MRS and
functional MRI studies suggest that the magnetic
resonance techniques may indeed be useful in this
regard. However, it remains to be determined whether
the information gained from routine clinical MR
evaluation justifies the expense of higher field magnets and the associated risks to the patient in such a
high magnetic field environment.
14.6
References
Agus, Z. A., Kellepouris, E., Dukes, I. and Morad, M. (1989) Cytosolic

magnesium modulates calcium channel activity in mammalian ventricular
cells. American Journal of Physiology, 256, C452C455.
Altura, B. M., Altura, B. T., Gebrewold, A. et al. (1984) Magnesium deficient
diets and microcirculatory changes in situ. Science, 223, 13151317.
Brooke, N. S. R., Ouwerkerk, R., Adams, C. B. T. et al. (1994) Phosphorus-31
magnetic resonance spectra reveal prolonged intracellular acidosis in the
brain following subarachnoid hemorrhage. Proceedings of the National
Academy of Sciences of the USA, 91, 19031907.
Cady, E. B., Dawson, M. J., Hope, P. L. et al. (1983) Non-invasive investigation
of cerebral metabolism in newborn infants by phosphorus nuclear magnetic
resonance spectroscopy. Lancet, i, 10591062.
Cao, Y., Vikingstad, E. M., Huttenlocher, P. R. et al. (1994) Functional magnetic
resonance studies of the reorganization of the human hand sensorimotor
area after unilateral brain injury in the perinatal period. Proceedings of the
National Academy of Sciences of the USA, 91, 96129616.
Cohen, M. S. and Bookheimer, S. Y. (1994) Localization of brain function using
magnetic resonance imaging. Trends in the Neurological Sciences, 17,
268277.
DeSalles, A. A. F., Kontos, H. A., Becker, D. P. et al. (1986) Prognostic
significance of ventricular CSF lactic acidosis in severe head injury. Journal
Detre, J. A., Zhang, W. G., Roberts, D. A. et al. (1994) Tissue specific perfusion
imaging using arterial spin labeling. Nuclear Magnetic Resonance in
Biomedicine, 7, 7582.
Ebel, H. and Gunther, T. (1980) Magnesium metabolism: a review. Journal of
Clinical Chemistry and Clinical Biochemistry, 18, 257270.
Fishman, R. A. (1965) Neurological aspects of magnesium metabolism.
Gadian, D. G. (1982) Nuclear Magnetic Resonance and Its Application to Living
Systems, Oxford University Press, New York.
267
Garfinkel, L. and Garfinkel, D. (1985) Magnesium regulation of the glycolytic

pathway and the enzymes involved. Magnesium, 4, 6072.
Gordon, R. E. (1985) Magnets, molecules and medicine. Physical Medicine and
Biology, 30, 741770.
Gunther, T., Hollriegl, V., Vormann, J. et al. (1994) Increased lipid peroxidation
in rat tissues by magnesium deficiency and vitamin E depletion. Magnesium
Bulletin, 16, 3843.
Gupta, R. K. and Gupta, P. (1984) NMR studies of intracellular metal ions in
intact cells and tissues. Annual Review of Biophysics and Bioengineering, 13,
2143.
Hanstock, C. C., Faden, A. I., Bendall, M. R. and Vink, R. (1994) Diffusionweighted imaging differentiates ischemic tissue from traumatized tissue.
Stroke, 25, 843848.
Hossmann, K. A. and Hoehnberlage, M. (1995) Diffusion and perfusion MR
imaging of cerebral ischemia. Cerebrovascular Brain Metabolism Review, 7,
187217.
Houkin, K., Kamada, K., Kamiyama, H. et al. (1993) Longitudinal changes in
proton magnetic resonance spectroscopy in cerebral infarction. Stroke, 24,
13161321.
Iles, R. A., Stevens, A. N. and Griffiths, J. R. (1982) NMR studies of metabolites
in living tissues. Progress in NMR Spectroscopy, 15, 49200.
Inao, S., Marmarou, A., Clarke, G. D. et al. (1988) Production and clearance of
lactate from brain tissue, cerebrospinal fluid and serum following experimental brain injury. Journal of Neurosurgery, 69, 736744.
Ishige, N., Pitts, L. H., Poliani, L. et al. (1987) The effect of hypoxia on
traumatic head injury in rats: Part 2. Changes in high-energy phosphate
metabolism. Neurosurgery, 20, 854858.
Ishige, N., Pitts, L. H., Berry, I. et al. (1988) The effects of hypovolemic
hypotension on high-energy phosphate metabolism of traumatized brain in
rats. Journal of Neurosurgery, 68, 129136.
Ito, J., Barzo, P., Fatouros, P. et al. (1995) Characterization of edema by
diffusion weighted imaging following closed head injury and secondary
insult in the rat. Journal of Neurotrauma, 12, 475.
Ito, J., Marmarou, A., Barzo, P. et al. (1996) Characterization of edema by
diffusion-weighted imaging in experimental traumatic brain injury. Journal
Lenkinski, R. E., Gennarelli, T. A., McIntosh, T. K. et al. (1995) Magnetic
resonance spectroscopy of the injured brain. Journal of Neurotrauma, 12, 351.
McIntosh, T. K., Faden, A. I., Bendall, M. R. and Vink, R. (1987) Traumatic
brain injury in the rat: alterations in brain lactate and pH as characterized
by 1H and 31P nuclear magnetic resonance. Journal of Neurochemistry, 49,
15301540.
Mayer, M. L., Westbrook, G. L. and Guthrie, P. B. (1984) Voltage-dependent
block by Mg2+ of NMDA responses in spinal cord neurons. Nature, 309,
261263.
Moseley, M. E., Cohen, Y., Mintorovitch, J. et al. (1990) Early detection of
regional cerebral ischemia in cats: comparison of diffusion- and
T2-weighted MRI and spectroscopy. Magnetic Resonance Medicine, 14,
330346.
Nigam, S., Averdunk, R., and Gunther, T. (1986) Alteration of prostanoid
metabolism in rats with magnesium deficiency. Prostaglandins, Leukotrienes
and Medicine, 23, 110.
Peden, C. J., Cowan, F. M., Bryant, D. J. et al. (1990) Proton MR spectroscopy of
the brain of infants. Journal of Computer Assisted Tomography, 14, 886894.
Peden, C. J., Rutherford, M. A., Sargentoni, J. et al. M. (1993) Proton
spectroscopy of the neonatal brain following hypoxic-ischaemic injury.
Developmental Medicine and Child Neurology, 35, 502510.
Petroff, O. A. C., Prichard, J. W., Behar, K. L. et al. (1985) Cerebral intracellular
pH by 31P nuclear magnetic resonance spectroscopy. Neurology, 35, 781788.
Rango, M., Lenkinski, R. E., Alves, W. M. and Gennarelli, T. A. (1990) Brain pH
in head injury: an image-guided 31P magnetic resonance spectroscopy
study. Annals of Neurology, 28, 661667.
Roberts, D. A., Detre, J. A., Bolinger, L. et al. (1994) Quantitative magnetic
resonance imaging of human brain perfusion at 1.5-T using steady-state
inversion of arterial water. Proceedings of the National Academy of Sciences of
the USA, 91, 3337.
Rothman, S. M. (1983) Synaptic activity mediates death of hypoxic neurons.
Science, 220, 536537.
Shulman, R. G., Blamire, A.M., Rothman, D. L. and McCarthy, G. (1993)
Nuclear magnetic resonance imaging and spectroscopy of human brain
function. Proceedings of the National Academy of Sciences of the USA, 90,
31273133.
Smith D. H., Okiyama K., Gennarelli T. A., and McIntosh T. K. (1993)
Magnesium and ketamine attenuate cognitive dysfunction following
experimental brain injury. Neuroscience Letters, 157, 211214.
Smith, D. H., Kimura, H., Lenkinski, R. E. et al. (1995) Identification of diffuse
axonal injury in pig brain using magnetization transfer imaging. Journal of
Sorensen, A. G., Wray, S. H., Weisskoff, R. M. et al. (1995) Functional MR of
brain activity and perfusion in patients with chronic cortical stroke.
American Journal of Neuroradiology, 16, 17531762.
Sutton, L. N., Wang, Z., Duhaime, A. C. et al. (1995) Tissue lactate in pediatric
head trauma: a clinical study using 1H NMR spectroscopy. Pediatric
268
Terasaki, M. and Rubin, H. (1985) Evidence that intracellular magnesium is

present in cells at a regulatory concentration for protein synthesis.
Proceedings of the National Academy of Sciences of the USA, 82,
73247326.
Vink, R. (1993) Nuclear magnetic resonance characterization of secondary
mechanisms following traumatic brain injury. Molecular Chemistry and
Neuropathology, 18, 279297.
Vink, R., Faden, A. I. and McIntosh, T. K. (1988) Changes in cellular
bioenergetic state following graded traumatic brain injury in rats: determination by phosphorus-31 magnetic resonance spectroscopy. Journal of
Vink, R., Golding, E. M. and Headrick, J. P. (1994) Bioenergetic analysis of
oxidative metabolism following traumatic brain injury in rats. Journal of
Vink, R. and McIntosh, T. K. (1990) Pharmacological and physiological effects
of magnesium on experimental traumatic brain injury. Magnesium Research,
3, 163169.
Vink, R., Portoghese, P. S. and Faden, A. I. (1991) Kappa-opioid antagonist
improves cellular bioenergetics and recovery after traumatic brain injury.
Vink, R., McIntosh, T. K., Demediuk, P. and Faden, A. I. (1987a) Decrease in
total and free magnesium concentration following traumatic brain injury in
rats. Biochemical and Biophysical Research Communications, 149, 594599.
Vink, R., McIntosh, T. K., Weiner, M. W. and Faden, A. I. (1987b) Effects of
traumatic brain injury on cerebral high energy phosphates and intracellular
pH: a 31P magnetic resonance spectroscopy study. Journal of Cerebral Blood
Flow and Metabolism, 7, 563571.
Vink, R., McIntosh, T. K., Demediuk, P. et al. (1988a) 31P NMR characterization
of graded traumatic brain injury in rats. Magnetic Resonance Medicine, 6,
3748.
Vink, R., McIntosh, T. K., Demediuk, P. et al. (1988b) Decline in intracellular
free magnesium concentration is associated with irreversible tissue injury
following brain trauma. Journal of Biological Chemistry, 263, 757761.
Warach, S., Dashe, J. F. and Edelman, R. R. (1996) Clinical outcome in ischemic
stroke predicted by early diffusion-weighted and perfusion magnetic
resonance imaging a preliminary analysis. Journal of Cerebral Blood Flow
Williams, S. R., Crockard, A. and Gadian, D. G. (1989) Cerebral ischemia
studied by nuclear magnetic resonance spectroscopy. Cerebrovascular Brain
Metabolism Review, 1, 91114.
Yoshida, K., Furuse, M., Izawa, A. et al. (1994) Dynamics of cerebral
metabolism in patients with chronic subdural hematoma evaluated with
phosphorus-31 MR spectroscopy before and after surgery. American Journal
of Neuroradiology, 15, 16811686.
Younkin, D. P., Delivoria-Papadopoulos, M., Leonard, J. C. et al. (1984) Unique
aspects of human newborn cerebral metabolism evaluated with phosphorus nuclear magnetic resonance spectroscopy. Annals of Neurology, 16,
581586.
Younkin, D. P., Delivoria-Papadopoulos, M., Maris, J. et al. (1986) Cerebral
metabolic effects of neonatal seizures measured with in vivo 31P NMR
spectroscopy. Annals of Neurology, 20, 513519.
Zimmerman, R. A., Bilanuik, L. T., Hackney, D. B. et al. (1986) Head injury:
early results of comparing CT and high- field MR. American Journal of
Roentgenology, 147, 12151222.
Part Three
TREATMENT
Introduction
A number of recent clinical studies have tested the
value of various components of head injury treatment
and studies of steroids, hyperventilation and barbiturates have led to changes in management. However, it
is clear from surveys both in USA and Australia that the
management of head injuries may vary widely from
one place to another. This divergence of practice and
the increasing number of clinical trials being reported
has led to the development of guidelines to replace ad
hoc management. The most comprehensive set of
guidelines published so far has been prepared by the
American Association of Neurological Surgeons in
conjunction with the Brain Injury Foundation. In this
evidence-based publication, degrees of certainty are
graded according to the strength of the data, i.e.
whether it is based on controlled trials, non-controlled

trials, case reports or expert opinion. Conclusions
derived from this information are then graded as to
whether they support a standard treatment, where the
evidence justifies a universal mode of care, a guideline,
where the evidence justifies a strong recommendation
for a mode of care, or an option, where the evidence
does not give strong direction and the decision must be
made by the treating clinician. A requirement of
guideline development is that they should be reviewed
regularly to take account of new evidence.
In this section, the scientific basis and evidence in
support of different components of treatment are
reviewed and, in appendices, concise management
details are given where appropriate. The individual
chapters accord closely with the American Association
of Neurological Surgeons guidelines.
15
FROM ACCIDENT SITE

TO TRAUMA CENTER
J. E. Gilligan
15.1
General aspects of trauma
The management of the head injury must be an

integral part of a general or holistic plan for the
management of the injured patient (Adams Cowley,
1984). Blunt trauma, especially from automobile
accidents, is usually a multiple system disorder and
an isolated head injury is relatively uncommon. For
treatment to be most effective, management systems
should be established that take into account the needs
of a particular geographical region.
In recent years, trauma management has become
codified as the Advanced Trauma Life Support Program (ATLS) of the American College of Surgeons (in
Australia termed the emergency management of
severe trauma or EMST). This codification gives all
health-care workers, neurosurgeons, other surgeons,
intensive care specialists, anesthetists, general practitioners, nurses and ambulance officers a common
terminology and approach to the management of
trauma. This chapter will outline the stages of management of a patient with a head injury, who is likely
to have other injuries, from the scene of an accident to
the stage of definitive treatment.
15.1.1
DEATHS FROM TRAUMA
Early data (Reinfurt et al., 1978, Trunkey, 1993)

indicated a broadly trimodal distribution:
very early (immediate or instantaneous deaths),

usually due to irremediable central nervous or
trunk injury (3450%);
early (< 4 h), usually from blood loss or airway
problems; this period includes the so-called golden
hour, in which patients may gain particular benefit
from circulatory and respiratory resuscitation in the
prehospital phase (1830%);

hospital deaths within 24 hours (21%; Reinfurt et al.,

1978);
deaths after days or weeks in hospital, often from
multiple system failure associated with sepsis or
from brain damage not severe enough to be
immediately fatal (2027%).
However, a more recent Scottish study (Wyatt et al.,

1995) has reported a lower (7%) proportion of early
deaths, possibly reflecting improved early trauma
care.
15.2
Management at the accident site
The care of the trauma patient involves an escalating

series of health-care workers, beginning at the accident site where bystanders may give first aid until
ambulance officers, including paramedics, are able to
institute more advanced measures.
The most immediate life-threatening effects of
injury are those that impair delivery of oxygen to
tissues and the most vulnerable patients are those
with impaired consciousness who are unable to
protect their airway. The first persons on the scene
may be passers-by who have only their wits, hands
and lungs with which to apply the basic principles of
resuscitation encapsulated in the acronym Dr ABC
where Dr = danger remove and A, B, C = airway,
breathing and circulation. That is:
remove the patient from a hazardous situation;
and establish

airway care: prevent upper airway obstruction and

lung soiling;
breathing: restore breathing by artificial respiration;
circulation: perform cardiopulmonary resuscitation
in the event of circulatory arrest.
272
15.2.1
FROM ACCIDENT SITE TO TRAUMA CENTER
THE RISK OF SPINAL INJURY

A patient with a head injury who has lost consciousness must be assumed to have a spinal injury until
proved otherwise (Albin, 1984; Wilder, 1984; Swain,
Dove and Baker, 1991). In a conscious patient the
presence of neck pain, paresthesia or poor head
control signifies the possibility of spinal injury.
In 23% of cervical spinal fractures, diagnosis was
delayed on admission to hospital (Reid et al., 1987)
and 10% developed neurological deficits while under
medical care (Rogers, 1957; Scher, 1977).
A rigid cervical collar and spinal board should be
attached to all patients who may have a spinal injury
as described above. The collar prevents uncontrolled
neck movement and the spinal support board or frame
gives rigid support to the thoracic and lumbar spines
(Grant, 1975). Emergency personnel should use these
aids before extracting patients whenever the mechanism of injury suggests that a spinal injury is possible
e.g. in all vehicular accidents and falls from greater
than standing height.
It may be necessary to turn a patient on to the side
to prevent aspiration from vomitus. Head rotation
independent of the body increases the risk of spinal
cord injury and a log-rolling technique, keeping the
head supported in the same relative position to the
trunk, should be used for turning. This principle
applies at all stages of care until spinal injury is
excluded by adequate radiographs (section 15.7.1).
15.2.2
ASSESSMENT ON SITE
The first clinical assessment (see also section 15.5.3) is

performed by emergency medical personnel often
while the patient is still in a vehicle. This assessment
will be necessarily brief and incomplete, because of
limited access, poor light and intrusive noise. However, a limited assessment of conscious state, limb
movement and sensation, respiratory status, sites of
pain, obvious fractures, pulse and other basic circulatory assessment can be achieved. This will be the
first of many assessments, repeated in the ambulance
during transit to hospital and regularly thereafter.
Continued reassessment is the keystone of successful
management, enabling the early detection of often
subtle signs of deterioration (Figure 15.1).
15.2.3 EXTRICATION FROM A VEHICLE. AND

DEPLOYMENT OF RESCUE SERVICES
Management at an accident site often involves other

emergency services, such as the fire department and
police. Hence the rescue team may involve:
the ambulance service for resuscitation skills;

the fire department for fire prevention and
extrication equipment and skills (sometimes provided by the police).
It may be necessary to remove parts of the vehicle to

create space around the victim and to minimize
further injury during extrication. Whenever possible,
extrication should be undertaken in a controlled and
unhurried fashion. However, if there is an acute fire
risk extrication may need to proceed rapidly and at
some risk to the patient. Even in such an urgent
situation, attention should be given, as far as possible,
to keeping the head and trunk immobilized and
aligned (Grant, 1975). The rescue squad should comprise personnel with skills appropriate to a particular
situation. They will range from a standard ambulance
crew to paramedics or a full disaster medical team.
The medical personnel at a major accident sites may
need to assist in difficult extrications or to amputate
entrapped limbs. They should be familiar with the
safety requirements at accident sites, such as protective clothing and headgear, and the defensive
positioning of emergency vehicles.
15.2.4
TIME SPENT AT THE SCENE
This should be kept to the minimum necessary to

establish airway control and adequate oxygenation,
control bleeding and splint fractures. Transport should
not be delayed by protracted efforts at resuscitation
(Chapter 17). Bickell et al. (1994) have cast doubt on
the value of routinely giving i.v. fluids at the accident
site after penetrating trauma in an urban environment,
since the amount of crystalloid that can be given via a
conventional i.v. line is small relative to likely blood
losses and vigorous supply of crystalloids may promote further hemorrhage.
Analgesia (gaseous or parenteral) should be given
according to emergency service protocols.
In cold climates and following partial immersion in
water, further heat loss should be prevented. Hypothermia (core temperature less than 33C) may
decrease consciousness and shivering will increase
oxygen requirements.
15.3 Transportation to and between

hospitals
Prehospital care and transport (Table 15.1), especially
in urban areas, is usually managed by ambulance
officers (emergency medical technicians, EMT). Those
with the highest skill levels have the title paramedic.
Primary transport (from the accident site) follows
stabilization and loading without delay. Some
THE LEVEL 1 TRAUMA CENTER
Table 15.1 Terminology applicable to movement of

patients to or between hospitals (this order of transfer is
likely to apply in rural situations; on occasion a
subsequent or tertiary move may be required)
Primary care
Performed at the accident site
Primary hospital
The hospital of first care
Primary transport
Movement to the first, usually

nearest, hospital
Secondary transport
Movement to a secondary center

with greater resources for
definitive care
patients, especially in movement from (often remote)

rural areas, may require specialist intensive care/
critical care teams for secondary (interhospital) transport (Gentleman, 1993; Gilligan et al., 1977).
The integrated escalation of care beginning at the
accident site continues to the neurosurgical and other
services located in a major trauma center. A level 1
trauma center must provide 24-hour on-site specialist
resources for injuries affecting all areas of the body,
including head, chest, abdominal, spinal, limbs and
craniofacial regions (Adams Cowley, 1984).
In large cities, emergency services usually deliver
trauma patients directly to the major medical centers.
In rural areas, patients may be taken at first to small
hospitals, where local physicians need to initiate triage
and emergency management without the support of
surgeons and trained trauma personnel.
When injuries occur in rural areas where there are
limited medical resources, the initial care, stabilization
and subsequent transfer to a definitive center can
become a complex undertaking. In general, 1015% of
trauma patients in rural areas will need to be transferred to a major trauma center (Adams Cowley, 1984).
Effective planning and communications has the
following requirements.
An emergency phone access to trauma centers with

24-hour specialist availability enables rural MOs to
seek advice on emergency management and to
request assistance, as small centers may have
limited resources and not possess critical care
transport expertise. In addition, long-distance transmission of radiological data such as CT scans
(teleradiology) may enable expert opinion to be
delivered to a small rural center.
Ambulance radio should be used in transit to alert
the trauma center of imminent arrival of an injured
patient (Adams Cowley, 1984).
The major trauma center should be in ready communication with the hospital of first contact and should
have access to a specialized aeromedical service able
to deploy fixed- and rotary-wing aircraft according to
273
distance and geography (Moylan, 1988; Gilligan et al.,

1977). Air medical transport of the critically injured
patient requires staff with a wide range of resuscitative skills, in particular intubation, cricothyrotomy,
tracheostomy or other airway access, artificial ventilation, infusion of narcotics and volume expanders,
pleural drainage, blood transfusion, measurement of
intra-arterial and central venous pressures, and urinary catheterization. Sometimes long distances must
be covered. For example, in Australia it may take over
10 hours traveling from some remote areas to reach a
major trauma center.
15.4
The Level 1 trauma center
To use community health resources most effectively,

the major trauma center should serve a large population in practice this places it in a major city.
Locating all necessary facilities on one campus
avoids the need for further transport between centers
and fragmentation of care.
The relevant specialties within the major trauma
center neurosurgical and other surgical disciplines,
intensivists and anesthetists should be integrated in
order to provide a standardized and cohesive
approach to multiple trauma (Adams Cowley, 1984).
15.4.1
THE TRAUMA TEAM
The first phase of management within the trauma

center should depend upon a defined trauma team.
The members of the trauma team must have a strong
commitment to acute trauma management. The following is a standard team structure.
Medical officers (MOs):

team coordinator or supervisor, ideally a surgeon;
intensive care or anesthesia specialists;
accident and emergency physicians;
other specialties as required (neurosurgical,
orthopedic, general surgical); these should be
either consultant staff in-house or, in their
absence, senior registrars or residents.
Nursing staff (registered nurses, RNs):
three experienced nurses, usually from accident
and emergency or ICU, should be available on
call.
15.4.2
DIVISION OF TASKS
The division of tasks within the trauma team needs to

be clearly established. For example, as follows.

Overall assessment, integration of operations: MO.

Respiratory resuscitation: RN and MO.
Circulatory resuscitation (i.v. lines) and monitoring:
RN and MO.
Recording/procedures: RN and MO.
274
Figure 15.1
Service.)
Ambulance officers case report form. (Source: reproduced by courtesy of the South Australian Ambulance
THE LEVEL 1 TRAUMA CENTER
275
276
Table 15.2 Suggested criteria for assembly of the trauma

team (Source: after American College of Surgeons
Committee on Trauma, 1988)
Shock
Pale sweaty skin; empty veins
BP(systolic) < 90 mmHg
Pulse < 50 or > 130 beats per minute
(ii) fall: height of fall

(iii) projectile injuries: type and caliber of
weapon
(c) Environment: temperature (cold or hot), presence of fire or smoke (which may cause carbon
monoxide poisoning)
(d) Other persons injured or dead.
2. Neurological state: Essential information, which
may be obtained even from untrained bystanders,
includes loss of consciousness, amnesia, fitting,
limb movement, sensation and the location of any
pain.
3. Previous health and medications: see section
15.8.
Neurological abnormality
Diminished conscious state
Limb paresthesia or weakness
Spinal injury
The ambulance service report form should be

designed to allow this information to be recorded
easily and clearly (Figure 15.1).
Other injury
Crush injury of the legs or trunk
Amputation or ischemia or a limb
Severe faciomaxillary injury
Moderately severe injury to two or more areas of: head,
neck, chest, abdomen, pelvis, back or femur
Multiple fractures of long bones or the pelvis
(b)
These are based on either historical or examination

criteria.
Examination criteria
Respiratory distress
Airway obstruction (partial or complete)
Respiratory rate < 9 or > 30 breaths per minute
Irregular or shallow respiration
Complaint of breathing difficulty
Historical criteria
Vehicular accident involving:
impact of more than 60 km/h
a death
ejection from a vehicle
severe vehicular damage
Pedestrian or cyclist: impact at a speed of more than
30 km/h
Fall: more than 5 meters
Standard precautions, such as gloves, protective glasses, gowns and other protective clothing should be
used to minimize the risk of accidental transmission of
contagious disease to the health worker.
This sequence for examination and management is
derived from the ATLS.
The stages are as follows.
Table 15.2 sets out suggested criteria for assembling

the trauma team.
15.5
Care of the patient in the hospital
15.5.1
CLINICAL ASSESSMENT (see also Chapter 8)
(a)
History of the accident
Data collection must begin at the accident site.

Information needed by the hospital staff to accurately
assess the patient with a head injury and plan
definitive care includes the following.
1. Circumstances of the injury
(a) Time, date, location
(b) Mechanism of the accident
(i) vehicular accidents: type of vehicle (bicycle,
car, etc.), speed of travel, position within the
vehicle, restraint devices used, helmet use;
whether the victim was ejected
Examination and management
Primary survey: this is an assessment of the status

of the airway, breathing, circulation and presence of
any neurological disability), as summarized in the
acronym ABCD.
Resuscitation: as described in section 15.4.1, the
trauma team (Table 15.2) requires several physicians
and nurses so that assessment, resuscitation and
other measures can occur concurrently. Radiological
studies should begin with resuscitation, the urgent
films being those of chest, cervical spine and
pelvis.
Secondary survey: this secondary, or ongoing survey, may be supplemented by more detailed radiology such as CT and i.v. contrast studies (Crosby
and Lui, 1990). Peritoneal lavage may also be
indicated at this stage (section 15.6.1).
The relevant specialties are consulted once the primary assessment, resuscitation and secondary survey
have been completed. Any urgent problem, such as a
suspected extradural hematoma, decrees that the
relevant specialty be called immediately. However it is
also necessary to ensure that one specialty does not
surge prematurely into detailed investigations (e.g. a
cranial CT scan) in a patient whose ventilatory and
circulatory status has not yet been stabilized.
The hemodynamic and neurological status must be
recorded at intervals throughout these stages. Only
after resuscitation is established can definitive exam-
CARE OF THE PATIENT IN THE HOSPITAL
ination and specific treatment begin (Adams Cowley,

1984). This represents a reversal of the traditional
sequence of history-taking, detailed physical examination, laboratory tests, X rays, diagnosis and
treatment.
15.5.2

high flow oxygen (12 l/min; Katsurada, 1973);

where necessary endotracheal intubation and intermittent positive pressure ventilation (IPPV; Demling and Reissen, 1990; Chapter 17);
at least two wide-bore intravenous (i.v.) lines; blood
taken for group and match, Hb, coagulation studies,
baseline biochemistry and others as required by
community law, such as blood alcohol;
urinary catheter (see below);
nasogastric tube;
electrocardiographic (ECG) monitoring for detection of serious dysrhythmias;
an arterial line for continuous blood pressure (BP)
display is useful if facilities are available; automatic
non-invasive blood pressure (NIBP) devices may be
unreliable in the shocked patient (Rutten et al., 1986);
pulse oximetry, although this may also be unreliable
in a patient with poor skin perfusion (Clayton et al.,
1991).
Nasogastric intubation will drain gastric content and

decompress the stomach, which may become distended following bag mask ventilation. However the
hazards to be avoided are tracheal placement and
perforation of a fractured cribriform plate. When an
anterior fossa fracture is suspected oral intubation
should be performed.
15.5.3
(a)
PRIMARY SURVEY IN DETAIL
Airway patency
Blockage of the upper airway (from lips to larynx) in

hypoxic and restless head-injured patients may be
caused by the tongue, masseteric spasm, vomitus or
blood. The Guedel airway aims to keep the teeth apart
and allow the passage of air above the tongue, but it is
seldom adequate for more than the short term.
Endotracheal intubation may be required.
(b)
Table 15.3 Forms of artificial (controlled) ventilation

This may take the following forms according to
circumstances:
Mouth-to-mouth or mouth-to-face-mask ventilation,

the simplest form of artificial ventilation.
Bag-mask ventilation (usually following placement of a

Guedel airway).
Laryngeal mask ventilation: this can be used as a

short-term technique when intubation may not be
possible (Grantham, 1994).
Endotracheal intubation, required in those in whom the

above measures cannot secure adequate airway patency
or to facilitate intermittent positive pressure ventilation
(IPPV) when prolonged ventilatory inadequacy is likely.
Intubation is indicated for patients with GCS less than 8
(e.g. inappropriate verbal response, flexion to pain, eye
opening to pain).
Cricothyrotomy, when other maneuvers have failed, is

the simplest way to obtain emergency airway access in
the presence of total airway obstruction, as in some
severe compound face and jaw wounds.
RESUSCITATION AND MONITORING
Resuscitation during the primary survey may require

the following:
Breathing
Controlled ventilation
This is required if the patient is hypoventilating or
apneic. Even relatively mild head trauma may cause
277
transient respiratory arrest, which responds well to

controlled ventilation. Blood gas analysis of PaO2 and
PaCO2 should be undertaken early (Table 15.3)
Endotracheal intubation
Patients with GCS less than 9 commonly require
intubation, usually via the oral route. Nasotracheal
intubation is inadvisable in patients with suspected
anterior cranial fossa fracture because of the possibility of passage of the tube through the cribriform plate.
In restless patients, or those with increased muscle
tone, and airway obstruction, safe endotracheal intubation may require a rapid sequence involving muscle
relaxant and sedative/anesthetic agents (Doolan and
OBrien, 1985; Talucci, Shaikh and Schwab, 1988;
Grande, Barton and Stene, 1988). This is a complex
maneuver requiring adequate resources and skilled
personnel (Table 15.4).
In a patient with suspected cervical spine injury,
neck flexion or extension during intubation should be
minimized to reduce the possibility of cervical vertebral displacement and cord injury. Studies suggest
that all airway procedures produce small amounts of
spinal movement (Aprahamian et al., 1984) but there is
no data to suggest that such movements necessarily
represent an added risk to the spinal cord (Crosby and
Lui, 1990). An experienced extra assistant to maintain
in-line neck stabilization is advisable (Turner, 1989;
Joyce, 1988). Alternatively, the patient may remain in
the cervical collar applied in the field. Fiberoptic
intubation may be considered, but is an elective
278
Table 15.4

Requirements for endotracheal intubation
Drugs*
short-acting i.v. induction agents (e.g. thiopentone,
ketamine)
non-depolarizing (rapid acting) muscle relaxant
suxamethonium
(in some centers rapid-acting depolarizing agents
such as vecuronium are recommended)
Effective oxygen system, for preoxygenation; medical
suction capable of coping with vomitus
Reservoir bag, anesthesia face mask, selection of Guedel
airways
Two laryngoscopes, checked (in case one fails)
Range of endotracheal tubes
Bougies to facilitate placement of the tube
Ties to secure the tube once placed
Alternative techniques in the event of failure to intubate
(e.g. cricothyrotomy equipment, laryngeal mask)
Monitoring: ECG, BP, pulse oximeter, end-tidal CO2
monitor
Stethoscope to check air entry and detect inadvertent
endobronchial intubation
* Medical officers should be aware of the increased

sensivity of shocked patients to i.v. sedative and
anesthetic agents
An adult requires a 28 French gauge tube for effective

blood drainage (Westaby, 1991). Needle thoracostomy,
sometimes advocated for tension pneumothorax,
should be regarded only as a temporary means of
relieving pressure before placing a catheter. It also
carries the risk of puncturing the underlying lung and,
in the event of a misdiagnosis, it may create the very
problem it was meant to relieve.
An open chest wound or sucking pneumothorax
should be managed by formally placing a pleural
drain and then applying an occlusive dressing over
the wound. Sealing the wound prior to this may result
in a tension pneumothorax (Weigelt, 1986).
(c)
Circulation
Initial volume replacement should be with i.v. crystalloid, usually Ringers solution, Hartmanns solution or physiological saline. If hypotension persists,
colloid should be introduced. Vasopressors are rarely
appropriate at this early stage (Chapter 17). Cardiac
arrest at the accident site is usually irreversible.
(d)
technique not suited to situations where blood and

vomitus may rapidly obscure vision. Surgical access
via cricothyrotomy has been advocated, but this may
not be easy: McGill, Clinton and Ruiz (1982) reported
difficulties in as many as 29%. The operator should
use the surgical technique with which he or she is
most familiar.
Disability (neurological)
At this preliminary neurological assessment, an abbreviated form of the Glasgow Coma Scale is recommended, remembered by the acronym A-V-P-U.

Awake?
Voice response?
Painful stimulus response?
Unconscious (unresponsive)?
Preparation for emergency endotracheal intubation

Trauma personnel need to be familiar with the rapidsequence anesthetic technique outlined above. An
assistant familiar with the procedure who may apply
cricoid pressure is essential and when cervical injury
is suspected it is prudent to have a third person to
maintain in-line stabilization of the head and neck.
Following intubation, IPPV should be monitored by
arterial blood gases. A chest X-ray is performed to
confirm optimal tube length and placement, and to
exclude a pneumothorax. Narcotics/sedatives (e.g.
morphine/midazolam), supplemented if necessary by
relaxants, need to be administered intermittently to
facilitate ventilation.
Pleural drainage
Pleural air and/or blood should be drained by
intercostal catheter. Following local anesthesia, a skin
incision and blunt dissection to the pleura are performed, and the catheter is introduced carefully to
avoid damage to the underlying lung by sharp stylets.
(e)
Exposure for assessment
At the accident site, removing clothing for purpose of

assessment may need to be tempered with the
practicalities of heat loss, lack of space for access and
awareness of the patients sensibilities. In the relative
privacy of the hospital, removal of all clothing is
necessary to determine the full extent of injury.
15.5.4
SECONDARY SURVEY IN DETAIL
When the emergency measures outlined above are

completed, and the patient is stable, a more detailed
head-to-toe appraisal is performed, starting at the
head and progressing over the rest of the body. It
includes:

scalp, eyes, ears and mouth;

cranial nerve function;
Glasgow Coma Score plus other neurological
findings;
assessment of spinal cord function.
CARE OF THE PATIENT IN THE HOSPITAL
In persons who have had a period of unconsciousness

the secondary survey may indicate the need for a CT.
Reappraisal should be frequent, in order to determine the progression of any signs and paying special
attention to conscious state.
Intravenous analgesia, fracture splintage, tetanus
immunization and i.v. antibiotic, if these are indicated
for compound fractures, should be initiated.
(a)
Neurological status
The Emergency Room physician should note the

Glasgow Coma Scale (Teasdale and Jennett, 1974), and
conduct a more detailed examination of sensory
status, reflex activity, motor power and coordination.
The neurosurgeon should then have available for
review:

the history of the event;

cardiorespiratory status;
conscious state assessments;
pupils and fundal examination;
lateralized extremity differences;
otoscopic examination.
(b)
Thorax
Observation may reveal the see-saw respiratory

pattern of a high spinal injury or upper airway
obstruction, or a paradoxical segment (flail chest),
usually due to the mechanical instability of extensive
rib-cage injury. A seat belt pattern on the chest
suggests the possibility of lung, heart or visceral
damage. Clinical signs of hemopneumothorax may
be evident. Subcutaneous emphysema may be associated with rib-cage injury or a breach of the bronchial tree. Myocardial contusion and cardiac tamponade may follow blunt chest injury but they are
difficult to differentiate clinically at this stage. Heart
sounds are often soft, and myocardial injury may
need to be excluded by 12-lead ECG, echocardiography or CK myocardial fraction assay. Cardiac
tamponade may require pericardiocentesis. Thoracic
cage impact may rarely cause tricuspid or other
valvular damage and the aorta may be torn, with
widening of the mediastinum and false aneurysm
formation. Aortography should be used when this is
suspected.
The back should be examined after turning with a
log-rolling technique, keeping head and trunk
aligned.
(c)
Abdomen
A comatose patient may suffer serious blunt trauma

to the abdomen with little external evidence One
279
study reported that 16.7% of comatose head injury

patients had significant intra-abdominal bleeding
(Butterworth et al., 1979). Seat-belt bruising may be
indirect evidence of intra-abdominal injury such as
visceral rupture (Cope and Stebbings, 1991). Pain
and tenderness of the abdomen may be modified by
shock and depression of conscious state due to head
injury, alcohol or other recreational substances, or by
drugs given for analgesia or to facilitate artificial
ventilation.
(d)
Pelvis and perineum
The iliac crest and pubes should be palpated and

compressed for mobility indicating an injury of the
pelvis. The penis, scrotum, vulva and anus should be
examined; meatal blood staining arouses suspicion of
a penetrating wound or urethral injury associated
with a pelvic fracture. Complete assessment requires
digital examination of the vagina and rectum, otherwise penetrating wounds may be missed. Priapism
and lax anal tone may be due to spinal cord injury.
With a pelvic fracture, upward displacement of the
prostate may be due to urethral rupture.
(e)
Urinary catheterization
This is necessary for several reasons.

A normal urine output is a major indicator of
volume restoration. Blood-stained urine may indicate
urinary-tract trauma. In patients resuscitated some
time after injury, the brownish urine of methemoglobinuria may indicate muscle damage or crush syndrome. Relieving a full bladder may reduce restlessness in the obtunded patient, and a catheter is
required before extended transportation.
However, catheterization may cause increased
trauma in a patient with urethral rupture associated
with a pelvic fracture: and therefore rectal examination should precede urethral catheterization. If there
are signs suggesting urethral injury, catheterization
should not proceed until the tract has been defined
with a urethrogram. In the event of urethral rupture,
suprapubic cystotomy may be necessary. In females
urethral damage with pelvic fracture is much less
common.
(f)
Thermoregulation
Hypothermia, detected by core temperature measurement, may contribute to a depressed conscious state.
Conversely a high environmental temperature and
excessive clothing can potentiate hyperthermia. A core
temperature of more than 40C can alter consciousness
and cause cerebral damage (Hamilton, 1976; Eichler,
McFee and Root, 1969).
280
(g)
Laboratory data
Baseline data should be reviewed early to identify any

other factors that may affect consciousness. These
include a high blood alcohol, unsuspected diabetes
mellitus, pre-existent renal failure or other metabolic
disorders. Abnormal coagulation studies (INR) may
be due to liver disease or anticoagulant medication.
15.6
Setting surgical priorities
Once circulatory and respiratory resuscitation have

been achieved, continued bleeding may require surgical intervention. At this point the leader of the
trauma team must allocate priorities between competing needs, for example in a patient with neurosurgical, abdominal and limb injuries. Diagnostic
peritoneal lavage or an abdominal CT scan is often
mandatory in the unconscious patient before deciding on laparotomy.
15.6.1
EMERGENCY LAPAROTOMY
Some authors advocate a low diagnostic threshold

for performing a laparotomy, especially in patients
who are intubated and on controlled ventilation and
sedation/relaxation (Hill, Schecter and Trunkey,
1988). In such patients, the risk factors associated
with induction of anesthesia have already been overcome and there is little additional risk incurred by
exploration. Definitive knowledge of the abdominal
status may be helpful, especially if craniotomy or
extended orthopedic procedures are planned in the
immediate future.
15.6.2 FURTHER INVESTIGATION OF THE ABDOMEN:
PERITONEAL LAVAGE
Peritoneal lavage is useful in determining the likelihood of major abdominal bleeding but it is probably
best performed at the center where there are resources
for immediate laparotomy if it is positive.
Indications for peritoneal lavage (Cope and Stebbings, 1991) include:

depressed conscious state (head injury, drugs, etc.)

causing difficulty in assessment;
equivocal clinical findings on abdominal examination;
multiple injuries, especially of chest, pelvis or
spinal cord;
persistent hypotension despite apparently adequate
fluid replacement;
stab wounds with a breach of the peritoneum.
Lavage is unnecessary when laparotomy is clearly

indicated on clinical grounds, as in a shocked patient
with progressive abdominal distension.
(a)
Technique of peritoneal lavage
A nasogastric tube and urine catheter are first placed,

according to the trauma management protocol. After
skin preparation and local anesthesia, a subumbilical
5 cm vertical incision is made in the midline. The
layers are identified and the linea alba is divided. The
peritoneum is identified and incised and a peritoneal
dialysis catheter without guide wire is introduced and
directed towards the pelvis. Any free fluid is aspirated. More than 5 ml of blood is regarded as an
indication for laparotomy. If blood is not aspirated, the
catheter is connected to an i.v. giving set, primed with
sterile physiological saline or Hartmanns solution
(37C); 1 liter is rapidly instilled, left for 3 minutes
then allowed to drain back into the now dependent i.v.
pack. Of this, 20 ml is used for laboratory assessment
(cell counts, microbiological and other examination).
(b)
Laparotomy after positive lavage
Positive lavage findings that indicate the need for

laparotomy are:

enteric fluid;
> 5 ml blood;
RBC > 100 000/mm3;
bacteria (Cope and Stebbings, 1991).
Laparotomy is also often undertaken because of:

unexplained shock;
a rigid silent abdomen;
abdominal penetrating wounds;
evisceration;
X-ray evidence of free intraperitoneal gas or ruptured diaphragm.
15.6.3 FURTHER INVESTIGATION IN MULTIPLE

TRAUMA
Once the patient is stable, special radiological investigation such as CT of head, chest or abdomen may be
needed. Many accident and emergency departments
have a CT scanner on site, thus avoiding the need for
the patient to be transported considerable distances
within the hospital.
15.6.4
INTRAHOSPITAL TRANSPORTATION
An investigation may take minutes only, but transporting and positioning a patient with attached
apparatus on the scanning platform may be time
consuming and risky.
If blood pressure can only be held stable by
extensive ongoing blood volume replacement, the
patient may decompensate suddenly and transport for
investigations may be ill-advised.
EFFECTS OF CONCURRENT MEDICAL CONDITIONS
Full monitoring should continue during transport

and investigations and the patient should be under the
supervision of senior medical and nursing staff at all
times.
15.6.5
(a)
SPECIFIC INVESTIGATIONS
The trunk
Abdominal injury
The CT scan may identify both qualitative and
quantitative evidence of visceral bleeding. However,
in considering the likelihood and significance of intraabdominal bleeding, the clinician needs to take into
account other signs such as the rate of i.v. volume
replacement required for circulatory stability and the
absence of other evidence of blood loss (e.g. from limb
fractures).
MRI scanning has no place in the evaluation of
acute abdominal trauma at present. A major limitation
is the need to avoid using devices containing ferromagnetic metal, such as electric motors and metal
attachments on resuscitation instrumentation, in the
vicinity of the MRI magnet system.
Thoracic injury
Major injury, e.g. to the aorta, may require urgent
radiological investigation.
(b)
The limbs
Neurosurgical emergencies such as an intracranial

hematoma in an unconscious or deteriorating patient,
clearly take priority over limb injuries. Although
important, these injuries are not usually immediate
threats to life. Bony injury may cause blood loss,
which is replaceable; deformity, which is controllable
by splintage; and soiled wounds, which can be
temporarily covered. However, rapid intervention
may be required for compartment syndromes and
some vascular injuries.
A lateral cervical spine X-ray does not absolutely

exclude a fracture. Display of C7 is essential, aided if
necessary by shoulder traction and a swimmers view.
Flexion and extension views may be needed and CT
scanning may be required later (Swain, Dove and
Baker, 1991).
The chest film may reveal a number of immediate
problems, such as rib-cage and thoracic spinal bony
injury, lung collapse, contusion, hemopneumothorax
and other barotrauma. Mediastinal widening from
hematoma may indicate aortic injury. The cardiac
outline may show chronic cardiac enlargement or
tamponade. Abnormality of diaphragmatic contour
may suggest rupture of a hemidiaphragm, phrenic
nerve lesion or hepatic hematoma. Chronic lung
conditions may also be noted.
The pelvic film, especially in the unconscious
patient in whom bony tenderness cannot be elicited,
may reveal fractures and the potential for major blood
loss.
15.7.2
15.8 The effects of concurrent medical

conditions
Pre-existent illness and medications (Tables 15.5, 15.6)
can adversely affect the response to injury and thus
Table 15.5 Some pre-existent conditions that can affect the

response to trauma (Milzman, Hinson and Magnant, 1993)
Radiological examination
As outlined above, this is conducted in two phases:

15.7.1
EMERGENCY STUDIES
After the primary survey, radiological examination

should be confined to:

lateral cervical spine, including C7 vertebra.;

AP chest;
pelvic views (Swain, Dove and Baker, 1991).
LATER STUDIES
These are conducted when circulation and respiration

are stable and may include CT scans of the head,
cervical spine, chest or abdomen.
Contrast studies, e.g. of the urinary tract, should
also be performed after ventilation, airway status and
volume restoration are under control, and when there
is no immediate need for laparotomy. Suspected limb
fractures may now be X-rayed, to complete the
orthopedic assessment.
15.7
281

Cardiac disease
Hypertension
Central nervous system disease, e.g. CVA, epilepsy
(Zwimpfer and Moulton, 1993)
Chronic pulmonary disease
Diabetes
Obesity
Immunological disorders, AIDS
Neoplastic conditions
Hematological disorders/coagulopathies
Epilepsy
Chronic renal disease
Chronic liver disease
Pregnancy
Allergy
282
Table 15.6 Common drugs and substances that may

affect the response to trauma (Milzman, Hinson and
Magnant, 1993)
CNS
sedatives, recreational substances (alcohol, narcotics,
hallucinogenic, stimulants and other substances
causing excitatory phenomena), carbon monoxide
poisoning (smoke inhalation)
anticonvulsants, antidepressants (the effect of
anesthetic agents such as halothane on the cerebral
vasculature will be considered elsewhere)
Hematological
anticoagulants and non-steroidal anti-inflammatory
agents such as aspirin (platelet effect) predispose to
bleeding
Cardiovascular
beta blockers (affect myocardial response to blood
loss), other hypotensives (ACE inhibitors and other
vasodilators) may reduce the effectiveness of the
vasomotor response
Urinary tract
diuretics (reduced blood volume; cause K+
deficiency).
Endocrine
prolonged steroid use may impair response to stress
insulin given for diabetic management may have
adversely affected blood sugar level
affect management (MacKenzie, Morris and Edelsten,

1989). Carers should take particular note of medical
emergency bracelets indicating chronic disorders. The
acute physiological and chronic health evaluation
score, version II (APACHE II; Knaus, Draper and
Wagner, 1985) is widely used to quantify the degree of
disturbance due to critical illness, but it underscores
severe trauma (Vassar et al., 1992). Increasing injury
severity score (ISS) and older age are major predictors
of mortality (Milzman, Hinson and Magnant, 1993).
15.8.1
CARDIOVASCULAR
Hypertension prolongs hospital stay (MacKenzie,

Morris and Edelsten, 1989), but previous coronary
artery disease and myocardial infarction increase risk
for mortality, particularly when surgery is required
(Goldman et al., 1977; American College of Surgeons
Committee on Trauma Management, 1988). Impaired
cardiac reserve may cause intolerance to sudden blood
loss and vigorous fluid loading. The use of pulmonary
capillary wedge pressure measurement to guide fluid
restoration has been urged in cardiac patients (Daranko, Klossner and Laaksoned, 1987). If inotropes are
required in the acute stages, catecholamine infusions
with a short half-life, such as adrenalin, are appropriate. Chronic anemia can reduce tissue oxygen
delivery in blood loss.
15.8.2
NEUROLOGICAL
Cerebrovascular disease, epilepsy and some psychiatric conditions (Drubach, Kelly and Dolif, 1994) may
occasionally be responsible for traumatic brain injury.
On the other hand, pre-existing neurological deficits
such as hemiplegia or a postictal state may be
confused with the acute effects of trauma.
15.8.3
RESPIRATORY
Asthma and chronic obstructive pulmonary disease

(COPD) complicate respiratory management, especially in patients with impaired consciousness.
Hypoxia occurs more readily, and chronically produced sputum may predispose to atelectasis (OBrien
and Criner, 1994), These effects may be greater in the
obtunded patient and with rib pain. Advancing age,
smoking and obesity compound respiratory problems
(Grande et al., 1991).
15.8.4
METABOLIC
Diabetics are more often involved in accidents (Rodriguez et al., 1992). Stress in a diabetic may produce
hyperglycemia, leading to osmotic diuresis, acidbase
disturbance and impaired consciousness. Temporary
stress hyperglycemia may also occur in non-diabetics
or prediabetics. Conversely, insulin may cause hypoglycemia, which also affects the conscious state.
Diabetic peripheral vascular disease may predispose
to wound sepsis. Obesity may predispose to falls
(Boulanger, Milzman and Rodriguez, 1994), make
endotracheal intubation, tracheostomy and placement
of vascular catheters more difficult and increase
postsurgical and anesthesia complications.
15.8.5
IMMUNOLOGICAL AND NUTRITIONAL
Impaired immunological responses may predispose to

sepsis; HIV infection, opportunistic infection and
coagulation deficits (Esposito, Aprahamian and
Campbell, 1993). Neoplasia may be associated with
impaired nutrition and susceptibility to fractures
involving metastatic deposits. Allergies to antibiotics,
may precipitate anaphylactic or anaphylactoid (first
time) reactions (Fisher, 1987).
15.8.6
SKELETAL
Cervical spondylosis may predispose to spinal injury

following impact.
15.8.7
RENAL
Chronic renal impairment requiring a sustained fluid

intake increases the mortality from trauma (Milzman,
TRANSPORT MODALITIES
Hinson and Magnant, 1993) since decreased renal

perfusion in acute hypovolemia may convert chronic
insufficiency to acute renal failure. Hourly urine
volumes are not an adequate marker of renal function
in such conditions and close biochemical and hemodynamic monitoring is essential.
15.8.8
DRUGS AND ALCOHOL
Misuse of narcotics, recreational drugs (such as

alcohol, hallucinogens and marijuana) and other nontherapeutic substances (Table 15.6) may predispose
the patient to trauma and complicate assessment of
consciousness. Furthermore, chronic misusers may
have an impaired response to injury through malnutrition, pulmonary disease, AIDS or chronic hepatitis
(the latter two constituting an infection hazard to
rescuers). Illicit substance use has been reported in
over 40% of head injury victims (Milzman et al.,
1992b). Alcohol remains the commonest substance
encountered. As well as the complicating effects of
acute intoxication, long term abuse may produce
chronic liver disturbance with coagulation defects and
this group of patients is prone to acute subdural
hematoma following relatively minor trauma. Drug
and alcohol withdrawal may complicate management
in the first few days.
15.8.9
CARBON MONOXIDE
Carbon monoxide poisoning from smoke inhalation

may impair conscious state (Myers, Snyder and
Emhoff, 1985) and complicate the assessment of the
effects of trauma.
15.8.10
CONGENITAL COAGULOPATHIES
Congenital coagulopathies (e.g. hemophilia) may

increase mortality, unless the disorder is controlled
with antihemophiliac globulin (factor VIII; Milzman,
Hinson and Magnant, 1993).
15.8.11
PREGNANCY
Pregnancy renders the patient and the fetus vulnerable to head and other injury (Barrett, 1984), especially
near term. Blood volume in pregnancy is increased by
50% and the cardiac output by 1.5 l/min. The hematocrit drops from 40% to about 32%. Hence hypovolemia may reduce the placental flow and deprive the
fetus. In late pregnancy there is relatively little
amniotic fluid to protect the fetus. Lax muscles give
less protection against blunt abdominal trauma. In late
pregnancy the patient should be transported on her
side since uterine compression of the superior vena
cava may produce the supine hypotensive syndrome
283
(Mighty, 1994). Minute ventilation rises about 40% as a

result of increased metabolic activity; and this needs to
be taken into account in setting the parameters for
artificial ventilation. Gastric emptying is delayed,
predisposing to inhalation of gastric content in the
obtunded state, affecting anesthesia and resuscitation.
Medication taken during pregnancy may also affect
management (Table 15.5).
15.9
Transport modalities
Depending on distances and geographic conditions,

helicopter, fixed wing aircraft or road ambulance may
be used (Figures 15.215.5; Table 15.7). Helicopter
services usually operate within a radius of about
200 km. Primary (from the site) evacuation by helicopter to a trauma center is available in some
widespread urban areas and large cities, such as
London. In rural areas helicopters are frequently used
for secondary (interhospital) evacuation. Beyond
200 km fixed-wing aircraft are usually employed.
15.9.1
AEROMEDICAL STAFF
An intensive care transport or retrieval team may be

provided by the trauma center or by a separate
service. Medical and nursing staff from intensive care,
anesthesia or emergency medicine are often used. A
minimum of two experienced persons is required.
Various combinations are may be used medical
officer/paramedic; paramedic/nurse; medical officer/
nurse. However the greatest skill range is likely to be
found with the medical officer/nurse combination,
and this is particularly important when the team must
work with the staff of a rural hospital in providing
resuscitation, surgery and anesthesia for emergency
procedures.
15.9.2
AIRBORNE NEUROSURGICAL SERVICES
Acute intracranial hematomas usually require rapid

surgical drainage and the time lag to treatment for
such patients in remote areas is a major concern
(Simpson et al., 1984). As an alternative to rapid
delivery of a deteriorating patient to a trauma center,
a neurosurgeon may be included in an airborne
medical team. The local MO can undertake the
immediate burr hole drainage before the team arrives,
and the neurosurgeon can complete the operation. In
such a situation, the decision to operate on site or after
transfer to the trauma center needs to be made in
consultation with the neurosurgical service at the
trauma center. In general, when the patient is deteriorating and transport time to the trauma center is likely
to be 2 hours or more, a neurosurgeon should be
included in the team (Simpson et al., 1988).
284
Figure 15.2 Bell 412 helicopter for emergency medical use: a twin-engine, four-blade rotary-wing craft with a cruising
speed of 210 km/h, capable of carrying two critically injured patients on ventilators or four stretcher patients in mass
casualty situations. The medical team is up to three persons. The common radius of operations is up to 210 km. The aircraft
has an endurance of 312hours and has instrument flight rules (IFR) and global positioning system (GPS) navigational aids
with single or dual pilot missions. (Source: reproduced by courtesy of SGIC State Rescue Helicopter Service and the Sunday
Mail newspaper, South Australia.)
Figure 15.3 Beechcraft KingAir B200 C series air ambulance operated by the Royal Flying Doctor Service. This twin-engine,
pressurized aircraft can carry two critically injured ventilator patients and has full monitoring facilities, DC and AC power,
a wide cargo door and inboard loading system to facilitate stretcher handling at remote locations. Common radius of
operations is up to 1700 km at a speed of around 450 km/h. (Source: reproduced by courtesy of the Royal Flying Doctor
Service, Central Sector, South Australia.)
285
Figure 15.4 Interior of jet air ambulance (BAe 125800)

capable of operating two ventilator patients with a range of
4500 km without refueling at a speed of approximately
500 mph (800 km/h). A medical team of up to three persons
is carried. This patient is ventilated, with multiple infusions
in progress and with ECG, intra-arterial BP and pulse
oximetry monitoring. Such aircraft are commonly used for
missions in the range beyond 1800 km, where their superior
speed is advantageous. (Source: reproduced by courtesy of
National Jet Systems, Adelaide, South Australia.)
Table 15.7 Transport modalities used in transporting

critically ill trauma patients, South Australia,19841995.
(Included are 28 patients in whom craniotomy for
extradural hematoma was performed in a rural hospital
before transport)
Modality
No.
Average distance
(km)
Road ambulance
223
56
Helicopter
254
105
Fixed-wing aircraft:
Turbo prop.
Executive jet
Airliner
559
18
18
353
2620
8700
Total
1072
Figure 15.5 Interior of Boeing 747 showing stretcher accommodation. With sufficient space (up to 15 seats) a stretcher
patient can be carried on intercontinental repatriation missions, with provision of oxygen, ventilation and monitoring
equipment. Airlines require medical teams to be fully independent and on long missions, which may take up 40 hours
hospital to hospital, a medical team of three may be needed.
286
15.9.3
PREPARATION FOR TRANSPORT
The patient should be stabilized before transport by

restoring blood volume, controlling ventilation and
establishing pleural drainage. Controlled ventilation
may be indicated by impending ventilatory insufficiency or airway obstruction (Gentleman, 1993), and
also by restlessness or uncontrollable behavior which,
in an aircraft, may be hazardous to the patient and to
others. Control of biochemical disorders, especially of
K+, blood sugar, body temperature and polyuria (due
to diabetes insipidus or hyperglycemia) should be
initiated before extended trips. If the patient is
effectively stabilized, the only in-flight measures need
be continued volume restoration and analgesia.
There should be two or more intravenous lines. If
central venous access (e.g. subclavian or internal
jugular) is used, radiological checking before an
extended trip is prudent. Adequate supplies of blood,
plasma expanders and other fluids for the trip are
obtained. Continuous BP display via arterial transducer is the most convenient monitor of the circulation. If an intracranial pressure device has been
inserted, ICP should be displayed during transport on
a multiple-channel pressure monitor. In general,
patients who are intubated and ventilated for
transport are moderately hyperventilated (PaCO2 =
2832 mmHg). If signs develop that suggest an
expanding mass lesion then mannitol may be given.
Mannitol should not be given routinely since it may
interfere with volume resuscitation and also mask the
signs of a mass lesion. A urinary catheter is required
(Table 15.8).
(a)
Analgesia
Analgesics by small i.v. increments or continuous i.v.

infusion give better control than i.m. doses. Gaseous
analgesia (N2O) given over a long period may cause
encysted gas spaces to enlarge through a counterdiffusion effect (Owen, 1978).
Regional analgesia is useful for limb fractures.
During prolonged air transport, fractured limbs may
swell within recently applied plasters, especially if the

limbs are dependent, and so plasters are often split
beforehand. In obtunded patients ischemic limb pain
will not be evident; hence observation of the distal
circulation is mandatory.
15.9.4 DETERIORATION IN CONSCIOUS STATE IN
TRANSIT
This may be from complications of the injury such as

intracranial bleeding, pneumothorax or uncontrolled
abdominal or other bleeding. However, a deterioration in conscious state should always prompt a rapid
check for technical causes such as ventilator disconnection or failure of the oxygen supply. If intracranial bleeding is considered likely, then hyperventilation and mannitol may be used.
15.9.5
Copies of the original records should be accompany

the patient. Observations must continue during
transit. In prolonged transport crossing various time
zones, charting should be based on destination time,
so that a continuous time sequence is maintained.
15.9.6
Table 15.9 Aircraft requirements for aeromedical transport

(Source: reproduced from Oh, 1990, with permission)
Table 15.8 Major interventions with critically injured

rural patients prior to transport to trauma centers, South
Australia, 19841995 (n = 1072)
Intervention
Venous cannulation
Central venous cannulation
Blood transfusion
Intubation, controlled ventilation
Pleural drainage
Surgery pretransport
Craniotomy
1072
159
160
463
123
54
29
(100)
(15)
(15)
(43)
(12)
(5)
(2.5)
SELECTION OF AIRCRAFT
While road transport may be used in short trips and

when circumstances such as adverse weather prevent
flight, aircraft play a major role in medical transport.
The requirements of any medical transport vehicle
include aspects of safety, space, fittings and communications (Table 15.9). The experience in South Australia is based on serving a population dispersed over
large distances (Figure 15.6).
n (%)
DOCUMENTATION AND CHARTING

Adequate safety
No abrupt movements in any axis
Sufficient room for at least one critically ill patient with
an attendant at the head end
Adequate supply of energy and gases for life support
system
Easy embarkation and disembarkation of the patient
Adequate lighting and climate control
Tolerable noise and vibration levels
Adequate speed
Minimal need for secondary transport (e.g. road
transport in air evacuation)
Good communication systems
Preferably pressurization to sea level pressure (in
fixed-wing aircraft)
Figure 15.6
19841995.
The number and radii of operations using fixed-wing and helicopter transport in South Australia from
Case history
Even in remote areas, effective resuscitation and maintenance, coupled with rapid availability of jet aircraft and
mobile medical teams, can achieve good results.
A 5-year-old child suffered a closed head injury from a
crushing accident. He required emergency intubation and
controlled ventilation in the local hospital because of
impaired consciousness. The nearest neurosurgical unit was
2600 km distant. Pending arrival of an ambulance jet
aircraft and also during the 4 hour flight to Adelaide he was
maintained on IPPV. He made an uneventful recovery.
(a)
287
Rotary-wing aircraft
Helicopters have the advantage of flexibility, enabling

landing in proximity to an incident or a hospital
(Figure 15.6). Their relative slowness (usually less than
200 km/h) and short range is usually acceptable
within the usual radii of operation (up to 200 km).
Some helicopters lack adequate space for patient care
and generate excessive noise and vibration.
Case history
A 23-year-old man was struck on the head by a cricket ball
in a town 200 km from Adelaide, South Australia. Shortly
afterwards he lost consciousness, fitted and developed a
dilated pupil. The LMO, under telephone instructions from
ICU and neurosurgical consultants, instituted anesthesia
and controlled ventilation and performed a burr hole to
drain an extradural clot while a retrieval team with a

neurosurgeon flew by helicopter to the town. The neurosurgeon completed the operation and the patient, on controlled
ventilation, was transported by the helicopter to Adelaide
for further care and recovery.
(b)
Fixed wing aircraft
These are used over longer distances, e.g. over 200 km.
Turbo prop aircraft such as the King Air (Figure 15.3)
are commonly used because of their reliability and
economy (Gilligan et al., 1996). Beyond 1200 km jet
aircraft become cost-efficient. Regular passenger airliners may be used, in which 1215 seats may be
required to accommodate a single patient (Figure 15.5;
Table 15.9).
Case history
A 30-year-old soccer player had a head-to-head collision
with another player. He was conscious but dazed and after
15 minutes complained of headache. Initially admitted to a
country hospital, his conscious state was reported as vague.
BP was 120/90 mmHg, pulse 5060/min. He was transferred to a regional center where he was noted to be confused
and had a right hemiparesis. CT demonstrated an extradural hematoma with no skull fracture. Following discussion with the trauma center in Adelaide, mannitol was
given and a retrieval team and neurosurgeon was flown to
the regional center (400 km). An extradural hematoma was
288
relieved by craniotomy and the ventilated patient was then

flown to Adelaide for ongoing care.
With no neurosurgical facility in the rural center, the
most rapid surgical relief was achieved by use of the
airborne team, and the patient was transferred for specialized postoperative management back at the state trauma
center.
(c)
Cabin pressurization
Pressurized aircraft fly at altitudes to 40 000 feet

(12 000 m). This enables flight above most episodes of
adverse weather. Most aircraft are now equipped
with global positioning system (GPS) navigation,
weather radar and other instrumentation for operational safety in marginal weather. Cabin altitude is
usually kept at 50008000 feet (15252600 m; Figure
15.5). Sea-level cabin pressure, however, can be achieved by flying at a lower level (e.g. 20 000 feet/
6000 m). However, this is feasible only in ambulance
aircraft, executive turbo prop or jet, or military
aircraft. The lower altitude is achieved at the expense
of greater fuel usage and sometimes greater turbulence. Commercial airliners are usually required to
maintain certain altitudes by regulation.
Long-distance medical evacuation (medevac) may
sometimes be justified on humanitarian and financial
grounds when convalescence in a foreign country is

likely to be protracted.
Case history
A 58-year-man developed signs of an intracerebral bleed
while working in South-east Asia. An intracerebral clot was
drained by craniotomy in a nearby hospital. He was then
transported by a medical team in an ambulance jet aircraft
10 000 km to a neurosurgical center in Australia for
convalescence.
(d)
The effects of altitude on transport
Reduction in partial pressure of oxygen with altitude

Increasing cabin altitude is associated with a reduction
in the partial pressure of oxygen in the air (Figure
15.7). Thus a cabin altitude of 10 000 feet (3000 m) is
associated with a total air pressure of 523 mmHg and
a partial pressure of oxygen (FIO2 0.21) of 110 mmHg,
versus approximately 160 mmHg at sea level.
Expansion of gas in cavities at altitude
Compliant cavities, with no external drainage, will
expand their contents in accord with the principles of
Figure 15.7 Graph of the relationship of altitude to barometric pressure. Note that at 5000 m (18 000 feet) the atmospheric
pressure is half that at sea level. (Source: reproduced from McNeil, 1983, with permission.)
SUMMARY
Boyles law. Thus an undrained pneumothorax or

closed lung cyst will expand as ambient pressure
reduces and this may result in clinical deterioration.
A pneumothorax of volume 500 ml at sea level will
have a pressure of 760 mmHg. At 10 000 feet (3000 m)
the volume would expand to 706 ml.
P1 V1 = P2 V2
i.e.
760 mmHg 500 ml = 538 mmHg V2 ml

therefore V2 = 760 500/538
= 706 ml.
This is achieved by compression of lung volume and

hence reduced oxygenation. For a spherical cyst, the
radius would increase by approximately 12%. The
volume of the lesion may compress more lung tissue
as ambient pressure in the bronchial tree reduces and
tension pneumothorax may result. A similar mishap
occurs more commonly due to a valve effect from torn
lung parenchyma which allows air into the pleural
cavity during inspiration, but closes during expiration, hence progressively collapsing the lung.
Rigid cavities remain at a constant volume, and
hence the same pressure. However, the ambient
pressure reduces with altitude. Thus, at 10 000 feet
(3000 m) the ambient pressure is 520 mmHg,
240 mmHg less than at sea level. Hence, nasal sinus
obstruction is a common cause of discomfort with
change in altitude. Blood pressure is expressed
in mmHg above ambient pressure so that a mean BP of
90 mmHg is 90 mmHg above the ambient pressure.
Thus a rigid cavity with entrapped gas (e.g. still at
760 mmHg) will undergo less perfusion. At first
venous outflow will be impeded, causing vascular
congestion. Later, as the pressure differential increases, arterial inflow may also be compressed. Thus
discomfort and nasal bleeding may result.
Gases in cavities that communicate with the ambient air (e.g. the bronchial tree, the stomach with a
gastric tube in place or pleura with a functioning
pleural drain) will drain according to the pressure
gradient.
15.9.7 ENTRAPPED INTRACEREBRAL AND
INTRAOCULAR AIR
The possibility of enclosed air (e.g. intraventricular,

subarachnoid or intraocular air) expanding at high
altitude and causing intracranial compression often
causes concern when transporting head-injured
patients by air. After placing intraocular air in experimental animals, intraocular tension rose to about
40 mmHg (Dieckert et al, 1986) at a simulated cabin
altitude of 8000 feet (2600 m). Ocular rupture at altitude has been reported in patients with intraocular air
289
(Colvin, 1980). Although a penetrating wound should

allow equal ingress and egress of air according to
pressure gradients, this may not occur, because of
valve effects. This is particularly likely when intracranial air is present.
Case history
A conscious British soldier was being repatriated by air
from the Falkland Islands, for medical care, after sustaining
a frontal fossa fracture and a CSF leak a week previously. In
mid-Atlantic, he developed severe headache and become
drowsy, then unconscious. The Hercules transport aircraft
made an emergency landing in Miami and the patient
immediately regained normal consciousness. CT scan
revealed about 5 ml of air in the frontal subarachnoid space
and the lateral ventricles. His repatriation was completed
and CSF leak repair was performed 2 weeks later, when
intracranial air was no longer present (R. Bullock, personal
communication).
Encysted air may expand in reduced ambient pressure, but denitrogenation can reduce such effects
(Hyldegaard and Madsen, 1994). When animals were
injected with subcorneal air, 100% oxygen administration, after a brief and insignificant rise in ocular
tension, led to by a much faster reduction in bubble
size than when air was breathed, and bubble disappearance by 200 minutes.
Continuous administration of high flow oxygen
well before (e.g. more than 1 hour) and during a flight,
is normal practice in the injured patient, and should
therefore reduce the risks of expansion of both
intracranial and intraocular air.
Hence it should be possible to transport most headinjured patients by air. However if there is a large
amount of intracranial air, it is prudent to maintain the
minimal achievable cabin altitude.
15.10
Summary
Effective management of the head-injury patient

requires an integrated management plan from the
scene to the neurosurgical center. The escalating
hierarchy of personnel may include first-aid workers,
ambulance staff of various skill levels, critical care
transport teams and medical and nursing staff of both
regional medical centers and Level 1 trauma
institutions.
The priorities of resuscitation and assessment
should be categorized according to the Advanced
Trauma Life Support (ATLS) program. Primary survey
and emergency resuscitation phases merge to ongoing
(secondary) survey.
Determining the probable severity of the head
injury requires knowledge of:
290

the neurological state on repeated assessment;

the factors involved in the accident;
the severity of other injuries;
any previous medical history;
the resuscitative measures required to maintain
stability.
Concurrent medical conditions, especially those of the

respiratory, cardiovascular, renal and central nervous
systems, can adversely affect morbidity and mortality
in all forms of trauma.
Transport teams of various levels of skills may be
required:

standard ambulance teams;

paramedics with specialized trauma skills;
airborne intensive care/critical care teams, commonly a medical officer/nurse combination.
A mobile neurosurgical service may also be required

for urgent decompression of suspected intracranial
hematomas when the patient is more than 2 hours
distant from the trauma center (Simpson et al., 1988).
Safe transport to secondary centers can be achieved
with mobile medical retrieval teams with appropriate
resuscitative skills and equipment. Key safety factors
include effective stabilization of ventilation, blood
volume and thermoregulation initiated before transportation and continuing during transit.
The choice of rotary- versus fixed-wing aircraft is
largely dependent on distance. Helicopter transport
range rarely exceeds 200 km. Fixed-wing aircraft are
best used for distances greater than 150 km.
Level 1 trauma centers require a dedicated in
house trauma team. A major function of the trauma
team is to allot priorities in management.
15.11
References
Adams Cowley, R. (1984) The problem, in Progress in Critical Care Medicine, vol.
1. Multiple Trauma, (ed. J. Wilder), S. Karger, Basel.
Albin, M. S. (1984) Acute spinal cord trauma, in Textbook of Critical Care, (eds
W. C. Shoemaker, W. L. Thompson and P. R. Holbrook), W. B. Saunders,
Philadelphia, PA, pp. 928936.
American College of Surgeons Committee on Trauma Management (1988)
Hospital and Prehospital Resources for Optimal Care of the Injured Patient,
American College of Surgeons, Chicago, IL.
Aprahamian, C., Thompson, B. M., Finger, W. A. and Darin, J. C. (1984)
Experimental cervical spine injury model: examination of airway management and splinting techniques. Annals of Emergency Medicine, 13, 584587.
Australian and New Zealand College of Anesthetists and Australian College
of Emergency Medicine (1992) Joint Statement on Minimum Standards for
Transport of the Critically Ill, Australian and New Zealand College of
Anesthetists, Melbourne, Victoria, p. 23.
Babcock, J. L. (1976) Cervical spine injuries. Diagnosis and classification.
Archives of Surgery, 111, 646.
Barrett, S. (1984) Trauma during pregnancy, in Progress in Critical Care
Medicine, vol. 1. Multiple Trauma, (ed. R. J. Wilder), S. Karger, Basel.
Bickell, W. H., Wall, M. J., Pepe, P. E. et al. (1994) Immediate versus delayed
fluid resuscitation for hypotensive patients with penetrating torso injuries.
New England Journal of Medicine, 17, 11051109.
Boulanger, B. R., Milzman, D. P. and Rodriguez, A. (1994) Obesity, in Trauma
and Pre-existing Disease, Part 2: Specific Management Concerns. Critical Care
Clinics, (eds A. Rodriguez, B. R. Boulanger and D. P. Milzman), W. B.
Saunders, Philadelphia, PA, vol. 10(3), pp. 613622.
Butterworth, J. F., Maull, K. I., Miller, J. D. and Becker, D. P. (1979) Detection

of occult abdominal trauma in patients with severe head injuries. Lancet, ii,
759762.
Clayton, D. G., Webb, R. K., Ralston, A. C. et al. (1991) A comparison of 20
pulse oximeters under conditions of poor perfusion. Anaesthesia, 46, 310.
Colvin, J. (1980) Penetrating eye injuries. Medical Journal of Australia, 29:
630.
Cope, A. and Stebbings, W. (1991) Abdominal trauma, in ABC of Major Trauma,
(eds D. Skinner, P. Driscoll and R. Earlam), British Medical Journal
Publications, London.
Crosby, E. T. and Lui, A. (1990) The adult cervical spine: implications for
airway management. Canadian Journal of Anaesthetics, 37, 7793.
Daranko, M. S., Klossner, J. A. and Laaksoned, V. O. (1987) Restoration of
volume by crystalloid vs colloid after coronary artery bypass. Critical Care
Medicine, 15, 559566.
Demling, R. and Reissen, R. (1990) Pulmonary dysfunction after cerebral
injury. Critical Care Medicine, 18, 768774.
Dieckert, J. P., OConnor, P. S. et al. (1986) Air travel and intraocular gas.
Ophthalmology, 93, 642645.
Doolan, L. A. and OBrien, G. M. (1985) Safe intubation in cervical spine
injury. Anaesthetics and Intensive Care, 13, 319324.
Drubach, D. A., Kelly, M. and Dolif, C. (1994) Traumatic injury in patients
with neurologic and psychiatric disease, in Trauma and Pre-existing Disease,
Part 2: Specific Management Concerns. Critical Care Clinics, (eds A. Rodriguez,
B. R. Boulanger and D. P. Milzman), W. B. Saunders, Philadelphia, PA.
Eichler, A. C., McFee, A. S. and Root, A. (1969) Heat stroke. American Journal
of Surgery, 118, 855863.
Esposito, T. J., Aprahamian, C. and Campbell, W. (1993) Human immunodeficiency virus and infectious considerations, in Trauma and Pre-existing disease,
Part 1: Critical Considerations. Critical Care Clinics, (ed. R. A. Carlson), W. B.
Saunders, Philadelphia, PA, pp. 741763.
Fisher, M. M. (1987) Anaphylaxis. Disease a Month, 33, 435479.
Gentleman, D. (1993) Clinical experience with the reduction of early
secondary insults after severe head injury. Proceedings of the 2nd International
Neurotrauma Symposium, July 1993.
Gilligan, J. E., McCleave, D., Nicholson, B. et al. (1977) Retrieval of the
critically ill in South Australia; a co-ordinated approach. Medical Journal of
Australia, 2, 849855.
Gilligan, J. E., Goon, P., Maughan, G. et al. (1996) An airborne intensive care
facility (fixed wing). Anaesthetics and Intensive Care, 24, 245253.
Goldman, L., Caldera, D. L., Nussbaum, S. R. et al. (1977) Multifactorial index
of cardiac risk in non cardiac procedures. New England Journal of Medicine,
297, 845.
Grande, C. M., Barton, R. B. and Stene, J. K. (1988) Appropriate techniques for
airway management of emergency patients with suspected spinal cord
injury. Anesthesia and Analgesia, 67, 714715.
Grande, D. C., Tissot, M., Bhatt, V. et al. (1991) Pre-existing compromising
conditions, in Trauma, Anaesthesia and Critical Care, (eds L. Capan, S. Miller
and J. Turndof), J. B. Lippincott, Philadelphia, PA, pp. 219257.
Grant, H. (1975) Vehicle Rescue, Prentice-Hall, Maryland.
Hamilton, D. (1976) The immediate management of heat stroke. Anaesthesia,
31, 270272.
Health commission of New South Wales (1977) The Granville rail disaster.
Medical review seminar. Lidcombe Hospital, 18 January, 1977 (transcript).
Health Devices Alerts (1985) Defibrillator/monitors, Battery Powered (11129),
ECRI, Plymouth Meeting, PA.
Hill, A. C., Schecter, W. P. and Trunkey, D. D. (1988) Abdominal trauma and
indications for laparotomy, in Trauma, (eds K. L. Mattox, E. E. Moore and D.
V. Feliciano), Appleton & Lange, Norwalk, CT, pp. 401416.
Hyldegaard, O. and Madsen, J. (1994) Effect of air, heliox and oxygen
breathing on air bubbles in aqueous tissues in the rat. Undersea and
Hyperbaric Medicine, 21(4), 413424.
Joyce, S. M. (1988) Cervical immobilisation during orotracheal intubation in
trauma victims. Annals of Emergency Medicine, 17, 88.
Knaus, W. A., Draper, E. A. and Wagner, C. P. (1985) APACHE II, a severity of
disease classification system. Critical Care Medicine, 13, 818.
McGill, J., Clinton, J. E. and Ruiz, E. (1982) Cricothyrotomy in the emergency
department. Annals of Emergency Medicine, 11, 361364.
MacKenzie, E. J., Morris, J. and Edelsten, S. (1989) Effect of pre-existing
disease on length of hospital stay in trauma patients. Journal of Trauma, 29,
6.
MacKenzie, E. J., Morris, J. and Edelsten, S. (1990) The effect of pre-existing
conditions on mortality in trauma patients. Journal of the American Medical
McNeil, E. L. (ed.) (1983) Airborne Care of the Ill and Injured, Springer-Verlag,
Berlin.
Meyer, A., Messich, J., Yarny, P. et al. (1992) Prospective comparison of clinical
judgment and APACHE II score in predicting the outcome in critically ill
surgical patients. Journal of Trauma, 32, 6.
Mighty, H. (1994) Trauma in pregnancy, in Trauma and Pre-existing Disease, Part
2: Specific Management Concerns. Critical Care Clinics, (eds A. Rodriguez, B. R.
Boulanger and D. P. Milzman), W. B. Saunders, Philadelphia, PA.
Milzman, D. P., Hinson, D. and Magnant, C. M. (1993) Overview and
outcomes, in: Trauma and Pre-existing disease, Part 1: Critical Considerations.
REFERENCES
Critical Care Clinics, (ed. R. A. Carlson), W. B. Saunders, Philadelphia, PA.
Milzman, D. P., Boulanger, B. R. and Rodriguez, A. (1992a) Does pulmonary
artery catheter use improve survival in trauma patients with myocardial
disease? Southern Medical Society Research symposium, San Antonio, TX.
Milzman, D. P., Boulanger, B. R., Jolin, S. et al. (1992b) Positive toxicology
screening predicts improved survival in severely injured neurotrauma
patients. Annals of Emergency Medicine, 21, 662 (abstract).
Moylan, J. A. (1988) Impact of helicopters in trauma care and clinical results.
Annals of Surgery, 208, 673678.
Myers, R. A. M., Snyder, S. K. and Emhoff, T. A. (1985) Subacute sequelae of
carbon monoxide poisoning. Annals of Emergency Medicine, 14, 11631167.
OBrien, G. M. and Criner, G. J. (1994) Chronic pulmonary disease in the
trauma patient, in Trauma and Pre-existing Disease, Part 2: Specific Management Concerns. Critical Care Clinics, (eds A. Rodriguez, B. R. Boulanger and
D. P. Milzman), W. B. Saunders, Philadelphia, PA.
Oh, T. (ed.) 1990 Intensive Care Manual, 3rd edn, Butterworths, Sydney,
NSW.
Owen, W. E. (1978) Tympanic membrane rupture with nitrous oxide
anaesthesia. Anesthesia and Analgesia, 57, 283.
Pugh, L. G. C. (1964) Deaths from exposure on Four Inns walking
competition, March 1415 1964. Lancet, i, 12101212.
Raptopoulos, V. (1994) Abdominal trauma. Emphasis on computed tomography. Radiologic Clinics of North America, 32(5), 969987.
Reid, D. C., Henderson, R., Saboe, L. and Miller, J. D. R. (1987) Etiology and
clinical course of missed spine fractures. Journal of Trauma, 27(9), 980986.
Reinfurt, D. W., Stewart, J. R., Hall, R. G. et al. (1978) Injury scaling research.
Report No. DOT-HS-701539. Highway Safety Research Center, University
of North Carolina.
Rodriguez, A., Milzman, D. P., Boulanger, B. R. et al. (1992a) Pre-existing
disease in trauma patients: a predictor of fate independent of age and ISS.
Rogers, W. A. (1957) Fractures and dislocations of the cervical spine and
etiological significance. End-result study. Journal of Bone and Joint Surgery,
39A, 341376.
Rutten, A. J., Ilsley, A. H., Skowronski, G. A. and Runciman, W. B. (1986)
Measurement of mean arterial blood pressure by automatic oscillometers,
arterial cannulation and auscultation. A comparative study. Anaesthesia and
Intensive Care, 14(1), 5865.
Scher, A. T. (1977) A plea for routine radiographic examination of the cervical
291
spine after head injury. South African Medical Journal, 51, 885887.
Simpson, D. A., North, J. B., Gilligan, J. E. et al. (1984) Neurological injury in
South Australia: the influence of distance on management and outcome.
Australian and New Zealand Journal of Surgery, 54, 2935.
Simpson, D. A., Heyworth, J. S., McLean A. J. et al. (1988) Extradural
haemorrhage: strategies for management in remote places. Injury, 19,
307312.
Swain, A., Dove, J. and Baker, H. (1991) Trauma of the spine and spinal cord,
in ABC of Major Trauma, (eds D. Skinner, P. Driscoll and R. Earlam), British
Medical Journal Publications, London.
Talucci, R. C., Shaikh, K. A. and Schwab, C. W. (1988) Rapid sequence
induction with oral endotracheal intubation in the multiply injured patient.
American Journal of Surgery, 54, 185187.
Teasdale, G. and Jennett, B. (1974) Assessment of coma and impaired
consciousness: a practical scale. Lancet, ii, 8183.
Thompson, W. (1990) Intracranial hypertension, in Intensive Care Manual, 3rd
edn, (ed. T. Oh), Butterworths, Sydney, NSW.
Trunkey, D. D. (1983) Trauma. Scientific American, 249(2), 2027.
Turner, L. M. (1989) Cervical spine immobilisation with axial traction: a
practice to be discouraged. Journal of Emergency Medicine, 7, 385386.
Vassar, M. J., Wilkerson, C. L., Duran, P. J. et al. (1992) Comparison of
APACHE II, TRISS and a proposed 24 hour post ICU point system for
prediction of outcome in ICU trauma patients. Journal of Trauma, 32(4),
490500.
Weigelt, J. (1986) Initial management of the trauma patient, in Critical
Management of the Trauma Patient. Critical Care Clinics, 4(2), W B Saunders,
Westaby, S. (1991) Thoracic trauma, in ABC of Major Trauma, (eds D. Skinner,
P. Driscoll and R. Earlam), British Medical Journal Publications,
London.
Wilder, R. J. (1984) Initial management of the critically injured patient, in
Progress in Critical Care Medicine, vol. 1. Multiple Trauma, (ed. R. J. Wilder), S.
Karger, Basel, p. 13.
Wyatt, J., Beard, D., Gray, A. et al. (1995) The time of death after trauma. British
Zwimpfer, T. J. and Moulton, T. J. (1993) Neurologic trauma concerns, in
Trauma and pre-existing disease, Part 1. Critical considerations. Critical Care
Clinics, (ed. R. A. Carlson), W. B. Saunders, Philadelphia, PA, pp.
727739.
16
FLUID, ELECTROLYTE AND

METABOLIC MANAGEMENT
Peter D. Thomas
16.1
Introduction
The medical management of the severely head-injured

patient involves manipulations of the intracranial
pressure (ICP), cerebral perfusion pressure (CPP),
seizure control, fluid and electrolyte therapy and
nutritional support. The goal of fluid management is
homeostasis, i.e. to provide appropriate parenteral
and/or enteral fluid to maintain intravascular volume, left ventricular filling pressure, cardiac output,
O2)
blood pressure and ultimately oxygen delivery (D
to tissues, when normal physiological functions are
often altered by surgical and traumatic stress and by
anesthetic agents. A specific aim is to prevent second O2 to the
ary neuronal damage due to inadequate D
brain; this requires adequate ventilation and oxygenation as well as cardiac output. The most obvious
consequences of inappropriate fluid resuscitation are
shock from insufficient volume replacement, and
pulmonary edema from overhydration.
Glucose, lipids, protein, vitamins and essential trace
elements are provided in order to retard the autocannibalism of visceral protein for gluconeogenesis.
Fluid planning must be individualized and take into
account the patients cardiac, renal, pulmonary, nutritional, functional and premorbid state.
Fluid and electrolyte abnormalities may result from
the brain injury itself, e.g. diabetes insipidus, from
therapy and from other injuries. The relationship
between systemic abnormalities of water, electrolytes
and acidbase metabolism and the brain can be
approached from two vantage points:

the effects of CNS lesions on renal function, water

and electrolytes;
the effects of metabolic abnormalities on the CNS.
16.2
Rationale of metabolic support
The outcome after a severe head injury is influenced

by age, pre-accident health, the severity of the primary
injury and any of six potentially preventable secondary insults: intracranial hematoma, raised ICP, seizures, meningitis, hypoxia and hypotension (Price,
1992). About 20% of patients with head injuries also
suffer injury to other systems. The injured brain is
highly vulnerable to hypoxia and ischemia, which
may exacerbate neuronal damage by a number of
mechanisms, including the release of excitotoxic neurotransmitters such as glutamate (Zwimpfer and
Moulton, 1993). Extracranial injury that does not result
in hypotension or hypoxia appears to have little
influence on outcome (Jennett, Teasdale and Braakman, 1979). Hence preventing and aggressively treating hypotension and hypoxia are essential parts of the
early management of all patients with head injury.
Approximately 50% of patients admitted to the
hospital in a coma following head injury will require
craniotomy (Butterworth and De Witt, 1989). Others
who do not require craniotomy may require anesthesia and surgery for other major injuries. The
anesthesia and surgery can both affect fluid and
electrolyte status.
Intravenous therapy is an integral part of both the
initial resuscitation phase and the later maintenance or
supportive phase of management. The primary aim of
resuscitation is to achieve satisfactory ventilation and
restore intravascular volume, cardiac output and tissue
perfusion, thereby preventing the sequelae of shock,
acute renal and multiple organ failure and secondary
neuronal damage. The secondary aim is correction of
any underlying disorder. In adults prolonged hypotension is rarely due to head injury alone, unless the injury
is massive and brain death is imminent, and concealed
hemorrhage must be sought. In infants and young
children, hypovolemic shock may result from blood
loss in a large extradural hematoma, especially if there
is additional bleeding under the scalp. Shoemaker has
O2 and
demonstrated that O2 transport variables (D
oxygen consumption VO2) can reliably predict

outcome in postoperative surgical patients. Conversely
when these variables are manipulated in a controlled
294
FLUID, ELECTROLYTE AND METABOLIC MANAGEMENT
prospective manner, outcome can be improved (Shoemaker et al. 1982, 1988). The determinants of oxygen
delivery are summarized in Figure 16.1. Fluid resuscitation augments preload, increases stroke output and
total systemic flow, thereby increasing oxygen delivery.
Care must be taken to avoid hemodilution, which may
simultaneously reduce arterial oxygen content and
O2 despite improvement in
prevent an increase in D
cardiac index. Oxygen saturation should be maintained
at 90% or greater in accordance with the recommendation of the American College of Chest Physicians
Consensus Conference on Mechanical Ventilation
(Slutsky, 1993).
Following resuscitation, fluid therapy aims to maintain euvolemia, provide normal daily requirements
and replace any continuing losses. It is usually
combined with nutritional support.
Cerebral autoregulation determines the relationship
between cerebral blood flow (CBF) and metabolic
needs (metabolic coupling) and systemic hemodynamics (pressure autoregulation). When metabolic
coupling is intact, cerebral oxygen delivery is determined by metabolic need and is unaffected by minor
changes in blood pressure or viscosity. Both aspects of
cerebral autoregulatory function may be impaired
following traumatic brain injury, so that cerebral
oxygen delivery is influenced by cerebral perfusion
pressure (CPP) and blood viscosity (Shoemaker et al.,
1988). Thus maintaining a normal CPP is a key aspect
of the management of the head-injured patient.
Most severely injured patients develop only small
changes in serum electrolyte concentrations, even
though there are large fluid and electrolyte shifts
between body compartments. Thus it is important to
understand the physiological basis of the common
O2 (ml/min) is
Figure 16.1 Systemic oxygen delivery D
determined by the product of the oxygen content of the
blood and total systemic blood flow. The major determinants
of each component are pictured. Hb = hemoglobin; % SAT =
percentage saturation of Hb; PaO2 = partial pressure of
oxygen in arterial blood; 0.0031 = solubility coefficient for
oxygen in plasma.
electrolyte problems, the factors controlling electrolyte

distribution and extracellular fluid volume and the
distribution of administered fluids since the choice of
intravenous fluids will influence brain metabolism
and volume (Sutin, Ruskin and Kaufman, 1992). From
this knowledge appropriate therapeutic strategies can
be developed.
16.3
Basic principles
16.3.1
FLUID DISTRIBUTION
The total body water content of humans is approximately 60% of body weight (Figure 16.2). Twothirds is located in the intracellular and one-third in
the extracellular compartment (Harrison, Darrow
and Yannet, 1936). The extracellular compartment
can be further subdivided into interstitial (proteinpoor) and intravascular (protein-rich) compartments.
Normally, 75% of the extracellular water is contained
within the interstitium and 25% in the vasculature.
There are marked differences in electrolyte content
between intracellular and extracellular fluids: potassium is predominantly intracellular, and sodium and
chloride extracellular (Gamble, Ross and Tisdall,
1923). The energy-consuming Na+ K+ pump is
Figure 16.2
liters).
Distribution of body water in a 70 kg man (45.5
BASIC PRINCIPLES
required to maintain these concentration gradients,

which are sustained during sodium overload or
depletion states. Distribution of water between the
intracellular and extracellular compartments is
determined by osmosis.
Extracellular fluid includes water attached to bone
and dense connective tissue, as well as free transcellular or third-space fluid formed by secretions
from epithelia and other membranes. Free transcellular fluid includes cerebrospinal fluid (CSF), urine,
gastrointestinal and other secretory fluid, which may
have strikingly different osmolalities from the two
major body fluid compartments (Shoemaker, 1982).
Although this third space may be greatly expanded in
disease states such as ascites, pleural effusion, paralytic ileus or bowel obstruction, it is separated from
other body fluid compartments by epithelium and
therefore does not communicate freely with them. The
third space can be thought of as parasitic it is
generated at the expense of the other two compartments but is not available to come to their rescue in
states of hypovolemia. Since most forms of expansion
of the extracellular fluid do not involve the third space
fluids and the amount of water attached to bone and
dense connective tissue is relatively fixed, the normal
body water distribution can be simplified as shown in
Figure 16.3.
Figure 16.3 Simplified distribution of body water in a 70 kg

man. The forces that influence fluid distribution between the
compartments are shown in italics.
295
Cell membranes are normally relatively impermeable to ions but freely permeable to water. Water,
electrolytes and energy substrates continually move in
and out of cells to maintain a dynamic steady-state
relationship and major fluid and electrolyte shifts
occurring with disease are often associated with only
minimal serum concentration changes. For example, a
decrease in extracellular sodium concentration and
osmolarity following an infusion of 5% dextrose in
water is followed by diffusion of water into the cells
until the osmotic pressure between the two compartments is equal. Conversely, an increase in extracellular
sodium concentration and osmolarity after infusion of
hypertonic saline is followed by movement of water
from the intracellular to the extracellular compartment. Infusion of an isotonic electrolyte solution does
not alter the osmolarity of the extracellular compartment and therefore does not lead to shifts of water into
or out from the intracellular compartment.
Water movement between the two components of
the extravascular space is determined primarily by
differences in protein concentration. Colloid osmotic
pressure (COP) is the net osmotic pressure across the
capillary membrane resulting from the impermeability of the endothelium to plasma proteins. Because
physical confinement of protein molecules produces
an osmotic imbalance, water moves from interstitial
space into the blood stream. Crystalloid molecules
cannot establish a pressure gradient because they
move freely across the capillary membrane.
Protein molecules are negatively charged and
attract a small number of positive ions, preventing
them from diffusing across the endothelium. These
retained ions produce an additional osmotic pressure
(called the Donnan equilibrium effect). Some 60% of
normal COP is produced by protein molecules (75%
by albumin, the rest mainly by globulins and fibrinogen); 40% is produced by the electrostatically held
cations. However, if the plasma protein concentration
rises, the influence of the Donnan equilibrium effect
on COP increases disproportionately, producing a
curvilinear relationship between total protein concentration and COP.
COP is measured by placing a protein solution on
one side of a semipermeable membrane and a proteinfree solution on the other side. Fluid moves through
the membrane until a pressure is generated that
prevents further osmosis. This pressure is defined as
the COP (Morissette, 1977). Plasma COP in normal
ambulatory patients is approximately 25 mmHg, in
supine patients 22 mmHg and in an Intensive Care
Unit patient 1820 mmH (Weil et al., 1974, 1981;
Rackow, Fein and Leppo, 1977). Although protein is
the primary determinant of osmotic pressure in
peripheral tissues, it contributes very little to the total
number of particles dissolved in plasma. Total plasma
296
osmolality is normally 285 mosmol/kg, of which

plasma protein accounts for less than 1 mosmol/kg.
Infusing an iso-oncotic solution such as 5% albumin
in isotonic saline expands the intravascular space
without producing a shift of water from other compartments. A hyperoncotic solution such as 25%
albumin expands blood volume to a extent greater
than the volume infused by drawing water and
electrolytes from the interstitium. However water
does not diffuse from the intracellular space if
extracellular osmolarity remains unchanged.
Starling defined the forces influencing the bulk
movement of water between the vascular and interstitial compartments (Starling, 1896). Pappenheimer
and Soto-Rivera derived a mathematical expression
for these forces, referred to as the Starling equation of
transcapillary exchange: (Pappenheimer and SotoRivera, 1948).
Qf = Kf [(Pc Pi) (c i)]
where Qf = total fluid flow across the capillary
membrane; Kf = fluid filtration coefficient; Pc =
capillary hydrostatic pressure; Pi = interstitial hydrostatic pressure; = osmotic reflection coefficient; c =
capillary oncotic pressure; i = interstitial oncotic
pressure.
The four pressures in this equation are called the
Starling forces. The net driving pressure favoring
filtration is Pc Pi. The hydrostatic pressure within the
capillary is the major force driving fluid into the
interstitium and is essentially unopposed by the
interstitial hydrostatic pressure. This is usually
slightly negative and approaches zero or becomes
slightly positive only when substantial amounts of
edema accumulate (Meyer, Meyer and Guyton, 1968;
Guyton, Granger and Taylor, 1971). The plasma COP
is thus the only force acting to retain fluid within the
intravascular space. The interstitial COP works in the
opposite direction; the net effect of these opposing
forces is described by c i
Kf, the filtration coefficient, has two components: Lp,
the hydraulic conductivity, which describes how
rapidly fluid can pass through the microvascular
exchange barrier (capillary membrane, interstitial gel
and terminal lymphatics) and S, the capillary surface
area available for filtration (Granger, 1979; Harms et
al., 1981; Demling et al., 1982). If either component of
Kf increases, e.g. through damage to the endothelial
membrane or dilatation of precapillary vasculature in
response to increased cardiac output (CO), the rate
and amount of fluid filtered increase independently of
changes in the Starling forces (Peters and Hargens,
1981).
The reflection coefficient defines the capacity of
the capillary membrane to prevent translocation of
proteins. If = 1, the membrane is totally impermeable and proteins are able to exert their full oncotic
force; if = 0, the membrane permits protein to pass
without impedance (Peters and Hargens, 1981).
varies in capillary membranes throughout the body,
being approximately 0.9 for systemic and 0.7 for lung
capillaries (Wittmers, Bartlett and Johnson, 1976). The
net effect of the Starling forces across capillaries is to
produce fluid movement from the intravascular to the
interstitial space. Accumulation of interstitial fluid in
the lung is much more life-threatening than peripheral
edema formation and there has been great interest in
evaluating factors that are likely to reduce the risk of
clinically significant pulmonary edema (Allen et al.
1987; Civetta, 1979; Crandall et al., 1983).
Modest increases in pulmonary capillary hydrostatic pressure or decreases in plasma oncotic pressure
do not result in pulmonary edema because the lung
has mechanisms for resisting accumulation of interstitial fluid. In the first place, the pulmonary interstitial hydrostatic pressure (Pi) is normally slightly
negative (Civetta, 1979; Levine et al., 1967; Taylor et al.,
1982). When fluid first begins to accumulate in the
interstitium, the pressure rapidly increases (i.e. compliance is low). This increase in hydrostatic pressure
opposes any further fluid entry by decreasing the
hydrostatic pressure gradient, Pc Pi (Allen et al.,
1987). However, as interstitial fluid pressure rises
above atmospheric, resistance to fluid transport
through the interstitium decreases markedly and
further increases in fluid volume cause only minimal
increases in interstitial hydrostatic pressure (Levine et
al., 1967; Goldberg, 1980). Pi appears to level off
between 1 and 5 mmHg in severe pulmonary edema
(Battacharya, Gropper and Staub, 1984).
A second mechanism which resists the accumulation of pulmonary interstitial fluid is a decrease in the
interstitial oncotic pressure resulting from the accumulation of protein-poor fluid (Granger,1979). Pulmonary interstitial oncotic pressure is normally about
75% of the plasma level and the oncotic gradient
across the pulmonary capillary membrane (c i) is
only 46 mmHg. If serum COP decreases, interstitial
oncotic pressure will decrease proportionately as fluid
enters the interstitial space, avoiding a change in the
oncotic gradient (Granger, 1979; Demling, 1980; Demling et al., 1979).
Thirdly, large protein molecules such as albumin are
normally excluded from a substantial part of the
interstitial fluid volume by the density of the interstitial
matrix. However, as the degree of hydration increases
the fibers in the interstitial matrix are stretched apart,
allowing protein molecules to enter previously
unavailable space (Granger, 1979). This lowers i,
widens the transcapillary oncotic gradient (c i), and
thus opposes further pulmonary edema formation.
BASIC PRINCIPLES
Lymphatic drainage is the fourth and probably

most important safety factor. Experimental studies
have shown that the lymph flow rate can increase
tenfold when interstitial fluid volume or pressure
increases.
The bloodbrain barrier is formed by a closely
woven mesh of capillary cells joined by a continuous
array of tight junctions (zonula occludens) with an
effective pore size of 0.70.9 nm. Because the blood
brain barrier is a lipid membrane with small pores, it
is easily permeated by water and very small or
lipophilic molecules. The bloodbrain barrier is relatively impermeable to electrolytes, water-soluble nonelectrolytes and plasma proteins. The endothelial cells
in peripheral tissues are joined by interspersed tight
junctions and the effective endothelial pore size is
about 45 nm. Because of the large pore size in
peripheral tissues, the endothelial membrane is freely
permeable to water and electrolytes, but only partially
permeable to plasma proteins. In the peripheral
tissues there is an active lymphatic system, which
helps reabsorb interstitial fluid. In the brain, however,
there are no lymphatics. When the bloodbrain barrier
is injured, the barrier becomes permeable to electrolytes and large plasma proteins such as albumin
(Zornow et al., 1988). Hypertonic fluids will only
remove interstitial and intracellular fluid from regions
of the brain where the bloodbrain barrier is intact,
and will have no effect on areas where the blood
brain barrier is permeable to electrolytes.
The difference in membrane function between the
bloodbrain barrier and the endothelium in peripheral tissues is of practical importance. For example, if
the intravascular concentration of sodium is increased
from 140 mmol/l to 150 mmol/l, there will only be a
small and short-lasting effect on the peripheral interstitial tissues, owing to the free movement of both
sodium and water across the endothelium. However
such a change in sodium concentration will have a
pronounced effect on the brain. The osmotic pressure
gradient will increase drawing water from the brain
interstitium to the intravascular space.
16.3.2
OSMOLALITY AND TONICITY
Osmolality is the number of dissolved particles in a

kilogram of solvent (water) and determines the
osmotic force of the solution (Shoemaker, 1982). Cell
membranes are highly permeable to water, which will
diffuse across them rapidly to maintain osmotic
equilibrium between the extracellular and intracellular fluid compartments (Brobeck,1973). Hence the
relative volumes of the compartments will be determined by their sodium and potassium content.
Plasma osmolality can be measured directly or
calculated approximately by means of the formula:
297
POSM = 2PNa+ + glucose + urea

where the concentration of sodium, glucose and urea
are all expressed in mmol/l (Worthley, Guerin and
Pain, 1987).
The plasma sodium (PNa+) level is multiplied by
two to account for anions. This rule of thumb does not
take into account the low-concentration cations such
as potassium, calcium and magnesium or organic
molecules such as creatinine, amino acids and lactate.
This error of exclusion is balanced by the presence of
multivalent anions such as sulfate and phosphate
which contribute fewer particles per equivalent than
do univalent ions.
Although urea is included in the calculation of
osmolality, it diffuses readily across cell membranes
and does not influence relative compartment volumes
(Brobeck,1973). Exogenous solutes such as mannitol or
alcohol may create an osmolar gap between the
measured and unmeasured osmolalities (Table 16.1).
This formula must be interpreted in conjunction
with the clinical assessment of fluid status as, for
example, an increased osmolality may result either
from excess sodium or a deficit of water.
Tonicity defines effective osmolality, i.e. the osmotic
force due to the particles which cannot permeate
between compartments. Because all body fluids must
be in osmotic equilibrium, changes in tonicity, regardless of their cause, will result in fluid shifts between
the ECF and the ICF in order to re-establish osmotic
equilibrium. The administration of hypertonic fluids
such as 50% dextrose or hypertonic saline will increase
the tonicity of the ECF and result in fluid shifts into
the extracellular compartment. Thus patients with
osmotic hypertonicity will have a contracted intracellular compartment.
This effect had been applied to the treatment of
patients with cerebral edema in whom intracellular
and extracellular volumes can be decreased by achieving hypernatremia. Conversely in hyponatremic states
the tonicity of the ECF is decreased and water must
shift into and expand the intracellular fluid to maintain osmolality. As brain expansion is limited by the
confines of the cranial vault, severe hyponatremia
Table 16.1 Osmotically active particles

Extracellular fluid
Intracellular fluid
Sodium
Glucose
(Mannitol)
Urea*
Ethanol*
Potassium
Magnesium
Inorganic phosphate
Urea*
Ethanol*
*These particles are distributed freely in body water. Other

particles are unable to permeate compartments.
298
may result in raised intracranial pressure (ICP). Urea

increases osmolality but not tonicity because it passes
freely through biological membranes, and will not
alter the intracellular volume,.
16.3.3 DISTRIBUTION OF ADMINISTERED FLUID IN
THE BODY
All infused fluids may be considered as being composed of plasma equivalents (colloidal solutions),
saline equivalents (crystalloid solutions) or water
equivalents. Plasma equivalents remain essentially
confined to the circulation as a result of their oncotic
pressure. Saline equivalents are distributed between
the interstitium and plasma, but remain restricted to
the extracellular fluid compartment because of their
sodium content. Water equivalents are distributed
according to the distribution of water in the body
(Figure 16.2).
(a)
Plasma equivalents
Plasma equivalents include blood, its components and

colloids. Blood may be administered either whole, or
in its component form; the latter is preferred, as
fractionation optimizes the availability of the various
blood components for special situations and improves
the survival of the stored blood elements. Its use is
limited by cost and availability, the risks of allergic
reactions and infection, the delays necessary for crossmatching and a limited shelf life.
Patients who are bleeding acutely are best treated in
part with whole blood since this also replaces clotting
components and platelets. Otherwise hypovolemia
can be treated effectively by colloid or crystalloid
solutions, and oxygen transport function can be
restored by red-cell concentrates, although these take
several hours to be fully effective since storage impairs
oxygen dissociation.
The absolute minimum acceptable hematocrit in the
trauma patient is difficult to determine. The goal of
blood transfusion is to maintain an adequate level of
Table 16.2
tissue oxygenation, but the focus in clinical medicine

has been on the level of serum hemoglobin or the
O2 is
hematocrit. It has been estimated that tissue D
optimal when the hematocrit is 33% (Sutin, Ruskin
O2 may not be
and Kaufman, 1992), but the effect on D
the same as the effect on oxygen utilization. A
consensus development conference on perioperative
red cell transfusion concluded that available evidence
did not support the recommended practice of maintaining hemoglobin levels at 10 g/dl or hematocrits at
30% (Consensus Development Conference, 1988). In
the United States, the National Blood Resource Education Program guidelines for red blood cell transfusion
state that adequate oxygen-carrying capacity can be
met by a hemoglobin of 7 g/dl (hematocrit of approximately 21%) or even less when the intravascular
volume is adequate for perfusion. The guidelines add
the caveat that the decision to transfuse a given
patient should be based on consideration of the
patients age, etiology and degree of anemia, hemodynamic stability and presence of coexisting cardiac,
pulmonary or vascular disease. While the normal
brain is very tolerant to anemia, the limits of isovolemic anemia are unknown. Because autoregulation
may be impaired after injury, it cannot be assumed
that the safe lower limit of the hematocrit is the same
as in the neurologically normal subject.
Colloidal solutions comprise preparations of plasma
or high-molecular-weight synthetic substances (Table
16.2). These fluids do not cross capillary membranes
readily and are initially confined mainly to the
intravascular space. If the colloidal solution has an
oncotic pressure that is higher than that of the plasma
interstitial fluid will move into the intravascular space
and expand the plasma volume by more than the
administered volume of colloidal solution, hence the
term plasma expander.
Plasma-protein derivatives
These include 5% normal serum albumin, which
contains human albumin 50 g/l and sodium
Colloidal solutions
Type
Normal serum
albumin
Concentrated
albumin
Dextran 70
(Macrodex)
Polygeline
(Haemaccel)
Volume
(ml)
Mean mol.
wt (kDa)
Protein
content
Sodium content
(mmol/bottle)
Oncotic
pressure
Other constituents
Half-life
Side effects
(%)*
Cost/bottle
(Aust$)
500
69
5% (25 g)
70
Iso-oncotic
510 d
0.02 (0.004)
60.00
100
69
20% (20 g)
10
5
510 d
0.01 (0.003)
85.00
500
70
6%
77 or 0
2.5
12 h
0.07 (0.017)
8.00
500
35
3.5%
72
Iso-oncotic
Potassium, 5 mmol/l;
calcium, 6 mmol/l
4h
0.15 (0.05)
12.50
*Percentage incidence of all reactions. Percentage incidence of major (grade-3 and grade-4) reactions in parentheses.
BASIC PRINCIPLES
140 mmol/l, concentrated (20%) albumin, which contains 200 g/l of albumin and less sodium (100 mmol/l).
The latter is usually restricted to hypoproteinemic
patients with either hypernatremia or fluid overload. A
500 ml infusion of 5% albumin expands the intravascular volume by 450500 ml (Lamke and Liljedahl,
1976). After 2 hours, under conditions of normal
capillary permeability, 90% remains within the intravascular space (Rainey and English, 1988). Eventually
the administered albumin is distributed throughout the
extra cellular space (Rainey and English, 1988). After
infusion of 100 ml of concentrated albumin (20%), the
plasma volume continues to increase over the next
3060 minutes to achieve a final blood volume
expansion of 400450 ml. Redistribution of 350 ml of
interstitial fluid to the intravascular space is necessary
for this to occur. In patients with extracellular or total
body water depletion, equilibration is slow and
incomplete (Beecher, 1945). Therefore, in acute hypovolemia, 5% albumin should be given rather than the
hyperoncotic form. The 20% albumin solution is
usually used in patients with concomitant hypovolemia and elevated extracellular fluid volume.
The duration of vascular retention and the hemodynamic effects of infused albumin solutions depend
greatly on the patients disease state. This variability
may result from leakage into the interstitium, preferential binding of albumin in the skin and wound, or
increased catabolism.
Albumin has several unique effects that differentiate
it from other colloids as well as from crystalloids
(Emerson, 1989). Albumin can bind reversibly with
both anions and cations, which allows it to regulate
the extracellular concentration of various substances
such as iron, lipids and bilirubin (Emerson, 1989).
Albumin also has the ability to act as a free radical
scavenger and may limit lipid peroxidation (Holt,
Ryall and Campbell, 1984; Wasil et al., 1987; Stocker,
Glazer and Ames, 1987; Pirisino et al., 1988). These
properties may in the future become the major reason
for choosing albumin as a resuscitative fluid.
Albumin may also play a role in maintaining
normal microvascular permeability to protein molecules. Endothelial cells contain pores through which
protein molecules may leave the vascular space.
Albumin may help regulate the permeability of these
pores. Hypoalbuminemia increases capillary permeability and restoring the albumin level reduces permeability to normal (Harms et al., 1981; Demling et al.,
1982). However, the intravascular albumin level necessary to maintain normal microvascular permeability is
not known.
Plasma-protein solutions are heated to 60C for 10
hours during processing and are free of transmissible
diseases (Rainey and English, 1988). Severe sideeffects, such as hypotension, which is probably due to
299
kinin activation, are rare. Because of their long half-life

care must be taken to avoid circulatory overload in
patients who have been resuscitated with plasmaprotein solutions and appropriate hemodynamic monitoring is required.
Synthetic colloidal solutions
Dextrans are polysaccharides composed of glucose
molecules polymerized into chains of various lengths.
They are classified according to molecular weight and
are available in isotonic saline or 5% dextrose. Because
of their shape dextran molecules are hyperoncotic
i.e. they have the water-attracting equivalent of
several individual molecules; thus they produce a
plasma-volume expansion of about 120% of the
infused volume (Scheinkestel et al., 1989).
Potentially life-threatening toxic effects may complicate dextran administration, including acute renal
failure, anaphylaxis and bleeding diathesis (Atik,
1969). Dextran 40 has been associated with acute renal
failure due to irreversible plugging of the renal
tubules; Dextran 70 is rarely associated with the
development of acute renal failure.
The incidence of anaphylactoid reactions after dextran administration was reported to be 0.032%. Dextran
40 produced fewer reactions than Dextran 70 and severe
reactions were uncommon (Ring and Messmer, 1977).
Dextran 70 and, to a lesser extent, Dextran 40
produce a dose-related hemostatic defect. This has a
number of mechanisms but is due primarily to a
reduction in platelet adhesion and aggregation normally mediated by factor VIIIR:antigen activity (Alexander et al., 1975). Dextran also lowers the levels of all
clotting factors by hemodilution, coats blood vessel
walls and cellular elements and impairs the elasticity
and tensile strength of fibrin clots (Atik, 1967; Weiss,
1967; Adelson, Crosby and Roeder, 1955; Karlson et al.,
1967; Muzaffar et al., 1973). Primary hemostasis is
affected and clinically the picture mimics von Willebrands disease. Bleeding is more likely in patients
with pre-existing coagulation abnormalities. To minimize this risk, the volume of dextran infused should
be limited to 20 ml/kg/d or 1.5 g/kg/d (Atik, 1967).
Some 5070% of infused dextran is excreted by the
kidneys and the rest is slowly metabolized.
Because of their interference with coagulation,
dextrans are rarely used in the management of
severely injured patients.
Polygeline (Haemaccel, Hoechst) consists of ureabridged gelatin, molecular weight range 500050 000;
mean 35 000, derived from cattle-bone gelatin and
suspended in saline. It is an effective plasma volume
expander in critically ill patients (Edwards et al., 1989;
Mishler, 1984). The low-molecular-weight gelatin portion distributes readily through the ECF and produces
300
Table 16.3
Composition of commonly prescribed electrolyte solutions

Crystalloid solutions
Glucose (g)
Calories (kcal)
Na+ (mmol/l)
K+ (mmol/l)
Cl (mmol/l)
Ca2+ (mmol/l)
Lactate (mmol/l)
Osmolarity (mosmol/l)
pH
Isotonic
(0.9%) saline
Hartmanns (Ringers
lactate) solution
5% dextrose
in water
4% dextrose
in 0.18% saline
150
150
300
6.0
131
5
111
2
29
279
5.1
50
205
277
3.56.5
40
164
30
30
282
3.56.5
a plasma volume expansion of only about 70% of the

infused volume. Renal excretion accounts for 85% of
elimination, fecal excretion for 10% and the remainder
is metabolized to non-essential amino acids. Polygeline contains potassium and calcium and this must be
remembered when large volumes are infused. Rapid
infusions cause histamine release, which usually takes
the form of urticaria, but anaphylaxis has been
reported (Scheinkestel et al., 1989). Gelatin products
are not associated with renal failure or coagulopathy.
(b)
Water-equivalent solutions
Crystalloid solutions
Crystalloids are isotonic mixtures of sodium chloride

and other physiologically active solutes (Table 16.3).
Sodium is the major component of crystalloid fluids

and the distribution of infused sodium will determine
the distribution of infused crystalloid fluids. Since
sodium is the major solute in the extracellular space
and 80% is extravascular (Edelman and Leibman,
1959), infused sodium will reside primarily in the
extravascular compartment.
Physiological saline and Hartmanns solution are
the two most frequently used crystalloids and their
volume-expanding effects are identical (Cervera and
Moss, 1975). Figure 16.4 shows the effect of colloid
and crystalloid infusions on blood volume in critically ill adults. Physiological saline and Hartmanns
solution are both freely permeable across the vascular membrane and distribute evenly throughout
Figure 16.4 The influence of colloid and crystalloid infusion on blood volume in critically ill adults. (Source: redrawn from
Shippy, Appel and Shoemaker, 1984.)
BASIC PRINCIPLES
the extracellular space. In normal individuals

approximately 25% of the infused volume remains
within the blood vessels when equilibrium is reached
(usually within 2030 minutes). In other words, for
every liter of infused crystalloid approximately
750 ml will pass into the interstitium and 250 ml will
remain in the plasma. Thus interstitial edema is a
necessary consequence of volume resuscitation with
crystalloids and should not, unless excessive, be
interpreted as evidence of fluid overload. However,
the volume of crystalloid used in resuscitating
patients with head injury should be tempered by the
possibility of worsening brain edema.
Crystalloids are well suited for replacing extracellular fluid losses (dehydration); they are also used to
replace blood loss, based on the notion that acute
hemorrhage (or hypovolemia) causes an interstitial
fluid deficit that must also be replaced. Indeed,
crystalloids have proven effective in resuscitating
patients following acute hemorrhage and continue to
be widely used for this purpose (Moss and Gould,
1988; Dodge and Glass, 1982; Tranbaugh and Lewis,
1985; Shackford, 1987; Horton, Landreau and Tuggle,
1985; Lowe et al., 1979; Virgilio et al., 1979).
Crystalloids are non-toxic and do not produce
anaphylactoid reactions.
(c)
Water-equivalent solutions
Solutions of 5% dextrose or 4% dextrose in 0.18%

isotonic saline are essentially water rendered isotonic
to prevent local red-cell lysis at the infusion site. At
most infusion rates insufficient dextrose is present to
alter blood glucose levels, so the fluid is distributed
evenly throughout total body water. Thus for every
liter of solution administered, two-thirds will enter the
intracellular space and one-third the extracellular
space. Three-quarters of the extracellular fluid will be
in the interstitium and one-quarter in the plasma.
Thus about 8% only of the infused volume remains in
the circulation. For this reason 5% dextrose solution is
the fluid of choice for patients with ischemic heart
disease or congestive cardiac failure, since it does not
expand the circulation and increase the cardiac workload.
Dextrose and cerebral ischemia
The observation that carbohydrates promote ischemic
damage in the central nervous system is not new but
seems forgotten (Voll and Auer, 1988). The central
nervous system relies on glucose for much of its
energy needs. When cerebral ischemia develops,
glucose infusion will promote anaerobic glycolysis
and produce large quantities of lactic acid, which,
301
accumulating locally, can reduce flow even further.

Thus dextrose infusion during experimental cardiopulmonary resuscitation was associated with a much
higher mortality (Lundy et al., 1987). At this stage, the
routine infusion of glucose is not recommended in
patients at risk of cerebral insufficiency.
Other fluids such as 1 liter of half-isotonic 0.45%
saline can be regarded as comprising 500 ml of isotonic
saline and 500 ml of free water.
(d)
Hypertonic saline
Hypertonic saline will increase the osmolality of the

ECF, causing a water shift from cells to re-establish
osmotic equilibrium. One liter of a 3% saline solution
contains 510 mmol each of sodium and chloride a
total of 1020 osmotically-active particles. These will be
remain in the ECF, thus increasing its number of
osmotically active particles by 25% and its volume by
7%. ECF osmolality will rise to 339 mmol/kg. To reestablish osmotic equilibrium, about 1.5 liters of water
will shift from the ICF into the ECF. Thus the 1 liter of
a 3% saline solution will increase the ECF volume by
about 2.5 liters in total. As in other states of ECF
expansion, three-quarters of this volume (1.9 l) is
distributed to the interstitium and one-quarter (0.6 l)
to the plasma, explaining the propensity of hypertonic
saline to cause pulmonary and peripheral edema
(Table 16.4).
Hypertonic saline is effective in restoring volume
after hemorrhage (Gala et al., 1991), endotoxic shock
(Horton and Walker, 1991), trauma (Mattox et al., 1991;
Vassar et al., 1991) and burn injury (Griswold et al.
1991; Monafo, Halverson and Schechtman, 1984).
Several clinical studies have suggested a better survival after resuscitation with hypertonic solutions
compared with isotonic fluids (Mattox et al., 1991;
Monafo, Halverson and Schechtman, 1984; Vassar et
al., 1991).
Hypertonic solutions have a greater volumeexpanding capacity than equal volumes of isotonic
crystalloid solutions. Approximately three times the
volume of 0.9% NaCl compared with 3.0% NaCl was
Table 16.4 Body compartment expansion according to

type of resuscitation fluid: compartment expansion (ml)
per 1000 ml of administered fluid
Fluid
Colloid
Isotonic (0.9%) saline
5% dextrose
Half isotonic (0.45%)
saline
3% saline
Plasma
Interstitium
Intracellular
1000
250
83
167
0
750
250
500
0
0
667
133
600
1900
1500
302
required to restore MAP in a canine hemorrhage

model, although the overall amount of sodium administered did not differ between the two experimental
groups. Because small infusion volumes cause relatively large increases in intravascular fluid volume,
blood volume restitution is rapid; however this effect
is short-lived (Gala et al., 1991).
Hypertonic solutions may reduce edema formation
in non-injured tissues (Battistella and Wisner, 1991;
Cross et al., 1988; Wisner, Schuster and Quinn, 1990).
This may be particularly useful in head-injured
patients. Several studies of resuscitation in experimental brain injury have found that hypertonic
solutions are more effective at lowering intracranial
pressure and decreasing edema in the uninjured
cerebral hemisphere than isotonic fluids (Battistella
and Wisner, 1991; Wisner, Schuster and Quinn, 1990).
In a study of trauma victims with concomitant head
injuries, hypertonic saline resuscitation was associated
with increased survival compared with Hartmanns
solution (Vassar et al., 1991).
At present hypertonic saline, with or without the
addition of colloid, is most commonly used for
volume resuscitation after trauma or burns. The
efficacy of small volumes and the rapidity of administration make its use convenient. Other reported
benefits of hypertonic compared to isotonic solutions
are improved pulmonary gas exchange (Boldt et al.,
1991), increased renal cortical and gut mucosal blood
flow (Behrman et al., 1991), decreased bacterial translocation (Reed et al., 1991a) and the induction of
natriuresis (Gala et al., 1991) and kaliuresis (Battistella
and Wisner, 1991).
Although these properties would be advantageous
in many clinical situations, the use of hypertonic
solutions is not without problems. In an experimental
model, increased bleeding, probably due to vasodilatation, was reported if hypertonic fluid was administered within 15 minutes of injury (Krausz et al.,
1992). If more than 10% of normal plasma volume is
replaced with hypertonic saline, increases in prothrombin time and activated partial thromboplastin
time and a decrease in platelet aggregation are
observed (Reed et al., 1991b). If large volumes of
hypertonic saline are used, potential complications
include hypernatremia, hyperchloremia, hyperosmolarity, hypokalemia, metabolic acidosis, intracellular
dehydration, cerebral hemorrhage, inhibition of lipolysis, hyperosmolar coma and central pontine myelinolysis (Carvajal and Parks, 1988; Cross et al., 1988;
Griffel and Kaufman,1992). These complications are
related directly to the solute load infused and
inversely to the patients volume of distribution.
When using hypertonic saline serum sodium levels
(less than 160 mmol/l) and serum osmolarity (less
than 320 mosmol/l) require close monitoring.
16.4
Fluid resuscitation
16.4.1
SHOCK
When hypovolemia is severe, tissue oxygenation

becomes impaired and the clinical and metabolic
features of shock appear. The need for prompt
restoration of adequate plasma volume under such
circumstances is well recognized (Shoemaker, 1976;
Shires and Canizaro, 1973). Hypotension has a markedly deleterious effect on the outcome from traumatic
head injury and indeed from other injuries (Price and
Murray, 1972; Siegel et al., 1991). Hence patients with
head injury should receive sufficient intravenous fluid
resuscitation to avoid hypovolemia and hypotension
(Buchman, Menker and Lipsett, 1991; Levison and
Trunkey, 1982). Few issues in critical care medicine
have prompted as much diversity of opinion as the
most appropriate asanguineous fluid to achieve this
goal (Virgilio, Smith and Zarins, 1979; Shoemaker and
Hauser, 1979; Velanovich, 1989).
The primary abnormality in hypovolemic shock is a
reduction of plasma volume. Absolute or relative
plasma volume deficits contribute to the disturbances
of tissue oxygenation in the other three types of
circulatory shock, cardiogenic, obstructive and septic,
and these also may occur in the head-injured patient.
Decreased plasma volume produces a decrease in left
ventricular end-diastolic volume and stroke volume.
Although sympathetic nervous system activation can
initially maintain cardiac output (CO) by inducing
tachycardia and arterial pressure by producing vasoconstriction, at some point compensatory limits are
reached and CO and systemic blood pressure may
suddenly fall.
There is clear agreement that the major goal of
treatment of circulatory shock associated with hypovolemia is rapid restoration of blood volume and
tissue oxygenation (Rackow et al., 1983; Weil and
Henning, 1979). The use of asanguineous fluids for
initial resuscitation provides better restoration of
capillary blood flow than does immediate transfusion.
Moderate hemodilution, for example, to a hemoglobin
level of 1012 g/dl, is well tolerated by most patients,
O2 if intravascular volume is
and does not lower D
maintained (Messmer, 1975).
16.4.2
CHOICE OF FLUID
The hemodynamic response to fluid infusion is influenced by the choice of fluid, vascular tone and cardiac
compliance. Several studies have directly compared
the plasma volume-expanding and hemodynamic
effects of colloids and crystalloids. In these studies, a
preselected volume of fluid was administered rather
than a volume adjusted on the basis of physiological
FLUID RESUSCITATION
response. In stable postoperative patients plasma

volume responses were compared to 1 liter infusions
of 6% Dextran 70, 6% hetastarch, 5% albumin, 3.5%
urea-bridged gelatin (Haemaccel) and physiological
saline. Significant increases in plasma volume occurred with all colloids but not with physiological saline.
The greatest increases occurred with Dextran 70
(790 ml) and 6% hetastarch (710 ml; Figure 16.5; Lamke
and Liljedahl, 1976).
In acutely ill postoperative patients Lazrove and
associates evaluated responses to infusion of 500 ml of
5% albumin and 6% hetastarch over 1 hour, using a
prospective, randomized crossover design (Lazrove et
al., 1980). Both solutions significantly increased
O2 and V
O2. Plasma volume
plasma volume, CO, D
expansion lasted for 15 hours with albumin and for
at least 3 hours with hetastarch. Similar hemodynamic
effects of 5% albumin and 6% hetastarch were reported in a series of critically ill patients by Puri and
associates (1983).
The relative effectiveness in critically ill patients of 1
hour infusions of 500 ml of whole blood, 5% albumin,
6% Dextran 70, 10% Dextran 40 and 1 liter of Ringers
lactate were compared (Shoemaker, 1976). Blood and
colloid infusion produced significant increases in
blood volume, CO, stroke volume, left ventricular
stroke work index, central venous pressure (CVP) and
pulmonary arterial wedge pressure (PAWP), but
improvements after crystalloid infusion were insignificant, even though twice the volume was infused
(Figure 16.6). In a later study by the same group
involving critically ill patients with ARDS, 100 ml of
25% albumin was compared with 1000 ml of Ringers
lactate (Hauser et al., 1980). Plasma volume increased
by an average of 465 ml with 25% albumin but by only
194 ml with Ringers lactate. Significant increases in
O2 occurred only in the albumin group.
O2 and V
D
Figure 16.5 Plasma volume expansion after infusion of 1

liter of (1) Dextran 70, (2) 6% hexastarch, (3) 5% albumin, (4)
modified fluid gelatin and (5) NS in postoperative patients.
(Source: reproduced from Lamke and Liljedahl, 1976, with
permission.)
303
Figure 16.6 Maximal changes in blood volume (ml) versus

change in cardiac output (l/min) after infusion of 500 ml of
whole blood or colloids and 1000 ml of RL in critically ill
patients. (Source: reproduced from Shoemaker, 1976, with
permission.)
Thus the more potent volume expanding properties

of colloids compared to crystalloids may permit more
rapid restoration of hemodynamic stability in hypotensive critically ill patients (Shoemaker et al., 1981;
Modig, 1986). Our current clinical practice is to use
both. In the resuscitative phase the emphasis is on
volume restoration and we use mainly colloid; in the
maintenance phase we use mainly crystalloid.
16.4.3
MONITORING
Changing vascular and cardiac compliances in the

shock patient make it difficult to rely on any single
hemodynamic measurement to define adequate resuscitation. The adequacy of volume resuscitation is
measured by blood pressure, heart rate and urine
output, which reflect a summation of many different
processes and are only gross estimates of end-organ
perfusion (Lowell et al., 1990). Mean arterial pressure
(MAP), heart rate, central venous pressure (CVP) and
pulmonary arterial wedge pressure (PAWP) do not
reflect intravascular volumes in a linear fashion
(Dawidson et al., 1991; Shippy, Appel and Shoemaker,
1984). CVP may fall during fluid infusion if blood
volume expansion results in decreased autonomicnervous-system-induced arteriolar and venous vasoconstriction (Baek et al., 1975). Changes in left ventricular compliance during fluid resuscitation may
also limit the usefulness of PAWP as a measure of left
ventricular end-diastolic volume (preload; Calvin,
Driedger and Sibbald, 1981).
Even with sophisticated monitoring techniques,
such as SwanGanz catheterization and oxygen saturation monitoring, the adequacy of resuscitation and
304
Table 16.5 Guidelines for fluid challenge utilizing pulmonary arterial diastolic or pulmonary arterial wedge pressure
monitoring: 7/3 rule for fluid challenge. PAWP = pulmonary arterial wedge pressure (mmHg); PADP = pulmonary
arterial diastolic pressure (mmHg). (Source: adapted from Weil and Henning, 1979)
PAWP/PADP
For 10 min before challenge
Fluid infusion rate

200 ml 10 min
100 ml 10 min
50 ml 10min
< 12
< 16
16
During 10 min infusion
Change > 7
Change 3
Stop
Continue infusion without interruption
Immediately after 10 min infusion
Change > 3, 7
Wait 10 min
After 10 min wait
Change > 3
Change 3
Stop
Repeat fluid challenge
oxygen delivery to individual tissue beds in vital end

organs cannot be determined. Poole et al. demonstrated this in an experimental canine shock model.
Despite restoration of MAP and cardiac output (CO),
cerebral blood flow did not return to preshock levels
(Poole et al., 1986).
Nevertheless for clinical purposes the hemodynamic response to fluid resuscitation should be
assessed by using the fluid challenge technique
originally described by Weil and Henning (Table 16.5;
Weil and Henning, 1979). When this is combined with
sequential measurements of CO and measurements of
O2 and mixed
O2, V
tissue oxygenation (lactate, D
venous oxygen saturation) the risks of under- and
over-resuscitation are minimized (Shoemaker and
Czer, 1979).
16.4.4
CRYSTALLOIDS VERSUS COLLOIDS
The basic disagreement that initiated the colloid versus

crystalloid controversy was a conceptual one. Those
favoring crystalloids believe that an extracellular fluid
deficit plays a primary role in the pathophysiology of
hypovolemic shock and therefore repletion of this
deficit is essential. Colloid proponents believe that
decreased blood volume and reduced oxygen transport are the critical pathophysiological factors in
hypovolemic shock; therefore, rapid and complete
repletion of the intravascular compartment is critical
for resuscitation.
In a series of experimental and clinical studies
initiated in the early 1960s, Shires and associates
showed that severe hemorrhagic shock was associated
with a large decrease in extracellular volume, caused
predominantly by intracellular shift of interstitial fluid
in addition to external losses or shift in the intravascular space (transcapillary refill; Shires et al., 1964,
1972; Carrico, Canizaro and Shires, 1976). In a canine
hemorrhagic shock model, animals resuscitated with
shed whole blood alone or whole blood plus 10 ml/kg
plasma had a significantly higher mortality than
animals resuscitated with whole blood and 50 ml/kg

Ringers lactate (Figure 16.7; Shires et al., 1964). In a
baboon model of hemorrhagic shock, microelectrodes
were used to monitor transmembrane potential gradients across individual skeletal muscle cells. Sustained cell depolarization, a 49% decrease in extracellular water, and 6% increase in intracellular water
increased intracellular sodium and decreased intracellular potassium were all recorded, suggesting a
failure of the ATPase-dependent Na+ K+ pump
(Shires et al., 1972). Resuscitation with balanced salt
solution produced a return of the potential difference
to normal together with a decrease in intracellular and
an increase in extracellular volume. Similar alterations
in water distribution were described in patients with
circulatory shock and those undergoing major operative procedures (Shires, 1965; Shires, Williams and
Brown, 1960). These changes also resolved with
infusion of balanced salt solutions.
Figure 16.7 Survival after resuscitation of dogs from

hemorrhagic shock. Each group consisted of 10 dogs, bled
into shock by the Wiggers method. The RL group received
fluid equivalent to 5% of body weight followed by return of
shed blood. The plasma group received 10 ml/kg of donor
plasma plus shed blood. The blood group received shed
blood alone. (Source: reproduced from Shires et al., 1964,
1964, American Medical Association, with permission.)
EFFECTS OF INTRAVENOUS FLUIDS ON THE BRAIN
There is however, considerable controversy regarding the methods used by Shires and colleagues to
measure extracellular fluid volume (Shoemaker, 1976;
Roth, Lax and Malone, 1969). Using different techniques and models of hemorrhagic shock, other groups
have found either a minimal decrease in extracellular
volume accountable by the plasma volume loss or else
an increase in extracellular volume (Roth, Lax and
Malone, 1969; Serkes and Lang, 1966; Moore et al.,
1966). There is similar debate about the changes in
extracellular volume after major surgery, where no
significant change was found in patients after cholecystectomy and increases were found in patients after
cardiac surgery (Roth, Lax and Malone, 1969).
Despite this there is considerable evidence, from
experimental models of severe hemorrhagic shock and
from clinical studies of traumatic shock, that subjects
can be successfully resuscitated with balanced salt
solutions and blood (either whole blood or packed red
blood cells) without the need for colloidal solutions
(Lowe et al., 1979; Moss et al., 1981; Weaver et al., 1978).
In the study of Lowe et al. (1979), patients with
traumatic abdominal injuries requiring laparotomy
were resuscitated randomly with either Ringers
lactate or 4% albumin together with washed red blood
cells for the entire emergency room and intraoperative
period. Crystalloids alone were used postoperatively.
Clinical criteria (systemic blood pressure, pulse and
urine output) were used as the end points for
resuscitation. Both groups of patients were resuscitated
successfully and overall mortality was low (4.3%).
There was no difference between the two groups in the
need for postoperative ventilatory support. Since most
patients in this study were not in shock on admission,
the results might not be applicable to patients who are
in severe hemorrhagic shock. Moss et al. (1981) repeated
the study in a group of patients with traumatic shock,
defined as systolic arterial pressure of 80 mmHg or the
need for five or more red blood cell transfusions before
surgery, with similar results. Pulmonary edema did not
complicate the use of colloid or crystalloid in either of
these studies. In both studies similar volumes of colloid
and crystalloid were needed for resuscitation. This
surprising finding may be due to the use of clinical
rather than physiological end points for fluid resuscitation. Whether the successful outcome with crystalloid
resuscitation in these studies is related to replacement
of an interstitial fluid deficit or to the limited amount of
the infused solution that remains in the circulation has
not been ascertained. Despite the numerous studies
comparing crystalloids to colloids, none has unequivocally demonstrated a distinct survival advantage with
either therapy.
The current recommendations of the American
College of Surgeons for initial resuscitation of patients
with traumatic injury (hemorrhagic shock) include
305
rapid infusion of up to 2 liters of Ringers lactate until

hemodynamic stability is restored. If the patient
remains unstable packed red blood cells are then
infused. The use of two to four large-bore intravenous
lines can compensate for the much more limited
volume expansion produced by crystalloids. Complications of this approach to fluid infusion are minimal
(Moss et al., 1981).
The choice of fluids in patients with massive
bleeding that is difficult to control, or those with preexisting medical problems, is more controversial. In
these situations fluid administration may result in
fluid overload and subsequent morbidity. The risk of
pulmonary edema has been central to the debate over
the relative merits of crystalloids versus colloids. As
mentioned earlier, capillary membrane permeability,
capillary hydrostatic pressure, colloid osmotic pressure (COP) and pulmonary lymph flow are all
important factors in the development of pulmonary
edema. Despite claims to the contrary (Virgilio et al.,
1979), weight gain and systemic edema due to fluid
infusion are not benign problems (Falk, 1991; Lowell et
al., 1990). Decreased chest wall compliance due to
tissue edema may occur following hydration with
crystalloid (Brinkmeyer et al., 1981). Peripheral edema,
particularly in a debilitated patient, can cause
decreased mobility and subsequent skin breakdown.
Tissue oxygenation is decreased when edema is
present (Heughan, Niinikoski and Hunt, 1972) and
wound healing may be impaired (Falk, 1991; Mangalore and Hunt, 1972; Niinikoski, Hengan and Hunt,
1972).
Edema of the gastrointestinal tract may result in
ileus and intolerance for enteric alimentation (Falk,
1991). The potential for bacterial translocation and the
development of systemic sepsis and multiple systems
organ failure may also be increased (Baker et al., 1988;
Wilmore et al., 1988).
In addition, systemic edema can increase time on
mechanical ventilation, require diuresis or dialysis
and lengthen the stay in the ICU
16.5 Effects of intravenous fluids on the

brain
Parenteral fluid therapy, particularly in patients with
head injury, may increase brain swelling, intracranial
pressure and neurological dysfunction and this may
be reflected in increased mortality and morbidity
(Vassar et al., 1991; Battistella and Wisner, 1991; Falk,
1991; Fein et al., 1982). However, there is little
information available on the relative effects of colloid
or crystalloid administration on cerebral edema formation in normal subjects or in patients with brain
injury and decreased intracranial compliance (Zornow
et al., 1988).
306
16.5.1 THE EFFECTS OF CRYSTALLOIDS AND

COLLOIDS ON THE NORMAL BRAIN
Early studies in animals demonstrated brain shrinkage

in response to infusion of hypertonic solutions and
expansion in response to hypotonic solutions (Weed
and McKibben, 1919a, b). In general, hypertonic fluids
decrease both brain interstitial and intracellular volume and thus decrease ICP, and hypotonic fluids have
the opposite effects. It was not possible to determine
from these studies whether the increase in ICP and
cerebral water content with hypotonic solutions was
due to a decrease in plasma osmolality or a decrease in
colloid oncotic pressure (COP). To clarify this issue,
Zornow, Todd and Moore (1987) examined the acute
effects of changes in plasma osmolality and plasma
COP in normal rabbits with an intact bloodbrain
barrier. In one group COP was decreased without a
change in plasma osmolarity, whereas in a second
group a hypo-osmolar condition was produced without change in the COP. Cerebral edema, estimated by
changes in brain specific gravity, developed only in the
hypo-osmolar group, suggesting that changes in COP
do not influence intracranial water distribution in the
absence of brain injury (Figure 16.8). These and other
studies demonstrate that when the bloodbrain barrier
is intact, brain volume is strongly influenced by plasma
osmotic pressure and is unaffected by changes in
plasma oncotic pressure (Hindman et al., 1990). The
brain, like a red blood cell, swells in a hypotonic
solution and contracts in a hypertonic solution.
The effects of fluid resuscitation in experimental
hemorrhagic shock on ICP and cerebral edema in
animals with intact bloodbrain barriers have been
evaluated in several studies (Poole et al., 1986; Prough et
al., 1985; Gunnar et al., 1986). In different models, fluid
Figure 16.8 Acute cerebral effects of changes in plasma

osmolality and oncotic pressure in normal rabbits. The
lower the specific gravity, the greater the brain water
content **p < 0.01 versus other groups. (Source: from Zornow, Todd and Moore, 1987, with permission.)
resuscitation with hypertonic crystalloid solutions was

consistently associated with a lower ICP than fluid
resuscitation with Ringers lactate, physiological saline
or 10% Dextran 40 (Prough et al., 1985; Gunnar et al.,
1986, 1988). Resuscitation with 6% hetastarch was
associated with significantly lower ICP than resuscitation with Ringers lactate (Poole et al., 1986).
The unique structural characteristics of the cerebral
capillaries (the bloodbrain barrier) limit the effects
of changes in COP on cerebral water distribution.
The intercellular pores of non-cerebral capillaries are
approximately 65 in diameter. Ionic solutes such as
sodium and chloride are able to migrate freely
between the interstitium and the intravascular space
and therefore have no influence on water distribution (Zornow, Todd and Moore, 1987). High-molecular-weight compounds like albumin and ureabridged gelatin are retained in the intravascular
space and produce an oncotic gradient that tends to
retain water in the capillaries. Because of the small
pore size (8 ) of cerebral capillaries, changes in
serum electrolyte content (osmolarity) have significant effects on water movement. An osmotic gradient
of 1 mosmol/l produces a hydrostatic gradient of
19.3 mmHg. Because there are so few protein molecules compared with the number of inorganic ions,
their effect is minimal (Albright, Latchaw and Robinson, 1984). A 10 mm decrease in COP during crystalloid resuscitation has the same effect on cerebral
water distribution as a decrease in osmolality of only
0.5 mosmol/l. Clearly the influence of changes in
osmolarity on cerebral water distribution dwarfs the
effects of alteration of COP.
16.5.2 THE EFFECTS OF CRYSTALLOIDS AND
COLLOIDS ON THE BRAIN AFTER INJURY
Patients with severe traumatic injuries often require

resuscitation with large volumes of intravenous fluid
to restore hemodynamic stability. When a head injury
is also present, the choice of fluid for resuscitation may
influence the development of brain edema. Animal
studies confirm that damage to the bloodbrain
barrier (e.g. by cryogenic injury) increases the permeability of capillaries to both proteins and electrolytes so
that osmotic and oncotic gradients cannot be established (Zornow et al., 1988; Weed and McKibben,
1919a). In these circumstances capillary hydrostatic
pressure determines the rate at which edema forms.
When hemorrhagic shock was combined with cryogenic brain injury, brain water fell in uninjured brain
after resuscitation with hypertonic saline but not in
areas of brain injury where the bloodbrain barrier
was damaged (Wisner, Schuster and Quinn, 1990).
Other animal studies evaluating the effects of fluid
resuscitation in hemorrhagic shock on neurological
METABOLIC RESPONSE TO INJURY
function have shown that, although ICP in animals

resuscitated with hypertonic saline was significantly
lower, cerebral perfusion pressure was significantly
higher in those receiving 6% hetastarch compared
with those receiving hypertonic saline or physiological saline. Neurological function was also best in
the hetastarch group, suggesting that restoration of
cerebral perfusion pressure is a more important goal
than change in ICP alone. The vasodilatory effect of
hypertonic saline may have resulted in a lower mean
arterial pressure compared with hetastarch.
These experimental results reaffirm that hypertonic
saline will shift water from the intracellular to the
extracellular space. Ringers lactate is a hypo-osmolar
solution and 6% hetastarch is iso-osmolar; hence the
differing effects of these solutions on ICP can be
explained by their osmolarities.
The lower ICP after resuscitation with hypertonic
crystalloid solutions may assist in restoring of cerebral
O2, but there is not universal
blood flow and D
agreement on this issue (Prough et al., 1986).
The days of keeping the patient with a head injury
dry are over. There is no good evidence supporting
fluid restriction as a means of limiting cerebral edema
after brain injury (Morse et al., 1985). Dehydration
increases sympathetic stimulation, metabolism and O2
demand (Beckstead et al., 1978). Indeed, the therapeutic aim is now to maintain euvolemia and normal
physiological indices, especially cerebral perfusion
pressure. Animal studies support this approach
(Smith et al., 1982; Ito et al., 1979).
16.6
Metabolic response to injury
16.6.1
OVERVIEW
Major injury, whether surgical or accidental, evokes

predictable metabolic, hormonal and hemodynamic
responses (Buckingham, 1985; Table 16.6). Changes in
carbohydrate metabolism include increased endogenous hepatic glucose production (gluconeogenesis)
and reduced glucose clearance (insulin resistance),
which results in hyperglycemia. Fat becomes the
major body fuel; therefore lipolysis is increased and
lipogenesis is retarded. Changes in protein metabolism are manifested by negative nitrogen balance
Table 16.6

Metabolic response to injury
Altered protein homeostasis

Hypermetabolism
Altered carbohydrate metabolism
Sodium and water retention
Increased lipolysis
307
reflecting accelerated net protein breakdown (catabolism). The magnitude of these changes is proportional
to the extent of the injury (Weissman, 1990).
16.6.2
STARVATION
Starvation, i.e. lack of nutrient input, often accompanies injury, while nutrient utilization is normal at
the cellular level. During starvation, metabolic adaptation occurs, with the aim of conserving essential
tissues. There is little or no activation of metabolic
mediators and the system is still able to respond to
normal physiological stimuli. Glycogen reserves are
only small, approximately 200400 g; thus liver glycogenolysis is only useful for 24 hours. Decrease in
insulin and increase in glucagon secretion result in
protein and fat breakdown to provide energy. Amino
acids are converted to pyruvic acid, acetyl CoA and
tricarboxylic acid (Krebs) cycle intermediates. Glycerol from fat is fed into the glycolytic pathway and
fatty acids form acetyl CoA, some of which enters the
Krebs cycle inside the mitochondria; some is converted to ketones, which are used by skeletal muscle
and brain. Protein catabolism results in increased
urinary loss of urea, sodium, potassium, magnesium
and calcium. As the new energy pathways become
established, protein breakdown decreases and fat
becomes the chief energy source, supplying 7590% of
the calories. The basal metabolic rate decreases as a
result of decreasing lean body mass, decreased physical activity, and decreased thyroxine production.
These changes are summarized in Figure 16.9.
In the absence of sepsis or injury, the starved patient
can usually be rapidly converted to the anabolic state
by administering moderate amounts of nutritional
substrate.
16.6.3
METABOLIC STRESS
When surgery, trauma or sepsis occur, a new process is

activated (Cerra, 1986). Injured or dead tissue, dividing organisms, severe perfusion deficits and resolving
hematomas activate mediator systems that affect endorgan function, producing the clinical, physiological
and metabolic manifestations of the response to stress.
This process is dynamic, and begins with a lag or ebb
phase during which there is little metabolic activity.
This is followed by a flow phase during which
metabolic activity increases and peaks, usually on day
3 or 4 after injury. The processes then abate over
another 34 days, unless a complication ensues. In the
event of a new stress, the process reactivates and
reaches a new peak.
The resting energy expenditure rises to an amount
that depends on the type and severity of the stress.
At high levels of stress, 30% or more of the increased
308
Figure 16.9 Acute non-stressed fasting metabolism. The initial flow of substrate is designed to provide glucose for obligate
and non-obligate glucose users. With time, ketones and fats become the main energy substrates, resulting in daily urinary
nitrogen excretion.
energy expenditure appears to come from the oxidation of amino acids, 3040% from glucose and
3040% from fat. With activation of metabolic mediators by trauma, gluconeogenesis becomes less
responsive to exogenous nutritional substances (section 16.6.4(b)). Thus, administered glucose has progressively less inhibiting effect on lipolysis, proteolysis and gluconeogenesis. Glucose calories, as well as
other calories, which are in excess of the existing
demands promote lipogenesis, with excess CO2 production and hepatic dysfunction.
Proteolysis increases and amino-acid flux attempts
to meet the demands of energy production and
protein synthesis. Ureagenesis is increased and more
nitrogen is present in the urine. Gluconeogenesis is
increased; peripheral glucose uptake is normal, but
much of the glucose is recycled as lactate and
alanine. Thus, increased plasma glucose and lactate
levels are a normal part of the response. Ketosis is
relatively depressed. These changes are summarized
in Figure 16.10.
Malnutrition usually develops rapidly in posttraumatic patients, taking days instead of the weeks
needed in simple starvation. It is difficult to exclude
the effects of bedrest (disuse) on some of the parameters. Indeed, in patients with isolated closed-head
injuries many reports have described persistent
excretion of large amounts of urinary nitrogen up to
weeks after severe head injury. However some studies indicate that the response abates in 57 days
unless a complication ensues. The early nitrogen
excretion appears to reflect the stress response,

whereas the persistent nitrogen excretion may be due
to bedrest and disuse in a largely young and muscular group of patients.
In some instances under new stimuli the process
reactivates and the state of persistent hypermetabolism with or without the development of multiple
system organ failure ensues (Cerra, 1986). Risk factors include severe perfusion shock, severe septic
shock, persistent septic sources, and recurrent septic
episodes.
16.6.4
(a)
MEDIATORS OF THE RESPONSE TO INJURY
The neuroendocrine axis
The mechanisms initiating, regulating and sustaining

this response are not all identified. It is known that
injured patients have elevated levels of the antiinsulin hormones: cortisol, glucagon, and the catecholamines. Insulin levels are usually elevated, but
not sufficiently to prevent the commonly noted
hyperglycemia. Growth hormone, aldosterone and
ADH are also elevated. These elevations may in part
be neurally mediated via the hypothalamus (Hume
and Egdahl, 1959). Patients with severe head injury
may develop abnormalities of fluid and electrolyte
balance because of damage to the neuroendocrine
axis (e.g. diabetes insipidus following pituitary damage; Figure 16.11).
309
Figure 16.10 Summary of moderate to severe stress metabolism. The neuroendocrine and intrinsic mediator systems
modulate the metabolic machinery to mobilize fat and amino acid stores, which provide a continuous supply of fuel to meet
increased energy demands and provide an increased supply of substrate for the production of stress proteins. Increased
urinary nitrogen excretion is one result of the process. BCAA = branched-chain amino acids.
(b)
Cytokines
Several non-endocrine factors play important roles in

the response to stress. These include interleukin-1 (IL1), tumor necrosis factor (TNF), IL-2 and gammainterferon (IFN). The reader is referred to the extensive
review of this subject by Weissmann (1990).
16.6.5 EFFECT OF STRESS ON FLUID AND
ELECTROLYTE MOVEMENTS
The changes in electrolyte concentrations and fluid

distribution after trauma and critical illness have been
extensively studied. In the normal subject neural,
hormonal, hemodynamic and renal mechanisms function in a highly integrated manner to preserve sodium
and water homeostasis. There are two main objectives.
The first is to keep the concentration of sodium in the
ECF within a very narrow range. Together with its
associated anions, sodium constitutes more than 90%
of the total solute in the ECF and controls the
distribution of water between the cells and the
extracellular space. Large deviations from normal in
ECF sodium concentration cause cells to shrink or
swell, which may have serious consequences for brain
function. Sodium concentration is kept constant by
finely adjusting the water content of the ECF through
the secretion of antidiuretic hormone (ADH, vasopressin; Figure 16.12). At least four independent physiological stimuli release ADH into the blood stream:
osmotic, hemodynamic (hypotension and/or hypovo-
lemia), nausea and emesis, and hypoglycemia. Of

these, osmotic stimulation is the most important for
fine control of ADH secretion and maintenance of
water balance. A number of drugs used in critically ill
patients affect ADH secretion. Halothane and morphine in high doses increase ADH release, while
phenytoin and chlorpromazine inhibit ADH release.
Positive pressure ventilation also increases secretion.
The thirst mechanism assists by controlling fluid
intake to some extent in the conscious patient.
The second objective is to keep the total sodium
content of the ECF within normal limits and thus
maintain a normal ECF volume. Since sodium is the
major cation of the ECF and the body adjusts the water
around it to maintain a normal sodium concentration,
the total number of sodium ions in the ECF will
determine the ECF volume. Large deviations from
normal in the ECF sodium content cause fluctuations
in the circulating blood volume. Both volume contraction and expansion can have serious effects on brain
function, particularly in the presence of pre-existing
brain damage (Aubry and Nankin, 1965). Normally,
ECF sodium is regulated by several closely coordinated mechanisms that adjust the amount of sodium
lost through the kidneys.
The relative fullness of the ECF is sensed by

receptors located in low- (intrathoracic) and highpressure (intra-arterial) areas of the cardiovascular
system. When changes in ECF volume occur, renal,
neural and hormonal mechanisms modulate renal
310
(a)
(b)
Figure 16.11 Hypothalamic control of pituitary secretion. (a) Vasopressin and oxytocin are secreted in the paraventricular
and supraoptic nuclei of the hypothalamus and transported in the axons of the supraoptico-hypophyseal tract to the
posterior lobe of the pituitary gland; these hormones are then released into the circulation. (b) Hypothalamic releasing
hormones are formed in the hypothalamus and are transported to a capillary plexus formed by the superior hypophyseal
arteries around the median eminence and infundibular stem of the hypothalamus; these hormones there enter into the
vascular system and are taken by the portal veins to the adenohypophysis. (Source: reproduced from Brown, David and
Reilly, 1995, with permission.)
311
Figure 16.12 Control of body fluids. This shows the pituitary gland at the top of the diagram, the heart in the center and
the two kidneys. The factors affecting ACTH, ADH and catecholamine release are shown and the effects on the kidney of
these two hormones are simplified at the bottom of the diagram. The possible results of the kallikrein system, natriuretic
hormone and angiotensin are also included.
sodium excretion such that renal sodium and water

excretion fall when absolute or relative ECF volume
is low and increase when ECF volume is high. The
renal mechanisms influencing sodium excretion in
response to changes in ECF volume are listed in
Table 16.7.
Physical factors in the peritubular capillary environment are one of the principal mechanisms influencing Na+ excretion in response to changes in ECF
volume. When renal perfusion pressure falls or
oncotic pressure in the peritubular capillary increases, the balance of Starling forces favors sodium and
water reabsorption in the proximal tubule. Conversely, if the hydrostatic pressure in the peritubular capillary is increased, sodium excretion
occurs. Changes in the resistance of the glomerular

afferent or efferent arteriole affect renal sodium
reabsorption independent of changes in renal perfusion pressure or oncotic pressure. Vasodilatation
(e.g. with low-dose dopamine) increases capillary
Table 16.7 Renal mechanisms influencing sodium

excretion in response to changes in ECF volume

Physical factors in peritubular capillary environment

(i.e. hydrostatic and oncotic pressure)
Changes in resistance of afferent or efferent arteriole
Renal adrenergic nerve fibers
Hormonal mechanismsreninangiotensinaldosterone;
atrial natriuretic peptide (ANP)
312
hydrostatic pressure and renal sodium excretion,

while vasoconstriction (e.g. heart failure) reduces
hydrostatic pressure and renal sodium excretion.
Adrenergic nerve fibers innervate the renal tubules
and may influence renal sodium excretion independent of changes in renal hemodynamics. Stimulation enhances sodium reabsorption, while denervation reduces renal Na+ reabsorption.
The major hormonal mechanism influencing renal
sodium excretion is the reninangiotensinaldosterone system. The juxtaglomerular apparatus
senses a reduction in effective blood volume, renin
is released from the kidney, ultimately increasing
angiotensin II and aldosterone levels. Aldosterone
increases distal tubular sodium reabsorption and
potassium excretion. Angiotensin II also increases
systemic vascular resistance and proximal tubular
sodium reabsorption by increasing resistance in the
efferent arteriole.
ANP is produced both in atrial myocytes and within

the CNS is also important in body sodium regulation
(Needleman and Greenwald, 1986; Samson, 1987). Its
actions tend to oppose those of ADH and angiotensin,
and so it appears to play a role in regulating fluid and
electrolyte balance. ANP may play a role in local
control of brain volume as well as electrolyte and
water content. ANP is released into the systemic
circulation from the atria, stimulated by increased
intravascular volume or by increased atrial pressure
independent of intravascular volume, and under saltloading conditions as with infusion of hypertonic
saline. ANP decreases renal vascular resistance and
increases glomerular filtration rate. It induces natriuresis and diuresis and inhibits the reninangiotensin
aldosterone axis. ANP also reduces systemic vascular
resistance and blood pressure.
Fluid balance disturbances are common among
patients with head injury and are often multifactorial.
Hyponatremia has often been attributed to a cerebral
salt-wasting state or to inappropriate secretion of
antidiuretic hormone (SIADH; as noted earlier, a
moderate degree of ADH secretion is part of the
response to stress). However, hyponatremia may in
fact often be iatrogenic (Bouzarth et al., 1982; Nelson et
al., 1981). Hypernatremia and hyperosmolarity are
present in many head-injured patients as a consequence of the fluid restriction and hyperosmolar
agents used to manage the increases in intracranial
pressure. Hypothalamic dysfunction can cause diabetes insipidus with hypernatremia.
The major fluid/electrolyte changes immediately
after acute injury are:
release of K+ by the cells, leading to transient

relative hyperkalemia, increased K+ excretion and
subsequent reduction of total body K+;
retention of Na+ and water; water is retained

relatively more than Na+, leading to dilutional
hyponatremia.
These changes are maximum on the first and second

day after injury and persist for 47 days. They are
more pronounced after severe trauma (Verney, 1954).
16.7 Fluid therapy in uncomplicated

postoperative and post-traumatic states
16.7.1
FLUID AND ELECTROLYTE REQUIREMENTS
Abnormal fluid losses occur through:

external routes, such as hemorrhage, GI losses,

wounds and burns; CSF otorrhea in young children
can sometime be clinically significant;
internal shifts into the interstitial water or extracellular water;
local edema due to surgical wounds, crushing
injuries, burns, venous thrombosis;
fluid accumulation in the transcellular space, e.g.
pleural effusion, ascites and paralytic ileus.
Fluid replacement must take into account basal losses,

additional extrarenal losses and distributional changes, based on an understanding of the underlying
physiological and metabolic changes.
Once a stable circulation has been established, water
replacement in adult patients with no oral intake
should equal urinary losses, usually 10001500 ml/d,
plus insensible water losses, usually 600800 ml/d,
minus the volume of endogenous water released by
tissue breakdown during semistarvation. In uncomplicated, afebrile postoperative patients, the latter is
usually only 100200 ml but may be as much as
1000 ml in a critically ill trauma patient with sepsis.
Thus, replacement volume is approximately
15003000 ml/d, depending on weight, age, sex and
overall status. Elderly patients with underlying cardiac or renal disease usually require less.
Water requirements increase by about 300 ml/d for
each 1C elevation in body temperature. Fluid requirements are greatly increased by shock, multiple
trauma, sepsis, postoperative complications and other
critical illnesses. Under these conditions, criteria for
fluid therapy are based primarily on hemodynamic
factors and adequate oxygen delivery.
Basal sodium requirements are met by the daily
administration of 500 ml of physiological saline
(75 mmol/d) plus the estimated volume of extrarenal
fluid losses (e.g. nasogastric suction, diarrhea, CSF
and other fistulae, and pleural or peritoneal drains).
In an uncomplicated postoperative patient, 40 mmol
potassium per day is usually sufficient, unless the
patient is suffering from renal failure or upper GIT
FLUID THERAPY IN UNCOMPLICATED STATES
bleeding. Such patients may develop hyperkalemia

and thus usually need less K+ replacement. Fluid
therapy of brief duration does not usually require
magnesium, but debilitated patients who do not have
renal failure may benefit from 1015 mmol Mg2+
daily.
Nutritional support plays an integral role in the
management of metabolic stress. The time to initiate
nutritional support is not completely settled. It is
generally agreed that oxygen transport must be
restored, stabilized and maintained prior to beginning
nutritional therapy. In a previously well nourished
patient, our practice is to wait a few days, whereas in
a severely malnourished patient, therapy should begin
as soon as O2 transport is stabilized.
16 7.2
NUTRITIONAL REQUIREMENTS
Nutrition may be administered via the enteral or

parenteral route. The essential requirements are as
follows.
(a)
Calories
Energy as carbohydrate or fat is given to meet the

metabolic needs of the patient (i.e. basal metabolic rate
plus additional demands). Sufficient energy of nonprotein origin must be given to prevent infused
protein being used as an energy source. Energy is
measured in kilocalories (kcal) or kilojoules (kJ) one
kilocalorie = 4.184 kilojoules. A satisfactory ratio of
non-nitrogen kilocalories per gram of nitrogen is 150:1
(each 6.25 g protein yields 1 g nitrogen).
Although numerous equations and nomograms
have been used to estimate metabolic rate and
nitrogen requirements, none predict accurately which
caloric or nitrogen substrate will be utilized most
effectively. Furthermore, substrate utilization varies
during the course of an illness.
The increase in metabolism during the acute phase
of injury is short-lived, and nutritional requirements
fall rapidly within 25 days as shock, pain and sepsis
resolve and recovery begins. During convalescence
very few hospital patients expend more than
15002000 kcal (63008400 kJ) per day.
Glucose
Glucose and lipid solutions are the standard intravenous energy substrates. In a normal person a daily
infusion of 4 mg/kg/min of glucose suppresses gluconeogenesis maximally Approximately one-third of the
infused glucose is oxidized immediately and the
remainder is stored as glycogen or lipid. Increasing the
infusion rate increases the amount of glucose stored,
thus increasing O2 utilization and CO2 production. A
313
greater caloric intake is unnecessary in most patients

and may be detrimental in those with severe respiratory failure, particularly if infusion rates of greater than
7 mg/kg/min (equivalent to 700 g of dextrose or
2800 kcal/d in a 70 kg subject) are used. Excessive
glucose administration may lead to hepatic steatosis,
hyperbilirubinemia and an elevated alkaline
phosphatase.
In patients requiring parenteral nutrition, glucose
should be used as the major caloric source and infused
at a rate no greater than 30 kcal/kg/d (equivalent to
500 g of dextrose or 2000 kcal/d in a 70 kg subject).
Each gram of glucose provides approximately 4 kcal.
Lipid
Intravenous lipids are available as emulsions similar
in particle size to chylomicrons, and are cleared as
such by the body. Chylomicrons are acted upon by
plasma lipoprotein lipase to form free fatty acids and
glycerol, which are either oxidized or stored as lipid.
Although there is an increasing trend toward the
routine use of lipid solutions as an energy source,
these expensive substances have no advantages over
dextrose. Lipid solutions are only necessary to provide
the essential fatty acids linolenic and linoleic. In the
absence of essential fatty acids, linoleic and hence
arachidonic acid levels fall and skin lesions (desquamatous dermatitis) may result. Approximately 9 kcal
is provided for each gram of fat used by the body.
(b)
Protein
The normal adult only needs an oral intake of 20 g of

protein to meet daily protein requirements. Protein
ingested in excess of this amount is broken down to
provide energy and nitrogenous wastes. In the acutely
ill patient, the minimal quantity of protein required is
unknown, but it is unlikely to exceed 50 g/d in a
patient who is not losing protein externally. The
usually quoted requirement is 12 g/kg body weight
per day.
Normal adults require 20 L-amino acids for protein
synthesis, although only leucine, isoleucine, valine,
lysine, threonine, phenylalanine, methionine and tryptophan cannot be synthesized and are therefore
essential. In patients requiring parenteral nutrition,
histidine, arginine and cysteine may also be
required.
For effective nitrogen utilization, a ratio of essential
to total amino acids of 2:5 is advisable, and should
reflect the amino acid profile of a normal diet. To
reduce the oxidation of amino acids as an energy
substrate, a source of alternate energy should be
administered simultaneously in a ratio of 150 kcal for
each 1 g of nitrogen.
314
(c)
Table 16.8 Recommended daily allowance (RDA) of

vitamins for adults (MVI-12 contents)
Vitamins and trace elements
Vitamins are dietary compounds that act as cofactors

for intermediary metabolism. Patients may develop a
deficiency of the water-soluble C and B group vitamins
and vitamin K within 4 weeks of parenteral nutrition;
hence they should be administered to all patients
receiving intravenous nutrition at doses equivalent to
normal daily requirement (Table 16.8). Vitamin deficiency may rarely affect neurological assessment.
The essential trace elements are those present in the
body in amounts of less than 50 g/g tissue, and are
required in the diet in milligram quantities or less per
day. They are usually defined by their deficiency states,
i.e. a reproducible functional and/or structural abnormality associated with a specific biochemical change
which is prevented or reversed by the element.
With the exception of zinc, the body stores of the
essential trace elements, copper, iodine, iron, manganese, cobalt, selenium, chromium and molybdenum
are usually adequate for patients requiring parenteral
nutrition for less than 3 months. A total of 2.5 mg of zinc
per day will normally maintain the zinc balance,
although, in patients with diarrhea, up to five times this
amount may be required. Table 16.9 summarizes
parenteral nutrient intake for adults. Trace element
deficiency can result in neurological manifestations,
which may interfere with the neurological assessment
of the head-injured patient and wound healing, as is the
case with some vitamin deficiency syndromes. There is
little data regarding the optimal amounts of provision
of trace elements to critically ill patients. Guidelines
will only be established when more accurate methods
are developed for assessing and monitoring trace
element status in patients in intensive care.
Gastric atony is usually present during states of
metabolic stress and responds poorly to motility
enhancers. Gastric feeding is associated with a very
high incidence of aspiration and great care must be
taken to avoid this. These problems are alleviated by
using sites for feeding distal to the pylorus and in our
view enteral nutrition should be used whenever
Table 16.9
Vitamin
A
D
E
K
B1 (thiamin)
B2 (riboflavin)
B6 (pyridoxine)
Niacin
Folic acid
B12
Biotin
Pantothenic acid
Ascorbic acid
RDA (mg)
MVI-12(mg)
1
0.005
10
0.1
1.5
1.6
2.2
18
0.4
0.003
0.05
5
60
1
0.005
10
3
3.6
4
40
0.4
0.005
0.06
15
100
MVI-12 does not contain vitamin K
possible. In many patients the gut provides an

adequate portal of entry for nutritional support at a
much reduced risk. By keeping bacteria and toxins
within its lumen, the gut has a role in host defense. In
states of metabolic stress, especially coupled with gut
starvation, mucosal permeability increases, allowing
movement of enteric bacteria to the regional lymph
nodes and thence to the systemic circulation.
Proven benefits of enteral feeding include maintenance of an intestinal barrier and better immune
function, improved liver blood flow following shock
and better wound healing. Gut feeding can be started
early even in the presence of a mild ileus, although
early feeding does not attenuate the metabolic
response to blunt trauma (Eyer et al., 1993). Exercise,
even when done passively, is a potent anabolic
stimulus and retards nitrogen mobilization in critically ill and injured patients. While vigorous nutritional support appears to affect mortality favorably
and speed recovery from head injury, nutrient administration alone cannot override the metabolic response
to head injury (Clifton et al., 1984; Young et al., 1985).
Despite adequate caloric administration, severe head
Usual daily dose range of parenteral nutrient intake for the adult
Nutrient
Water
Calories
Glucose
Protein equivalent (as amino acids)
Na+
K+
Ca2+
Mg2+
PO4
Amount
Comment
25003500 ml
25003500
600900 g
60130 g
100120 mmol
80120 mmol
410 mmol
1215 mmol
1015 mmol
3040 ml/kg/d
3040 kcal/kg/d
12 g/kg/d
with GI losses; in elderly, CCF
with renal failure
with GI losses
when glucose alone is given; with renal failure
DISORDERS OF WATER AND ELECTROLYTE BALANCE
injury is followed by negative nitrogen balance,

weight loss and depression of immune status and of
plasma protein levels.
Table 16.10 Pathogenesis of hyponatremia (ADH =

antidiuretic hormone; GI = gastrointestinal)
(d)
Nutrition and outcome
Since protein-calorie malnutrition can adversely affect

the structure and function of many organs, including
the lungs (Arora and Rochester, 1982), heart (Heymsfield et al., 1978), gastrointestinal tract (James, 1971)
and musculoskeletal system (Jeejeebhoy, 1986) as well
as wound healing (Haydock and Hill, 1987) and
immune function, nutritional support might be expected to have a beneficial effect on mortality and
morbidity following head injury.
However, although there are studies that suggest
that nutritional support is beneficial in supporting
general metabolism following head injury (McMahon
et al., 1993), a significant relationship between nutrient
intake and outcome has not been proved.
16.8 Disorders of water and electrolyte

balance
Hypovolemia (from aggressive use of diuretics),
hyperosmolarity (from osmotic diuretics, hyperglycemia resulting in glycosuria, and diabetes insipidus)
and disorders of sodium metabolism (especially the
syndrome of inappropriate ADH secretion SIADH)
are common problems in head-injured patients. These
and other disorders may cause secondary brain injury
through their effects on cerebral perfusion, intracranial pressure and neuronal function.
16.8.1
HYPONATREMIA
Hyponatremia (plasma Na+ < 135 mmol/l) is seen in

512% of patients with severe head injury (Steinbok
and Thompson, 1978). It is especially important,
since the osmotic buffering capacity of the brain may
be impaired following primary brain injuries. ECF
osmolality is reduced in all cases of hyponatremia in
the absence of abnormal solute accumulation, and
even small decreases in osmolality may cause brain
edema, impairing brain function and increasing ICP.
The pathogenesis of hyponatremia is summarized in
Table 16.10.
Hyponatremia may occur in association with different levels of body fluid tonicity (Table 16.11) and
so be classified as isotonic, hypertonic or hypotonic,
based on the measured plasma osmolality. Hyponatremia following head injury is hypotonic and is
usually thought to be due to SIADH. The risk of
hyponatremia seems greater in those with severe
head injuries, chronic subdural hematoma and basal
skull fracture (Steinbok and Thompson, 1978; Doczi
et al., 1982). There are numerous reports alleging that

315
Shift of water from the cell (e.g. hyperglycemia,

mannitol infusion)
Shift of sodium into the cell (e.g. potassium loss)
Retention of water (e.g. syndrome of inappropriate
ADH secretion, edematous states)
Loss of sodium (e.g. through kidneys, GI tract, skin)
this type of hyponatremia is due to renal salt wasting

caused by an unknown humoral substance released
in response to cerebral injury (cerebral salt wasting);
however, the evidence is not strong and the reported
cases most probably represent SIADH (Oh and Carroll, 1992). Deterioration in level of consciousness,
new focal deficits, myoclonus, seizures or increasing
ICP should alert the physician to the possibility of
hyponatremia and hypo-osmolality in the patient
with a severe brain injury.
(a)
Isotonic hyponatremia
An apparent decrease in serum sodium concentration

(measured by flame photometry) occurs when the
mass of the non-aqueous components of serum is
increased by severe hyperlipidemia or hyperproteinemia. This is usually referred to as pseudo-hyponatremia and can be readily identified by finding a
normal serum osmolality (Weisberg, 1989).
(b)
Hypertonic hyponatremia
In patients with an increased amount of an impermeant solute (i.e. one that is unable to cross the blood
Table 16.11
Common causes of hyponatremia
Isotonic
Pseudohyponatremia
Hyperlipidemia
Hyperproteinemia
Hypertonic
Hyperglycemia
Mannitol, glycerol or sorbitol excess
Hypotonic
Water retention
Syndrome of inappropriate antidiuretic hormone
secretion
Syndrome of inappropriate antidiuresis
Psychogenic polydipsia
Transurethral resection of the prostate (TURP)
syndrome
Salt depletion
Adrenocortical failure
Diuretic excess
Potassium depletion
316
brain barrier), such as glucose, mannitol, glycerol or

sorbitol, osmotic equilibration occurs by water shifting from the ICF to the ECF, thus diluting the ECF
sodium. In such circumstances, hyponatremia is often
associated with an elevated measured osmolality. For
every 1 mmol/l rise in glucose above a plasma glucose
level of 5.6 mmol/l, the plasma sodium decreases by
0.288 mmol/l (Katz, 1973; Crandall, 1974; Jenkins and
Larmore, 1974; Robin et al., 1979). The corrected
sodium value is derived from the formula:
Corrected Na value = measured plasma Na+
(plasma glucose value
5.6) 0.288.
(c)
Hypotonic hyponatremia
Hyponatremia is almost always caused by an excess of

total body water. This may be due to excessive
infusion of hypotonic or water-generating intravenous
fluids such as 5% dextrose, 1.5% glycine or 0.45%
saline, particularly when administered rapidly or in
the presence of high circulating ADH levels. Hyponatremia may rarely be due to loss of exchangeable
sodium or potassium (Fuisz, 1963).
Since hyponatremia may be associated with an
alteration in both total body water and total body
solute the ECF volume may be increased (hypervolemia), decreased (hypovolemia) or unchanged
(isovolemia; Humes, 1986).
In health, a fluid intake of up to 1520 liters may be
tolerated before water is retained. However, less water
is needed to produce hyponatremia in individuals in
whom there is also an increase in non-osmotic
stimulation of ADH by hypovolemia, hypotension,
pain, nausea or postoperative stress (Hayes, 1968;
Chung et al.,1986; Sinnatamby et al., 1974), or a
reduction in urinary osmotic excretion, as in patients
with a reduced urea production, or a reduction in
renal capacity to respond to ADH.
The proposal that disease may cause a sick cell
syndrome (Flear and Singh, 1973; Flear, Gill and Burn,
1981) in which ECF sodium leaks into cells and
produces hyponatremia is a misconception (Leaf,
1974; Bichet and Schrier, 1982).
Loss of potassium may cause hyponatremia as
sodium shifts into the cell in exchange for K+.
Syndrome of inappropriate ADH secretion (SIADH)
This syndrome is defined as hyponatremia due to an
elevated level of ADH inappropriate for the prevailing
osmotic or volume stimuli (Robinson, 1985; Bartter
and Schwartz, 1967). ADH secretion from the neurohypophysis is no longer under normal regulatory
influences. The criteria for the diagnosis of SIADH are
listed in Table 16.12.
Table 16.12 Criteria for the diagnosis of the syndrome of

inappropriate secretion of antidiuretic hormome (SIADH)

Hypotonic hyponatremia
Urine osmolality greater than plasma osmolality
Urine sodium excretion greater than 20 mmol/l
Normal renal, hepatic, cardiac, pituitary, adrenal and
thyroid function
Absence of hypotension, hypovolemia, edema and
drugs
Correction by water restriction
SIADH is a form of dilutional hyponatremia; ECF

volume is usually increased by 34 liters and peripheral edema does not occur. Because of the expanded
ECF volume, glomerular filtration rate is increased
and the reninangiotensinaldosterone mechanism is
suppressed, resulting in a decrease in the renal
absorption of sodium.
Syndrome of inappropriate antidiuresis (SIAD)
This syndrome is defined as hyponatremia due either
to increased ADH secretion initiated by stimulation of
high- or low-pressure baroreceptors, or to a reduction
in non-ADH renal water excretion mechanisms
(Robertson, 1989). The causes of SIAD are listed in
Table 16.13.
(d)
Clinical features of hyponatremia
Hypotonic hyponatremia is almost always caused by

excess total body water, and the symptoms are due
Table 16.13 Causes of syndrome of inappropriate

antidiuresis
Baroreceptor-mediated
Hypovolemia
Addisons disease
Hepatic failure
Nephrotic syndrome
Drugs: frusemide, thiazides
Cardiac failure
Cortical influences
Pain, anxiety and nausea
Postoperative
Trauma
Drugs
Narcotics, barbiturates, chlorpropamide, phenothiazines,
carbamazepine, tricyclics, nicotine derivatives, clofibrate,
vincristine, vinblastine, cyclophosphamide
Reduction in renal response to antidiuretic hormone
Renal failure
Hypothyroidism
Addisons disease
Drugs: indomethacin, frusemide, thiazides
to the cerebral water excess (cerebral edema; Wasterlain and Posner, 1968). Normally the brain partially
adapts to the hypo-osmolality within 24 hours,
reducing the cerebral water excess by losing or
inactivating intracellular osmotically active solutes
(Arieff, 1984, 1985; Arieff and Guisado, 1976; Arieff,
Liach and Massry, 1976).
If the patient develops hyponatremia of 125 mmol/l
(osmolality 260 mosmol/kg) acutely, that is within 3
days, then symptoms of cerebral edema usually occur.
These include headache, anorexia, nausea, weakness,
lethargy, confusion, disorientation, blurred vision,
muscle cramps, coma and seizures (Arieff, 1984;
Arieff, Liach and Massry, 1976; Plum and Posner,
1980).
Chronic hyponatremia, on the other hand, with
plasma sodium values well below 125 mmol/l, may be
well tolerated because of cerebral osmotic compensation (Arieff, 1984). Thus the mortality associated with
hyponatremia is usually related to its rate of development rather than its severity. Chronic hyponatremia
has a mortality of less than 10% (Sterns, 1987),
whereas a mortality up to 50% has been reported with
acute hyponatremia (Robertson, 1989).
(e)
Diagnosis of hyponatremia
The evaluation of hyponatremia with hypo-osmolality

begins with a clinical assessment of volume status and
measurement of urinary indices (Table 16.14). However, volume status can be difficult to assess clinically
in a critically ill patient. ECF volume may be affected
by blood loss, the amount and type of fluid administered and the use of diuretics. Urinary Na+ measurements are affected by the use of osmotic and nonosmotic diuretics. Figure 16.13 summarizes our
diagnostic approach to hyponatremia.
Acute hyponatremia is usually due to excess water,
as in dextrose solutions administered inappropriately
to patients in the postoperative period. Diagnosis is
based on measurement of plasma osmolality, osmolar
gap and urinary sodium. The osmolar gap is the
Table 16.14 Extracellular volume (ECV) and urinary

sodium (uNa+) in hyponatremia and hypo-osmolality
ECV
uNa+
(mmol/l)
Dehydration
Decreased
< 10
Congestive heart failure, cirrhosis
Increased
< 10
Renal disease, adrenal insufficiency,

diuretics
Decreased
> 25
SIADH
Normal or
increased
> 25
Diagnosis
317
difference between the measured osmolality and the

calculated osmolality (2 Na+ + glucose + urea, where
plasma sodium, glucose and urea are measured in
millimoles per liter; Worthley, Guerin and Pain, 1987).
When there is an excess of total body water, urinary
sodium is usually greater than 40 mmol/l and plasma
urea is low. In the rare case of hyponatremia due to
salt depletion, urinary sodium is usually less than
20 mmol/l and the plasma urea and uric acid are often
high (Humes, 1986). In patients with head injuries
hyponatremia will almost always be associated with a
urinary osmolality of greater than 100 mosmol/kg.
The presumptive diagnosis of SIADH can be confirmed if simply restricting fluids results in a reduction of urinary Na+ losses and correction of hyponatremia. The clinical picture most often confused with
SIADH is a patient with isotonic fluid loss, usually
due to diuretics, who is treated with hypotonic fluid
replacement. In this situation, the body preserves
volume at the expense of tonicity a form of
hypotonic dehydration. It should be stressed that the
absence of an increase in ECF volume excludes a
diagnosis of SIADH.
Chronic hyponatremia accompanies other diseases
and it is important to differentiate SIADH from
sodium depletion due to prolonged vomiting, adrenal
insufficiency or hypopituitarism.
(f)
Treatment of hyponatremia
Treatment aims to remove excess water and treat of

the underlying cause.
Chronic hyponatremia (> 3 days duration)
If the patient is asymptomatic and hyponatremia has
been present for more than 3 days, fluid restriction
and removal of any precipitating factor is often all that
is required (Cluitmans and Meinders, 1990). Fluid
should be restricted to less than 500 ml/d. The rate of
correction should be no greater than 12 mmol/l/d
(0.5 mmol/l/h) and is continued only until the plasma
sodium is 130 mmol/l (Laureno and Karp, 1988;
Sterns, Riggs and Schochet, 1986; Plum and Posner,
1980; Cluitmans and Meinders, 1990; Norenberg and
Papendick, 1984). If fluid deprivation is difficult to
sustain in patients with SIADH or SIAD then patients
with hyponatremia and chronic CCF may benefit from
an ACE inhibitor added to frusemide (furosamide).
This will inhibit the stimulation of thirst and ADH
release by angiotensin II (Dzau and Hollenberg, 1984;
Packer, Medina and Yushak, 1984). However, in these
patients a direct ADH inhibitor such as phenytoin may
be of greater value (Mulinari et al., 1990). This has been
used to reduce ADH release from the hypophysis in
318
Figure 16.13 Diagnostic approach to hyponatremia. Arrows indicate the direction of change; single and double arrows
indicate the magnitude of change. Iso = isotonic; N = normal; V = variable. In hypertonic hyponatremia due to
hyperglycemia, the corrected plasma Na+ = plasma Na+ + (plasma glucose 10)/3 mmol/l.
patients with SIADH due to CNS disorders including

head injury (Martinez-Maldonado, 1980).
Complications of treatment of chronic hyponatremia
In chronic hyponatremia the brain is less able to
tolerate rapid correction and may develop central
pontine and extrapontine myelinolysis if hypertonic
saline is used. The lesions of central pontine myelinolysis (CPM) are caused by the destruction of myelin
sheaths in the center of the basilar portion of the pons
and may extend from the midbrain to the lower pons.
The clinical features range from coma, flaccid quadriplegia, facial weakness and pseudobulbar palsy to
minor behavioral changes without focal findings. The
onset may be from one to several days after the

hyponatremia has been corrected and may require
MRI to confirm the diagnosis (Brunner et al., 1990).
While an association with CPM and correction of
hyponatremia has been reported, the relationship
between CPM and the rapidity of correction of
hyponatremia has not been firmly established in
clinical practice. In 170 cases of CPM, hyponatremia
occurred in only 28% of patients, most of whom were
corrected slowly (Ayus, Krothapalli and Arieff, 1985),
an observation that has been confirmed by others
(Tormey, 1990). Nevertheless, there is persuasive
experimental evidence for the association between
rapidity of correction of chronic hyponatremia and
CPM (Norenberg and Papendick, 1984).
319
Acute hyponatremia (< 3 days duration)
Complications of treatment of acute hyponatremia
In this condition, the rate of reduction of plasma

sodium is 0.5 mmol/l/h or greater and/or the hyponatremia is of less than 3 days duration (Cluitmans
and Meinders, 1990). It is often associated with
inappropriate intravenous fluid administration and
with the postoperative period (Arieff, 1986). The total
water excess in adults ranges from 510 liters.
The rate of correction of acute hyponatremia should
be no greater than 2 mmol/l/h of sodium until the
plasma level has increased to 120 mmol/l or by a
maximum of 20 mmol/l during the first 24 hours
(Arieff and Guisado, 1976; Ayus, Krothapalli and
Arieff, 1987, 1988).
This is achieved initially by intravenous administration of hypertonic saline given at 5070 mmol/h. It is
important to note that the purpose of using hypertonic
saline is not to correct a saline deficit, since there is in
fact not a total Na+ deficiency, but rather the hypertonic saline draws water into the intravascular compartment and reduces brain edema. A spontaneous or
frusemide-induced diuresis (e.g. by 2040 mg intravenously) is then required to excrete the water excess.
When the plasma sodium has increased to 120 mmol/
l, precautions should be taken to prevent the plasma
sodium rising to greater than 130 mmol/l over the
next 24 hours. If the hyponatremia presents with
convulsions, then urgent correction of the cerebral
edema using 250 mmol of hypertonic sodium chloride
over 10 minutes (equivalent to 500 ml of 20% mannitol), has been used. This will immediately elevate
the plasma sodium in adults by about 7 mmol/l (Table
16 15; Worthley and Thomas, 1986).
The complications reported with the use of hypertonic

saline include congestive cardiac failure, cerebral
hemorrhage intracerebral and subdural, presumed
due to tearing of bridging veins and CPM. To reduce
the incidence of congestive cardiac failure, CVP or
PAWP should be monitored throughout saline administration. Cerebral hemorrhage will only occur if there
is an inappropriate administration of hypertonic
saline in normonatremic states (Finberg, Luttrell and
Redd, 1959).
Table 16.15 Treatment of severe hyponatremia (ACE =

angiotensin-converting enzyme; ADH = antidiuretic
hormone)
Acute hyponatremia
Fluid restrict to 500 ml/d or less
Hypertonic saline (5070 mmol/h)
Diuresis of 160 ml/h or greater
Rate of correction
First day: no greater than 20 mmol/l/d (2 mmol/l/h)
until plasma sodium 120 mmol/l
Second day: attempt to keep plasma sodium between
130 and 135 mmol/l
Seizures: 100250 mmol hypertonic saline over 10 min
Chronic hyponatremia
Fluid restrict to 500 ml/d or less
Rate of correction:
No greater than 12 mmol/l/d (0.5 mmol/l/h) until
the plasma sodium is 130 mmol/l
Drugs
Lithium, demeclocycline
Phenytoin
ACE inhibitor
ADH inhibitor
16.8.2
HYPERNATREMIA
Hypernatremia is defined as a plasma sodium greater

than 145 mmol/l. It is always associated with hyperosmolality and may be caused by water depletion,
excessive administration of sodium salts or both
(Table 16.16).
In water-depleted states, the thirst mechanism
normally stimulates a conscious individual to drink
water. Thus water loss only produces hypernatremia if
the patient is unconscious or otherwise physically
unable to drink. Excessive administration of sodium
salts is a rare cause of hypernatremia and usually only
occurs through therapeutic misadventure.
Hypernatremia is the usual cause of hyperosmolar
states in neurosurgical patients, although marked
Table 16.16
Causes of hypernatremia
Water depletion
Extrarenal loss
Exposure
GIT losses
Renal loss
Osmotic diuresis
Urea, mannitol, glycosuria
Diabetes insipidus
Central (neurogenic)
Post-traumatic, postoperative
Metastatic tumors, craniopharyngioma, pinealoma,
cysts
Meningitis, encephalitis
Fat embolism
Granulomas (TB sarcoid)
Idiopathic
Nephrogenic
Congenital
Hypercalcemia
Hypokalemia
Drugs: lithium, amphotericin B
Renal diseases: chronic pyelonephritis; medullary
sponge kidney; polycystic kidney; analgesic
nephropathy; postobstructive uropathy; multiple
myeloma; amyloid; sarcoid
Salt gain
Hypertonic saline or sodium bicarbonate
320
hyperglycemia may contribute to hyperosmolality in

states of insulin resistance, such as in sepsis, or as a
result of inadequate insulin administration with parenteral nutrition.
(a)
Clinical features
Hypernatremia is usually symptomatic if plasma

sodium is 155160 mmol/l or greater (equivalent to an
osmolality of 330 mosmol/kg or greater). This equates
to a water depletion of 6 liters or greater in a 70 kg
man (Arieff, 1984). Clinical features include pyrexia,
restlessness, weakness, paralysis, irritability, drowsiness, lethargy, confusion, tremor, hyper-reflexia, seizures and coma (Ross and Christie, 1969). The signs of
hypernatremia are due mainly to increased effective
osmolality, which reduces brain cell volume (Oh and
Carroll, 1992). As already noted, hypernatremia is
much better tolerated if it develops slowly. Increase of
the brain volume may be accomplished by intracellular shift of sodium, potassium and chloride, and
the accumulation of organic solutes, mainly taurine
and other amino acids (Lohr, McReynolds and Grimaldi, 1988). Normalizing brain volume in chronic
hypernatremia therefore requires an increase in the
total solute content of the brain. A sudden lowering of
the osmolality of the ECF may cause a shift of water
into the brain cells with resultant cerebral edema.
(b)
Diabetes insipidus
Diabetes insipidus (DI) is characterized by polyuria

(315 l/d) and polydipsia due to the complete or
partial failure of ADH secretion (central or neurogenic
DI), or decrease in the renal response to ADH
(nephrogenic DI; Table 16.16). Both situations result in
an inability to concentrate urine and to conserve
Figure 16.14
solute-free water in the face of appropriate stimuli.

Head injury is a common cause of central DI. A very
early onset is characteristic of major hypothalamic
damage and is associated with a high mortality (Seckl,
Dunger and Lightman, 1987).
Head-injured patients with fractures involving the
base of the skull and sella turcica appear to be at
increased risk of DI (Crompton, 1971; Edwards and
Clark, 1986). Diabetes insipidus is common in
patients prior to brain death. Traumatic DI is thought
to be caused by a shearing injury to the pituitary
stalk. The time of onset is variable, but it usually
appears after 510 days (Kern and Meislin, 1984). In
most cases of post-traumatic DI the onset is characterized by polyuria, hypernatremia and plasma
hyperosmolality, sometimes as early as 1224 hours
after injury. If the damage is limited to the pituitary
or lower pituitary stalk, the DI may only be transient, and this is usually the case. High stalk lesions
or injury to the hypothalamus usually cause permanent DI. Since most patients in ICU are unable to
control their own water intake, hemodynamic instability may occur with untreated DI.
Nephrogenic DI is often mild because the patient
can usually produce an isotonic urine; thus the
diuresis is limited to 35 l/d.
(c)
Diagnosis
Every patient with hypernatremia may be considered

to have an inadequate water intake, either absolute or
relative. Excessive sodium administration as the cause
of hypernatremia is usually obvious from the history.
Whether hypernatremia is due to insufficient water
intake or to excessive water loss can be determined by
measurements of urine osmolality (Figure 16.14).
Normal concentration of urine (urine osmolality
Differential diagnosis of hypernatremia. Posm = plasma osmolality; DI = diabetes insipidus.
> 700 mosmol/l) suggests that the main problem is

insufficient water intake, with or without excessive
extrarenal water loss. Urine osmolality between
700 mosmol/l and the osmolality of plasma suggests
partial central DI, osmotic diuresis, diuretic therapy,
acquired (partial) nephrogenic DI or renal failure. A
high (> 30 ml/kg/h) urine volume may be the first sign
of diabetes insipidus, but is also seen in osmotic
diuresis. A urine osmolality below that of plasma
suggests either complete central DI or nephrogenic DI.
Laboratory investigations show hypernatremia, persistently hypotonic urine and low urine osmolality,
usually less than 200 mosmol/l. In milder cases urine
osmolality may not be below that of plasma, and may
be 300600 mosmol/l. Patients with central DI are
unable to increase urinary osmolality during fluid
restriction, but remain sensitive to exogenous ADH.
Patients with partial DI have urinary volumes much
less than those with complete DI. Other causes of
polyuria which should be considered in the headinjured patient include osmotic diuretics (mannitol,
iodinated contrast-medium), severe hyperglycemia
and fluid overload. In contrast to DI, a solute diuresis
is usually accompanied by a higher urinary osmolality
(between 250 and 320 mosmol/l; Shucart and Jackson,
1976).
(d)
Treatment of hypernatremia
Hypotension due to a reduction in intravascular fluid

volume should be treated by isotonic saline infusions
to replace the intravascular volume and improve
tissue perfusion before hypotonic solutions are used.
Pure water depletion is treated by water administration. (In a conscious patient with normal gastrointestinal function, oral ingestion of water is encouraged.) If intravenous fluid is required, 5% dextrose or
hypotonic saline solutions (e.g. 0.45% saline) are often
used, as sterile water through a peripheral vein causes
hemolysis (Krumbhaar, 1914). In rare cases, where
saline and dextrose solutions are deemed inadvisable;
for example, in a patient who is volume-depleted and
severely hypernatremic, central venous administration of sterile water can be used without causing
hemolysis (Worthley, 1986).
The rate of increase in osmolality should be no
greater than 2 mosmol/kg/h (Arieff, 1984; Arieff and
Guisado, 1976). Hypernatremia due to excess sodium
is treated by water and diuretic administration
(Addleman, Pollard and Grossman, 1985). The management of DI depends on the cause (Table 16.17).
(e)
Hyperosmolar therapy
Intravenous 20% mannitol (1098 mosmol/l) a sixcarbon sugar similar to glucose, is often used to
321
Table 16.17 Treatment of diabetes insipidus (DI) in the

critically ill ICU patient
Remarks
1 Confirm diagnosis
Altered consious state or inability

to detect thirst or take oral fluids.
Hypertonicity (> 295 mosmol/kg),
hypernatremia, inappropriately
dilute urine. DI may be transient
or permanent; permanent form
seen in one-third of patients with
trauma or neurosurgery. Presence
of associated anterior pituitary
hypofunction not uncommon in
patients with severe head trauma.
2 Replace free water

deficit
Free H2O deficit (liters) =

0.6 (body weight [kg]) ([Na+/
140] 1)
Parenteral replacement of half of
deficit immediately; rest over
2436 hours. Use 5% dextrose, not
saline solutions
3 Give maintenance
fluids and replace
urine output
hourly with
dextrose solutions
Labor-intensive; inaccurate fluid

balance may occur with high
volumes of urine
4 Daily weight
Not always available in adult ICU
5 Close monitoring
of electrolytes and
fluid balance
6 Aqueous
vasopressin (AVP)
If urine output remains excessive

(> 200250 ml/h) in the absence of
diuretics or, if maintenance of fluid
balance is difficult or
hyperosmolality is present give
510 U subcutaneously or i.m.
7 Desmopressin
(DDAVP) for
patients with
severe DI
More recent regimen. Synthetic

analog of AVP with longer half-life
and fewer vasoconstrictive effects.
Give 12 g subcutaneously or i.v.
every 1224 h or by nasal
insufflation of 520 g every 12 h.
Watch for water intoxication
(hyponatremia and elevated
urinary osmolality) during
therapy; withhold DDAVP
periodically to document
persistence of DI.
lower ICP. The aim is not to induce hypovolemia,

which may be detrimental, but rather to maintain a
normovolemic state while increasing serum sodium,
osmolality and tonicity, thus decreasing the extracellular fluid volume within the brain. A serum
sodium level of 145155 mmol/l and serum osmolality of 300320 mosmol/l are often generated, and
excessive dehydration may precipitate prerenal renal
failure. Fluids are administered to maintain CVP
322
between 5 and 10 mmHg. In young, otherwise fit

head-injured patients, pulmonary artery catheterization is rarely indicated.
Mannitol is administered as an intermittent intravenous bolus (0.250.5 g/kg body weight) up to
4-hourly. Where the bloodbrain barrier is intact,
mannitol remains within the intravascular and extracellular spaces (Sutin, Ruskin and Kaufman, 1992). It
has several effects, including an immediate increase in
circulating blood volume and arterial blood pressure
(Wise and Chater, 1962; Roberts et al., 1987). It also
reduces blood viscosity by hemodilution (Burke et al. ,
1981; Muizelaar et al., 1983) and by increasing red
blood cell deformability. It causes osmotic dehydration of the brain, which reduces the volume of
extracellular free water (Roberts et al., 1987), thus
decreasing brain volume and ICP as well as causing a
urinary osmotic diuresis. Repeated administration of
mannitol may result in a systemic hyperosmolar state
and dehydration, which should be avoided since
cerebral perfusion may decrease and renal failure may
occur. Repeated administration of mannitol may result
in its eventual movement through the bloodbrain
barrier into the extracellular space, where the
increased oncotic pressure may retain free water and
worsen edema. Thus mannitol use should be restricted
to the minimum amount needed to control ICP and
CPP. Our practice is to measure serum osmolality
daily and cease mannitol if serum osmolality exceeds
320 mosmol/l.
Mannitol induces diuresis primarily by elevating
the osmotic pressure of the glomerular filtrate to such
an extent that the tubular reabsorption of water and
solutes is hindered. It promotes the excretion of
sodium; however, proportionately more water than
sodium is excreted. Excretion of potassium, chloride,
calcium, phosphorus, magnesium, urea and uric acid
are also increased. Mannitol is only very slightly
metabolized, if at all, to glycogen in the liver. It is
freely filtered by the glomeruli, with less than 10%
tubular reabsorption, and is not secreted by tubular
cells. The elimination half life in adults is about 100
minutes. Approximately 80% of a 100 g dose is
excreted unchanged in the urine within 3 hours. In the
presence of renal disease in which glomerular function
is impaired or in conditions that impair small vessel
circulation, such as congestive cardiac failure, cirrhosis with ascites, shock or dehydration, mannitol
clearance is lower than normal. Mannitol does not
cross the intact bloodbrain barrier unless very high
concentrations are present in the plasma or the patient
has acidosis.
Hypertonic saline (7.5%) has also been used to
lower ICP in animal models of head injury (Freshman
et al., 1993). It was as effective as 20% mannitol in
treating elevated ICP caused by a space-occupying
lesion. Hypertonic saline has the additional benefit of

rapid small volume resuscitation of associated hemorrhagic shock (section 16.5.2).
16.8.3
HYPOKALEMIA
Hypokalemia is defined as a serum potassium of less

than 3.5 mmol/l or plasma potassium less than
3.0 mmol/l. It may be due to a reduction in total
body potassium, caused by a decreased oral intake,
increased renal or gastrointestinal loss, or to a compartmental shift of potassium from the ECF to the
ICF (Tannen, 1986). In contrast to Na+ and C1,
which are predominantly in the ECF and regulated
by the kidneys, K+ is predominantly in the ICF; its
movements are primarily determined by nutritional
and metabolic factors (e.g. pH) and are mediated by
neural and hormonal influences (Schultze and Nissenson, 1980).
Hypokalemia may be caused by osmotic diuretics,
diabetes insipidus, steroids, catabolic losses in
trauma and secondary hyperaldosteronism (commonly caused by hypovolemia or cardiac failure).
Excessive renal K+ loss also occurs in the presence of
alkalemia and hypomagnesemia. Gastrointestinal K+
losses may be due to vomiting, nasogastric suction or
diarrhea. Intracellular shift of K+ occurs in metabolic
or respiratory alkalosis, exogenous insulin administration and endogenous or exogenous catecholamine administration.
(a)
Clinical features
The clinical features of hypokalemia are listed in

Table 16.18. These include ventricular and supraventricular arrhythmias, ileus, enhancement of digitalis
toxicity and hypokalemic nephropathy; nephropathy
occurs if hypokalemia is severe and prolonged and
results in impaired renal water conservation and
polyuria. If K+ depletion is severe neuromuscular
manifestations include weakness, hyporeflexia and
even paralysis. The muscular weakness can interfere
with attempts to wean patients from ventilatory
support in the ICU.
(b)
Treatment
Treatment of hypokalemia is usually commenced

when plasma K+ is less than 3.0 mmol/l. While there is
normally a linear relationship between serum K+ and
total body K+ (Sterns et al., 1981), acidbase disturbances and other causes of compartmental K+ shift
make estimates of total body K+ difficult in severely ill
patients. Thus K+ therapy must be monitored by
frequent estimations of serum K+. Serum magnesium
levels should be measured in patients with refractory
Table 16.18
Clinical features of hypokalemia
Cardiac
Predisposition to digoxin toxicity
Ventricular and atrial tachycardias
Torsades de pointes
ECG changes: PR prolongation, T-wave inversion,
prominent U waves
Skeletal muscle
Weakness, hypotonicity
Ascending paralysis, ventilatory failure
Cramps, rhabdomyolysis
compartment shift, or may be artifactual, due to

hemolysis after collection.
Hyperkalemia is uncommon in patients with an
isolated head injury unless overzealous use of osmotic
diuretics has caused renal failure. Apparent hyperkalemia may be seen with hemolysis or acidosis. In
patients with additional trauma, rhabdomyolysis and
massive transfusion of old blood may result in
hyperkalemia.
Gastointesinal
Constipation, ileus
(a)
Central nervous system

Depression, coma
Hyporeninemic hypoaldosteronism
Renal
Nephrogenic diabetes insipidus (polyuria, polydipsia)
Metabolic alkalosis
Interstitial nephritis
Predisposition to urinary tract infection
Endocrine
Glucose intolerance
hypokalemia, since both electrolytes may be disturbed

together and hypokalemia may be difficult to correct
without concomitant correction of hypomagnesemia
(Tannen, 1986). Hypokalemia protects against hypocalcemic tetany; thus hypocalcemia should be corrected before or together with the correction of
hypokalemia.
When hypokalemia is associated with metabolic
alkalosis the standard treatment is intravenous or oral
potassium chloride. If the patient has renal tubular
acidosis and hypokalemia, then potassium acetate is
given. If phosphate depletion is also present, potassium phosphate may be administered. Phosphate can
only be infused slowly (at a maximum rate of
24 mmol/h/70 kg ) and is usually given to correct a
phosphate deficiency rather than potassium deficiency; thus potassium chloride is often given
concomitantly.
Intravenous replacement of potassium should not
normally exceed 40 mmol/h. Plasma potassium levels
should be monitored at 26-hourly intervals (Tannen,
1986; Stockigt, 1977). Spironolactone, amiloride or
ACE inhibitors are more useful for preventing renal
potassium loss than for correcting an existing deficit
(Tannen, 1986).
16.8.4
HYPERKALEMIA
Hyperkalemia is defined as a serum potassium greater

than 5.0 mmol/l or plasma potassium greater than
4.5 mmol/l. It may be due to excessive intake, severe
tissue damage, decreased excretion or body fluid
323
Hypoaldosteronism (Barratt, 1981)
This usually occurs in patients who have non-insulindependent diabetes mellitus and mild renal failure,
and may be provoked by cyclo-oxygenase inhibitors.
The patient is hyperkalemic with metabolic acidosis
(Type 4 renal tubular acidosis, RTA) and has an
elevated plasma creatinine and urea nitrogen. The
etiology is not known, although hyporeninemia, prostacyclin deficiency and excess atrial natriuretic hormone have all been implicated. Non-steroidal antiinflammatory drugs (NSAIDs), calcium-channel
blockers and beta-adrenergic blockers should not be
administered as they will contribute to the hyporeninemic state (Williams, 1986).
Primary hypoaldosteronism
This is characterized by normal or high renin levels
and is due to an enzymatic defect in the aldosterone
pathway.
Pseudo-hypoaldosteronism
This is characterized by normal or high levels of
aldosterone with a renal tubular resistance to the
action of the hormone. This disorder is commonly
caused by spironolactone and amiloride. The suppression of aldosterone production with hyperkalemia
(Edes and Sunderrajan, 1985) described with heparin
use is due to the antiseptic in commercial heparin
solutions and not to heparin (Jaques, 1985).
(b)
Clinical features of hyperkalemia
The clinical features of hyperkalemia include tingling,

paresthesias, weakness, flaccid paralysis and hypotension. ECG changes include T-wave peaking, P-wave
flattening, PR prolongation (until sinus arrest with
nodal rhythm occurs), QRS widening, shortened QT
interval and deep S wave. Finally, a sine-wave ECG
pattern develops, which deteriorates to asystole,
occurring at plasma potassium levels of 7 mmol/l or
greater.
324
(c)
Treatment
+
For mild hyperkalemia (i.e. plasma K > 4.5 and

< 5.5 mmol/l) treatment usually involves reducing the
potassium intake, correcting renal failure, and treating
the cause (e.g. ceasing NSAIDs or spironolactone).
Treatment of severe hyperkalemia (i.e. plasma
K+ > 5.5 mmol/l) often involves the administration of
intravenous glucose and insulin, sodium bicarbonate
and calcium chloride (Table 16.19; Tannen, 1986),
although nebulized salbutamol (1020 mg) may also
be useful as it can reduce the plasma potassium by
0.61.0 mmol/l after 30 minutes and will last for 2
hours (Williams, 1986). Resonium A has the capacity
to bind 3.1 mmol of potassium per gram by releasing
3.1 mmol of sodium. Calcium chloride (or gluconate)
does not reduce the plasma potassium and is used
only to counteract the toxic cardiac effects of hyperkalemia. Since all the methods of reducing the
potassium level are only temporary, treatment should
also be directed at the underlying cause, and may also
include dialysis.
16.8.5
HYPOCALCEMIA
Hypocalcemia occurs in up to 70% of patients in ICU

(Desai, Carlson and Geheb, 1988; Zaloga and Chernow, 1986). It is defined as a total plasma calcium less
than 2.2 mmol/l and may be caused by either a
reduced ionized or protein-bound calcium. In critically ill patients calcium loss due to increased tissue
sequestration of calcium may be associated with
pancreatitis, septicemia, burns or toxic shock (Chernow et al., 1982a; Zaloga, 1991). However, hypocalcemia is only important clinically when the physiologically active, ionized fraction (40% of total) is
reduced. Ionized calcium is lowered by alkalosis, by
high circulating free fatty acids, as in pancreatitis or
sepsis, and by agents that chelate calcium, such as
albumin infusions, bicarbonate administration and
high citrate loads administered with massive blood
transfusion (Zaloga and Chernow, 1986). Ionized
hypocalcemia (less than 1.15 mmol/l) occurs in
Table 16.19
1520% of patients in the ICU (Zaloga, Chernow and

Cook, 1985) but cannot be predicted by estimation of
either total serum calcium or of ionized calcium
corrected for pH and serum albumin concentration
(Desai, Carlson and Geheb, 1987).
Common causes of hypocalcemia in head-injured
patients include respiratory alkalosis, sepsis, hypomagnesemia and rhabdomyolysis in the polytrauma
patient. Hypomagnesemia impairs both parathyroid
hormone (PTH) release and action. Critically ill
patients who are septic or who have unexplained
ionized hypocalcemia may be vitamin-D-deficient,
have an impairment of vitamin D metabolism or
action, or have impaired release of PTH (Desai,
Carlson and Geheb, 1987; Zaloga et al., 1984; Knochel,
1977). Hyperphosphatemia in severe rhabdomyolysis
causes hypocalcemia by calcium precipitation or
impaired vitamin D metabolism.
(a)
Clinical features
Clinical manifestations of ionized hypocalcemia are

generally not evident until ionized calcium is less than
0.7 mmol/l. Signs include hypotension, impaired ventricular function, bradycardia, bronchospasm and
laryngospasm. Weakness, tetany, hyper-reflexia, agitation, confusion and seizures may occur, but are less
common than in patients with hypoparathyroidism.
The response to digoxin and catecholamines may also
be impaired by hypocalcemia.
(b)
Treatment
Treatment is only undertaken if the patient is symptomatic. Acute hypocalcemia may be corrected by
intravenous 10% calcium chloride (0.7 mmol calcium
per milliliter). Before initiating treatment, hypomagnesemia should be excluded, as hypomagnesemic
patients respond poorly to calcium. Calcium chloride
10%, 5 ml intravenously (i.e. 3.4 mmol) or 10 ml of
calcium gluconate 10% (i.e. 2.3 mmol) may be administered over 35 minutes. Rapid infusion of calcium
may result in unpleasant flushing, hypertension,
Treatment of life-threatening hyperkalemia
Agent
Dose
Onset (min)
Duration (h)
1015
Sodium bicarbonate
Calcium chloride (10%)
50 g dextrose
20 IU insulin
50100 mmol
510 ml (3.46.8 mmol)
60240
15
12
12
Oral or rectal resonium A
50 g
120240
312
Nebulized salbutamol
1020 mg
30
Intravenous
Dextrose and insulin
bradycardia and atrioventricular nodal blockade, and

may precipitate digitalis toxicity. Persistent unexplained hypocalcemia requires further evaluation to
exclude hypoparathyroidism and vitamin D
deficiency.
16.8.6
HYPOPHOSPHATEMIA
Hypophosphatemia is defined as a fasting plasma

phosphate level of less than 0.8 mmol/l. Moderate
hypophosphatemia exists if the phosphate level is
between 0.32 and 0.8 mmol/l (Bohannon, 1989).
Severe hypophosphatemia exists when the plasma
level is less than 0.32 mmol/l (Hayek and Eisenberg,
1989). It may be caused by decreased intake, increased
excretion or intracellular redistribution.
Parenteral or enteral nutrition induces a transcellular
shift of phosphate from the ECF to the ICF via insulinmediated intracellular transport of phosphate with
glucose (refeeding syndrome; Solomon and Kirby,
1990; Aubier et al., 1985). Respiratory alkalosis causes
intracellular shift of phosphate by stimulating glycolysis and intracellular phosphate trapping. Hypophosphatemia induced by transcellular shift is associated
with hypophosphaturia and the normal body stores of
phosphate are maintained. Hypokalemia and hypomagnesemia may also promote hypophosphatemia by
impairing renal phosphate reabsorption.
necessary. Generally, however, in the critically ill

patient, the phosphate level should be kept above
0.8 mmol/l, to ensure adequate respiratory, cardiac and
intracellular function. Phosphate at 4080 mmol/24 h
may be administered orally or intravenously as sodium
or potassium phosphate (Chernow et al., 1989). If an
intravenous solution is administered as the dihydrogen
salt, an equimolar amount of Na+ or K+ and 80% of the
molar amount as H+ is administered with the phosphate. If it is administered as the monohydrogen salt,
then twice the molar amounts of sodium or potassium
are administered with the phosphate.
16.8.7
Clinical features
Clinical manifestations of hypophosphatemia are not

usually seen unless levels are markedly reduced, i.e. to
less than 0.3 mmol/l. Patients who are hyperventilated or malnourished, alcoholic, septic or have
diabetic ketoacidosis are at increased risk of severe
hypophosphatemia, and diaphragmatic weakness and
respiratory failure may occur (Lau, 1986).
(b)
Investigations
Patients with hypophosphatemia in the absence of

alkalosis and a urinary phosphate greater than
25 mmol/d,are suffering excessive renal loss. If the
urinary loss is less than 5 mmol/d, then renal loss is
excluded (Knochel, 1981). If hypophosphatemia is
associated with hypercalcemia, hyperparathyroidism
should be considered.
(c)
Treatment
Hypophosphatemia is usually asymptomatic. If hypophosphatemia is due to compartmental shift resulting

from respiratory alkalosis or to catecholamine infusions, even though phosphate levels may decrease to
0.4 mmol/l, treatment with phosphate may not be
HYPOMAGNESEMIA
Hypomagnesemia is defined as a plasma level of less

than 0.7 mmol/l and is associated with a 24-hour urine
magnesium of less than 1 mmol/l in the absence of
abnormal renal magnesium losses. This is a common
electrolyte disturbance in ICU patients (Chernow et al.,
1982b; Kingston, Al-Siba and Skooge, 1986; Paymaster,
1976) and is caused by decreased intake or increased
loss. The causes are similar to those for hypophosphatemia. Common causes in patients with head injury
include diuretics (osmotic and non-osmotic), diabetes
insipidus, respiratory alkalosis, sepsis and certain
drugs including aminoglycosides.
(a)
(a)
325
Clinical features
Clinical features tend only to occur when the plasma

level is less than 0.5 mmol/l (Table 16.20; Paymaster,
1976; Fishman, 1965; Whang et al., 1984). The clinical
diagnosis of magnesium depletion is difficult because
only 1% of total body magnesium is contained in
plasma and because total plasma magnesium does not
accurately reflect the ionized fraction (55% of the
total), which is the physiologically active form. The
remaining plasma magnesium is bound to protein
(33%) or is chelated. Measurement of ionized magnesium is not available. Hypomagnesemia may be
associated with hypokalemia (due to renal loss),
Table 16.20
Clinical features of hypomagnesemia
Neurological
Confusion, irritability, delirium, convulsions
Ataxia, athetoid movements
Weakness, tremors, cramps, tetany
Cardiovascular
Tachyarrhythmias (torsades de pointes)
Enhanced digoxin toxicity
Biochemical
Resistant hypokalemia, hypocalcemia, renal tubular
acidosis
326
hypocalemia (due to impaired release of PTH and

impaired peripheral action of PTH) and renal tubular
acidosis (RTA). In one study, more than 40% of
hypokalemic patients revealed an associated hypomagnesemia (Gums, 1987).
(b)
Investigations
With magnesium deficiency, the renal loss is usually

minimal (0.050.10 mmol/d). An intravenous dose of
magnesium is normally excreted in the urine in 24
hours; thus 30 mmol administered over 812 hours
and a 24-hour collection of urine can be used as an
estimate of magnesium balance. Normally more than
60% of the dose is excreted; if less than 50% is excreted
the patient has a magnesium deficiency (Salem,
Munoz and Chernow, 1991).
(c)
(a)
Treatment
Symptomatic hypomagnesemia (e.g. causing refractory arrhythmias or seizures) is treated by intravenous

bolus injection of 510 ml of 49.3% MgSO4 solution
(2 mmol/ml magnesium; Cohen, Gambill and Eggers,
1977). Less severe degrees of hypomagnesemia are
treated by adding magnesium salts to the enteral or
parenteral solutions.
16.9 Special fluid and electrolyte

problems in the neurosurgical patient
16.9.1
NEUROGENIC PULMONARY EDEMA
Patients with a variety of central nervous system

disorders, including head injury, may develop acute
pulmonary edema in the absence of cardiopulmonary
disease or other apparent causes. This neurogenic
pulmonary edema may follow head trauma (Ducker,
Simmons and Martin, 1969), intracranial tumors,
subarachnoid hemorrhage, stroke, seizures (Bonbrest,
1965), Guillain-Barre syndrome and meningitis.
Although the cause is unclear, it is thought that
neurogenic pulmonary edema may be due to massive
alpha-adrenergic discharge (Wray and Nicotra, 1978).
The sympathetic discharge increases systemic and
pulmonary vascular resistances. This transiently
increases left atrial pressure and pulmonary artery
wedge pressure (PAWP), and pulmonary capillary
disruption occurs. An animal model has also implicated endothelial injury in its development (Minnear
et al., 1987). The altered pulmonary capillary permeability results in the accumulation of pulmonary edema
fluid, which has a high protein content.
A previously healthy 24-year-old male was admitted to ICU following a gunshot wound to the head
(Figure 16.15(a)). Frothy pulmonary edema fluid
(b)
Figure 16.15 A 24-year-old man suffered a gunshot wound
to the head. Shortly after admission to the ICU he developed
marked frothy sputum and oxygen desaturation. (a) The CT
scan shows the missile track and a tight brain. (b) The chest
X-ray shows fluffy exudates bilaterally. There is a (Swan
Ganz) catheter. Hemodynamic measurement confirmed a
non-cardiogenic basis for the pulmonary edema. Mechanical
ventilation and diuresis resulted in a successful outcome.
issued from the endotracheal tube and marked oxygen

desaturation was noted on pulse oximetry. He
required ventilation with 100% O2 and positive endexpiratory pressure (PEEP); chest X-ray confirmed the
presence of pulmonary edema (Figure 16.15(b)). The
CVP was 10 mmHg. A SwanGanz catheter was
inserted to facilitate fluid management in this patient,
who subsequently made a satisfactory recovery.
The treatment of neurogenic pulmonary edema is
similar to the treatment of adult respiratory distress
syndrome (ARDS) from any cause. In addition to
REFERENCES
ventilatory support with positive-pressure ventilation

and positive end-expiratory pressure (PEEP), fluid
management is of the utmost importance. Fluid intake
should be crystalloid solution, rather than proteincontaining solutions, and should be minimal. If PAWP
is monitored, it should be kept low (57 mmHg) to
minimize further extravasation of proteinacious fluid.
If mean arterial pressure or perfusion is inadequate,
inotropes should be used to maintain cerebral and
peripheral tissue perfusion. For a detailed account of
the respiratory management of neurogenic pulmonary
edema see Chapter 17.
16.9.2 NON-KETOTIC HYPERGLYCEMIC
HYPEROSMOLAR COMA
Non-ketotic hyperglycemic hyperosmolar coma

(NHHC) may occur following head injury, intracerebral hemorrhage, brain tumor or cerebral infarction (Park, Meacham and Netsky, 1976). About twothirds of patients who develop NHHC have no history
of diabetes mellitus. Many have intercurrent infections, or are taking drugs that alter glucose tolerance,
such as steroids, phenytoin and thiazides, or are
dehydrated.
The diagnosis of NHHC is based on the findings of
hyperglycemia, glycosuria and a plasma osmolality
greater than 330 mosmol/l in the absence of ketosis.
Patients who have NHHC and are dehydrated
because of the osmotic diuresis are often potassiumdepleted. They may also have evidence of prerenal
renal failure which, if untreated, can progress to acute
renal failure.
Treatment is primarily correction of dehydration
and hypertonicity. Physiological saline is used for
volume replacement, together with careful monitoring
of serum and urine electrolytes and osmolality. After
sodium deficits have been corrected with physiological saline, half physiological saline or dextrose/
saline (4% dextrose in N/5 saline) should be used to
replace water deficits. The hyperglycemia is usually
responsive to insulin.
16.9.3
SPINAL-CORD-INJURED PATIENTS
Cervical cord injury may cause spinal shock due to

anatomical or functional transection of the cord and
this may last for 13 weeks. The interruption to
sympathetic outflow results in vasodilatation, pooling
of blood in peripheral vascular beds and hypotension.
Reflex bradycardia also occurs, because the cardiac
accelerator nerves arise from segments T1T4. The
hypotensive phase is usually preceded by a rise in
systemic pressure over 34 minutes as a result of
widespread initial sympathetic activation. This may
be particularly important in the presence of brain
327
injury, because cerebral autoregulation may be overcome and cerebral edema exacerbated.
Because the heart cannot compensate for overtransfusion or increased venous return, careful monitoring is mandatory. Circulating volume is critical,
and observation of the CVP or PAWP to fluid
challenge may be used to assess the volume status of
the patient.
Electrolyte abnormalities may develop as a result of
immobility and absent muscle activity. Mobilization of
calcium may increase, causing hypercalcemia and
hypercalciuria, from about 10 days after injury. The
increase in serum calcium levels may predispose to
ventricular arrhythmias during anesthesia. Thus calcium levels should be measured preoperatively in
these patients.
16.10
Conclusions
The goal of fluid therapy in the critically ill patient

is to restore and maintain normal O2 delivery to the
tissues before the sequelae of shock (including
secondary neuronal damage) occur.
Abnormalities of fluid, electrolyte and acidbase
metabolism may result in alterations in neurological
status and thus interfere with neurological assessment, as well as contributing to secondary neuronal
damage and thus worsening outcome.
Maintenance of normal cerebral perfusion pressure
(CPP) is a key aspect of the management of patients
with head injuries, since impairment of cerebral
autoregulation following traumatic brain injury
causes cerebral oxygen delivery to be influenced by
CPP.
Correction of hypovolemia is best achieved with
colloid, physiological saline or blood transfusion (if
necessary). At least two animal studies have indicated that the amount of sodium infused does not
adversely affect intracranial pressure (Ramming et
al., 1994). In one of the studies, infusion of a
dextrose solution in amounts similar to the infused
sodium load significantly decreased serum sodium,
increased brain edema and decreased neurological
outcome (Shapira et al., 1995).
16.11
References
Addleman, M., Pollard, A. and Grossman, R. F. (1985) Survival after severe

hypernatremia due to salt ingestion by an adult. American Journal of
Medicine, 78, 176178.
Adelson, E., Crosby, W. H. and Roeder, W. H. (1955) Further studies of a
hemostatic defect caused by intravenous dextran. Journal of Laboratory and
Clinical Medicine, 45, 441.
Albright, A. L., Latchaw, R. E. and Robinson, A. G. (1984) Intracranial and
systemic effects of osmotic and oncotic therapy in experimental cerebral
edema. Journal of Neurosurgery, 60, 481.
Alexander, B., Odake, K., Lawlor, D. et al. (1975) Coagulation, haemostasis and
plasma expanders: a quarter century enigma. Federal Proceedings, 34, 1429.
Allen, S. J., Drake, R. E., Williams, J. P. et al. (1987) Recent advances in
pulmonary oedema. Critical Care Medicine, 15, 963.
328
Arieff, A. I. (1984) Central nervous system manifestations of disordered

sodium metabolism. Clinics in Endocrinology and Metabolism, 13, 269294.
Arieff, A. I. (1985) Effects of water, acidbase and electrolyte disorders on the
central nervous system: hyponatremia, in Fluid, Electrolyte and Acidbase
Disorders, (eds A. I. Arieff and R. A. De Fronzo), Churchill Livingstone, New
York, pp. 9921007.
Arieff, A. I. (1986) Hyponatraemia, convulsions, respiratory arrest and
permanent brain damage after elective surgery in healthy women. New
England Journal of Medicine, 314, 15291535.
Arieff, A. I. and Guisado, R. (1976) Effects on the central nervous system of
hypernatremic and hyponatremic states. Kidney International, 10, 104116.
Arieff, A. I., Liach, F. and Massry, S. G. (1976) Neurological manifestations and
morbidity of hyponatraemia: correlation with brain water and electrolytes.
Medicine, 55, 121129.
Arora, N. S. and Rochester, D. F. (1982) Respiratory muscle strength and
maximal voluntary ventilation in undernourished patients. American
Review of Respiratory Disease, 126, 58.
Atik, M. (1967) Dextran 40 and dextran 70. Archives of Surgery, 94, 664.
Atik, M. (1969) The uses of dextran in surgery: a current evaluation. Surgery,
65, 548.
Aubier, M., Murciano, D., Lecocguic, Y. et al. (1985) Effect of hypophosphatemia on diaphragmatic contractility in patients with acute respiratory
failure. New England Journal of Medicine, 313, 420424.
Aubry, R. H. and Nankin, H. R. (1965) Measurement of the osmotic threshold
for vasopressin release in human subjects, and its modification by cortisol.
Journal of Clinical Endocrinology and Metabolism, 25, 14811490.
Ayus, J. C., Krothapalli, R. K. and Arieff, A. I. (1985) Changing concepts in
treatment of severe symptomatic hyponatremia. Rapid correction and
possible relation to central pontine myelinolysis. American Journal of
Medicine, 78, 897902.
Ayus, J. C., Krothapalli, R. K. and Arieff, A. I. (1987) Treatment of symptomatic
hyponatremia and its relation to brain damage. New England Journal of
Medicine, 317, 11901195.
Ayus, J. C., Krothapalli, R. K. and Arieff, A. I. (1988) Correction of
hyponatraemia and its relation to brain damage. New England Journal of
Medicine, 318, 13361337.
Baek, S. M., Makabali, G. G., Bryan-Brown, C. W. et al. (1975) Plasma
expansion in surgical patients with high central venous pressure (CVP): the
relationship of blood volume to hematocrit, CVP, pulmonary wedge
pressure and cardiorespiratory changes. Surgery, 78, 304.
Baker, J. W., Deitch, E. D., Ma, L. M. et al. (1988) Hemorrhagic shock induces
bacterial translocation from the gut. Journal of Trauma, 28, 896.
Barratt, L. J. (1981) Potassium handling by the kidney. Hyperkalaemia.
Australian and New Zealand Journal of Medicine, 11, 1622.
Bartter, F. C. and Schwartz, W. B. (1967) The syndrome of inappropriate
secretion of antidiuretic hormone. American Journal of Medicine, 42,
790806.
Battacharya, J., Gropper, M. A. and Staub, N. S. (1984) Interstitial fluid
pressure gradient measured by micropuncture in excised dog lung. Journal
of Applied Physiology, 56, 271.
Battistella, F. D. and Wisner, D. H. (1991) Combined haemorrhagic shock and
head injury: effects of hypertonic saline (7.5%) resuscitation. Journal of
Trauma, 31, 182.
Beckstead, J. E., Tweed, W. A., Lee, J. et al. (1978) Cerebral blood flow and
metabolism in man following cardiac arrest. Stroke, 9(6), 569.
Beecher, H. K. (1945) Preparation of battle casualties for surgery. Annals of
Surgery, 121, 769.
Behrman, S. W., Fabian, T. C., Kudsk, K. A. et al. (1991) Microcirculatory flow
changes after initial resuscitation of haemorrhagic shock with 7.5%
hypertonic saline/6% dextran 70. Journal of Trauma, 31, 589.
Bichet, D. and Schrier, R. W. (1982) Evidence against concept of hyponatraemia and sick cells. Lancet, i, 742.
Bohannon, N. J. V. (1989) Large phosphate shifts with treatment for
hyperglycaemia. Archives of Internal Medicine, 149, 14231425.
Boldt, J., Zickmann, B., Ballesteros, M. et al. (1991) Cardiorespiratory
responses to hypertonic saline solution in cardiac operations. Annals of
Thoracic Surgery, 51, 610.
Bonbrest, H. C. (1965) Pulmonary oedema following an epileptic seizure.
American Review of Respiratory Disease, 91, 97100.
Bouzarth, W. F., Shenkin, H. A. et al. (1982) Is cerebral hyponatraemia
iatrogenic? Lancet, i, 10611062.
Brinkmeyer, S., Safar, P., Motoyama, E. et al. (1981) Superiority of colloid over
electrolyte solution for fluid resuscitation (severe normovolemic hemodilution). Critical Care Medicine, 9, 369.
Brobeck, J. R. (ed.) (1973) Best and Taylors Physiological Basis of Medical Practice,
9th edn, Williams & Wilkins, Baltimore, MD, pp. 41134-123; 1151-29.
Brunner, J. E., Redmond, J. M., Haggar, A. M. et al. (1990) Central pontine
myelinolysis and pontine lesions after rapid correction of hyponatraemia: a
prospective magnetic resonance imaging study. Annals of Neurology, 27,
6166.
Buchman, T. G., Menker, J. and Lipsett, P. A. (1991) Strategies for trauma
resuscitation. Surgery, Gynecology, and Obstetrics, 172, 812.
Buckingham, J. C. (1985) Hypothalamo-pituitary responses to trauma. British
Medical Bulletin, 41, 203211.
Burke, A. M., Quest, D. O., Chein, S. and Cerri, C. (1981) The effects of
mannitol on blood viscosity. Journal of Neurosurgery, 55, 550553.
Butterworth, J. F. and De Witt, D. S. (1989) Severe head trauma: pathophysiology and management, in Critical Care Clinics, vol. 5(4), (ed. D. S. Prough), W.
B. Saunders, Philadelphia, PA, pp. 807820.
Calvin, J. E., Driedger, A. A. and Sibbald, W. J. (1981) The hemodynamic effect
of rapid fluid infusion in critically ill patients. Surgery, 90, 61.
Carrico, C. J., Canizaro, P. C. and Shires, G. T. (1976) Fluid resuscitation
following injury: Rationale for the use of balanced salt solutions. Critical
Care Medicine, 4, 46.
Carvajal, H. F. and Parks, D. K. (1988) Optimal composition of burn
resuscitation fluids. Critical Care Medicine, 16, 695.
Cerra, F. B. (1986) The role of nutrition in the management of metabolic stress,
in Critical Care Clinics, vol. 2(4), (ed. A. A. Meyer), W. B. Saunders,
Cervera, A. L. and Moss, G. (1975) Dilutional re-expansion with crystalloid
after massive haemorrhage: saline versus balanced electrolyte solution for
maintenance of normal blood volume and arterial pH. Journal of Trauma, 15,
498.
Chernow, B., Zaloga, G., McFadden, E. et al. (1982a). Hypocalemia in critically
ill patients. Critical Care Medicine, 10, 848851.
Chernow, B., Smith, J., Rainey, T. G. et al. (1982b). Hypomagnesemia:
Implications for the critical care specialist. Critical Care Medicine, 10,
193196.
Chernow, B., Bamberger, S., Stoiko, M. et al. (1989) Hypomagnesemia in
patients in postoperative intensive care. Chest, 95, 391397.
Chung, H. M., Kluge, R., Schrier, R. W. and Anderson, R. J. (1986)
Postoperative hyponatraemia. A prospective study. Archives of Internal
Medicine, 146, 333336.
Civetta, J. M. (1979) A new look at the Starling equation. Critical Care Medicine,
7, 84.
Clifton, G. L., Robertson, C. S., Grossman, R. G. et al. (1984) The metabolic
response to severe head injury. Journal of Neurosurgery, 60, 687696.
Cluitmans, F. H. M. and Meinders, A. E. (1990) Management of severe
hyponatremia: Rapid or slow correction. American Journal of Medicine, 88,
161166.
Cohen, H. B., Gambill, A. F. and Eggers, G. W. N. (1977) Acute pulmonary
oedema following head injury: two case reports. Anesthesia and Analgesia,
56, 136139.
Consensus Development Conference (1988) Perioperative red blood cell
transfusion. Report of consensus development conference on perioperative
red cell transfusion. Journal of the American Medical Association, 260,
27002703.
Crandall, E. D. (1974) Serum sodium response to hyperglycemia. New England
Journal of Medicine, 290, 465.
Crandall, E. D., Staub, N. C., Goldberg, H. S. et al. (1983) Recent developments
in pulmonary oedema. Annals of Internal Medicine, 99, 808.
Crompton, M. R. (1971) Hypothalamic lesions following closed head injury.
Brain, 94, 165172.
Cross, J. S., Griber, D. P., Burchard, K. W. et al. (1988) Hypertonic saline fluid
therapy following surgery: a prospective study. Journal of Trauma, 29, 817.
Dawidson, I. J. A., Willms, C. D., Sandor, Z. F. et al. (1991) Ringers lactate with
or without 3% dextran-60 as volume expanders during abdominal aortic
surgery. Critical Care Medicine, 19, 36.
Demling, R. H. (1980) Lung fluid and protein dynamics during haemorrhagic
shock, resuscitation and recovery. Circulatory Shock, 7, 149.
Demling, R. H., Manohar, M., Will, J. A. et al. (1979) The effect of plasma
oncotic pressure on the pulmonary microcirculation after haemorrhagic
shock. Surgery, 86, 323.
Demling, R. H., Harms, B., Kramer, G. et al. (1982) Acute versus sustained
hypoproteinaemia and post traumatic pulmonary oedema. Surgery, 92, 79.
Desai, T. K., Carlson, R. W. and Geheb, M. A. (1987) Parathyroid-vitamin D
axis in critically ill patients with unexplained hypocalcaemia. Kidney
International, 32(22), S225.
Desai, T. K., Carlson, R. W. and Geheb, M. A. (1988) Prevalence and clinical
implications of hypocalcaemia in acutely ill patients in a medical intensive
care setting. American Journal of Medicine, 84, 209214.
Doczi, T., Tarjanyi, J., Huszka, E. and Kiss, J. (1982) Syndrome of inappropriate
secretion of antidiuretic hormone (SIADH) after head injury. Neurosurgery,
10, 685688.
Dodge, C. and Glass, D. D. (1982) Crystalloid and colloid therapy. Seminars in
Ducker, T. B., Simmons, R. L. and Martin, A. M. (1969) Pulmonary oedema as
a complication of intracranial disease. American Journal of Diseases of
Children, 118, 638641.
Dzau, V. J. and Hollenberg, N. K. (1984) Renal response to captopril in severe
heart failure: role of frusemide in natriuresis and reversal of hyponatremia.
Annals of Internal Medicine, 100, 777782.
Edelman, I. S. and Leibman, J. (1959) Anatomy of body water and electrolytes.
American Journal of Medicine, 27, 256263.
Edes, T. E. and Sunderrajan, E. V. (1985) Heparin-induced hyperkalemia.
Archives of Internal Medicine, 145, 10701072.
Edwards, O. M. and Clark, J. D. A. (1986) Post-traumatic hypopituitarism.
Medicine, 65, 281290.
REFERENCES
Edwards, J. D., Nightingale, P., Wilkins, R. G. et al. (1989) Hemodynamic and
oxygen transport response to modified fluid gelatin in critically ill patients.
Critical Care Medicine, 17, 996.
Emerson, T. E. (1989) Unique features of albumin: a brief review. Critical Care
Medicine, 17, 690.
Eyer, S. D., Micon, L. T., Konstantinides, F. N. et al. (1993) Early enteral feeding
does not attenuate metabolic response after blunt trauma. Journal of Trauma,
34(5), 639644.
Falk, J. L. (1991) Fluid resuscitation and colloid-crystalloid controversy: new
thoughts on an old debate. Critical Care Medicine, 19, 451.
Fein, I. A., Rackow, E. C., Sprung, C. L. et al. (1982) Relation of COP to arterial
hypoxaemia and cerebral oedema during crystalloid volume loading of
patients. Annals of Internal Medicine, 96, 570.
Finberg, L., Luttrell, C. N. and Redd, H. (1959) Pathogenesis of lesions in the
nervous system in hypernatraemic states. Pediatrics, 23, 4653.
Fishman, R. A. (1965) Neurological aspects of magnesium metabolism.
Flear, C. T. G., Gill, G. V. and Burn, J. (1981) Hyponatraemia: mechanisms and
management. Lancet, ii, 2631.
Flear, C. T. G. and Singh, C. M. (1973) Hyponatraemia and sick cells. British
Freshman, S. P., Battistella, F. D., Matteucci, M. et al. (1993) Hypertonic saline
(7.5%) versus mannitol: a comparison for treatment of acute head injuries.
Journal of Trauma, 35(3).
Fuisz, R. E. (1963) Hyponatremia. Medicine, 42, 149168.
Gala, G. J., Lilly, M. P., Thomas, S. E. et al. (1991) Interaction of sodium and
volume in fluid resuscitation after haemorrhage. Journal of Trauma, 31, 545.
Gamble, J., Ross, G. and Tisdall, F. (1923) The metabolism of fixed base during
fasting. Journal of Biological Chemistry, 57, 633.
Goldberg, H. S. (1980) Pulmonary interstitial compliance and microvascular
filtration coefficient. American Journal of Physiology, 239, H189.
Granger, J. H. (1979) Role of the interstitial matrix and lymphatic pump in
regulation of transvascular fluid balance. Microvascular Research, 18, 207.
Griffel, M. I. and Kaufman, B. S. (1992) Pharmacology of colloids and
crystalloids. Critical Care Clinics, 8, 235.
Griswold, J. A., Anglin, B. L., Love, R. T. et al. (1991) Hypertonic saline
resuscitation: efficacy in a community-based burn unit. Southern Medical
Journal, 84, 692.
Gums, J. G. (1987) Clinical significance of magnesium: a review. Drug
Intelligence and Clinical Pharmacology, 21, 240246.
Gunnar, W., Merlotti, G. J., Jonasson, O. et al. (1986) Resuscitation from
hemorrhagic shock: Alterations of the intracranial pressure after normal
saline, 3% saline and dextran 40. Annals of Surgery, 204, 686.
Gunnar, W., Jonasson, O., Merlotti, G. et al. (1988) Head injury and
hemorrhagic shock: Studies of the blood brain barrier and intracranial
pressure after resuscitation with normal saline, 3% saline solution and
dextran-40. Surgery, 103, 398407.
Guyton, A. C., Granger, H. J. and Taylor, A. E. (1971) Interstitial fluid pressure.
Physiology Review, 51, 527.
Harms, B. A., Kramer, G. C., Bodal, B. I. et al. (1981) Effect of hypoproteinaemia on pulmonary and soft tissue oedema formation. Critical Care Medicine,
9, 503.
Harrison, N., Darrow, D. and Yannet, H. (1936) The total electrolyte content of
animals and its probable relation to distribution of body water. Journal of
Biological Chemistry, 113, 515.
Hauser, C. J., Shoemaker, W. C., Turpin, I. et al. (1980) Oxygen transport
responses to colloids and crystalloids in critically ill surgical patients.
Surgery Gynecology and Obstetrics, 150, 811.
Haydock, D. A. and Hill, G. L. (1987) Improved wound healing response in
surgical patients receiving intravenous nutrition. British Journal of Surgery,
74, 320323.
Hayek, M. E. and Eisenberg, P. G. (1989) Severe hypophosphataemia
following the institution of enteral feedings. Archives of Surgery, 124,
13251328.
Hayes, M. A. (1968) Water and electrolyte therapy after operation. New
Heughan, C., Niinikoski, J. and Hunt, T. K. (1972) Effect of excessive infusion
of saline on tissue oxygen transport. Surgery, Gynecology and Obstetrics, 135,
257.
Heymsfield, S. B., Bethel, R. A., Ansley, J. D. et al. (1978) Cardiac abnormalities
in cachectic patients before and during nutritional repletion. American Heart
Journal, 95, 584594.
Hindman, B. J., Funatsu, N., Cheng, D. C. H. et al. (1990) Differential effect of
oncotic pressure on cerebral and extracerebral water content during
cardiopulmonary bypass in rabbits. Anesthesiology, 73, 951.
Holt, M. E., Ryall, M. E. and Campbell, A. K. (1984) Albumin inhibits human
polymorphonuclear leucocyte luminol-dependent chemiluminescence: evidence for oxygen radical scavenging. British Journal of Experimental
Pathology, 65, 231.
Horton, J., Landreau, R. and Tuggle, T. (1985) Cardiac response to fluid
resuscitation from haemorrhagic shock. Surgery, Gynecology, and Obstetrics,
160, 444452.
Horton, J. W. and Walker, P. B. (1991) Small volume hypertonic saline dextran
resuscitation from canine endotoxin shock. Annals of Surgery, 214, 64.
329
Hume, D. M. and Egdahl, R. H. (1959) The importance of the brain in the

endocrine response to injury. Annals of Surgery, 150, 697712.
Humes, H. D. (1986) Disorders of water metabolism, in Fluids and Electrolytes,
(eds J. P. Kokko and R. L. Tannen), W. B. Saunders, Philadelphia, PA, pp.
118149.
Ito, U., Ohno, K., Nakamura, R. et al. (1979) Brain edema during ischaemia
and after restoration of blood flow. Stroke, 10(5), 542.
James, W. P. T. (1971) Effects of protein-calorie malnutrition on intestinal
absorption. Annals of the New York Academy of Science, 176, 224261.
Jaques, L. B. (1985) The new understanding of the drug heparin. Chest, 88,
751757.
Jeejeebhoy, K. N. (1986) Muscle function and nutrition. Gut, 27(Suppl.),
2539.
Jenkins, P. G. and Larmore, C. (1974) Hyperglycemia-induced hyponatremia.
Jennett, B., Teasdale, G. and Braakman, R. (1979) Prognosis of patients with
Karlson, K. E., Garzon, A. A., Shaftan, G. W. et al. (1967) Increased blood loss
associated with administration of certain plasma expanders: Dextran 75,
dextran 40 and hydroxyethyl starch. Surgery, 62, 670.
Katz, M. A. (1973) Hyperglycemia-induced hyponatraemia calculation of
expected serum sodium depression. New England Journal of Medicine, 289,
843844.
Kee, D. B. and Wood, J. H. (1984) Rheology of the circulation. Neurosurgery,
15(1), 125131.
Kern, K. and Meislin, H. (1984) Diabetes insipidus: occurrence after minor
head trauma. Journal of Trauma, 24, 6972.
Kingston, M. E., Al-Siba, M. B. and Skooge, W. C. (1986) Clinical manifestations of hypomagnesemia. Critical Care Medicine, 14, 950954.
Knochel, J. P. (1977) The pathophysiology and clinical characteristics of severe
hypophosphataemia. Archives of Internal Medicine, 137, 203220.
Knochel, J. P. (1981) Hypophosphatemia. Western Journal of Medicine, 134,
1526.
Krausz, M. M., Landau, E. H., Klin, B. et al. (1992) Hypertonic saline treatment
of uncontrolled hemorrhagic shock at different periods from bleeding.
Archives of Surgery, 127, 93.
Krumbhaar, E. B. (1914) Hemolysis due to intravenous injection of distilled
water. Journal of the American Medical Association, 62, 992996.
Lamke, L. O. and Liljedahl, S. O. (1976) Plasma volume changes after infusion
of various plasma expanders. Resuscitation, 5, 93100.
Lau, K. (1986) Phosphate disorders, in Fluids and Electrolytes, (eds J. P. Kokko
and R. L. Tannen), W. B. Saunders, Philadelphia, PA, pp. 398471.
Laureno, R. and Karp, B. I. (1988) Pontine and extrapontine myelinolysis
following rapid correction of hyponatraemia. Lancet, i, 14391441.
Lazrove, S., Waxman, K., Shippy, C. et al. (1980) Hemodynamic, blood volume
and oxygen transport responses to albumin and hydroxyethyl starch
infusions in critically ill postoperative patients. Critical Care Medicine, 8, 302.
Leaf, A. (1974) Hyponatraemia. Lancet, i, 11191120.
Levine, O. R., Mellins, R. B., Senior, R. M. et al. (1967) The application of
Starlings law of capillary exchange to the lungs. Journal of Clinical
Investigation, 46, 934.
Levison, M. and Trunkey, D. D. (1982) Initial assessment and resuscitation
Surgical Clinics of North America, 62, 916.
Lohr, J. W., McReynolds, J. and Grimaldi, T. (1988) Effect of acute and chronic
hypernatraemia on myoinositol and sorbitol concentration in rat brain and
kidney. Life Science, 271276.
Lowe, R. J., Moss, G. S., Jilek, J. et al. (1979) Crystalloid vs colloid in the
etiology of pulmonary failure after trauma. A randomized trial in man.
Lowell, J. A., Schifferdecker, C., Driscoll, D. F. et al. (1990) Postoperative fluid
overload: Not a benign problem. Critical Care Medicine, 18(Suppl.), S193.
Lundy, E. F., Kuhn, J. E., Kwon, J. M. et al. (1987) Infusion of 5% dextrose
increases mortality and morbidity following six minutes of cardiac arrest in
resuscitated dogs. Journal of Critical Care, 2, 414.
McMahon, M. M., Farnell, M. B. and Murray, M. J. (1993) Nutritional support
of critically ill patients. Mayo Clinic Proceedings, 68, 911920.
Mangalore, P. P. and Hunt, T. K. (1972) Effect of varying oxygen tensions on
healing open wounds. Surgery, Gynecology, and Obstetrics, 135, 756.
Martinez-Maldonado, M. (1980) Inappropriate antidiuretic hormone secretion
of unknown origin. Kidney International, 17, 554567.
Mattox, K. L., Maningas, P. A., Moore, E. E. et al. (1991) Pre-hospital
hypertonic saline/dextran infusion for post-traumatic hypotension. Annals
of Surgery, 213, 482.
Messmer, K. (1975) Hemodilution. Surgical Clinics of North America, 55, 659.
Meye, B. J., Meyer, A. and Guyton, A. (1968) Interstitial fluid pressure. Clinical
Research, 22, 263.
Minnear, F. L., Kite, C., Hill, L. A. and Hoyte, Z. (1987) Endothelial injury and
pulmonary congestion characterize neurogenic pulmonary oedema in
rabbits. Journal of Applied Physiology, 63(1), 335341.
Mishle, J. (1984) Synthetic plasma volume expanders their pharmacology,
safety and clinical efficacy. Clinical Haematology, 13, 75.
Modig, J. (1986) Effectiveness of dextran 70 versus Ringers acetate in
traumatic shock and adult respiratory distress syndrome. Critical Care
Medicine, 14, 454.
330
Monafo, W. W., Halverson, J. D. and Schechtman, K. (1984) The role of

concentrated sodium solutions in the resuscitation of patients with severe
burns. Surgery, 95, 129.
Moore, F. D., Dagher, F. J., Boyden, C. M. et al. (1966) Hemorrhage in normal
man. 1. Distribution and dispersal of saline infusions following acute
blood loss: clinical kinetics of blood volume support. Annals of Surgery,
163, 485.
Morissette, M. P. (1977) Colloid osmotic pressure: its measurement and
clinical value. Canadian Medical Association Journal, 116, 897.
Morse, M. L., Milstein, J. M., Haas, J. E. et al. (1985) Effect of hydration on
experimentally induced cerebral edema. Critical Care Medicine, 13(7), 563.
Moss, G. S. and Gould, S. A. (1988) Plasma expanders. American Journal of
Surgery, 155, 425434.
Moss, G. S., Lowe, R. J., Jilek, J. et al. (1981) Colloid or crystalloid in the
resuscitation of haemorrhagic shock: a controlled clinical trial. Surgery, 89,
434.
Muizelaar, J. P., Wei, E. P., Kontos, H. A. and Becker, D. P. (1983) Mannitol
causes compensatory cerebral vasoconstriction and vasodilation in
response to blood viscosity changes. Journal of Neurosurgery, 59, 822828.
Mulinari, R. A., Gavras, I., Wang, Y. X. et al. (1990) Effects of a vasopressin
antagonist with combined antipressor and antidiuretic activities in rats
with left ventricular dysfunction. Circulation, 81, 308311.
Muzaffar, T. Z., Stalker, A. L., Bryce, W. A. J. et al. (1973) Quantitative studies
on fibrin formation and effects of dextran. Bibliotheca Anatomica, 12, 340.
Needleman, P. and Greenwald, J. E. (1986) Atriopeptin: a cardiac hormone
intimately involved in fluid, electrolyte, and blood-pressure homeostasis.
Nelson, P. B., Seif, S. M., Maroon, J. C. and Robinson, A. G. (1981)
Hyponatraemia in intracranial disease. Perhaps not the syndrome of
inappropriate secretion of antidiuretic hormone (SIADH). Journal of
Niinikoski, J., Hengan, C. and Hunt, T. K. (1972) Oxygen tensions in human
wounds. Journal of Surgical Research, 12, 1977.
Norenberg, M. D. and Papendick, R. E. (1984) Chronicity of hyponatraemia as
a factor in experimental myelinolysis. Annals of Neurology, 15, 544547.
Oh, M. S. and Carroll, H. J. (1992) Disorders of sodium metabolism:
hypernatraemia and hyponatraemia. Critical Care Medicine, 20(1), 94103.
Packer, M., Medina, M. and Yushak, M. (1984) Correction of dilutional
hyponatremia in severe chronic heart failure by converting-enzyme
inhibition. Annals of Internal Medicine, 100, 782789.
Pappenheimer, J. R. and Soto-Rivera, A. (1948) Effective osmotic pressure of
the plasma proteins and other quantities associated with the capillary
circulation in the hind limbs of cats and dogs. American Journal of Physiology,
152, 471.
Park, B. E., Meacham, W. F. and Netsky, N. G. (1976) Nonketotic hyperglycaemic hyperosmolar coma: report of neurosurgical cases with a review
of mechanisms and treatment. Journal of Neurosurgery, 44, 409417.
Paymaster, N. J. (1976) Magnesium metabolism: a brief review. Annals of the
Royal College of Surgeons of England, 58, 309314.
Peters, R. M. and Hargens, A. R. (1981) Protein vs electrolytes and all of the
Starling forces. Archives of Surgery, 116, 1293.
Pirisino, R., Di Simplicio, P., Ignesti, G. et al. (1988) Sulfhydryl groups and
peroxidase-like activity of albumin as scavenger of organic peroxides.
Pharmacology Research Community, 20, 545.
Plum, F. and Posner, J. B. (1980) The Diagnosis of Stupor and Coma, 3rd edn, F.
A. Davis, Philadelphia, PA.
Poole, G. V., Johnson, J. C., Prough, D. S. et al. (1986) Cerebral haemodynamics
after haemorrhagic shock: effects of the type of resuscitation fluid. Critical
Poole, G. V., Miller, J. D., Agnew, S. G. and Griswold, J. A. (1992) Lower
extremity fracture fixation in head injured patients. Journal of Trauma, 32,
654659.
Price, D. J. (1992) The intensive care of head-injured patients, in Care of the
Critically Ill Patient, 2nd edn, (eds J. Tinker and W. M. Zapol), SpringerVerlag, London, pp. 831872.
Price, D. J. E. and Murray, A. (1972) The influence of hypoxia and hypotension
on recovery from head injury. Injury, 3, 218224.
Prough, D. S., Johnson, J. C., Poole, G. V. et al. (1985) Effects on intracranial
pressure of resuscitation from haemorrhagic shock with hypertonic saline
versus lactated Ringers solution. Critical Care Medicine, 13, 407.
Prough, D. S., Johnson, J. C., Stump, D. A. et al. (1986) Effects of hypertonic
saline versus lactated Ringers solution on cerebral oxygen transport
during resuscitation from hemorrhagic shock. Journal of Neurosurgery, 64,
627.
Puri, V. K., Howard, M., Pidipaty, B. B. et al. (1983) Resuscitation in
hypovolemia and shock: a prospective study of hydroxyethyl starch and
albumin. Critical Care Medicine, 11, 518.
Rackow, E. C., Falk, J. L., Fein, I. A. et al. (1983) Fluid resuscitation in
circulatory shock: a comparison of the cardiorespiratory effects of albumin,
hetastarch and saline solutions in patients with hypovolaemic and septic
shock. Critical Care Medicine, 11, 839.
Rackow, E. C., Fein, I. A. and Leppo, J. (1977) Colloid osmotic pressure as a
prognostic indicator of pulmonary oedema and mortality in the critically ill.
Chest, 72, 709.
Rainey, T. G. and English, J. F. (1988) Pharmacology of colloids and

crystalloids, in The Pharmacologic Approach to the Critically Ill Patient, (ed. B.
Chernow), Williams & Wilkins, Baltimore, MD, p. 219.
Ramming, S., Shackford, S. R., Zhuang, J. and Schmoker, J. D. (1994) The
relationship of fluid balance and sodium administration to cerebral oedema
formation and intracranial pressure in a porcine model of brain injury.
Reed, L. L., Manglano, R., Martin, M. et al. (1991a) The effect of hypertonic
saline resuscitation on bacterial translocation after haemorrhagic shock in
rats. Surgery, 110, 685.
Reed, R. L., Johnston, T. D., Chen, Y. et al. (1991b) Hypertonic saline alters
plasma clotting times and platelet aggregation. Journal of Trauma, 31, 8.
Ring, J. and Messmer, K. (1977) Incidence and severity of anaphylactoid
reactions to colloid volume substitutes. Lancet, i, 466.
Roberts, P., Pollay, M., Engles, L. et al. (1987) Effect on intracranial pressure of
furosemide combined with varying doses and administration of mannitol.
Robertson, G. L. (1989) Syndrome of inappropriate antidiuresis. New England
Robin, A. P., Ing, T. S., Lancaster, G. A. et al. (1979) Hyperglycemia-induced
hyponatremia: a fresh look. Clinical Chemistry, 25, 496497.
Robinson, A. G. (1985) Disorders of antidiuretic hormone secretion. Clinics in
Endocrinology and Metabolism, 14, 5588.
Ross, E. J. and Christie, S. B. M. (1969) Hypernatremia. Medicine, 48,
441472.
Roth, E., Lax, L. C. and Malone, J. V. (1969) Ringers lactate solution and
extracellular fluid volume in the surgical patient: a critical analysis. Annals
Salem, M., Munoz, R. and Chernow, B. (1991) Hypomagnesemia in critical
illness: a common and clinically important problem. Critical Care Clinics,
7(1), 225252.
Samson, W. K. (1987) Atrial natriuretic factor and the central nervous system.
Endocrinology and Metabolism Clinics of North America, 16, 145161.
Scheinkestel, C. D., Tuxen, D. V., Cade, J. F. and Shann, F. A. (1989) Fluid
management of shock in critically ill patients. Medical Journal of Australia,
150, 508517.
Schultze, R. G. and Nissenson, A. R. (1980) Potassium: physiology and pathophysiology, in Clinical Disorders of Fluid And Electrolyte Metabolism, 3rd edn,
(eds M. H. Maxwell and C. R. Kleeman), McGraw-Hill, New York, pp.
113143.
Seckl, J. and Dunger, D. (1989) Postoperative diabetes insipidus. British
Seckl, J. R., Dunger, D. B. and Lightman, S. L. (1987) Neurohypophyseal
function during early postoperative diabetes insipidus. Brain, 110,
737746.
Serkes, K. D. and Lang, S. (1966) Changes in extracellular fluid volume after
hemorrhage and tourniquet trauma. Surgery Forum, 17, 58.
Shackford, S. R. (1987) Fluid resuscitation of the trauma victim, in Trauma
Problems in Critical Care, (eds S. R. Shackford and A. Perel), J. B. Lippincott,
Philadelphia, PA, vol. 1 pp. 576587.
Shapira, Y., Artae, A. A., Qassam, N. et al. (1995) Brain oedema and neurologic
status with rapid infusion of 0.9% saline or 5% dextrose after head trauma.
Journal of Neurosurgical Anesthesia, 7,1725.
Shippy, C. R., Appel, P. L. and Shoemaker, W. C. (1984) Reliability of clinical
monitoring to assess blood volume in critically ill patients. Critical Care
Medicine, 12, 107.
Shires, G. T. (1965) The role of sodium containing solution in the treatment of
oligemic shock. Surgical Clinics of North America, 45, 365.
Shires, G. T. and Canizaro, P. C. (1973) Fluid resuscitation in the severely
injured. Surgical Clinics of North America, 53, 1341.
Shires, G. T., Williams, J. and Brown, F. (1960) Simultaneous measurement of
plasma volume, extracellular fluid volume and red blood cell mass in man
utilizing I131, S35, O4, Cr51. Journal of Laboratory and Clinical Medicine, 55,
776.
Shires, T., Coln, D., Carrico, J. et al. (1964) Fluid therapy in haemorrhagic
shock. Archives of Surgery, 88, 688.
Shires, G. T., Cunningham, J. N., Baker, C. R. F. et al. (1972) Alterations in
cellular membrane function during hemorrhagic shock in primates. Annals
Shoemaker, W. C. (1976) Comparison of the relative effectiveness of whole
blood transfusions and various types of fluid therapy in resuscitation.
Shoemaker, W. C. (1982) Fluids and electrolytes in the adult, in Critical Care:
State of the Art, (eds W. C. Shoemaker and W. L. Thompson), vol. 3, Society
of Critical Care Medicine, Fullerton, CA.
Shoemaker, W. C. and Czer, L. S. (1979) Evaluation of the biologic importance
of various hemodynamic and oxygen transport variables: which variables
should be monitored in postoperative shock? Critical Care Medicine, 7,
424.
Shoemaker, W. C. and Hauser, C. J. (1979) Critique of crystalloid versus
colloid therapy in shock and shock lung. Critical Care Medicine, 7, 117.
Shoemaker, W. C., Schluchter, M., Hopkins, J. A. et al. (1981) Comparison of
the relative effectiveness of colloids and crystalloids in emergency
resuscitation. American Journal of Surgery, 142, 73.
REFERENCES
Shoemaker, W. C., Appel, P. L., Waxman, K. et al. (1982) Clinical trials of
survivors. 7 cardiorespiratory patterns as therapeutic goals in critically ill
post-operative patients. Critical Care Medicine, 10, 398403.
Shoemaker, W. C., Appel, P., Kram, H. et al. (1988) Prospective trial of supra
normal values of survivors as therapeutic goals in high risk surgical
patients. Chest, 94, 11761186.
Shucart, W. A. and Jackson, I. (1976) Management of diabetes insipidus in
neurosurgical patients. Journal of Neurosurgery, 44, 6571.
Siegel, J. H., Gens, D. R., Mamantov, T. et al. (1991) Effect of associated injuries
and blood volume replacement on death, rehabilitation needs, and
disability in blunt traumatic brain injury. Critical Care Medicine, 19,
12521265.
Sinnatamby, C., Edwards, C. R. W., Kitau, M. and Irving, M. H. (1974)
Antidiuretic hormone response to high and conservative fluid regimens in
patients undergoing operation. Surgery, Gynecology, and Obstetrics, 139,
715719.
Slutsky, A. S. (1993) American College of Chest Physicians Consensus
Conference. Mechanical ventilation. Chest, 104, 183359.
Smith, S. D., Cone, J. B., Bowser, B. H. et al. (1982) Cerebral edema following
acute haemorrhage in a murine model: the role of crystalloid resuscitation.
Journal of Trauma, 22(7), 588.
Solomon, S. M. and Kirby, D. F. (1990) The refeeding syndrome: a review.
Journal of Parenteral and Enteral Nutrition, 14, 9097.
Starling, E. H. (1896) On the absorption of fluids from the connective tissue
spaces. Journal of Physiology (London), 9, 312.
Steinbok, P. and Thompson, G. B. (1978) Metabolic disturbances after head
injury: Abnormalities of sodium and water balance with special reference to
the effects of alcohol intoxication. Neurosurgery, 3(1), 915.
Sterns, R. H. (1987) Severe symptomatic hyponatraemia: treatment and
outcome. Annals of Internal Medicine, 107, 656664.
Sterns, R. H., Cox, M., Felig, P. U. et al. (1981) Internal potassium balance and
the control of the plasma potassium concentration. Medicine, 60, 339354.
Sterns, R. H., Riggs, J. E. and Schochet, S. S. Jr (1986) Osmotic demyelination
syndrome following correction of hyponatraemia. New England Journal of
Medicine, 314, 1535.
Stocker, R., Glazer, A. N. and Ames, B. N. (1987) Antioxidant activity of
albumin-bound bilirubin. Proceedings of the National Academy of Sciences of
the USA, 84, 5918.
Stockigt, J. R. (1977) Potassium metabolism. Anaesthesia and Intensive Care, 5,
317325.
Sutin, K. M., Ruskin, K. J. and Kaufman, B. S. (1992) Intravenous fluid therapy
in neurologic injury, in Critical Care Clinics vol. 8(2), (ed. B. S. Kaufman), W.
B. Saunders, Philadelphia, PA, pp. 367408.
Tannen, R. L. (1986) Potassium disorders, in Fluids and Electrolytes, (eds J. P.
Kokko and R. L. Tannen), W. B. Saunders, Philadelphia, PA, pp.
1502228.
Taylor, A. E., Parker, J. C., Kvietys, P. R. et al. (1982) The pulmonary
interstitium in capillary exchange. Annals of the New York Academy of Science,
384, 146.
Tormey, W. P. (1990) Central pontine myelinolysis and changes in serum
sodium. Lancet, 335, 1169.
Tranbaugh, R. F. and Lewis, F. R. (1985) Crystalloid fluid, in Controversies in
Trauma Management (Clinics in Emergency Medicine), (eds R. N. Dailey and
M. Callaham), Churchill Livingstone, pp. 121133.
Vassar, M. J., Perry, C. A., Gannaway, W. L. et al. (1991) 7.5% sodium chloride/
dextran for resuscitation of trauma patients undergoing helicopter transport. Archives of Surgery, 126, 1065.
Velanovich, V. (1989) Crystalloid versus colloid fluid resuscitation: a metaanalysis of mortality. Surgery, 105, 65.
Verney, E. B. (1954) Water diuresis (John Malet Purser Lecture). Irish Journal of
Medical Science, 345377.
Virgilio, R. W., Smith, D. E. and Zarins, C. K. (1979) Balanced electrolyte
solutions: Experimental and clinical studies. Critical Care Medicine, 7, 98.
Virgilio, R. W., Rice, C. L., Smithe, D. E. et al. (1979) Crystalloid vs colloid
resuscitation. Is one better? Surgery, 85, 129139.
Voll, C. L. and Auer, R. N. (1988) The effect of post ischaemic blood glucose
levels on ischaemic brain damage in the rat. Annals of Neurology, 24,
638646.
Wasil, M., Halliwell, B., Hutchison, D. C. et al. (1987) The antioxidant action of
human extracellular fluids. Biochemistry Journal, 243, 219.
331
Wasterlain, C. G. and Posner, J. B. (1968) Cerebral oedema in water

intoxication: I. Clinical and chemical observations. Archives of Neurology, 19,
7178.
Weaver, D. W., Ledgerwood, A. M., Lucas, C. E. et al. (1978) Pulmonary effects
of albumin resuscitation for severe hypovolaemic shock. Surgery, 113, 387.
Weed, L. H. and McKibben, P. S. (1919a) Experimental alteration of brain bulk.
Weed, L. H. and McKibben, P. S. (1919b) Pressure changes in the cerebrospinal
American Journal of Physiology, 48(4):512.
Weil, M. H. and Henning, R. J. (1979) New concepts in the diagnosis and fluid
treatment of circulatory shock. Anesthesia and Analgesia, 58, 124.
Weil, M. H., Morissette, M., Michaels, S. et al. (1974) Routine plasma colloid
osmotic pressure measurements. Critical Care Medicine, 2, 229.
Weil, M. H., Michaels, S., Puri, V. K. et al. (1981) The stat laboratory: facilitating
blood gas and biochemical measurements for the critically ill and injured.
American Journal of Clinical Pathology, 76, 34.
Weisberg, L. S. (1989) Pseudohyponatraemia: a reappraisal. American Journal of
Medicine, 86, 315318.
Weiss, H. J. (1967) The effect of clinical dextran on platelet aggregation,
adhesion and ADP release in man: in vivo and in vitro studies. Journal of
Laboratory and Clinical Medicine, 69, 37.
Weissman, C. (1990) The metabolic response to stress: an overview and
update. Anaesthesiology, 73, 308327.
Whang, R., Oei, T. O., Aikawa, J. K. et al. (1984) Predictors of clinical
hypomagnesemia. Hypokalemia, hypophosphatemia, hyponatremia and
hypocalcemia. Archives of Internal Medicine, 144, 17941796.
Williams, G. H. (1986) Hyporeninemic hypoaldosteronism. New England
Journal of Medicine, 10411042.
Wilmore, D. W., Smith, R. J., ODwyer, S. T. et al. (1988) The gut a central
organ following surgical stress. Surgery, 104, 917.
Wise, B. L. and Chater, N. (1962) The value of hypertonic mannitol solution in
decreasing brain mass and lowering cerebrospinal fluid pressure. Journal of
Wisner, D. K., Schuster, L. and Quinn, C. (1990) Hypertonic saline resuscitation of head injury: Effects on cerebral water content. Journal of Trauma, 30,
75.
Wittmers, L. E., Bartlett, M. and Johnson, J. A. (1976) Estimation of capillary
permeability coefficient of inulin in various tissues of the rabbit. Microvascular Research, 11, 67.
Worthley, L. I. G. (1986) Hyperosmolar coma treated with intravenous sterile
water. A study of three cases. Archives of Internal Medicine, 146, 945947.
Worthley, L. I. G., Guerin, M. and Pain, R. W. (1987) For calculating osmolality,
the simplest formula is the best. Anaesthesia and Intensive Care, 15,
199202.
Worthley, L. I. G. and Thomas, P. D. (1986) Treatment of hyponatraemic
seizures with intravenous, 29.2% saline. British Medical Journal, 292,
168170.
Wray, N. P. and Nicotra, M. B. (1978) Pathogenesis of neurogenic pulmonary
oedema. American Review of Respiratory Disease, 118, 783786.
Young, B., Ott, L., Norton, J. et al. (1985) Metabolic and nutritional sequelae in
the nonsteroid-treated head injury patient. Neurosurgery, 17, 784791.
Zaloga, G. P. (1991) Hypocalcaemic crisis. Critical Care Clinics, 7, 191200.
Zaloga, G. P. and Chernow, B. (1986) Hypocalcemia in critical illness. Journal
of the American Medical Association, 256(14), 19241929.
Zaloga, G. P., Chernow, B. and Cook, D. (1985) Assessment of calcium
homeostasis in the critically ill patient: the diagnostic pitfalls of the McLean
Hastings nomogram. Annals of Surgery, 202, 587594.
Zaloga, G. P., Chernow, B., Cook, D. et al. (1984) The importance of measured
serum ionized calcium levels in critically ill patients. Critical Care Medicine,
12(3), 236.
Zornow, M. H., Scheller, M. S., Todd, M. M. et al. (1988) Acute cerebral effects
of isotonic crystalloid and colloid solutions following cryogenic brain
injury in the rabbit. Anesthesiology, 69, 180.
Zornow, M. H., Todd, M. M. and Moore, S. S. (1987) The acute cerebral effects
of changes in plasma osmolality and oncotic pressure. Anaesthesiology, 67,
936.
Zwimpfer, T. J. and Moulton, R. J. (1993) Neurologic trauma concerns, in
Critical Care Clinics, vol. 9(4), (eds D. P. Milzman, B. R. Boulanger and A.
Rodriguez), W. B. Saunders, Philadelphia, PA, pp. 727739.
17
Respiratory and cardiovascular

support
John A. Myburgh
17.1
Introduction
Our survival depends on the production of energy in

the form of adenosine triphosphate by the process of
oxidative metabolism. This in turn depends on a
continual supply of substrates, particularly oxygen
and glucose, to vital tissue beds such as the brain and
myocardium. In a critically ill patient, the supply of
oxygen may be compromised while at the same time
metabolic demand is increased. Inadequate oxygen
delivery may become the factor mitigating against
recovery or even survival of the patient (Schumacker
and Cain, 1987). In a patient with brain injury, cerebral
metabolic reserve is already reduced by the primary
insult and subsequent oxygen lack is a potent cause of
avoidable secondary brain injury (Chesnut et al.,
1993).
This chapter will address fundamental physiological principles of oxygen delivery from the atmosphere to the cells, the pathophysiological effects of
head injury on oxygen delivery and principles of
resuscitation and the maintenance of cardiopulmonary homeostasis.
17.1.1
THE PATIENT AT RISK
Although severe head injury is most common in

healthy young adults, particularly males, a proportion
of patients may suffer from pre-existing medical
conditions, which may be exacerbated by the stress of
neurotrauma and complicate its management (Luce,
1993; Kaufman et al., 1993). Ischemic heart disease and
hypertension mandate meticulous attention to mean
perfusion pressures of brain and myocardium; these
patients may require higher pressures than the normal
population to ensure optimal oxygen delivery. Control
of excessive centrally mediated sympathetic activity is
important in reducing both cerebral and myocardial
oxygen demand. Respiratory problems such as
asthma or chronic obstructive pulmonary disease may
affect mechanical ventilation.
Pre-existing nervous system disorders such as epilepsy, cerebrovascular or neuromuscular disease may
make assessment of effects of trauma and prognosis
more difficult.
Renal and hepatic dysfunction, particularly if there
is pre-existing impairment, may alter pharmacokinetics and drug interactions, leading to either toxic or
subtherapeutic effects. As an example, phenytoin,
frequently administered for prophylaxis of post-traumatic seizures, is a potent inducer of hepatic enzymes
and will alter the pharmacokinetic profiles of lipid
soluble drugs such as -adrenergic blockers and
opiates (Runciman, Myburgh and Upton, 1990).
Operative intervention may potentiate the acute
stress response, even in the young previously fit
patient (section 15.8).
17.1.2
THE ROLE OF INTENSIVE CARE
Management of head-injured patients in the Intensive

Care Unit (ICU) aims to optimize oxygen and substrate delivery to vital tissues and to detect events
that may harm the injured brain before they do
damage.
This requires prompt detection of potentially harmful changes by accurate monitoring and early intervention (Borel et al., 1990; Andrews, 1993). All aspects
of the patients state must be taken into account, with
particular attention to the cardiorespiratory, renal,
metabolic and nutritional systems (Bloomfield, 1989).
Although ICUs were established to preserve life and
improve the quality of survival, they are frequently
the setting for complications that increase morbidity
and mortality. Complications may result from intravascular catheters or from mechanical ventilators.
Nosocomial infection (Rodriguez et al., 1991) is a major
source of increased morbidity and prolonged ICU stay
(Hsieh et al., 1992). The acute respiratory distress
syndrome and multiple organ system failure may
occur as a result of the initial insult (Piek et al.,
1992).
334
RESPIRATORY AND CARDIOVASCULAR SUPPORT
It is therefore important that the risk of each

procedure or drug administered is balanced by a net
benefit to the patient. To minimize complications and
to maximize patient benefit, all clinicians involved in
management must understand cardiopulmonary and
metabolic pathophysiology and the particular effects
of head injury. This applies both to neurosurgeons and
intensivists.
The main determinants of alveolar oxygen tension

(PAO2) are inspired oxygen tension and alveolar
ventilation.
17.2
Inspired oxygen tension (PIO2) is determined by the

fractional concentration of oxygen (FiO2) and the
ambient barometric pressure (BP) corrected for saturated water vapor pressure:
Physiology
Optimal cardiorespiratory function exists when oxygen delivery from the atmosphere to the cell matches
oxygen demand (Figure 17.1; Nunn, 1987a).
After a severe head injury, the brain is very
vulnerable to secondary insults, and remains so for a
period of perhaps 23 weeks. The oxygen cascade
may be threatened at a single link in the pathway, for
example by airway obstruction, but more often several
links are threatened simultaneously. The threat may be
primarily to the respiratory system (e.g. by chest
trauma or pneumonia) or to the circulatory system
(e.g. by hypovolemia or myocardial failure). Continued impairment of oxygen delivery may lead to
multiple organ failure, especially after traumatic injuries and sepsis.
Understanding the physiological determinants of
the various stages of the oxygen cascade may provide
guidance for optimizing cellular oxygen tension.
(West, 1985; Nunn, 1987a).
17.2.1 DISTRIBUTION OF OXYGEN FROM THE

ATMOSPHERE TO THE ALVEOLI
(a)
Inspired oxygen tension
PIO2 = FiO2 (BP Pwat)

where Pwat represents the saturated water vapor
pressure (47 mmHg).
PIO2 is proportional to FiO2 if barometric pressure and
saturated water vapor pressure remain constant.
Ambient air has an oxygen concentration of 21%, so
that by administering 100% oxygen with a closed
circuit PIO2 may be raised from 150 mmHg in air to
700 mmHg. This is achieved by alveolar denitrogenation since ambient air contains 80% nitrogen.
Hence, since PIO2 can be effectively and markedly
raised by the administration of supplemental oxygen,
all severely injured patients should receive supplemental oxygen at the highest percentage available as
early as possible, to maximize the supply of oxygen to
the cells. (The effect of altitude on PIO2 is discussed in
Chapter 15.)
Figure 17.1 Transfer of oxygen from ambient air to mitochondria and summary of factors that influence oxygenation at
different levels in the oxygen cascade. PEO2 = end-expired gas. See text for details. (Source: modified from Nunn, 1987a,
p. 243, with permission.)
PHYSIOLOGY
Table 17.1
systems.
Capacities of available oxygen delivery
Apparatus/device
Oxygen flow
(l/min)
Inspired oxygen
concentration (%)
2
4
6
4
6
8
10
12
4
8
48
28
35
45
35
50
55
60
65
35
40, 60
4080
Variable
Variable
21100
21100
Nasal catheters
Semi-rigid masks (e.g.
Hudson, Edinburgh,
Harris, Mary Catterell)
Venturi type mask (e.g.
Ventimask, Accurox)
Reservoir plastic masks
(e.g. Pneumask,
Polymask)
Anesthetic circuits
Ventilators
Table 17.1 shows the oxygen delivery capacities of

available oxygen circuits.
(b)
Alveolar ventilation
Adequacy of ventilation is reflected by the PaCO2,

normally 40 mmHg at sea level. Hypercapnic respiratory failure is defined as a PaCO2 greater than
50 mmHg at sea level unless there is a compensatory
metabolic alkalosis. (Serum bicarbonate concentration
indicates whether hypercapnia is acute or chronic, as
renal compensation for chronic hypercapnia takes 23
days.)
PAO2 is related directly to alveolar ventilation and
inversely to arterial carbon dioxide tension (PaCO2).
This relationship is expressed in the alveolar gas
equation:
PAO2 = PIO2 PaCO2/RQ
where RQ is the respiratory quotient and reflects the
ratio of oxygen consumption to carbon dioxide production; normally 0.8.
The PAO2 in normal adults breathing air at sea level is
150 mmHg.
Hypercapnia may reduce PAO2 by displacing oxygen from the alveoli. Hypercapnia-induced respiratory acidosis may also reduce the consciousness level.
However this acidosis rapidly resolves on restitution
of effective alveolar ventilation. The acute stress
response to injury also increases basal metabolic rate
and carbon dioxide production which may further
reduce PAO2. The respiratory quotient usually remains
within the normal range due to an associated increase
O2).
in oxygen consumption (V
335
The other determinant of PaCO2 is carbon dioxide

CO2); normally 200 ml/minute in a
production (V
healthy person. An increase in carbon dioxide produc A)
tion or decrease in effective alveolar ventilation (V
will increase PaCO2 as expressed in the relationship:
CO2/V
a.
PaCO2 = V
Adequacy of ventilation is assessed by arterial blood
gas analysis obtained from repeated arterial punctures
or preferably from indwelling arterial catheters. A
continuous approximation of PaCO2 may be gained
from the end-tidal carbon dioxide tension measured
by a capnograph. Continuous end-tidal CO2 tension
may not represent PaCO2 in ventilated patients, particularly if there is coexistent lung injury or disease, and
should be checked frequently against PaCO2 (Liu, Lee
and Bongard, 1992). End-tidal carbon dioxide is
usually slightly less than PaCO2, but wide limits of
agreement between these parameters have been demonstrated (Russell, Graybeal and Strout, 1990;
Myburgh, 1991, Pansard et al., 1992).
Ventilatory variables such as respiratory rate (f),
tidal volume (VT) and their product, minute volume
E), may be accurately measured with most modern
(V
A) is difficult to
ventilators. Alveolar ventilation (V
determine and cannot be measured directly, although
E and dead
alveolar ventilation may be inferred if V
D) are calculated:
space ventilation (V
A + V
D.
E = V
V
Dead space refers to those airways that are ventilated
but take no part in gaseous exchange and consists of
both conducting airways (anatomical dead space) and
non-perfused alveoli (physiological dead space;
Nunn, 1987a). The normal ratio of dead space to tidal
volume (VD/VT ratio) is 0.3. This may be calculated in
patients whose PaCO2 and expired carbon dioxide
tensions (PECO2) are known, using the modified Bohr
equation:
VD
VT
PaCO2 PECO2
PaCO2
This ratio is an index of effective ventilation and is

helpful in determining the time to wean from mechanical ventilation. The VD/VT ratio in mechanically
ventilated patients with normal lungs is approximately 0.5. A ratio of 0.6 or greater is associated with
inadequate alveolar ventilation and hypercapnia.
In acute head injury, impaired conscious state is
associated with a loss of protective airway reflexes
and hypoventilation. Abnormal breathing patterns
may be exacerbated by alcohol, drugs, respiratory
injury or disease, resulting in alveolar hypoventilation and a rise in PaCO2. (North and Jennett, 1974;
Demling and Riessen, 1990). Hypoxia then rapidly
336
ensues and is potentiated by an increase in oxygen

consumption (Pfenninger and Lindner, 1991). This
reinforces the need for an adequate airway, ventilation and supplementary oxygen in these patients.
(c)
Lung mechanics
In addition to respiratory rate, tidal and minute

volume, other aspects of lung mechanics, in particular
lung compliance and airways resistance, are important
for effective alveolar ventilation in both spontaneously breathing and mechanically ventilated
patients (Bone, 1985; Nunn, 1987b).
17.2.2 DETERMINATION OF THE ARTERIAL OXYGEN
TENSION
Arterial oxygen tension (PaO2) is determined by the

diffusion capacity of the alveolarcapillary membrane,
pulmonary shunting and the relationship between
ventilation and perfusion. PaO2 is frequently a major
limiting step in the transfer of oxygen to the tissues.
Diffusion of oxygen across the short diffusion
distances of the alveolar capillary membrane normally
results in a negligible reduction in PaO2.
Ventilationperfusion scatter or mismatching of
ventilation and perfusion (normally by 4 mmHg)
occurs through shunt and increased dead space
ventilation and reduces both oxygen and carbon
dioxide transfer.
Shunt is defined as perfusion of unventilated areas
of the lung. Normally a small shunt of approximately
4 mmHg exists as a result of the bronchial and cardiac
venous (Thebesian) circulations. Important features of
true shunting are that hypoxia due to shunting cannot
be abolished by increasing the FiO2, and the degree of
resulting hypoxia is proportional to the shunt
fraction.
Mismatch is increased in patients with underlying
chronic airways disease or pulmonary hypertension.
Neurally induced ventilationperfusion mismatch has
been described as an uncommon complication of acute
head injury (Schumacker et al., 1979). This may be
exacerbated by acute intercurrent pulmonary pathology such as contusion, aspiration or infection, or
simply by placing patients on mechanical ventilators
(Hubmayr, Abel and Rehder, 1990). Pulmonary thromboembolism may cause severe hypoxia through both
ventilationperfusion mismatch and reduction in cardiac output (Demling and Riessen, 1990).
These three factors diffusion gradient, shunting
and ventilation perfusion mismatch are responsible
for a normal alveolararterial oxygen gradient of
approximately 10 mmHg, determined by subtracting a
measured PaO2 from PAO2 calculated by the alveolar
gas equation.
Potentiation of any of these physiological causes of

reduced arterial oxygen tension by pathological processes will result in hypoxia.
An increased alveolararterial oxygen gradient
after head injury is most often due to alveolar
obstruction or collapse caused by contusion, aspiration of gastric content or blood or consolidation from
infection. Neurally induced ventilationperfusion
mismatch is rare.
17.2.3 DELIVERY OF OXYGEN FROM THE ALVEOLI
TO THE TISSUES
O2) to individual cells involves

Oxygen delivery (D
convective and diffusive oxygen transport (Leach and
Treacher, 1992).
Convective oxygen transport refers to the bulk
movement of oxygen in the blood. It depends primarily on the distribution of cardiac output to individual
organs and the mechanisms regulating tissue
microcirculation.
Diffusive oxygen transport is the process of uptake
and extraction of oxygen from blood by the tissues
(Leach and Treacher, 1992).
O2 may be calculated as the product of arterial
D
oxygen content (CaO2) and cardiac output (Q):

O2 = Q
CaO2.
D
O2 is 9001300 ml/min.
The normal value for D
(a)
Oxygen content
Arterial oxygen content (CaO2) is the total oxygen per

milliliter of blood and includes both oxygen combined
with hemoglobin and dissolved oxygen. CaO2 (normal
value 1620 ml O2/100 ml blood) may be measured
directly with oxygen content analyzers or calculated
by the following equation:
CaO2 = (Hb 1.34 SaO2) + (PaO2 0.003)
where Hb is hemoglobin (g/dl); 1.34 is the oxygen
carrying capacity of hemoglobin (ml/g); SaO2 is
arterial oxygen saturation of hemoglobin (%); 0.003
represents the solubility of oxygen in blood in plasma
at 37C.
The chemical relationship between oxygen and hemoglobin is represented by the oxygenhemoglobin
dissociation curve. The idealized curve assumes a
normal hemoglobin and P50 (oxygen tension with 50%
hemoglobin saturation). A number of logarithmic
equations have been designed to determine oxygen
saturation from a single PaO2 measurement (Breuer et
al., 1989). However, P50, which is influenced by pH
and 2,3-diphosphoglycerate, a marker of intracellular
metabolism and utilization of oxygen, is frequently
reduced in critically ill patients (Myburgh, Webb and
PHYSIOLOGY
Worthley, 1991). Wide limits of agreement between

calculated and measured saturations have been demonstrated. Therefore oxygen saturations, either arterial
or venous, should be measured directly by oximetric
analysis when calculating oxygen delivery or consumption (Myburgh, 1992). This also applies to the
measurement of jugular venous bulb oxygen saturation (see below).
The major determinant of arterial oxygen content is
hemoglobin. Dissolved oxygen is a minor component
of the total. Anemic or bleeding patients have a
reduced oxygen transport capacity; hence transfusion
of blood is an integral part of resuscitation in trauma
patients.
There is contention on the optimal hemoglobin level
in severely head-injured patients with coexisting
extracranial injuries (Dhainaut et al., 1990; Greenberg,
1990) and consideration must be given to the risk and
benefits of small blood transfusions intended to raise
the hemoglobin level by only 12 g/dl (Dietrich et al.,
1990). There is increasing evidence that a fall in
hemoglobin is part of the early response to trauma
(Crosby, 1992), and that single unit transfusion does
not significantly increase tissue oxygen utilization and
may cause splanchnic ischemia (Silverman and Tuma,
1992). Impairment of immune function and the risk of
transmitting infection are additional arguments
against single unit transfusion in critically ill
patients.
Optimal tissue oxygen extraction from hemoglobin
to the brain and myocardium occurs at a hematocrit
of around 0.32 (Dantzker, 1989). It is not possible to
recommend a specific hemoglobin or hematocrit for
all clinical situations. However a hemoglobin of 8 g/dl
is regarded as the transfusion limit for young and
otherwise healthy trauma victims.
(b)
Cardiac output
Cardiac output is the product of heart rate (HR) and

stroke volume (SV). Any reduction in cardiac output
will reduce oxygen delivery.
Heart rate, normally 70 beats per minute, is determined by autonomic influences on intrinsic cardiac
pacemakers and is the principal factor maintaining
cardiac output in the traumatized patient (Schumacker and Cain, 1987; Guyton, 1991). The heart rate
response may be blunted in patients on beta-blockers
or in patients with high spinal injuries causing cardiac
denervation (Luce, 1985).
Stroke volume, normally 70 ml/beat, is determined
by three factors:
preload, defined as end diastolic cardiac muscle

fiber length and representing the volume status of
the patient;

337
contractility,
the
inotropic
state
of
the
myocardium;
afterload, the impedance to ventricular ejection.
Preload is the most important determinant of stroke

volume after trauma (Dhainaut et al., 1990). Reduced
preload, i.e. hypovolemia, must be aggressively treated to ensure adequate tissue oxygen delivery.
Impaired contractility may result from direct myocardial injury or pre-existing ischemic cardiomyopathy. Drugs (inotropes) are being increasingly used to
maintain adequate contractility once hypovolemia
has been corrected, not only to promote oxygen
delivery but also to maintain cerebral and myocardial
perfusion (Schumacker and Samsel, 1989; Dhainaut et
al., 1990).
Mean arterial pressure (MAP) is the product of
cardiac output and total peripheral resistance, derived
as systemic vascular resistance (SVR) from pulmonary artery catheter measurements.
17.2.4 PERIPHERAL EXTRACTION AND UTILIZATION
OF OXYGEN
The uptake and extraction of oxygen from blood by

the tissues are dependent primarily on diffusive
oxygen transport (Leach and Treacher, 1992). This
refers to the transfer of oxygen from blood through the
extracellular matrix along a capillaryintracellular
oxygen tension gradient; the process depends on
arterial oxygen tension. Alterations in microvasculature tone by neurohumoral (myogenic) and metabolic means adjust local oxygen delivery to changing
metabolic demands. Microvascular beds in different
tissues respond to increased oxygen requirements
either by increasing flow or by recruiting additional
vessels. For example, skeletal muscle can substantially
increase the number of functioning capillaries in
response to the increased metabolic demand, whereas
the myocardium and brain are more dependent on
changes in pressure and flow.
In acute neurotrauma, it is essential to maintain
adequate cardiac output and cerebral perfusion pressure in order to sustain flow-dependent oxygen transport. Interstitial edema, by increasing the distance for
diffusion, also jeopardizes oxygen transfer (Schumacker and Samsel, 1989).
Assessment of oxygen utilization by tissues is
difficult and in general only global oxygen consumption and delivery can be measured easily. Global (i.e.
whole body) oxygen consumption (normally
175300 ml/min) is best measured by indirect calorimetry of expired respiratory gases using metabolic
monitors. The accuracy of these monitors is reduced at
high inspired oxygen concentrations (FiO2 > 0.8) and
by air leaks (Takala et al., 1989). Oxygen consumption
may be determined by the Fick equation, which
338
expresses the relationship between cardiac output (Q),
oxygen consumption (VO2) and the arteriovenous

oxygen content difference:
O2 = Q
(CaO2 CvO2).
V
The measurement and manipulation of oxygen delivery and extraction have been extensively debated
(Vincent, 1990; Dantzker, Foresman and Gutierrez,
1991; Shoemaker, Appel and Kram, 1992; Reed, 1993).
When oxygen delivery is reduced under physiological
conditions, oxygen consumption is maintained until
very low levels of oxygen delivery through increasing
oxygen extraction. Below a critical oxygen delivery
level oxygen consumption falls even though oxygen
extraction continues to increase (Schumacker and
Cain, 1987).
Under pathological conditions, such as trauma and
sepsis, the relationship between oxygen delivery and
consumption is more linear and a critical oxygen
delivery level cannot be identified. Recent studies
indicate that outcome in patients is improved if
oxygen consumption matches delivery, if necessary
through cardiovascular support (Shoemaker et al.,
1988; Reinhart, Hannemann and Kuss, 1990).
The clinical measurement of both oxygen consumption and delivery remains difficult and is compounded by mathematical errors (Archie, 1987; Reed,
1993). Elevated serum lactate does not necessarily
imply anaerobic metabolism, and its value as an index
of regional oxygen demand is limited (Silverman,
1991; Mizock and Falk, 1992).
17.3 Effects of head injury upon oxygen

delivery
Severe head injury may impair oxygen delivery
through the development of abnormal breathing patterns, activation of a systemic inflammatory response
and centrally mediated sympathetic overactivity.
17.3.1 BREATHING PATTERNS IN HEAD-INJURED
PATIENTS
Variable breathing patterns may result from head

trauma and other insults to the central nervous system
(North and Jennett, 1974). CheyneStokes or periodic
respiration is associated with bilateral hemispheric
supratentorial lesions; central neurogenic hyperventilation or apneustic respiration with pontine lesions,
and ataxic respiration or apnea with medullary or
midbrain lesions. However, in practice, the diffuse
nature of the primary and secondary effects of trauma
upon central brain structures means that these patterns are seldom of diagnostic value.
Abdominal paradox, an inward movement of the
abdominal wall during inspiration, indicates abnor-
mal diaphragmatic contraction and may be seen in

patients with high cervical spinal cord injury above or
involving the phrenic nerve nuclei (level C3C5; Luce,
1985). Chest trauma causing flail chest and pulmonary
contusion results in tachypnea and intercostal muscle
retraction (Demling and Riessen, 1990; Luce, 1993).
17.3.2 SYSTEMIC PHYSIOLOGICAL RESPONSE TO
ACUTE INJURY
The systemic physiological response to acute injury

and infection causes a potent, mediator-induced, nonspecific syndrome termed the systemic inflammatory
response syndrome (SIRS). This syndrome has been
extensively studied (Bone, 1991a, b; Bone et al., 1992a;
Vincent and Bihari, 1992) and includes one or more of
the following clinical manifestations:

body temperature greater than 38C or less than

36C;
heart rate greater than 90 beats/min;
tachypnea, with a respiratory rate greater than
20 breaths/min or hyperventilation, as indicated by
a PaCO2 less than 32 mmHg;
white blood cell count greater than 12 000/mm3 or
less than 4000/mm3, or containing more than 10%
immature neutrophils (bands) representing an acute
alteration from baseline in the absence of other
known causes such as chemotherapy, induced
neutropenia and leukopenia.
Infection may produce a systemic inflammatory

response syndrome. Non-infectious pathological
events such as multiple trauma, hemorrhagic shock,
ischemia and pancreatitis result in an identical syndrome (Bone et al., 1992a).
SIRS is mediated through a number of cytokines
such as tumor necrosis factor (Strieter, Kunkel and
Bone, 1993; Tracey and Cerami, 1993), interleukins
(Dinarello, 1984; Platanias and Vogelgang, 1990) and
arachidonic acid metabolites (Lewis, Austen and
Soberman, 1990). The organ system responses, particularly the cardiovascular, respiratory and renal responses, may either regress or progress to well-defined
clinical conditions such as acute lung injury, septic
shock, acute renal failure and multiple organ dysfunction syndrome (Clarke, 1985; Cerra, 1992). Hemodynamic hallmarks of this response are a high cardiac
output state due to reduced systemic vascular resistance and global blood flow redistribution. Cardiac
output and stroke volume are usually well maintained during the initial acute stages of this response
and may reflect a mediator-induced increase in oxygen
delivery to vital tissue beds.
It is important to recognize the systemic inflammatory response syndrome and to actively seek infectious and/or non-infectious causes. If infection occurs,
EFFECTS OF HEAD INJURY UPON OXYGEN DELIVERY
339
SIRS may progress to septic shock, maldistribution of

blood flow, myocardial depression and end organ
failure. (Bersten and Sibbald, 1989).
cess. Beta-blocking drugs appear to be preferable,

particularly when ICP is elevated (Robertson et al.,
1983), but are seldom needed.
17.3.3
(b)
NEUROHUMORAL RESPONSE
In addition to the acute systemic inflammatory

response, sympathetically mediated responses following acute head injury result in three clinical
symptom complexes: neurogenic hypertension, myocardial ischemia and neurogenic pulmonary edema.
(a)
Neurogenic hypertension
Neurogenic hypertension is common after head injury.

It appears to be sympathetically mediated and is
directly proportional to the degree of catecholamine
release (Graf and Rossi, 1978). Brain areas implicated
include the hypothalamus, the nucleus of the tractus
solitarius, the area postrema and areas A1 and A5 of
the medulla (Kirland and Wilson, 1991). Increased ICP,
brain-stem compression or medullary ischemia may
cause severe systemic hypertension, accompanied or
preceded by vagally mediated bradycardia (Graf and
Rossi, 1978). This is the classical Cushing response but
in practice it is not always seen; the hypertensive
response can be masked by coexisting hypotension
due either to hypovolemia or to a hyperdynamic
systemic inflammatory response.
In as many as 25% of patients with neurogenic
arterial hypertension, there is no associated increase
in ICP (Simard and Bellefleur, 1989).
Unlike essential hypertension, neurogenic hypertension may be quite labile, with varying cardiac
output and systemic vascular resistance, probably due
to mediator-induced vasoresponsiveness. When or
whether to treat neurogenic hypertension is still
controversial, since an elevated systemic arterial
pressure may represent an intrinsic compensation for
impaired cerebrovascular autoregulation and low
cerebral perfusion pressure. Cerebral autoregulation
may be abolished or impaired in the injured brain,
either regionally or globally, so that the normal
relationship between CPP and blood flow is disrupted
(Obrist et al., 1979) and cerebral blood flow becomes
perfusion pressure-dependent (Simard and Bellefleur,
1989). Theoretically, elevating systemic arterial pressure may benefit ischemic areas, but may also increase
hydrostatic pressure and vasogenic cerebral edema.
Conversely, low arterial pressures may increase ischemia, cellular hypoxia and cytotoxic cerebral edema
(Kong et al., 1991).
Neurogenic hypertension should only be treated
when mean arterial pressure approaches 150 mmHg
and sedation or analgesia with opiates and/or benzodiazepines has already been employed without suc-
Myocardial injury
Myocardial injury, in the absence of coronary artery

disease, may occur in up to 50% of head-injured
patients. The lesions are similar to those seen after
acute myocardial infarction, subarachnoid hemorrhage, pheochromocytoma or prolonged catecholamine infusions, where subendocardial hemorrhages
are commonly found at autopsy. Inhibition of both
catecholamine release (Talman, 1985) with adequate
sedation and beta-adrenergic receptor blockade (NeilDwyer et al., 1978; Cruikshank et al., 1987) have been
shown to prevent or ameliorate myocardial injury
after head injury.
Although the brain effector site is unknown, there
may be an association between hypothalamic lesions
and myocardial damage. In patients with myocardial
injury, cardiac isoenzymes may be elevated and
electrocardiogram changes consistent with an infarction pattern can appear. These changes may be
potentiated by pre-existing ischemic heart disease or
hypertension. It has been suggested that electrocardiogram changes may indicate patients with more
serious brain injury and thus a higher mortality
(Kirland and Wilson, 1991).
Similarly, a wide range of cardiac dysrhythmias
may occur as a result of centrally mediated sympathetic and vagal discharge. The true incidence of
dysrhythmias is unknown and all patients with a
severe head injury should have electrocardiograph
monitoring. Tachyarrhythmias are predominantly
supraventricular and include atrial fibrillation or
flutter and paroxysmal or multifocal atrial tachycardias (Cruikshank et al., 1987). Ventricular tachyarrhythmias are uncommon. Reducing sympathetic
tone with adequate sedation or occasionally with
beta-blockade is usually sufficient treatment for
tachyarrhythmias. In patients with poor ventricular
function persistent hemodynamically significant
tachyarrhythmias are best treated with class 3 agents
such as amiodarone. Nodal and sinus bradycardias
are quite commonly associated with sustained intracranial hypertension and rarely require treatment.
Atropine, beta-agonists or, rarely, transvenous pacing
may be indicated if hemodynamic compromise
results.
These cardiovascular conditions are seen frequently
during the initial stages of neurotrauma. However,
hypertension may persist into the rehabilitation phase.
In a study of head-injured patients in a rehabilitation
center, 11% had systemic hypertension yet only one
had a history of essential hypertension and 68% were
340
less than 30 years of age. At discharge, 63% still

required antihypertensive medications. A total of 21%
of these patients had non-specific STT wave changes
on their electrocardiograms (Kalizky et al., 1985).
(c)
Neurogenic pulmonary edema
The most dramatic of these sympathetically mediated

processes is neurogenic pulmonary edema. It was
reported many years ago to be present in most
patients who died of intracranial hemorrhage (Weismann, 1939) and in patients with severe isolated head
injury (Simmons et al., 1969), but was much less
common in patients who survived (Singbartl, Cunitz
and Hamrouni, 1982). The common locus of damage is
the brain stem. The mechanism is a large sympathetic
discharge (Demling and Riessen, 1990), which leads to
a redistribution of blood from the systemic to the
pulmonary vasculature and a rapid increase in pulmonary vascular pressures. This is a transient phenomenon but it may damage the pulmonary endothelium,
causing a persistent permeability defect (Theodore
and Robin, 1976)
This mechanism alone does not explain all cases of
neurogenic pulmonary edema. Other mechanisms
that may be involved include: neurogenic depletion
of pulmonary surfactant (Beckman, Bean and Baslock, 1974); direct sympathetic control of the number
and size of pulmonary endothelial pores (Rosell,
1980); constriction of the pulmonary lymphatics by
the sympathetic discharge (McHale and Roddie,
1983); lung microembolization from activation of the
clotting cascade, possibly caused by release of brain
tissue thromboplastin into the systemic circulation
(Kaufman et al., 1980; Simpson, Speed and Blumbergs, 1991); and the direct effects of increased
sympathetic activity on lung vasculature resulting in
pulmonary hypoperfusion and hypoxia (Balk and
Bone, 1983).
Typically, neurogenic pulmonary edema has an
immediate onset and becomes clinically evident 212
hours after injury. Occasionally clinical onset may be
delayed from 12 hours to several days. It is usually
self-limiting and, if the patient survives, it will resolve
in hours to days. A prolonged episode of pulmonary
edema would suggest other causes of either cardiogenic or non-cardiogenic origin.
The clinical findings in patients with neurogenic
pulmonary edema are tachypnea, hypoxia, decreased
pulmonary compliance, diffuse alveolar and interstitial fluffy infiltrates on chest X-ray and low
filling pressures measured by right heart catheters (cf.
acute respiratory distress syndrome).
Treatment is primarily supportive and aims at
ensuring adequate oxygenation and ventilation. This
usually requires endotracheal intubation and mechan-
ical ventilation with positive end-expired pressure

(PEEP).
Sympathetic overactivity is usually abolished with
adequate sedation and beta-blockade is usually
unnecessary. Diuretics are effective but must be
titrated against the volume status of the patient so that
cerebral perfusion pressure is not compromised.
17.4
Management: resuscitation
The resuscitation phase is defined as the period from

on-site resuscitation and evacuation to stabilization.
Priority must be given to the most life-threatening
problems and this is the principle of basic and
advanced trauma life support systems (Buss, Nel and
van den Heever, 1990; Trauma Committee, Royal
Australasian College of Surgeons, 1992).
17.4.1
AIRWAY MANAGEMENT
Loss of airway patency and hypoxia at any stage of

head injury management is a life-threatening event. It
is the main cause of secondary brain injury and
significantly increases morbidity and mortality (Gentleman, 1992; Chesnut et al., 1993).
All severely head-injured patients are at risk from
inadequate airway protection. The degree of airway
intervention necessary will depend on the level of
consciousness and adequacy of protective glottic
reflexes.
The mouth and upper airway must be inspected for
foreign bodies, bleeding or ill-fitting dentures and
cleared under direct vision using a rigid sucker.
Simple maneuvers such as chin lift and jaw thrust and
the use of oropharyngeal airways may be enough to
allow efficient oxygenation in self-ventilating patients.
Nasopharyngeal airways must be used with caution if
cribriform plate fracture is suspected since they have
been passed directly into the cranial cavity. They may
also cause trauma to the nasal mucosa with bleeding
into the nasopharynx and upper airway. The same
cautions apply to the use of nasotracheal and nasogastric tubes.
Emergency endotracheal intubation is indicated
whenever the airway is threatened, especially in the
comatose head-injured patient. It is also indicated in
faciomaxillary trauma or upper airway obstruction
from direct laryngeal trauma. This is best performed
where there is expertise in anesthesia and specific
airway techniques such as rapid sequence induction
and intubation, blind nasal intubation, fiberoptic
laryngoscopy, cricothyroidotomy and tracheostomy.
However, urgency may dictate intubation in less
optimal circumstances.
Indications for early endotracheal intubation after
head injury are:
MANAGEMENT: RESUSCITATION

coma;
inability to obey commands with evidence of
respiratory compromise:
tachypnea > 30/min
sternal recession
upper respiratory bleeding;
blood
gas
analysis
PaO2 < 70 mmHg;

PaCO2 < 45 mmHg.
Table 17.2 Priorities of rapid sequence induction of

anesthesia for emergency intubation in neurotrauma. See
text for further details. ETCO2 = end-tidal carbon dioxide
tension
Optimize conditions
for intubation
Skilled assistant
Adequate light
Suction
Neutral position
Equipment
Oropharyngeal airways
Self inflating hand ventilating
assembly and mask
100% oxygen
Two laryngoscopes
Magill forceps
Introducer/bougie
Endotracheal tubes (mm):
7.0, 8.0
8.0, 9.0
Cricothyroidotomy equipment:
Scalpel
6.0 mm cuffed endotracheal tube
Drugs
Induction agent
Suxamethonium (12 mg/kg)
Atropine (0.61.2 mg)
Adrenaline (10 ml 1:10 000
solution)
Sedation (fentanyl, diazepam,
morphine, midazolam)
Monitoring
Pulse oximetry
Capnography
Arterial blood pressure
Electrocardiograph
Procedure
Preoxygenation 100% oxygen for

34 minutes
Preload with 250500 ml colloid
intravenously
Manual in-line cervical spine
immobilization
Cricoid pressure applied
Induction agent + suxamethonium
Direct visualization of vocal cords
Oral endotracheal intubation
Inflation of cuff until sealed
Confirmation of end-tidal CO2 and
chest auscultation
Cricoid pressure released
Secure tube at correct length
Ensure adequate sedation muscle
relaxant
Consider naso- or orogastric
intubation
Chest X-ray
Confirm blood gas analysis
Indication for cricothyroidotomy or tracheostomy

are:

coma and inability to intubate;

trauma, bleeding or fractures involving the upper
airway.
Comatose patients (Glasgow Coma Score less than 8)

should be intubated electively as soon as possible to
prevent aspiration of gastric contents, to allow effective oxygenation, ventilation and control of PaCO2 and
definitive imaging with computed tomography (CT;
Pfenninger and Lindner, 1991). Similarly, head-injured
patients who are agitated or combative may best be
managed by early intubation and controlled ventilation until diagnostic and therapeutic interventions are
completed (Redan et al., 1991). These patients should
be intubated orally using a rapid sequence induction
of anesthesia after preoxygenation and with the
application of cricoid pressure as described in Table
17.2 (Walls, 1993). Patients with significant extracranial injuries causing ventilatory failure must also be
intubated early. Patients with moderately depressed
or fluctuating conscious states may be observed
provided adequate glottic reflexes are present and the
cardiorespiratory state is stable.
All unconscious head-injured patients must be
regarded as having cervical spine injuries and intubation should be performed with in-line cervical spine
immobilization with the head in the neutral position
so that neck flexion and extension is avoided. Patients
may be intubated with rigid collars in place and these
should be left in situ until definitive radiological
views of the cervical spine are obtained (Crosby and
Lui, 1990).
The best method of assuring correct placement of an
endotracheal tube is by directly visualizing the passage of the tube through the vocal cords (Szekely et al.,
1993). End-tidal carbon dioxide and a respiratory
waveform on capnography (Lillie and Roberts, 1988),
auscultation of the chest and misting on the endotracheal tube during exhalation give further confirmation of correct placement, although false positives
have been described following inadvertent esophageal
intubation (Holland, Webb and Runciman, 1993).
Advancing the tube down the right main bronchus is
a common error especially in children. It is best
prevented by repeated auscultation and chest X-ray
341
examination. Tubes must be secured at the appropriate

length (normally 21 cm at the teeth in women and
23 cm in men). Adhesive tape is the most effective
means of securing endotracheal tubes.
Tracheostomy is only indicated acutely in patients
who cannot be intubated; e.g. those with laryngeal or
342
tracheal trauma or some patients with high cervical

spinal cord injury.
17.4.2
VENTILATION
Once the airway is secured, all injured patients must

receive oxygen at the highest concentration available
(Oh and Duncan, 1988). The decision to continue
ventilation is primarily a clinical one and is based on
the severity and pattern of head injury, associated
injuries and pre-existing lung function. Arterial blood
gas analysis of PaO2, PaCO2 and pH are important but
should not delay this decision.
For self-ventilating patients, face mask oxygen
using the delivery systems outlined in Table 17.1 are
suitable. For intubated patients, manual or mechanical
ventilating assemblies can reliably deliver an FiO2 of
1.0. Hand-held, self-inflating assemblies can effectively ventilate both intubated and non-intubated
patients and also give an impression of lung compliance during inflation, thereby reducing the risk of
disconnection from either the endotracheal tube or the
face mask (Table 17.1).
Emergency assessment of oxygenation may be
difficult when an impaired conscious state may be due
to head injury, alcohol, drugs or sedatives. Pulse
oximetry has greatly aided this assessment and is now
mandatory in trauma patients. A saturation greater
than 95% generally indicates a PaO2 of at least
75 mmHg and is recommended.
Pulse oximeters are less reliable in hypoperfused,
hypothermic and agitated patients (Runciman et al.,
1993). Nevertheless, even under conditions of poor
perfusion, a recorded saturation of less than 95% with
a pulse waveform indicates hypoxia (Clayton et al.,
1991). This must be corrected by delivering the highest
possible oxygen concentration, reassessing the airway
and ventilation, and excluding potential causes of
unexplained hypoxia such as tension pneumothorax
or cardiac tamponade. Arterial blood gas analysis
should be performed as soon as possible, as this is the
only means of calibrating a pulse oximeter.
For 3040 years hyperventilation has been advocated for the initial reduction of intracranial pressure.
Global and regional cerebral blood flow may be
reduced by as much as 5060% during the first several
hours following injury. Acute hypocapnia may further
reduce cerebral blood flow in areas of the brain where
vasoresponsiveness to hypocapnia is relatively preserved. This may or may not reduce ICP, yet the
reduction in blood flow may worsen ischemia. Early
hyperventilation is now only recommended in
patients with clinical signs of herniation or rapid
neurological deterioration, prior to imaging or measurement of ICP and providing adequate circulatory
resuscitation is complete. The level of PaCO2 required
to reduce cerebral blood flow and achieve optimal ICP

is not well defined but, on balance, moderate hypocapnia of PaCO2 to not less than 30 mmHg is recommended. In the patient without clinical evidence of
intracranial hypertension or brain shift, ventilation
should be set to achieve normocapnia.
17.4.3
CIRCULATION
Hypotension after severe head injury is independently

associated with significant increases in morbidity
and mortality. Resuscitation protocols for braininjured patients should aim assiduously at avoiding
hypovolemic shock. (Buss, Nel and van der Heever,
1990; Chesnut et al., 1993; Hill, Abraham and West,
1993).
Restoring circulating blood volume is the basis of
hemodynamic resuscitation in trauma patients. The
initial assessment of circulatory status may be difficult
since blood pressure may be well maintained by
sympathetic activity. Tachycardia and reduced capillary return are cardinal signs of hypovolemia. However
tachycardia is common after trauma and peripheral
perfusion may be difficult to assess in hypothermic
patients. Hence, 12 liters of balanced salt solution
(lactated Ringers or physiological saline solution)
should be infused initially in all patients through largebore peripheral venous access until definitive monitoring of blood pressure and volume status is established.
Extracranial sources of hemorrhage must be identified
and rapidly treated. This requires a coordinated
trauma team approach where priorities are given to
the most life-threatening injuries in the shortest
possible time (Trunkey, 1991; Trauma Committee,
Royal Australasian College of Surgeons, 1992).
Hypotensive patients should be given colloid solutions to expand the intravascular volume and restore
preload. Normal (5%) serum albumin solution or
synthetic polygeline are suitable solutions. The use of
hypertonic solutions in neurotrauma remains controversial and is discussed in Chapter 16. Blood
transfusion must be administered to patients who
have lost more than 2030% blood volume or where
significant hemorrhage is anticipated. Patients who
require anesthesia for intubation should have adequate preload and arterial blood pressure monitoring
in situ as sudden ablation of sympathetic tone by
anesthetic agents may unmask a compensated hypovolemic state and induce hypotension.
As a general principle vasopressors should not be
used in hypovolemic patients until volume resuscitation is completed.
Causes of refractory hypotension in these patients
include acute spinal injury, tension pneumothorax,
cardiac tamponade and severe myocardial contusion.
These must be excluded early.
MAINTENANCE OF CARDIOPULMONARYCEREBRAL HOMEOSTASIS
17.4.4
SEDATION AND ANALGESIA
In the acute phase, sedation for intubated and ventilated patients must be titrated against the hemodynamic stability. High-dose narcotics (1525 g/kg
fentanyl) and non-depolarizing muscle relaxation
(pancuronium 810 mg) provide sufficient sedation
and allow control of ventilation for 12 hours while
imaging and other investigations are performed. This
combination provides hemodynamic stability; moreover tachycardias due to vagal blockade by pancuronium are reduced with high doses of fentanyl
(Salmenpera et al., 1983). Frequent reassessment of
conscious state is essential as excessive sympathetic
activity may potentiate raised intracranial pressure or
myocardial ischemia in a paralyzed but awake patient
(a state that should be carefully avoided). Sedation
may need to be supplemented with intermittent doses
of opiates or benzodiazepines.
17.4.5
(a)
MONITORING
Arterial pressure
During the initial resuscitation phase, a clinical

impression of cardiac output and tissue perfusion can
be gained from the strength of peripheral pulses and
the time required for superficial capillaries to refill
after they have been blanched (normally 23 s).
Accurate measurement of mean arterial blood pressure and heart rate is essential.
Non-invasive blood pressure devices are inaccurate
outside the normal range, tending to over-read in
states of hypotension and under-read in hypertension
(Rutten et al., 1986; Cockings et al., 1993). Radial,
brachial or femoral arterial catheterization is recommended. The arterial impulse may be connected via a
fluid-filled catheter to a transducer. This is the
standard method for measuring systemic blood pressure in the ICU and may also be used to obtain
samples for blood gas analysis (Runciman, Ilsley and
Rutten, 1988). Local vascular responses at the radial
arterial level may result in false elevations of or
reductions in blood pressure (Pauca et al., 1992).
Mean arterial pressure correlates more closely with
organ perfusion than systolic pressure and is less
susceptible to the errors of measurement that occur
with suboptimally damped transduced measuring
systems. Optimal mean arterial pressure is usually
6080 mmHg, although patients with pre-existing
hypertension or raised intracranial pressure may
require higher mean arterial pressures.
Complications of arterial catheterization such as
local hematoma formation, ischemia distal to the site
of insertion or local infection may be minimized by
slow continual flushing of these catheters with dilute
343
solutions containing heparin and by sterile catheter

insertion and maintenance techniques.
(b)
Electrocardiography
Electrocardiography is routine, but since changes in

heart rate in response to hypoxia and hypotension
occur late, the ECG is a poor marker of vital organ
oxygenation (Ludbrook et al., 1993).
(c)
Central venous pressure
Central venous pressure monitoring is a useful guide

to volume status, but the insertion of these catheters
should not delay the restoration of volume. Catheters
should be placed under sterile conditions by skilled
operators on a semi-elective basis and pressures
should be measured with electronic transducers (see
section 17.5.2(b)).
The response to fluid resuscitation is best assessed
by improvements in heart rate, capillary return and
mean arterial blood pressure, and by a maintained
urine output of 0.51 ml/kg/h.
17.4.6
NEUROSURGICAL DIAGNOSIS
Once patients have been adequately resuscitated the

definitive diagnosis of intracranial pathology, usually
by CT scanning, is essential. In a patient who shows
signs of brain-stem compression, evacuation of an
intracerebral hematoma is a matter of great urgency.
Resuscitation should therefore be conducted pari passu
with the CT scan, or the patient should be taken
directly to the operating room. Following CT diagnosis and/or evacuation of intracerebral hematomas,
ICP and other monitoring devices such as jugular bulb
oxygen catheters should be inserted
Accurate, lightweight and portable monitors allow
continuous display of ECG, heart rate, oxygen saturation and arterial pressure and are recommended
during transport. This phase of treatment is covered in
more detail in Chapters 15 and 18.
17.5 Management: maintenance of

cardiopulmonarycerebral homeostasis
After acute resuscitation and any necessary operative
procedures comes the phase of management in which
the chief aim is to prevent secondary hypoxicischemic
brain injury. This phase continues until the patient no
longer requires homeostatic control and support.
During this period patients are usually managed in the
Intensive Care Unit, with mechanical and/or pharmacological cardiorespiratory and metabolic support. The
period of vulnerability to secondary brain injury is not
344
clearly defined and varies from patient to patient, but it

may last up to 1014 days after injury and beyond the
time when intracranial pressure is elevated. Management of the patient requires close liaison between
intensivists, neurosurgeons and other involved
specialists.
All observations should be accurately recorded on
24-hour charts that correlate all vital signs and
cumulative hourly fluid balance. This facilitates early
identification of trends, a more important indicator of
progress than single isolated recordings.
17.5.1
(a)
RESPIRATORY MANAGEMENT
Airway management
Patients should be intubated until ICP is controlled

(i.e. consistently below 20 mmHg) and respiratory
function indices are satisfactory (see below). Endotracheal intubation may be performed either translaryngeally, through the nose or mouth, or transtracheally, via a tracheostomy. The development of
polyvinylchloride tubes with low-pressure, high-volume cuffs has reduced the incidence of tracheal
mucosal damage and ulceration from nasal or oral
endotracheal tubes. As a result, translaryngeal intubation may continue as long as the head injury and lung
function require it (Stone and Bogdonoff, 1992).
Nasal intubation gives good support to the endotracheal tube and patients often tolerate nasal tubes
with little sedation. These tubes are less prone to move
in the trachea than oral tubes and therefore cause less
tracheal mucosal damage. Patients are able to swallow
their secretions and oral hygiene is more easily
maintained than with an oral tube. However suction
through nasal tubes is more difficult because of their
narrower diameter and greater length. Nasal intubation is contraindicated in patients with basal skull
fracture especially when the cribriform plate is
damaged.
Oral intubation is preferred in an emergency. It
allows passage of a tube with a larger diameter (8 mm
or more) and easier bronchial toilet. However these
tubes are more prone to movement and more difficult
to secure. The incidence of nosocomial sinusitis is
significantly lower with oral intubation (Rouby et al.,
1994).
Tracheostomy is generally indicated in patients
where prolonged ventilation due to respiratory failure
is needed or where a slow return of consciousness is
anticipated, e.g. of more than 710 days.
These patients usually have poor protective glottic
reflexes and inadequate clearance of secretion, and are
prone to aspiration and nosocomial pneumonia. Tracheostomy tubes are more comfortable than translaryngeal tubes and permit easier tracheobronchial toilet.
Enteric feeding via a soft, small-bore nasoenteric tube

is safer after tracheostomy. Tubes with a device that
directs a stream of air retrogradely through the larynx
above the cuff site allow some patients to talk.
Tracheostomy should only be performed by experienced operators. The operation should not be done
until pulmonary reserve is adequate to tolerate a
procedure on the airway. In ventilated patients this
means that an FiO2 of 0.5 or less achieves a PaO2
greater than 80 mmHg and peak inflation pressures are
less than 30 cmH2O. Coagulation status should be
normal.
Surgical tracheostomy may be performed in the
operating room or in the ICU. Percutaneous dilatational tracheostomy techniques (Ciaglia, Firsching and
Syniec, 1985; Griggs et al., 1990) can be performed in
the ICU. They are quick and allow the passage of a
standard 8.0 or 9.0 mm tube. Tracheal stenosis is a
recognized complication of tracheostomy, although
the relative incidence of stenosis following the percutaneous procedure compared to surgical tracheostomy
is not yet known (McFarlane et al., 1994).
Irrespective of the route of intubation, cuffed tubes
must be inflated with sufficient volume to seal the
airway and prevent aspiration. Cuffs may develop
leaks as a result of changing cuff compliance, tracheal
dilatation with prolonged intubation or tube movement (Kearl and Hooper, 1993). Excessive cuff pressure may cause tracheal ulceration and later stenosis.
Frequent volumetric cuff checks and cuff pressure
measurements should be performed to ensure an
adequate and safe seal (Rischbieth, Blight and
Myburgh, 1994).
(b)
Principles of ventilation
Ventilation must provide adequate oxygenation and

control of arterial carbon dioxide tension.
Morbidity and mortality of patients supported by
mechanical ventilation, especially those receiving prolonged support, remain very high. There has been
increasing concern that iatrogenic problems relating to
mechanical ventilatory support may be important
contributing factors (MacIntyre, 1993).
Accidental disconnection from mechanical ventilators is the most common and potentially lethal
complication of mechanical ventilation. All ventilators must have a dedicated disconnection alarm.
Other complications of ventilation are patient
ventilator dyssynchrony, alveolar overdistention and
adverse effects of intrinsic positive end-expiratory
pressure (auto PEEP) .
Patientventilator dyssynchrony occurs either
when the breath delivered by the ventilator is out of
cycle with the spontaneous effort by the patient or
when the response of the ventilator is not adequate to
345
meet the flow demands of the patients effort. This

results in an increase in the pressure load placed on
the respiratory muscles and a substantial increase in
oxygen consumption and carbon dioxide production.
This may augment respiratory muscle fatigue and
failure (Mador, 1991; Oh et al., 1991). This is frequently
and incorrectly described as the patient fighting the
ventilator. The reverse is more appropriate: in reality
the ventilator is fighting the patient. The resulting
sympathetic overactivity may elevate ICP (Ersson et
al., 1990). Similar effects occur as a result of coughing
induced by endotracheal suctioning or of Valsalva
maneuvers in restless, agitated patients (Walsh et al.,
1989).
Optimizing the patientventilator interface requires
adequate sedation and the use of a mode of ventilation
that reduces the work of breathing.
Sedation is best achieved initially with a rapid onset
sedative/narcotic combined with a muscle relaxant,
such as fentanyl with pancuronium. Thereafter, combined infusions of a narcotic with a benzodiazepine
(e.g. morphine and midazolam) should be titrated
against the conscious state of the patient (Fisher and
Raper, 1991). The degree of sedation required may
vary widely according to the severity of the underlying head injury and individual pharmacokinetic
differences. The total dose of drug infused is not
important, but rather the titrated effect.
Propofol, a short-acting anesthetic induction agent
given by infusion, is becoming popular for the
treatment of head-injured patients. It has the theoretical advantage of a rapid and potent titrated sedation
without prolonging ventilation. Consciousness
returns rapidly on cessation of the drug because of its
rapid distribution and metabolism. (Farling, Johnston
and Coppel, 1989a, b). Propofol is a potent myocardial
depressant and cerebral perfusion should be monitored carefully. Definitive studies analyzing the longterm use of this agent as a sedative have yet to be
carried out.
Alveolar overdistention
There is now a clearer understanding of the pathophysiological consequences of alveolar overdistention in
mechanically ventilated, critically ill patients. Previously, patients were ventilated with tidal volumes
much larger (1015 ml/kg) than those occurring in
spontaneously breathing individuals. This is the
standard method of ventilating patients during general anesthesia and was advised in order to prevent
atelectasis by promoting recruitment of collapsed
alveoli, to negate the adverse effects of increased dead
space and to increase functional residual capacity.
There is increasing evidence that this technique is
dangerous in critically ill patients with injured lungs
or in those who require prolonged (> 24 h) ventilation.
The consequences include initiating or exacerbating
an acute lung injury (Tsumo, Prato and Kolobow, 1990;
MacIntyre, 1993), barotrauma to the lung and impairment of cardiac output (Dreyfuss and Saumon, 1992).
Pulmonary edema due to increased permeability has
also been demonstrated in human and animal studies
after several hours of ventilation with high tidal
volumes (Dreyfuss and Saumon, 1993). This phenomenon, termed volutrauma, is related to the degree of
alveolar volume and to distention rather than to peak
airway pressure.
In patients and animals with an acute lung injury,
large tidal volumes increase edema accumulation (and
extravascular lung water), leading to increased shunt,
impairment of oxygenation and a significant reduction
in pulmonary compliance (Carlton et al., 1990; Sohma
et al., 1992; Tsumo, Prato and Kolobow, 1990).
The inflating airway pressure is transmitted along
interstitial bronchovascular planes and may result in
pneumothorax, pneumomediastinum, pneumopericardium or pneumoperitoneum and increase in intrathoracic pressure. This will reduce venous return to
the heart, increase the impedance to ventricular
ejection and reduce cardiac output (Marini, 1990).
Neuromuscular blockade
(c)
The use of non-depolarizing muscle relaxants in

critically ill, ventilated patients should be questioned
for two reasons. Firstly a paralyzed but awake patient
is more susceptible to autonomic sympathetic swings
which will exacerbating raised ICP and predispose to
increased oxygen consumption. Secondly, a polyneuropathy has been demonstrated in ventilated
patients who have received regular non-depolarizing
muscle relaxants with or without steroids. This axonal
degeneration significantly prolongs ventilation and is
associated with increased extracranial complications
and morbidity (Gooch et al., 1991; Hansen-Flaschen,
Cowen and Raps, 1993; Hsiang et al., 1994).
Positive pressure ventilation is the routine form of

mechanical ventilation in current practice. Gas is
delivered under positive pressure to either a preset
pressure or preset volume. The mode selected for any
patient should be based upon the lung mechanics of
that patient (Bone and Eubanks, 1991; Sassoon, 1991;
Burchardi, Sydow and Criee, 1994). For most patients
with normal lungs a volume control mode will suffice.
Volume-controlled ventilation delivers a preset
tidal volume at a preset rate. This mode of ventilation
is suitable in patients with normal lung mechanics but
may cause barotrauma and alveolar overdistention in
patients with poorly compliant lungs. The mean
Modes of mechanical ventilation
346
inspiratory pressure may be reduced (pressure-limited

volume control ventilation) to prevent barotrauma,
but at the expense of the tidal volume. This mode
ignores spontaneous respiratory efforts by the patient
and should be used only in paralyzed, deeply unconscious or anesthetized and sedated patients. Most
portable transport ventilators are of this type and
patients being transported usually need muscle paralysis and increased sedation to optimize ventilation
(Phillips and Skowronski, 1986).
Assisted mechanical ventilation of various types is
designed to optimize the patientventilator interface
by using a volume control mode and allowing the
patient to trigger the ventilator to deliver a preset tidal
volume. This is accomplished by generating a small
negative airway pressure or flow that activates the
inspiratory cycle, the duration of which is determined
either by the ratio of inspiration to expiration (I:E
ratio) or by reaching a pressure limit. The sensitivity of
the triggering mechanism must be such that the
inspiratory efforts of the patient are detected without
imposing an increased work of breathing and oxygen
consumption, but not so sensitive that the ventilator
will cycle in response to fluctuations in airway
pressure.
Synchronized intermittent mandatory ventilation
(SIMV) is the standard form of assisted volume
controlled ventilation. This mode sets the mandatory
minute volume delivered to the patient while reducing dyssynchrony by only delivering ventilated
breaths at the end of a spontaneous expiration. The
degree of mandatory ventilation is adjusted according to the spontaneous minute volume of the patient.
Initially the ventilator respiratory rate may be set
high enough to provide most, if not all, of the
patients minute volume and then reduced as sedation is reduced or the underlying pathology
improves. The potential benefits of SIMV include less
patient ventilator dyssynchrony, lower sedation
requirements, reduced mean airway pressure by
combining spontaneous and ventilated breaths, and
improved respiratory muscle function by allowing
patients to breathe spontaneously (Bone and
Eubanks, 1991). Disadvantages may include insufficient mandatory minute volume in unstable patients
and the possibility of respiratory muscle fatigue in
patients with inappropriately low ventilator rates. As
with any volume control mode of ventilation, alveolar overdistention is a potential hazard; this can be
minimized by using an upper peak inspiratory pressure limit above which the ventilator will not deliver
a breath (usually 3040 cmH2O).
It has become routine to combine inspiratory
pressure support ventilation (PSV) with SIMV in
order to reduce the work of breathing through
ventilator and endotracheal tubing during the sponta-
neous breaths of the patient (Bersten et al., 1989). In

this mode, spontaneous inspirations are augmented
with a level of positive airway pressure that is preset
(usually to 20 cmH2O) to achieve a desired tidal
volume. As patients are weaned from SIMV by
reducing the ventilator rate and/or tidal volume, the
inspiratory pressure support will continue so that
patients set their own minute volume without incurring increased respiratory effort. Pressure support
ventilation alone is used chiefly to ensure an adequate
minute volume during the weaning phase and in
patients with adequate respiratory drive and minimal
sedation. It is not suitable for the acute phase of injury,
where control of minute ventilation and particularly
of PaCO2 is desired (Brochard et al., 1989; Santak et al.,
1991; Sassoon et al., 1991).
Pressure controlled ventilation delivers gas until a
preset peak inspiratory pressure is reached at a given
ventilator rate (Blanch et al., 1993a, b). The tidal
volume is determined by the patients compliance and
the I:E ratio. Increasing the inspiratory time in this
mode has been advocated so that the I:E ratio is
reduced (pressure control inverse ratio ventilation
PC-IRV) to progressively recruit collapsed alveoli.
Inspiratory pressure therefore remains constant in
order to maximize alveolar ventilation at the lowest
inflation pressure, not only through the whole period
of inspiration, but also when lung compliance
improves. The shortened expiratory time is intended
to prevent alveoli collapsing again and usually induces a higher auto PEEP. This mode of ventilation
remains controversial and manipulations of the I:E
ratio are best combined with frequent reassessment of
both lung mechanics and auto PEEP and the effects on
cardiac output and oxygen transport (Marcy and
Marini, 1991; Mercat et al., 1993).
(d)
Positive end-expiratory pressure (PEEP)
Auto PEEP
In addition to the factors outlined above, a subtle yet
potentially hazardous effect is the presence of an
elevated positive pressure on exhalation known as
intrinsic or auto PEEP (Schmidt and Wood, 1993).
Auto PEEP is seen typically in patients with airflow
obstruction, but it can also occur without obstruction
when minute ventilation is very high. Patients susceptible to the development of auto PEEP are those with
asthma, acute on chronic respiratory failure or acute
lung injury, where a prolonged expiratory time is
necessary.
Auto PEEP was first described in ventilated patients
(Pepe and Marini, 1982), but it can also occur in
spontaneously breathing patients as a result of exhalation against the glottis, varying in magnitude from
110 cmH2O. These patients usually have degrees of

airflow obstruction and the physiological effects of
auto PEEP are those of extrinsically applied PEEP,
namely alveolar recruitment to reduce shunt and
increase functional residual capacity. The work of
breathing to overcome auto PEEP may be significant.
When such patients are ventilated, auto PEEP together
with extrinsically applied PEEP may result in air
trapping at end expiration and the risk of barotrauma,
reduced cardiac output and increased physiological
dead space (Fessler et al., 1991; Pinsky, Desmet and
Vincent, 1992). The level of extrinsic PEEP applied
must be adjusted in accordance with the level of auto
PEEP so that alveolar overdistention and increased
work of breathing do not occur.
Auto PEEP can be dangerous if undetected. It is
therefore essential that auto PEEP be identified and
measured in ventilated patients by stopping air flow
at end expiration just before the next breath, allowing
the pressure in the airways and the ventilator tubing
to equilibrate and reading the pressure from a ventilator pressure gauge. Many new ventilators have the
facility to monitor and display the level of auto PEEP
(Gottfried, Reissman and Ranieri, 1992).
Extrinsic PEEP can be used to maintain alveolar
recruitment and functional residual capacity throughout the ventilatory cycle in order to improve arterial
oxygenation (Schmidt and Wood, 1993). Low levels of
PEEP have also been shown to significantly reduce the
work of breathing imposed by ventilators and endotracheal tubes, particularly when used with inspiratory pressure support (Bersten et al., 1989; Banner,
Blanch and Kirby, 1993). Consequently, PEEP of
510 cmH2O is usually added to most ventilatory
modes. Higher levels (more than 10 cmH2O) may
reduce cardiac output and cause barotrauma; they are
only used in patients with severe hypoxia who need
high levels of inspired oxygen.
The application of PEEP above 10 cmH2O should be
titrated against oxygen delivery to provide optimal
PEEP (Lin and Oh, 1990). Although the efficacy of
PEEP in preventing atelectasis is not conclusively
proven, it may be of some benefit in neurosurgical
patients who cannot be positioned with head elevation. PEEP can also be used in spontaneously breathing patients, either through a tightly fitting face mask
or an endotracheal tube, in which case it is called
continuous positive airway pressure (CPAP). CPAP is
useful in weaning patients with critical levels of auto
PEEP or with reduced ventricular function. The
withdrawal or reduction of positive pressure ventilation or PEEP from these patients may worsen oxygenation and require hours or even days of therapy to
reverse (Zwissler et al., 1991).
The use of PEEP in the head-injured patient has
been controversial. The rationale for avoiding PEEP
347
was that the increased intrathoracic pressure, transmitted via the venous system to the sagittal sinus and
cerebral veins, would raise intracranial pressure.
However there is evidence that PEEP only increases
intracranial pressure if it reduces systemic arterial
pressure (Frost, 1977; Shapiro and Marshall, 1978).
The mechanism suggested is arteriolar vasodilatation
induced by reduced cerebral perfusion pressure. On
balance, PEEP probably does not adversely effect
cerebrovascular dynamics unless there is an associated reduction in cerebral perfusion pressure, and
PEEP should be used primarily to improve oxygenation in the head-injured patient.
(e)
Control of carbon dioxide
Hyperventilation is of great value when ICP rises

acutely and requires urgent reduction such as during
resuscitation or prior to evacuation of intracranial
hematomas.
Urgent hyperventilation must be employed carefully because vigorous manual inflation (bagging) or
marked elevation of the rate of mechanical ventilation
may decrease mean arterial pressure or increase mean
airway pressures (Ersson et al., 1990). The temporary
benefits derived from hypocapnia-induced vasoconstriction may be lost if systemic arterial pressure and
hence cerebral perfusion pressure are reduced. Furthermore excessive vasoconstriction due to hyperventilation may result in a reduction of oxygen
delivery to ischemic levels. Hyperventilation during
early diagnosis and management should be used for as
short a time as possible. If diagnostic studies show no
mass lesion, and intracranial pressure is low, PaCO2
should gradually be returned to normal.
Prolonged reduction of PaCO2 to 25 mmHg or lower
has been advocated for the last two decades without
specific evidence that it reduces mortality or morbidity. Indeed, clinical studies have suggested that prolonged hyperventilation may be associated with a
worse outcome (Jennett et al., 1980; Proctor et al., 1984;
Muizelaar et al., 1991). Potential detrimental effects of
prolonged hyperventilation include increased cerebral
lactic acidosis and uncoupling of oxidative phosphorylation. Profound reduction in the cytochrome a1 a3
system has been attributed to cerebral oligemia
induced by prolonged hypocapnia (Proctor et al., 1984;
Rosner, 1987). Furthermore the vasoconstriction
induced by hypocapnia is temporary. Animal and
human studies have shown that despite continuous
hyperventilation to maintain a PaCO2 of 25 mmHg
there is a steady loss of pial arteriolar constriction and
a return of vessel diameters to baseline by 20 hours.
During this period, arterial and cerebrospinal fluid pH
return to normal. Return to normocapnia will now
result in vasodilatation, an increase in cerebral blood
348
volume and rise in intracranial pressure (Muizelaar et

al., 1988; 1991; Bouma and Muizelaar, 1992).
Hence continuous hyperventilation is probably only
effective for 2024 hours and subsequent attempts to
return to normal PaCO2 may increase intracranial
pressure. Therefore if hyperventilation is used initially, it should only be continued until the level of
intracranial pressure can be determined. Thereafter it
should be used to treat acute elevations in intracranial pressure while other measures such as osmotherapy, ventricular drainage and the evacuation of
mass lesions are being initiated (Pitts and Andrews,
1992). Routine prolonged hyperventilation, still practiced in many units, is difficult to justify.
(f)
Management of ventilated patients
Once mechanical ventilation is established, meticulous

attention to oxygenation and ventilation is essential. A
nursepatient ratio of 1:1 is desirable. In an unstable
patient with multiple injuries, or multiple organ
failure, a ratio of up to 2:1 may be justified. The
methods and principles of respiratory function monitoring described in section 17.2.1 should continue for
the duration of ventilation.
Humidification
In all ventilated patients, inspired gases must be
artificially humidified, since the humidifying functions of the nasopharynx are bypassed by endotracheal intubation. Normal respiratory humidification is important for conserving heat and water.
Humidification and warming continue along the
airways so that alveolar gas is 100% saturated at 37C.
Effective mucociliary function depends on a relative
humidity of 75100% irrespective of temperature. In
ventilated patients, the relative humidity of inspired
gases falls to 50% or less and rapid heat loss can result
in hypothermia. Cold, dry gases in the trachea and
bronchi increase mucous viscosity and inspissation,
leading to mucosal ulceration, depressed ciliary function and microatelectasis due to obstruction of small
airways. Humidification should provide 75100%
saturated gas in the trachea at a constant temperature
(3236C) without increasing the work of breathing,
dead space or resistance to either spontaneous or
controlled ventilation. Water-bath humidifiers and
aerosol nebulizers or atomizers are commonly used
but they may lead to infection from bacterial contamination of water reservoirs, to overhydration,
overheating and electrical hazards. Most of these
complications can be avoided by using heat and
moisture exchangers attached to the endotracheal
tube. These give safe and efficient humidification and
may be combined with a bacterial filter to maintain
sterile gases (Shelly, Lloyd and Park, 1988).
Patients with reactive airways due to pre-existing

asthma, chronic airway disease or acute bronchospasm from infection or pulmonary edema may
benefit from nebulization of 2-agonists (e.g. salbutamol) to improved gaseous exchange and reduced
inspired airway pressure. The addition of nebulized
anticholinergics (e.g. ipratropium bromide) may benefit patients with severe airflow obstruction, particularly in the weaning phase. The routine use of these
agents in ventilated patients, however, has not been
shown to be of benefit (ODriscoll et al., 1989).
(g)
Weaning from positive-pressure ventilation
In the head-injured patient, mechanical ventilatory

support can generally be withdrawn when intracranial pressure is consistently less than 25 mmHg and
there is no significant mass effect on CT scan.
Extracranial injuries should also be stable and should
not need inotropic support or preload augmentation.
Lung function, gaseous exchange and lung compliance should approach that of the patients presumed premorbid status. The patient should be placed
on assisted ventilation (either synchronized intermittent mechanical ventilation or pressure support) so
that spontaneous efforts can eventually take over
respiration. Sedation and analgesia should be minimal
so that cough and glottic reflexes are present.
As a general rule, weaning is not commenced until
PaO2 is 7080 mmHg with a FiO2 not greater than 0.4.
The PaO2/FiO2 ratio is a useful measure of the alveolar
arterial oxygen gradient (Table 17.3). Similarly, PaCO2
should be normal, as discussed in section 17.5.1(e).
Pulse oximetry has greatly facilitated assessment of
oxygenation, although the limitations outlined in
section 17.4.2 must be considered; frequent blood gas
analyses should be performed, particularly after a
change in ventilator settings or FiO2 (Linton, 1993).
Numerous indices, listed in Table 17.3, have been
used to predict success in weaning patients from
mechanical ventilation but these should be interpreted
within the specific clinical context. They are accurate
guides, providing the work of breathing through the
endotracheal tube and ventilator tubing is compensated by inspiratory pressure support of approximately 1015 cmH2O (Banner et al., 1993b). Alternatively, lung volumes can be assessed by connecting the
endotracheal tube to a Wrights respirometer. Of the
indices of lung mechanics, the ratio of respiratory rate
to expired tidal volume (f/VT) is easily determined
and values greater than 100 may accurately predict
successful weaning (Yang and Tobin, 1991). Lung
mechanics can also be assessed by the simple bedside
demonstration of a strong expulsive cough on tracheal
suction. This implies adequate forced vital capacity,
cough and glottic reflexes.
Table 17.3 Criteria for weaning head injured patients

from mechanical ventilation. SIMV = synchronized
intermittent mandatory ventilation, PaO2 = arterial oxygen
tension, FiO2 = fractional inspired oxygen concentration,
PaCO2 = arterial carbon dioxide tension, PEEP = positive
end expiratory pressure, VD/VT = dead space/tidal
volume ratio, f/VT = respiratory rate/tidal volume ratio,
PAO2 = alveolar oxygen tension
Clinical markers for
initiating weaning
Appropriate conscious state;

cooperative patient
Appropriate peripheral motor
function
Clinically and radiologically
resolving lung disease
Adequate analgesia
Hemodynamic stability
Metabolic, acidbase stability
Neurosurgical
factors
Stability of extracranial injuries

Control of cerebral perfusion
pressure
Ventilation
parameters
Patient trigger mode (SIMV,

pressure support)
PaO2 > 80 mmHg, FiO2 < 0.4
PaCO2 < 45 mmHg
PEEP < 10 cmH2O
Lung mechanics
and function
Respiratory rate < 35 breaths/

minute
Tidal volume > 4 ml/kg
Minute volume < 15 l/min
Maximum vital capacity > 1015
ml/kg
Maximum inspiratory pressure <
20 cmH2O
Static compliance > 35 ml/cmH2O
PaO2/FiO2 ratio > 150
Alveolararterial oxygen gradient
< 250 mmHg
VD/VT < 0.6
s/Q
t) < 20%
Shunt fraction (Q
Predictors of
weaning success
f/VT ratio < 100

Tidal volume > 4 ml/kg
Static compliance > 35 ml/cmH2O
PaO2/PAO2 ratio > 0.35
T-piece weaning
Patients with normal lungs who have been ventilated
for only a few days may be connected from full
ventilation to a humidified T-piece (blow-by system)
or open ventilator circuit for 2030 minutes. This is the
usual method of weaning following general anesthesia. In neurotrauma it is suitable for patients with
minor head injuries who have required intubation and
ventilation as part of resuscitation and assessment.
During the period of T-piece weaning, lung mechanics, in particular the f/VT ratio, and gas exchange
should be constantly monitored. If these parameters
are stable the patient can be left on this circuit until
extubation is appropriate.
349
Intermittent T-piece weaning is suitable for

patients with underlying airflow obstruction or
mechanical muscle weakness. The patient is connected
to a T-piece for trial periods, of perhaps 530 minutes
every 30120 minutes. Each trial period is increased
by 510 minutes until the patient maintains adequate
mechanics and oxygenation without mechanical ventilation. A balance must be achieved between allowing
the patient to exercise and causing respiratory muscle
fatigue with increased oxygen consumption. T-piece
weaning has been largely superseded by pressure
support ventilation using newer ventilators such as
the Siemens Servo 900C, Engstrom Erica or Puritan
Bennett 7200a (Tomlinson et al., 1989; Bone and
Eubanks, 1991; Oh et al., 1991).
Weaning from synchronized intermittent mandatory ventilation may be accomplished by progressively reducing the ventilator rate and tidal
volume until the patient can maintain an adequate
minute volume with spontaneous respiration. Inspiratory pressure support should be added as soon as the
patient begins spontaneous efforts and then maintained throughout the weaning period. Once the
patient is breathing adequately on pressure support
ventilation alone this may be gradually reduced to
10 cmH2O and the patient extubated or, if tracheostomized, placed on a humidified T-piece.
Extubation may be considered if the level of
consciousness has improved and the pharyngeal and
tracheal reflexes are able to protect the airway from
aspiration of secretions or gastric contents.
With recovery of physical and mental status, a patient
with a tracheostomy may progress from a cuffed to a
non-cuffed or fenestrated tube and then be extubated.
(h)
Complications: nosocomial pneumonia
Ventilated patients are susceptible to respiratory infections. Early pneumonia, defined as respiratory infection already present or developing soon after intubation, may arise from aspiration of gastric contents or
infection of collapsed lobes. Late or true ventilatorassociated (nosocomial) pneumonia is defined as
infection occurring 48 hours after intubation and not
present or incubating at time of intubation (Meduri,
1993).
True ventilator-associated pneumonia is associated
with a significant morbidity and mortality, particularly in those patients who progress to develop an
acute lung injury (section 17.5.1) In patients with
severe head injury who require positive pressure
mechanical ventilation for more than 3 days, the
incidence of pneumonia has been reported to be
3570%. Nosocomial pneumonia significantly prolongs ventilation and ICU stay and incurs a reported
mortality of up to 50% (Rodriguez, 1991).
350
Diagnosis of nosocomial pneumonia

The diagnosis and treatment of ventilator-associated
nosocomial pneumonia has been extensively debated
(Torres, Gonzalez and Ferrer, 1991; Meduri and Chastre, 1992; Meduri, 1993) and several criteria have been
proposed. Clinical criteria are fever, leukocytosis,
purulent tracheobronchial secretions and radiographic
evidence of a new or progressive pulmonary infiltrate.
Microbiological criteria are a sputum Gram stain
showing more than 25 polymorphonuclear leukocytes
and more than ten squamous epithelial cells per lowpower field with recovery of a significant pathogen
(by stain or culture). These findings in a previously
healthy person almost invariably indicate pneumonia
and appropriate antibiotics should begin as soon as
sensitivities have been established.
However, purulent secretions are invariably present
in patients during prolonged mechanical ventilation
and are usually not caused by pneumonia. Secretions
may originate from the sinuses, stomach or oropharynx and may accumulate above the endotracheal tube
cuff; these secretions can be aspirated by minor
manipulation of the tube. In addition, the proximal
airways of ventilated patients are colonized early by
potentially pathogenic organisms originating from
hospital cross-infection.
As a result, no combination of clinical variables is
completely accurate in diagnosing pneumonia in
ventilated patients. Postmortem studies have shown
that pneumonia is underdiagnosed in patients with
acute lung injury, but over diagnosed in acute respiratory failure from other causes. Although appropriately
selected antibiotics have been shown to improve
survival, empirical use of broad-spectrum antibiotics
in patients without infection is potentially harmful,
since it facilitates colonization and superinfection by
highly virulent organisms.
In one study, the mortality rate in patients who
received antimicrobial therapy before the onset of
pneumonia was 83% compared with a mortality of
48% in the group who did not receive prior antibiotics
(Meduri, 1993). A culture taken from lung tissue is the
gold standard in establishing the diagnosis, but is
often not feasible. In most patients quantitative cultures of tracheal aspirates, protected specimen brushing and bronchoalveolar lavage are recommended. It
is important to stress that systemic sources of infection
should be actively sought and treated.
Management
Management of ventilator-associated pneumonia is
twofold. The first essential step is aggressive pulmonary toilet, postural drainage and prevention of gastric
aspiration. Increased sedation may be necessary during
tracheal suctioning to prevent sympathetically mediated swings in blood pressure and intracranial pressure
(Ersson et al., 1990). There is increasing evidence that
ventilated patients nursed at 30 head elevation have a
lower incidence of pulmonary aspiration and improvement in PaO2 as a result of reduced shunting. The
positioning of head-injured patients has been the topic
of some controversy. This is discussed further in section
17.5.2(c). The second step is appropriate antibiotic
cover, using the least toxic bactericidal agent in
appropriate doses, guided by blood level monitoring if
necessary. Antibiotics should only be used when the
criteria for nosocomial pneumonia are fulfilled and the
bacteriological cultures are positive. Sensitivity testing
is the best guide to antibiotic choice.
(i) Complications: acute respiratory distress
syndrome/acute lung injury
The acute respiratory distress syndrome (ARDS) has
been defined as a sudden clinical pathophysiological
state characterized by severe dyspnea, hypoxia, diffuse bilateral pulmonary infiltrations and stiff lungs
following massive acute lung injury, usually in persons with no previous lung injury (Ashbaugh et al.,
1967). The hallmark of this syndrome is increased
permeability of the alveolarcapillary endothelium.
The constellation of clinical, radiological and pathophysiological findings, resulting from diffuse injury to
the lung parenchyma, that defines this syndrome has
been the subject of intense debate (Petty, 1990; Oh,
1992; Repine, 1992; McNaughton and Evans, 1992;
Kolleff and Schuster, 1995). The terms acute respiratory distress syndrome or acute lung injury are
generally used interchangeably, but they have recently
been defined in terms of the severity of insult,
specifically the degree of hypoxia; ARDS is now
regarded as the more severe form of acute lung injury
(Bernard et al., 1994).
The insult is not uniform and spares some areas of
lung parenchyma (Figure 17.2). In damaged areas, the
lung is atelectatic, edematous and hemorrhagic.
Hypoxia occurs as a result of increased intrapulmonary
shunting of blood caused by the loss of functional
alveoli and hence reduction in functional residual
capacity. Lung compliance is reduced (< 30 ml/
cmH2O) and physiological dead space is increased,
leading to hypercapnia. Diffuse alveolar and interstitial
infiltrates appear progressively on chest radiographs
and may be quantified in accordance with the severity
of injury (Murray et al., 1988). Microscopic examination
reveals intra-alveolar collections of proteinaceous
fluid, red blood cells and inflammatory cells. Microthrombi or white cell aggregates may be seen in small
vessels. After 2448 hours, hyaline membranes line the
alveoli. These are formed by fibrin that has escaped
Figure 17.2 Chest X-ray of a patient with acute respiratory

distress syndrome (ARDS), characterized by bilateral alveolar and interstitial infiltrates. Note the pulmonary artery
catheter inserted via the right internal jugular vein, which
has advanced too far into a segment of the right pulmonary
artery.
through the capillary walls. Subsequently, as repair of

the injury occurs, fibrosis may ensue.
Some but not all patients with ARDS develop
dysfunction or failure of one or more organ systems,
either sequentially or simultaneously. By contrast,
other patients develop multiple organ failure without
ARDS, although they may have less severe degrees of
parenchymal lung injury. This suggests that ARDS is
the respiratory manifestation of multiple organ failure syndrome, just as septic shock is the cardiovascular manifestation (Bone et al., 1992b).
The patient with head injury is at risk of developing
ARDS as a result of neurotrauma per se neurogenic
pulmonary edema is a form of ARDS or of
extracranial factors such as lung contusion, multiple
transfusion, burns, pneumonia, pancreatitis, aspiration of gastric contents and sepsis.
Treatment of ARDS
Treatment of patients with ARDS is primarily supportive and aims to maintain oxygen delivery to all organ
systems. Most patients with ARDS who die do so from
351
multiple organ failure or sepsis rather than respiratory

impairment (MacNaughton and Evans, 1992). Hence
the underlying cause must be treated and suspected
sites of sepsis need to be managed aggressively with
appropriate antibiotics and surgical drainage. Careful
fluid management is necessary to maintain euvolemia
without increasing lung water (Humphrey et al., 1990;
Hudson, 1992). Pulmonary artery catheters may be
used to facilitate the measurement of oxygen delivery
(Russell, Graybeal and Strout, 1990; Mitchell, 1992).
Pressure-limited ventilation and PEEP, as outlined in
section 17.5.1(d), form the mainstay of respiratory
support.
Permissive hypercapnia has been shown to reduce
mortality in ARDS (Hickling, 1990). This employs
pressure-limited ventilation at the expense of tidal
volume and a rise in PaCO2. In these studies PaCO2 was
allowed to rise to greater than 80 mmHg, without
adverse effects, despite incurring a moderate respiratory acidosis. However, this strategy may not be feasible
in head-injured patients, where control of PaCO2 is
essential for controlling ICP and CPP (section 17.5.1(e)).
The prognosis of ARDS depends mainly on the
cause and the presence of multiple organ failure. Fat
embolism and neurogenic pulmonary edema may
cause a severe but transient ARDS, which often has a
good outcome. Gram-negative sepsis and shock states
have a higher incidence of multiple organ failure and
a higher mortality.
Despite improvements in the understanding of
pathophysiological pathways and in mechanical support for patients with ARDS and the multiple organ
failure syndrome, no definitive magic bullet treatment is available and mortality remains high at
4070% (Kraus et al., 1993).
(j)
Complication: pulmonary thromboembolism
Fatal pulmonary embolus occurs in 23% of patients

with head and multiple injuries. It is usually caused
by thrombosis of the iliofemoral veins. Deep venous
thrombosis in the legs may occur in 3040% of these
patients but has a lower incidence of embolism.
High-risk factors include a previous history of
thromboembolism, major orthopedic (especially pelvic) injuries, other major trauma, prolonged coma and
hemiparesis. Prophylactic measures includes early
mobilization, graduated compression stockings, pneumatic calf stimulators and low-dose anticoagulants
such as heparin, low-molecular-weight heparin or
warfarin. Although bleeding is an uncommon complication of low-dose anticoagulants, it is prudent to
monitor coagulation profiles in patients with neurotrauma, giving anticoagulants only during the convalescent phase because of the risk of intracranial
bleeding (Moser, 1990; Sudlow et al., 1992).
352
In high-risk patients or those with proven thromboembolism in whom anticoagulation is contraindicated, early placement of vena caval filters has been
advocated (Persson, Davis and Villavicencio, 1991).
(k) Effects of pre-existing physical limitations of
the chest on weaning
Patients with massive obesity and kyphoscoliosis may
require pressure-limited ventilation. These patients
and those with respiratory muscle weakness due to
myopathies, neuropathies or paralysis may need
prolonged weaning. Pleural and parenchymal pathology limits alveolar ventilation by restricting lung
expansion and by increasing dead space. These disorders also increase the work of breathing, thereby
increasing oxygen consumption and respiratory muscle fatigue.
Drainage of the pleural space, usually by means of
tube thoracostomy, is indicated in patients compromised by pneumothorax, hemothorax or pleural
effusion.
Nutrition, introduced early and preferably by the
enteral route, will improve respiratory muscle function to a limited extent. It is associated with a
reduction in nosocomial pneumonia and preservation
of gut integrity (Chapter 16).
17.5.2 HEMODYNAMIC MANAGEMENT OF THE
ACUTE HEAD-INJURED PATIENT IN THE ICU
Cardiovascular stability in patients with neurotrauma

is of particular importance because of the need to
maintain an adequate cerebral perfusion as well as
myocardial, renal and splanchnic perfusion (Fessler
and Diaz, 1993).
Vital organ perfusion depends on an adequate
vascular volume, hence accurate monitoring of vascular volume and systemic blood pressure is essential.
As discussed in section 17.4.5(a), mean arterial pressure should be monitored continuously, usually with
an intra-arterial catheter.
(a)
Volume status monitoring
Right atrial pressure (central venous pressure) will

suffice to monitor volume status in most patients but
more specific monitoring of cardiac output may
sometimes be indicated. As with arterial pressure
measurement, this is most reliable with electronic
transducers.
Right ventricular preload is measured by inserting
a catheter into the superior vena cava and measuring
mean right atrial pressure when the tricuspid valve
is open. The pressure measurement provides an
approximation of right ventricular preload if ventricular compliance is normal, as it usually is in

young patients.
The response of right atrial pressure to a fluid
challenge will yield useful information regarding the
volume status. This is particularly helpful when right
atrial pressure is less than normal (approximately
5 mmHg) as pressure in the right ventricle is seldom
much higher than in the left. In this context, a low
right atrial pressure suggests decreased intravascular
volume.
Factors that influence the accuracy of right atrial
pressure measurements include right heart dysfunction, superior vena caval obstruction, tricuspid valve
disease, cardiac tamponade and raised intrathoracic
pressure (right atrial pressure may be elevated when
left ventricular pressure is normal).
Left ventricular end-diastolic pressure and volume
may be assessed with right atrial pressures if right
and left ventricular pressures are similar in diastole.
This is usually the case in patients with normal left
ventricular compliance. However, in patients with
poor left ventricular compliance or mitral valve
disease, left ventricular pressure may be elevated
when right ventricular pressure is normal. In these
patients, in older patients and in those with multiple
organ dysfunction, direct left-sided measurements
may be appropriate for assessing left ventricular
pressure.
Pulmonary artery pressure monitoring is widely
used in the management of critically ill patients.
However there has been strong opinion expressed
against its routine use (Robin, 1985). Conversely,
protagonists claim improved outcome in selected
patients (Shoemaker et al., 1988; Mitchell, 1992).
Left ventricular preload may be estimated by
passing a balloon-tipped catheter into the right-sided
pulmonary circulation. With the balloon inflated, the
catheter traverses the right atrium, right ventricle
and main pulmonary artery until it occludes a
branch of the pulmonary artery (Figure 17.2). Flow
distal to the balloon ceases and the pressure measured at the catheter tip distal to the balloon reflects
the pulmonary artery occlusion or wedge pressure
downstream. This approximates to left atrial pressure
and is equivalent to left ventricular end-diastolic
pressure when the mitral valve is open, assuming
that pulmonary venous pressure is not elevated and
that the mitral valve is normal. The left ventricular
end-diastolic pressure is assumed to approximate to
left ventricular end-diastolic volume and therefore to
left ventricular preload. With the balloon deflated, a
pulmonary arterial pressure waveform is continually
displayed. When the catheter is correctly placed right
atrial pressure is measured simultaneously from a
port in the right atrium. The pulmonary artery
353
Table 17.4 Hemodynamic indices obtained from pulmonary artery catheters. MAP = mean arterial pressure, SBP =
systolic blood pressure, DBP = diastolic blood pressure, CO = cardiac output, BSA = body surface area, RAP = right
atrial pressure, MPAP = mean pulmonary artery blood pressure, PAOP = pulmonary artery occlusion pressure.
Variable
Formula
Mean arterial pressure (MAP)
MAP = DBP +
Cardiac index (CI)
CI =
Stroke volume index (SVI)
SVI =
Systemic vascular resistance index (SVRI)

Pulmonary vascular resistance index (PVRI)
Left ventricular stroke work index (LVSWI)
Right ventricular stroke work index (RVSWI)
Indications for pulmonary artery catheterization in a

head-injured patient (Table 17.4)
There are no specific indications for pulmonary artery
catheters in managing patients with head injury. They
are most often used for acute intercurrent problems.
Clearly they should only be used when the information derived from them will significantly improve
management and the benefits outweigh their risks.
Situations in which they may be used include:
CO
CO
HR BSA
CI
79.98
MPAP PAOP
measurement of right heart pressures (right atrial

pressure, pulmonary artery pressure)
ARDS
pulmonary hypertension
pulmonary embolism
cardiac tamponade;
estimation of left heart filling (pulmonary artery
occlusion pressure)
left ventricular failure
response to fluid loading;
hemodynamic measurement (cardiac output, stroke
volume, systemic vascular resistance)

6090 mmHg
2.53.5 l/min/m2
BSA
79.98
CI
LVSWI = (MAP PAOP) SVI 0.0136
RVSWI = (MPAP RAP) SVI 0.0136
PVRI =
catheter may also be used to measure cardiac output

by the thermodilution method using a thermistor at
the distal end. New rapid-response catheters are able
to determine continuous right ventricular ejection
fraction and cardiac output (Dorman et al., 1992). A
wide range of cardiorespiratory variables (Table 17.4)
can be derived from these measurements, some of
which have been discussed in section 17.2.3, and
they play an essential part in providing optimal
oxygen delivery.
SBP DBP
MAP RAP
SVRI =
Normal values
3347 ml/beat
17002400 dyn.s/cm5/m2
150250 dyn.s/cm5/m2
4560 g.m/m2/beat
712 g.m/m2/beat
quantification of shock states (cardiogenic, septic,

hypovolemic)
assessment of response to treatment in the
above;
O2, D
O2);
derivation of oxygen variables (V
measurement of intracardiac shunt (acute ventricular septal defect).
(b)
Clinical use of central venous catheters
In interpreting central venous and pulmonary artery

catheter measurements, account must take of the
factors which alter the gradient between the pulmonary artery occlusion pressure and left ventricular end
diastolic pressure (Nadeau and Noble, 1986). Intravascular pressures must be referenced to the extravascular pressures around them.
Pulmonary artery occlusion pressure is lower than
true ventricular filling pressure in a spontaneously
breathing patient during inspiration when pleural
pressure may be greatly negative. Conversely, pulmonary artery occlusion pressure is higher than the true
filling pressure in a mechanically ventilated patient
during an inspiration with positive pressure. To avoid
erroneous interpretation, the pulmonary artery occlusion pressure should be measured in end-expiration.
Ventilated patients should remain on the ventilator
during the measurement.
Other pulmonary factors that affect the pressure
measurements include catheter-tip location in regions
of the lung where alveolar pressure exceeds arterial
pressure, PEEP and raised intrapleural pressure. Cardiac factors include mitral or aortic valve disease and
poor left ventricular compliance.
354
(c) Contraindications to central venous

catheterization
These include:

lack of suitable vascular access;

untreatable bleeding disorders;
patient instability preventing adequate time for the
procedure.
(d)
Complications of central venous catheters
These include:

trauma to structures at insertion (pneumothorax,

arterial
laceration,
hematoma,
hemothorax,
chylothorax);
arrythmias on passage of the catheter through the
heart;
perforation of great vessels and atrium;
migration into a distal vessel, which may cause
vessel rupture or lung infarction.
These complications may be avoided by determining

the position of the catheter tip on the chest radiograph
(Figure 17.2), by continuous monitoring of the pulmonary artery pressure waveform and by inflating the
balloon with only 11.5 ml of air to obtain a pulmonary artery occlusion pressure. (Hagley et al., 1992).
The optimal placement for right atrial pressure catheters is within the superior vena cava 2 cm above the
pericardial reflection (McGee et al., 1993).
Central venous catheters are a major source of
nosocomial infection and carry a significant morbidity
and mortality. In general catheters inserted during
resuscitation should be changed as soon as possible
unless strict asepsis was used in their insertion. The
incidence of infection increases markedly after 5 days
and varies with the site of the catheter, increasing
from subclavian through to internal jugular to femoral venous catheters (Norwood et al., 1991; Cobb et
al., 1992; Mermel and Maki, 1994).
A persistent or new pyrexia, leukocytosis and an
inflamed insertion site strongly suggest catheter
sepsis and warrant a new catheter. Culture of the
intradermal portion and tip of the catheter may
confirm an infection, although removal of the catheter
may be all that is necessary for treatment (Goldmann
and Pier, 1993).
(e)
Effects of therapy on cerebrovascular function
Until recently, head injury management was focused

on reducing ICP and cerebral blood volume, particularly by osmotherapy, hyperventilation and barbiturate coma. Each of these methods may in part reduce
ICP by reducing mean arterial pressure and therefore
CPP. Recent studies have questioned their efficacy and
emphasis is now being placed more on maintaining

CPP (Stone, 1989; Paulson et al., 1990; Bouma et al.,
1992; Miller et al., 1992; Fessler and Diaz, 1993;
Myburgh and Lewis, 1996).
The hemodynamic effects of ventilation on ICP and
CPP must be considered. Increased intrathoracic
pressure may reduce venous return and hence preload. At the same time reduction of ventricular
transmural pressure may reduce afterload. Thus in
hypovolemic patients ventilation may increase hypotension and reduce cerebral perfusion. Conversely,
patients with cardiogenic or neurogenic pulmonary
edema often improve dramatically with ventilation
because of the mechanical reduction in preload and
afterload and increase in cardiac output and cerebral
perfusion.
Impedance to right ventricular ejection may be
increased by high intrathoracic pressures or by preexisting or acute pulmonary hypertension, resulting in
interventricular septal shift, increased left ventricular
end diastolic volume and reduced stroke volume
(Zwissler et al., 1991; Pinsky, Desmet and Vincent,
1992). As discussed in section 17.5.1(b), the effects of
different modes of ventilation on auto PEEP must also
be considered.
CPP-based algorithms are discussed in Chapter 19;
they have altered the role of osmotherapy and elective
dehydration. When the bloodbrain barrier is intact,
mannitol, a six-carbon non-metabolic sugar similar to
glucose, remains within the intravascular space. It may
have several modes of action. These include reduction
in cerebrospinal fluid formation, reduction in brain
tissue volume (Donato et al., 1994), an immediate
increase in circulating blood volume and arterial blood
pressure (Roberts et al., 1987), reduction of blood
viscosity by hemodilution (Muizelaar, Lutz and Becker,
1984), increase in red blood cell deformability (Burke et
al., 1981) and scavenging of potentially toxic free
radicals (Fisher and Raper, 1988). Experimental studies
have shown that cerebral vasoconstriction occurs in
direct response to the decrease in blood viscosity after
mannitol infusion (Muizelaar et al., 1983). In other
words, a constant CBF is maintained by vasoconstriction if autoregulation is intact. Other studies have
supported the vasoconstrictor action of mannitol.
Intracranial pressure responded better to mannitol in
patients with a low initial cerebral perfusion pressure,
who are likely to be vasodilated, than those with a high
cerebral perfusion pressure where the cerebral vasculature is likely to be already fully vasoconstricted
(Rosner and Coley, 1987)
When the clinical goal is euvolemia and adequate
cerebral perfusion, osmotherapy may be seen as a
method of selectively reducing brain volume while
avoiding dehydration (Miller, Piper and Dearden,
1993).
Repeated administration of mannitol, especially in

combination with frusemide (furosamide), may result
in a systemic hyperosmolar state and dehydration
(Schettini, Stahurski and Young, 1982; Roberts et al.,
1987). Mannitol will eventually cross the bloodbrain
barrier into the extracellular space, increasing free
water and cerebral edema. The hemoconcentration
caused by dehydration may falsely suggest that red
cell mass is adequate. Hyperviscosity and microvascular shunting may adversely affect tissue oxygen
delivery.
A substantial part of the action of mannitol depends
on its ability to draw extravascular water into the
vascular space. If extravascular water is markedly
reduced mannitol will have only a minimal effect and
its toxicity will be increased. When dehydration is
well established, the patient will become refractory to
mannitol therapy.
The renal toxicity of virtually all drugs commonly
used in the neurosurgical patient, such as aminoglycoside antibiotics, mannitol, frusemide and contrast
media, as well as transfusion reactions, are increased
by dehydration. Dehydration or hypovolemia makes
patients more liable to unstable and inadequate
systemic and cerebral perfusion and to orthostatic
changes in the blood pressure.
Serum sodium should be kept at 145150 mmol/l
and serum osmolality should be no greater than
300 mosmol/l (although many authors will accept
320 mosmol/l). Osmolality should be measured
directly, as osmolalities calculated from sodium, urea
and glucose will underestimate true osmolality when
mannitol is used.
Mannitol doses should be limited to the minimum
amount needed to control ICP and CPP and should
not be given routinely with frusemide because of the
risk of excessive dehydration. Dehydrating agents
not only mannitol given appropriately can at times
be life-saving, but chronic dehydration is potentially
harmful.
(f)
355
ured at head level then CPP will not increase despite

the fall in ICP. When severe intracranial hypertension
exists, there is little cerebral venous volume available
for displacement and CPP may be higher with
patients nursed flat (Rosner, 1993). In one study, onethird of patients demonstrated an increase in ICP
with head elevation (Feldman et al., 1992). This may
be due to vasodilatation in response to decreased
CPP (Rosner, 1993).
Fluctuations in intracranial pressure and cerebral
perfusion pressure with head elevation may be minimized by maintaining euvolemia and the cardiorespiratory benefits of slight head elevation can then
be utilized.
Barbiturate therapy
There are few options available for treating patients
with raised ICP refractory to the above measures.
Sodium pentobarbital at doses causing burst suppression on the EEG will lower intracranial pressure.
However, these doses of barbiturates cause profound
myocardial depression and hypotension and reduce
cerebral perfusion pressure. Theoretically, the reduction in cerebral perfusion pressure and blood flow
may be offset by the reduction in cerebral oxygen
consumption induced by barbiturates. However two
randomized clinical trials have found no improvement in morbidity, mortality, or intracranial pressure
control when barbiturates where given (Ward et al.,
1985; Eisenberg et al., 1988). Barbiturates are still
sometimes used when all other measures have failed.
A recent concept of aggressively maintaining cerebral
perfusion pressure with inotropes and vasopressors
and using barbiturates to lower intracranial pressure
has yet to validated (Miller et al., 1992).
Barbiturate therapy has not been shown definitively to alter outcome in neurotrauma. In the group
of patients refractory to other means of treatment, the
persistent intracranial hypertension probably reflects
the severity of the primary injury (Chapter 19).
Head elevation
Head elevation to about 30 is widely used in nursing

critically ill patients with pulmonary congestion from
cardiac failure or pulmonary edema. Oxygenation is
improved by increasing functional residual capacity
and chest wall compliance and by reducing cephalad
movement of the diaphragm. It also reduces the risk of
aspiration of gastric contents.
In patients with head injury, head elevation may
reduce ICP by reducing cerebral venous pressure.
Rosner and Coley (1986) found that ICP was
710 mmHg higher in 5070% of patients nursed flat
compared with 3050 head elevation. However, if
there is an equivalent fall in arterial pressure meas-
(g)
Hemodynamic manipulation
In managing systemic arterial pressure it is important

to consider that systemic hypertension may be a
response to intracranial hypertension and therefore
necessary to maintain an adequate cerebral perfusion
pressure (Bouma et al., 1992).
Patient manipulations such as endotracheal and
nasotracheal suction, neurological examination, painful stimulation and even family visits, may increase
ICP via sympathetic stimulation. This is unlikely to be
harmful unless ICP is already raised. In this situation
the increases can be prevented by increasing sedation
and analgesia.
356
In current clinical practice optimal cerebral perfusion pressure is 6070 mmHg (Saul and Ducker, 1982;
Tsutsumi et al., 1985), although levels of
80100 mmHg have recently been proposed as providing better control of raised ICP. These levels of CPP
may be achieved by manipulating volume status and
by the use of vasopressors (Rosner, 1993).
Fluid management of head-injured patients has
focused for the last 20 years upon reducing intracranial pressure via strategies of elective dehydration
and osmotherapy. This was based on the concept that
less water available in the body meant less water
available to potentiate cerebral edema. There is,
however, no evidence to substantiate this hypothesis
and indeed the practice may caused more harm than
good. Attention has now shifted to maintaining brain
perfusion.
The primary goal of cerebral perfusion pressure
fluid management is euvolemia, but avoiding systemic overhydration (Marmarou, 1992). A normally
hydrated, euvolemic patient is more hemodynamically stable and has fewer blood pressure changes
associated with ventilator manipulation and less need
for vasopressor support. Furthermore, adequate perfusion of the kidneys minimizes the potentially
nephrotoxic effects of mannitol, frusemide and
sepsis.
Establishing euvolemia, however, requires great
care. The assessment of volume status depends on
clinical markers such as blood pressure, heart rate,
fluid balance and a measurement of preload (as
discussed in section 17.5.2(a)). Biochemical markers
such as serum sodium, urea, creatinine and osmolality
have been advocated but usually change late with
evolving volume depletion and are frequently a poor
guide to volume status. Reaching and maintaining
euvolemia depends on a number of factors, which
should be considered in concert.
Maintenance fluids
Balanced salt solutions maintain normal hydration and
replace insensible losses from pyrexia and ventilation.
When osmotic diuretics and inotropes are used (see
below), urine output may increase to 100200 ml/h,
resulting in dehydration. Fluids must be adjusted to
maintain a normal volume status as apparently
adequate fluid intake alone does not guarantee adequate vascular volume. It is possible for a patient to
have interstitial overhydration with an inadequate
intravascular volume.
Hypovolemic patients generally require colloid solutions. The best vascular volume expander in an anemic
patient is red blood cells. Serum albumin as either
normal (5%) or concentrated albumin is also useful.
Synthetic colloids such as polygeline are suitable,
although they have a short half-life and may interfere

with clotting. (The fluid management of head-injured
patients is discussed further in Chapter 16.)
Vasopressors
Augmenting mean arterial pressure and cardiac output by inotropes and vasopressors is standard practice
in managing critically ill patients with sepsis and
cardiac failure and is now more widely used to
maintain adequate cerebral perfusion in patients with
head injury (Chapter 19).
Euvolemia is an essential prerequisite for the use of
these agents. As discussed above, a hypovolemic
patient is susceptible to wide swings in blood pressure
and these may be accentuated by vasopressors. The
ideal inotrope and vasopressor is one that acts
rapidly and has a short half-life, thereby allowing
titration of dose against a desired end-point, which is
usually mean arterial pressure and/or oxygen
delivery.
Inotropes increase ventricular contractility by acting on myocardial alpha- and beta-receptors. They
include adrenalin (epinephrine), noradrenalin (norepinephrine), dopamine, dobutamine and isoprenaline
(Prichard et al., 1991; Runciman and Morris, 1993a).
The principal hemodynamic effect of inotropes is an
increase in cardiac output. Effects on peripheral
vasculature are variable and may be either vasoconstriction or vasodilatation. The catecholamines (adrenalin, noradrenalin and dopamine) are increasingly
regarded as first-line inotropes. They exert betaadrenergic effects predominantly at low doses( i.e.
increased heart rate, vasodilatation, inotropy and
bronchodilation) and alpha-adrenergic effects (vasoconstriction, inotropy) at higher doses (Moran et al.,
1993). Tachycardia and dysrhythmias are infrequent
(Runciman and Morris, 1993a).
Coronary perfusion and myocardial oxygen delivery
depend on aortic diastolic pressure and are at risk in
states of low systemic vascular resistance such as sepsis
(Schreuder et al., 1989). Patients with low cardiac
output and high systemic vascular resistance, as in
cardiogenic shock, are best treated initially with a
catecholamine to improve inotropy and maintain
coronary perfusion. The use of dobutamine and
isoprenaline to treat cardiogenic shock has been
increasingly questioned because of a high incidence of
tachycardia and reduction of aortic diastolic pressure.
The effects of low-dose dopamine, frequently
advocated to improve renal blood flow, are unsubstantiated. Any such effects are probably not specific
but mediated via increased cardiac output (Duke and
Bersten, 1992).
Vasoactive drugs include vasoconstrictors (phenylephrine, metaraminol) and vasodilators (sodium
CONCLUSION
nitroprusside, glyceryl trinitrate, dobutamine, isoprenaline). Vasoconstrictors may increase afterload on

the left and right ventricle. If right atrial and pulmonary artery occlusion pressure appear to be normal in
the face of vasoconstrictor therapy, vascular volume
may need to be increased.
Although the role of vasopressors in neurotrauma is
not fully established, the rationale for their use is the
loss of the capacity to maintain cerebral perfusion
pressure by cerebral autoregulation.
In severe brain injury cerebral perfusion may
become pressure-passive, i.e. directly dependent on
systemic arterial pressure. There may be marked
regional differences so that the relationship is difficult
to quantify. Vasoactive drugs should be titrated
against a selected cerebral perfusion pressure. The
degree to which the cerebral vasculature is directly
affected by these vasoactive agents is unclear.
Catecholamine infusions must be given through a
dedicated central venous line and patients receiving
these infusions must be appropriately monitored.
Adrenalin or noradrenalin may be infused using a
solution of 6 mg drug/100 ml saline and dopamine as
400 mg/100 ml. At these concentrations, 1 ml/h approximates 1 g/min. Infusion usually begins at
35 g/min and is then titrated against mean arterial,
intracranial and cerebral perfusion pressures.
Hypotensive agents
When cerebral autoregulation is lost, severe systemic
hypertension (i.e. mean arterial pressures greater than
120140 mmHg) may potentiate vasodilatation and
cerebral hyperemia, theoretically increasing cerebral
edema and intracranial hypertension (Paulson et al.,
1990). Thus control of severe systemic hypertension
may be necessary. This is best achieved initially by
sympathetic blockade with sedation or analgesia.
Hypotensive agents must be used with caution so as
to avoid hypotension and reduction in cerebral
perfusion.
Beta-adrenergic antagonists are effective in controlling sympathetically induced hypertension. They act
primarily at the beta-1 receptors in the heart, resulting
in negative chronotropy and inotropy. They inhibit
sympathetically-mediated release of renin from the
juxtaglomerular cells of the kidney, and this also
assists in lowering blood pressure. Thus beta-blockers
may be indicated when hypertension is due to
increased sympathetic tone with tachycardia and
increased cardiac output, as in severe head injury
(Robertson et al., 1983). Esmolol is ideally suited for
this. It has an extremely short half-life which allows
accurate titration by intravenous infusion at doses up
to 300 mg/kg/min (Hansson, 1991). All beta-blockers
may potentiate cardiac failure and bronchospasm and
357
must be used carefully in susceptible patients (Runciman and Morris, 1993b). Vasodilators are seldom
indicated in neurogenic hypertension.
Although glyceryl trinitrate and sodium nitroprusside are commonly used in other neurosurgical situations to control blood pressure, their propensity to
increase cerebral blood volume and intracranial pressure, and to cause preferential peripheral vasodilatation
over cerebral vasodilatation (cerebral steal), probably
contraindicates their use in head-injured patients, who
may have increased intracranial pressure (Anile et al.,
1981; Davis et al., 1981; Michenfelder and Milde, 1988).
17.5.3
FUTURE ADVANCES IN NEUROMONITORING
Neuromonitoring after head injury aims to provide

information about intracranial compliance and cerebral oxygen utilization (Lewis, Reilly and Myburgh,
1994). Measuring intracranial pressure alone does not
give any information regarding cerebral oxygen utilization (Miller et al., 1992), cerebral blood flow or
cerebral metabolic rate (Bouma et al., 1992).
Frequently, intracranial hypertension is present
without any mass lesion on CT scan, when the patient
has diffuse cerebral swelling. Hyperventilation and
other methods of reducing intracranial pressure have
been used empirically without there being any conclusive evidence that they improve outcome. Indeed,
they may cause harm.
Jugular venous oxygen saturation, measured by
fiberoptic catheters inserted into the jugular bulb,
allows calculation of arteriovenous differences and
oxygen uptake by the brain. Cerebral venous desaturation may occur as a result of intracranial hypertension, excessive hypocapnia, arterial hypoxia and
systemic hypotension. Preliminary studies suggest
that jugular bulb oximetry used in conjunction with
intracranial pressure and cerebral perfusion pressure
monitoring may allow more precise management of
cardiorespiratory variables so as to optimize cerebral
oxygen utilization (Sheinberg et al., 1992; Lewis, Reilly
and Myburgh, 1995).
17.6
Conclusion
At present little can be done to influence the outcome

from the primary brain injury. The duration of the
period of increased vulnerability to secondary ischemic and hypoxic insults remains unclear but may be
up to 10 days to 2 weeks in some patients. Secondary
insults must be prevented or rapidly detected and
treated. Adequate substrate delivery is the key to
preventing secondary insults. This requires assiduous
attention to cardiopulmonary homeostasis, cerebral
perfusion pressure and oxygen utilization. The interaction of these parameters must be recognized.
358
17.7
References
Andrews, B. T. (1993) The intensive care management of patients with head

injury, in Neurosurgical Intensive Care, (ed. B. T. Andrews), McGraw-Hill,
New York, pp. 227242.
Anile, C., Zanghi, F., Bracali, A. et al. (1981) Sodium nitroprusside and
intracranial pressure. Acta Neurochirurgica, 58, 203.
Archie, J. P. (1987) Mathematic coupling of data: a common source of error.
Annals of Surgery, 193, 296303.
Ashbaugh, D. G., Bigelow, D. B., Petty, T. L. et al. (1967) Acute respiratory
distress in adults. Lancet, ii, 319323.
Balk, B. and Bone, R. C. (1983) The adult respiratory distress syndrome.
Medical Clinics of North America, 67, 685700.
Banner, M. J., Blanch, P. B. and Kirby, R. R. (1993) Imposed work of breathing
and methods of triggering a demand-flow, continuous positive airway
pressure system. Critical Care Medicine, 21, 183190.
Banner, M. J., Kirby, R. R., Blanch, P. B. et al. (1993) Decreasing imposed work
of the breathing apparatus to zero using pressure-support ventilation.
Beckman, D. L., Bean, J. W. and Baslock, D. R. (1974) Neurogenic influence on
pulmonary compliance. Journal of Trauma, 14, 111115.
Bernard, G. R., Artigas, A., Brigham, K. L. et al. (1994) Report of the American
European consensus conference on ARDS: definitions, mechanisms, relevant
outcomes and clinical trial coordination. Intensive Care Medicine, 20, 225232.
Bersten, A. D., Rutten, A. J., Vedig, A. E. et al. (1989) Additional work of
breathing imposed by endotracheal tubes, breathing circuits, and intensive
care ventilators. Critical Care Medicine, 17, 671677.
Bersten, A. and Sibbald, W. J. (1989) Circulatory disturbances in multiple
systems organ failure. Critical Care Clinics, 5, 233254.
Blanch, P. B., Jones, M., Layon, A. J. et al. (1993a) Pressure-preset ventilation.
Part 1: Physiologic and mechanical considerations. Chest, 104, 590599.
Blanch, P. B., Jones, M., Layon, A. J. et al. (1993b) Pressure-preset ventilation.
Part 2: Mechanics and Safety. Chest, 104, 904912.
Bloomfield, E. L. (1989) Extracerebral complications of head injury. Critical
Care Clinics, 5, 881892.
Bone, R. C. (1985) Monitoring respiratory and hemodynamic function in the
patient with respiratory failure, in Mechanical Ventilation (eds R. R. Kirby, R.
A. Smith and D. A. Desautels), Churchill Livingstone, New York, pp.
137170.
Bone, R. C. (1991a) The pathogenesis of sepsis. Annals of Internal Medicine, 115,
457469.
Bone, R. C. (1991b) Lets agree on terminology: definition of sepsis. Critical
Care Medicine, 19(7), 973976.
Bone, R. C. and Eubanks, D. H. (1991) The basis and basics of mechanical
ventilation. Disease-a-Month, 6, 324406.
Bone, R. C., Balk, R. A., Cerra, F. B. et al. (1992a) Definitions for sepsis and
organ failure and guidelines for the use of innovative therapies in sepsis.
Journal, 55, 12345.
Bone, R. C., Balk, R., Slotman, G. et al. (1992b) Adult respiratory distress
syndrome. Sequence and importance of development of multiple organ
failure. Chest, 101, 32026.
Borel, C., Hanley, D., Diringer, M. N. et al. (1990) Intensive management of
severe head injury. Chest, 98, 180189. (Comment in Chest, 99(6) (1991),
15511552).
Bouma, G. J. and Muizelaar, J. P. (1992) Cerebral blood flow, cerebral blood
Neurotrauma, 9, S333S348.
Bouma, G. J., Muizelaar, J. P., Bandoh, K. et al. (1992) Blood pressure and
intracranial pressure-volume dynamics in severe head injury: relationship
with cerebral blood flow. Journal of Neurosurgery, 77, 1519.
Breuer, H-W. M., Groeben, H., Breuer, J. et al. (1989) Oxygen saturation
calculation procedures: a critical analysis of six equations for the determination of oxygen saturation. Intensive Care Medicine, 15, 385389.
Brochard, L., Harf, A., Lorino, H. et al. (1989) Inspiratory pressure support
prevents diaphragmatic fatigue during weaning from mechanical ventilation. American Review of Respiratory Disease, 139, 513521.
Burchardi, H., Sydow, M. and Criee, C. P. (1994) New ventilatory strategies in
severe respiratory failure, in Current Topics in Intensive Care (eds G. J. Dobb,
J. Bion, H. Burchardi and R. P. Dellinger), W. B. Saunders, London, pp.
81100.
Burke, A. M., Quest, D. O., Chein, S. et al. (1981) The effect of mannitol on
blood viscosity. Journal of Neurosurgery, 55, 170.
Buss, C. S., Nel, C. J. and van den Heever, C. M. (1990) Incidence of
extracranial injuries in fatal head injury patients. South African Journal of
Surgery, 28, 141144.
Carlton, D. P., Cummings, J. J., Scheerer, R. G. et al. (1990) Lung overexpansion
increases pulmonary microvascular protein permeability in young lambs.
Journal of Applied Physiology, 69, 577583.
Cerra, F. B. (1992) Multiple Organ Failure Syndrome. Disease-a-Month, 12,
845946.
Chesnut, R. M., Marshall, L. F., Klauber, M. R. et al. (1993) The role of the
secondary brain injury in determining outcome from severe head injury.
Ciaglia, P., Firsching, R. and Syniec, C. (1985) Elective percutaneous

dilatational tracheostomy: a new simple bedside procedure; preliminary
report. Chest, 87, 715719.
Clarke, G. M. (1985) Multiple System Organ Failure. Clinics in Anaesthesiology,
4, 10271053.
Clayton, D. G., Webb, R. K., Ralston, A. C. et al. (1991) A comparison of the
performance of 20 pulse oximeters under conditions of poor perfusion.
Cobb, D. K., High, K. P., Sawyer, R. G. et al. (1992) A controlled trial of
scheduled replacement of central venous and pulmonary-artery catheters.
Cockings, J. G. L., Webb, R. K., Klepper, I. D. et al. (1993) Blood pressure
monitoring applications and limitations: an analysis of 2000 incident
reports. Anaesthesia and Intensive Care, 21, 565569.
Crosby, E. T. and Lui, A. (1990) The adult cervical spine: implications for
airway management. Canadian Journal of Anaesthesia, 37, 7793.
Crosby, E. T. (1992) Perioperative haemotherapy: I. Indications for blood
component transfusion. Canadian Journal of Anaesthesia, 39, 695707.
Cruickshank, J. M., Degaute, J. P., Kuurne, T. et al. (1987) Reduction of stress/
catecholamine-induced cardiac necrosis by beta-selective blockade. Lancet,
ii, 585589.
Dantzker, D. (1989) Oxygen delivery and utilization in sepsis, in Critical Care
Clinics, (eds R. A. Balk and R. C. Bone), W. B. Saunders, Philadelphia, PA,
pp. 8198.
Dantzker, D. R., Foresman, B. and Gutierrez, G. (1991) Oxygen supply and
utilization relationships. American Review of Respiratory Disease, 143,
675679.
Davis, R. F., Douglas, M. E., Heenan, T. J. et al. (1981) Brain tissue pressure
measurement during sodium nitroprusside infusion. Critical Care Medicine,
9, 17.
Demling, R. and Riessen, R. (1990) Pulmonary dysfunction after cerebral
injury. Critical Care Medicine, 18, 768774.
Dhainaut, J. F., Edwards, J. D., Grootendorst, A. F. et al. (1990) Practical aspects
of oxygen transport: conclusions and recommendations of the Rountable
Conference. Intensive Care Medicine, 16, S179S180.
Dietrich, K. A., Conrad, S. A., Herbert, C. A. et al. (1990) Cardiovascular and
metabolic response to red blood cell transfusion in critically ill volumeresuscitated nonsurgical patients. Critical Care Medicine, 18, 9401044.
Dinarello, C. A. (1984) Interleukin-1 and the pathogenesis of the acute-phase
response. New England Journal of Medicine, 311, 14131418.
Donato T., Shapira Y., Artu A. et al. (1994) Effect of mannitol on cerebrospinal
fluid dynamics and brain tissue edema. Anesthesia and Analgesia, 78,
5866.
Dorman, B. H., Spinale, F. G., Kratz, J. M. et al. (1992) Use of a combined right
ventricular ejection fraction-oximetry catheter system for coronary bypass
surgery. Critical Care Medicine, 20, 16501655.
Dreyfuss, D. and Saumon, G. (1992) Barotrauma is volutrauma, but which
volume is the one responsible? Intensive Care Medicine, 18, 139141.
Dreyfuss, D. and Saumon, G. (1993) Role of tidal volume, FRC, and endinspiratory volume in the development of pulmonary edema following
mechanical ventilation. American Review of Respiratory Disease, 148,
11941203.
Duke, G. J. and Bersten, A. D. (1992) Dopamine and renal salvage in the
critically ill patient. Anaesthesia and Intensive Care, 20, 277302.
Eisenberg, H. M., Frankowski, R. F., Contant, C. F. et al. (1988) High-dose
barbiturate control of elevated intracranial pressure in patients with severe
Ersson, U., Carlson, H., Mellstrom, A. et al. (1990) Observations on intracranial
dynamics during respiratory physiotherapy in unconscious neurosurgical
patients. Acta Anaesthesiologica-Scandinavica, 34, 99103.
Farling, P. A., Johnston, J. R. and Coppel, D. L. (1989a) Propofol infusion for
sedation of patients with head injury in intensive care. A preliminary
report. Anaesthesia, 44, 222226.
Farling, P. A., Johnston, J. R. and Coppel, D. L. (1989b) Propofol infusion
compared with morphine and midazolam bolus doses for sedation of
patients with severe head injuries in the intensive care unit. Journal of Drug
Development, 2(Suppl. 2), 9798.
Feldman, Z., Kanter, M. J., Robertson, C. S. et al. (1992) Effect of head elevation
on intracranial pressure, cerebral perfusion pressure, and cerebral blood
flow in head-injured patients. Journal of Neurosurgery, 76, 207.
Fessler, H. E., Brower, R. G., Wise, R. A. et al. (1991) Effects of positive end
-expiratory pressure on the gradient for venous return. American Review of
Respiratory Disease, 143, 1924.
Fessler, R. D. and Diaz, F. G. (1993) The management of cerebral perfusion
pressure and intracranial pressure after severe head injury. Annals of
Emergency Medicine, 22, 9981003.
Fisher M. M. and Raper, R. F. (1988) The optimization of renal function in
acute illness. Medical Journal of Australia, 149, 546551.
Fisher, M. M. and Raper, R. F. (1991) Sedation in intensive care, in Update in
Intensive Care and Emergency Medicine 14, Update 1991, (ed. J. L. Vincent),
Frost, E. A. M. (1977) Effects of positive end-expiratory pressure on
intracranial pressure and compliance in brain-injured patients. Journal of
REFERENCES
Gentleman, D. (1992) Causes and effects of systemic complications among
severely head injured patients transferred to a neurosurgical unit. International Surgery, 77, 297302.
Goldmann, D. A. and Pier, G. (1993) Pathogenesis of infections related to
intravascular catheterization. Clinical Microbiology Reviews, 6, 176192.
Gooch, J. L., Suchyta, M. R., Balbierz, J. M. et al. (1991) Prolonged paralysis
after treatment with neuromuscular junction blocking agents. Critical Care
Medicine, 19, 11251130.
Gottfried, S. B., Reissman, J. and Ranieri, V. M. (1992) A simple method for the
measurement of intrinsic positive end-expiratory pressure during controlled and assisted modes of mechanical ventilation. Critical Care Medicine,
20, 621629.
Graf, C. J. and Rossi, N. P. (1978) Catecholamine response to intracranial
hypertension. Journal of Neurosurgery, 49, 862868.
Greenberg, A. G. (1990) To transfuse or not to transfuse that is the question!
Griggs, W. M., Worthley, L. I., Gilligan, J. E. et al. (1990) A simple percutaneous
tracheostomy technique. Surgery, Gynecology, and Obstetrics, 170, 543545.
Guyton, A. C. (1991) Cardiac output, venous return, and their regulation, in
Medical Physiology, W. B. Saunders, Philadelphia, PA, pp. 221233.
Hagley, M. T., Martin, B., Gast, P. et al. (1992) Infectious and mechanical
complications of central venous catheters placed by percutaneous venipuncture and over guidewires. Critical Care Medicine, 20, 1426.
Hansen-Flaschen, J., Cowen, J. and Raps, E. C. (1993) Neuromuscular
blockade in the intensive care unit. American Review of Respiratory Disease,
147, 234236.
Hansson, L. (1991) Review of state-of-the-art beta-blocker therapy. American
Journal of Cardiology, 67, 43B46B.
Hickling, K. G. (1990) Ventilatory management of ARDS: can it affect
outcome? Intensive Care Medicine, 16, 219226.
Hill, D. A., Abraham, K. J. and West, R. H. (1993) Factors affecting outcome in
the resuscitation of severely injured patients. Australian and New Zealand
Holland, R., Webb, R. K. and Runciman, W. B. (1993) Oesophageal intubation: an
analysis of 2000 incident reports. Anaesthesia and Intensive Care, 21, 608610.
Hsiang, J. K., Chesnut, R. M., Crisp, C. B. et al. (1994) Early, routine paralysis
for intracranial pressure control in severe head injury: is it necessary?
Hsieh, A. H., Bishop, M. J., Kubilis, P. S. et al. (1992) Pneumonia following
closed head injury. American Review of Respiratory Disease, 146, 290294.
Hubmayr, R. D., Abel, M. D. and Rehder K. (1990) Physiologic approach to
mechanical ventilation. Critical Care Medicine, 18, 103113.
Hudson, L. D. (1992) Fluid management strategy in acute lung injury.
American Review of Respiratory Disease, 145, 988989.
Humphrey, H., Hall, J., Sznadjer, I. et al. (1990) Improved survival in ARDS
patients associated with a reduction in pulmonary capillary wedge
pressure. Chest, 97, 11761180.
Jennett, B., Teasdale, G., Fry, J. et al. (1980) Treatment for severe head injury.
Kalisky, Z., Morrison, D. P., Meyers, C. A. et al. (1985) Medical problems
encountered during rehabilitation of patients with head injury. Archives of
Physiological Medicine Rehabilitation, 66, 2529.
Kaufman, H. H., Moake, J. L., Olson, J. D. et al. (1980) Delayed and recurrent
intracranial hematomas related to disseminated intravascular clotting and
Kaufman, H. H., Timberlake, G., Voelker, J. et al. (1993) Medical complications
of head injury. Medical Clinics of North America, 77, 4360.
Kearl, R. A. and Hooper, R. G. (1993) Massive airway leaks: an analysis of the
role of endotracheal tubes. Critical Care Medicine, 21, 518521.
Kirland, L. L. and Wilson, G. L. (1991) Extracranial effects of acute brain
injury. Problems in Critical Care, 5, 292306.
Kolleff, M. H. and Schuster, D. (1995) The acute respiratory distress syndrome.
Kong, D. L., Prough, D. S., Whitley, J. M. et al. (1991) Hemorrhage and
intracranial hypertension in combination increase cerebral production of
thromboxane A2. Critical Care Medicine, 19, 532538.
Kraus, P. A., Lipman, J., Lee, C. C. J. et al. (1993) Acute lung injury at
Baragwanath ICU. An eight-month audit and call for consensus for other
organ failure in the adult respiratory distress syndrome. Chest, 103,
18321836.
Leach, R. M. and Treacher, D. F. (1992) Oxygen transport: the relation between
oxygen delivery and consumption. Thorax, 47, 971978.
Lewis, A. R., Austen, K. F. and Soberman R. J. (1990) Leukotrienes and other
products of the 5lipoxygenase pathway. New England Journal of Medicine,
323, 645655.
Lewis, S. B., Reilly, P. L. and Myburgh, J. A. (1994) Developments in
intracranial pressure monitoring and investigation of head injured patients,
in Current Topics in Intensive Care (eds G. J. Dobb, J. Bion, H. Burchardi and
R. P. Dellinger), W. B. Saunders, London, pp. 119.
Lewis, S. B., Reilly, P. L. and Myburgh, J. A. (1995) Detection of cerebral
venous desaturation by continuous jugular bulb oximetry following acute
neurotrauma. Anaesthesia and Intensive Care, 23, 307314.
Lillie, P. E. and Roberts, J. G. (1988) Carbon dioxide monitoring. Anaesthesia
and Intensive Care, 16, 4144.
359
Lin, E. S. and Oh, T. E. (1990) Ventilation which mode?, in Clinical

Anaesthesiology, International Practice and Research, (ed. G. J. Dobb), Baillière
Tindall, London, pp. 441473.
Linton, D. M. (1993) Difficult ventilatory weaning. Current Opinion in
Anaesthesiology, 6, 347353.
Liu, S., Lee, T. and Bongard, F. (1992) Accuracy of capnography in
nonintubated surgical patients. Chest, 102, 15121515.
Luce, J. M. (1985) Medical management of spinal cord injury. Critical Care
Medicine, 13, 126.
Luce, J. M. (1993) Cardiopulmonary physiology and management in
neurosurgical intensive care, in Neurosurgical Intensive Care, (ed. B. T.
Andrews), McGraw-Hill, New York, pp. 130.
Ludbrook, G. L., Russell, W. J., Webb, R. K. et al. (1993) The electrocardiograph: applications and limitations an analysis of 2000 incident reports.
Anaesthesia and Intensive Care, 21, 558564.
McFarlane, C., Denholm, S. W., Sudlow, C. L. M. et al. (1994) Laryngotracheal
stenosis: a serious complication of percutaneous tracheostomy. Anaesthesia,
49, 3840.
McGee, W. T., Ackerman, B. L, Rouben, L. R. et al. (1993) Accurate placement
of central venous catheters: A prospective, randomized, multicenter trial.
McHale, H. G. and Roddie, I. C. (1983) Peripheral lymph flow during
intravenous noradrenaline infusion in sheep (abstract) Journal of Physiology
(London), 334, 50P.
MacIntyre, N. R. (1993) Clinically available new strategies for mechanical
ventilatory support. Chest, 104, 560565.
MacNaughton, P. D and Evans T. W. (1992) Management of Adult Respiratory
Distress Syndrome. Lancet, 339, 469472.
Mador, M. J. (1991) Respiratory muscle fatigue and breathing pattern. Chest,
100, 14301435.
Marcy, T. W. and Marini, J. J. (1991) Inverse ratio ventilation in ARDS. Chest,
100, 494504.
Marini, J. J. (1990) Paying the piper the linkage of alveolar ventilation to
alveolar pressure. Intensive Care Medicine, 16, 7374.
Marmarou, A. (1992) Increased intracranial pressure in head injury and
influence of blood volume. Journal of Neurotrauma, 9, S327S332.
Meduri, G. U. (1993) Diagnosis of ventilator-associated pneumonia. Infectious
Disease Clinics of North America, 7, 295329.
Meduri, G. U. and Chastre, J. (1992) The standardization of bronchoscopic
techniques for ventilator-associated pneumonia. Chest, 102, S557S564.
Mercat, A., Graini, L., Teboul, J. et al. (1993) Cardiorespiratory effects of
pressure-controlled ventilation with and without inverse ratio in the adult
respiratory distress syndrome. Chest, 104, 871875.
Mermel, L. A. and Maki, D. G. (1994) Infectious complications of SwanGanz
pulmonary artery catheters. Pathogenesis, epidemiology, prevention and
management. American Journal of Respiratory and Critical Care Medicine, 149,
10201036.
Michenfelder, J. D. and Milde, J. H. (1988) The interaction of sodium
nitroprusside hypotension and isoflurane in determining cerebral vasculature effects. Anesthesiology, 69, 870.
Miller, J. D., Piper, I. R. and Dearden, N. M. (1993) Management of intracranial
hypertension in head injury: matching treatment with cause. Acta Neurochirurgica (Supplement) (Vienna), 57, 152159.
Miller, J. D., Dearden, N. M., Piper, I. R. et al. (1992) Control of intracranial
pressure in patients with severe head injury. Journal of Neurotrauma, 9,
S317S326.
Mitchell, J. P. (1992) Improved outcome based on fluid management in
critically -ill patients requiring pulmonary artery catheterization. American
Review of Respiratory Distress, 145, 990998.
Mizock, B. A. and Falk, J. L. (1992) Lactic acidosis in critical illness. Critical
Moran, J. L., OFathartaigh, M. S., Peisach, A. R. et al. (1993) Epinephrine as an
inotropic agent in septic shock: a dose-profile analysis. Critical Care
Medicine, 21, 7077.
Moser, K. M. (1990) Venous thromboembolism. American Review of Respiratory
Disease, 141, 235249.
Muizelaar, J. P., Wei, E. P., Kontos, H. A. et al. (1983) Mannitol causes
compensatory cerebral vasoconstriction and vasodilation in response to
blood viscosity changes. Journal of Neurosurgery, 59, 822828.
Muizelaar, J. P., van der Poel, H. G., Li, Z. C. et al (1988) Pial arteriolar vessel
diameter and CO2 reactivity during prolonged hyperventilation in the
rabbit. Journal of Neurosurgery, 69, 923927.
Murray, J. F., Matthay, M. A., Luce, J. M. et al. (1988) An expanded definition
of the adult respiratory distress syndrome. American Review of Respiratory
Distress, 138, 720723.
Myburgh, J. A. (1991) The clinical utility of oximetry, capnography and
laboratory bicarbonate compared with blood gas analysis. Anaesthesia and
360
Myburgh, J. A. (1992) Derived oxygen saturations are not clinically useful for
the calculation of oxygen consumption. Anaesthesia and Intensive Care, 20,
460463.
Myburgh, J. A. and Lewis, S. B. (1996) Goal directed therapy in neurotrauma,
in Yearbook of Intensive Care and Emergency Medicine, (ed. J. L. Vincent),
Myburgh, J. A., Webb, R. K. and Worthley, L. I. G. (1991) The P50 is reduced in
critically ill patients. Intensive Care Medicine, 17, 355358.
Nadeau, S. and Noble, W. H. (1986) Misinterpretation of pressure measurements from the pulmonary artery catheter. Canadian Anaesthetists Society
Journal, 33, 352363.
Neil-Dwyer, G., Walter, P., Cruickshank, J. M. et al. (1978) Effect of propranolol
and phentolamine on myocardial necrosis after subarachnoid haemorrhage.
British Medical Journal, ii, 990992.
North, J. B. and Jennett, S. (1974) Abnormal breathing patterns associated with
acute brain damage. Archives of Neurology, 31, 338344.
Norwood, S., Ruby, A., Civetta, J. et al. (1991) Catheter-related infections and
associated septicemia. Chest, 99, 968975.
Nunn, J. F. (1987a) The oxygen cascade, in Applied Respiratory Physiology, 3rd
edn, Butterworths, London, pp. 100283.
Nunn, J. F. (1987b) Elastic forces and lung volumes, in Applied Respiratory
Physiology, 3rd edn, Butterworths, London, pp. 2341.
Obrist, W. D., Gennarelli, T. A., Segawa, H. et al. (1979) Relation of cerebral
blood flow to neurological status and outcome in head-injured patients.
ODriscoll, B. R., Taylor, R. J., Horsley, M. G. et al. (1989) Nebulised salbutamol
with and without ipratropium bromide in acute airflow obstruction. Lancet,
i, 14181420.
Oh, T. E. (1992) Defining adult respiratory distress syndrome. British Journal of
Hospital Medicine, 47, 350353.
Oh, T. E. and Duncan, A. W. (1988) Oxygen therapy. Medical Journal of
Australia, 149, 141146.
Oh, T. E., Bhatt, S., Lin, E. S. et al. (1991) Plasma catecholamines and oxygen
consumption during weaning from mechanical ventilation. Intensive Care
Medicine, 17, 199203.
Pansard, J. L., Cholley, B., Devilliers C. et al. (1992) Variation in arterial to endtidal CO2 tension differences during anesthesia in the Kidney Rest lateral
decubitus position. Anesthesia and Analgesia, 75, 506510.
Pauca, A. L., Wallenhaupt, S. L., Kon, N. D. et al. (1992) Does radial artery
pressure accurately reflect aortic pressure? Chest, 102, 11931198.
Paulson, O. B., Strandgaard, S., Edvinsson, L. et al. (1990) Cerebral
autoregulation. Cerebrovascular Brain Metabolism Review, 2, 161192.
Pepe, P. E. and Marini, J. J. (1982) Occult positive end-expiratory pressure in
mechanically ventilated patients with airflow obstruction. American Review
of Respiratory Disease, 126, 166170.
Persson, A. V., Davis, R. J. and Villavicencio, J. L. (1991) Deep vein thrombosis
and pulmonary embolism. Surgical Clinics of North America, 71,
11951209.
Petty, T. L. (1990) Acute respiratory distress syndrome. Disease-a-month, 1,
958.
Pfenninger, E. G. and Lindner, K. H. (1991) Arterial blood gases in patients
with acute head injury at the accident site and upon hospital admission.
Acta Anaesthesiologica Scandinavica, 35, 148152.
Phillips, G. D. and Skowronski, G. A. (1986) Manual resuscitators and
portable ventilators. Anaesthesia and Intensive Care, 14, 306313.
Piek, J., Chesnut, R. M., Marshall, L. F. et al. (1992) Extracranial complications
of severe head injury. Journal of Neurosurgery, 77, 901907.
Pinsky, M. R., Desmet, J. and Vincent, J. L. (1992) Effect of positive endexpiratory pressure on right ventricular function in humans. American
Review of Respiratory Disease, 146, 681687.
Pitts, L. H. and Andrews, B. T. (1992) Intracranial pressure monitoring and
treatment of intracranial hypertension, in Principles of Critical Care Medicine,
(ed. L. Wood), McGraw-Hill, New York, pp. 29853011.
Platanias, L. C. and Vogelgang, N. J. (1990) Interleukin-1: biology, pathophysiology, and clinical prospects. American Journal of Medicine, 89, 621629.
Prichard, B. N., Owens, C. W., Smith, C. C. et al. (1991) Heart and
catecholamines. Acta Cardiologica, 46, 309322.
Proctor, H. J., Cairns, C., Fillipo, D. et al. (1984) Brain metabolism during
increased intracranial pressure as assessed by niroscopy. Surgery, 96, 273.
Redan, J. A., Livingston, D. H., Tortella, B. J. et al. (1991) The value of
intubating and paralysing patients with suspected head injury in the
emergency department. Journal of Trauma, 31, 371.
Reed, R. L. II (1993) Oxygen consumption and delivery. Current Opinion in
Anaesthesiology, 6, 329334.
Reinhart, K., Hannemann, L., Kuss, B. (1990) Optimal oxygen delivery in
critically ill patients. Intensive Care Medicine, 16, S149S155.
Repine, J. E. (1992) Scientific perspectives on adult respiratory distress
syndrome. Lancet, 339, 466469.
Rischbieth, A. M., Blight, I. L. and Myburgh, J. A. (1994) Efficacy of
endotracheal tube cuff seal in three cuff management techniques. Anaesthesia and Intensive Care, 22, 219220.
Roberts, P., Pollay, M., Engles, C. et al. (1987) Effect on intracranial pressure of
furosemide combined with varying doses and administration of mannitol.
Robertson, C. S., Clifton, G. L., Taylor, A. A. et al. (1983) Treatment of

hypertension associated with head injury. Journal of Neurosurgery, 59,
455460.
Robin, E. D. (1985) The cult of the SwanGanz catheter. Annals of Internal
Medicine, 103, 445449.
Rodriguez, J. L., Gibbons, K. J., Bitzer, L. G. et al. (1991) Pneumonia: incidence,
risk factors, and outcome in injured patients. Journal of Trauma, 31, 907912.
Rosell, S. (1980) Neuronal control of microvessels. Annual Review of Physiology,
42, 359371.
Rosner, M. J. (1987) Cerebral perfusion pressure: the link between the cerebral
and systemic circulations, in Cerebral Blood Flow: Physiologic and Clinical
Aspects, (ed. J. H. Wood), McGraw-Hill, New York, pp. 425448.
Rosner, M. J. (1993) Pathophysiology and management of increased intracranial pressure, in Neurosurgical Intensive Care, (ed. B. T. Andrews),
McGraw-Hill, New York, pp. 57112.
Rosner, M. J. and Coley, I. (1986) Cerebral perfusion pressure, the ICP and
head elevation. Journal of Neurosurgery, 65, 636.
Rosner, M. J. and Coley, I. (1987) Cerebral perfusion pressure: a hemodynamic
mechanism of mannitol and the postmannitol hemogram. Neurosurgery, 21,
147156.
Rouby, J. J., Laurent, P., Gosnach, M. et al. (1994) Risk factors and clinical
relevance of nosocomial maxillary sinusitis in the critically ill. American
Journal of Respiratory and Critical Care Medicine, 150, 776783.
Runciman, W. B., Ilsley, A. H. and Rutten, A. J. (1988) Systemic arterial blood
pressure. Anaesthesia and Intensive Care, 16, 5457.
Runciman,W. B., Myburgh, J. A. and Upton, R. N. (1990) Pharmacokinetics
and pharmacodynamics in the critically ill, in Clinical Anaesthesiology,
International Practice and Research, (ed. G. J. Dobb), Baillière Tindall, London,
vol. 4, pp. 271303.
Runciman, W. B., Morris, J. L. (1993a) Adrenoceptor agonists, in Mechanisms of
Drugs in Anaesthesia, (eds S. A. Feldman, W. Paton and C. Scurr), Edward
Arnold, Sevenoaks, pp. 262291.
Runciman, W. B., Morris, J. L. (1993b) Adrenoceptor antagonists, in
Mechanisms of Drugs in Anaesthesia, (eds S. A. Feldman, W. Paton and C.
Scurr), Edward Arnold, Sevenoaks, pp. 291306.
Runciman, W. B., Webb, R. K., Barker, L. et al. (1993) The pulse oximeter:
applications and limitations an analysis of 2000 incident reports.
Russell, G. B., Graybeal, J. M. and Strout, J. C. (1990) Stability of arterial to end
tidal carbon dioxide gradients during postoperative cardiorespiratory
support. Canadian Journal of Anaesthesia, 37, 560566.
Russell, J. A., Ronco, J. J., Lockhat, D. et al. (1990) Oxygen delivery and
consumption and ventricular preload are greater in survivors than in
nonsurvivors of the adult respiratory distress syndrome. American Review of
Respiratory Distress, 141, 659665.
Rutten, A. J., Ilsley, A. H., Skowronski, G. A. et al. (1986) A comparative study
of the measurement of mean arterial blood pressure by automatic
oscillometers, arterial cannulation and auscultation. Anaesthesia and Intensive Care, 14, 5865.
Salmenpera, M., Peltola, K., Takkunen, O. et al. (1983) Cardiovascular effects
of pancuronium and vecuronium during high dose fentanyl anaesthesia.
Anesthesia and Analgesia, 62, 1059.
Santak, B., Radermacher, P., Sandmann, W. et al. (1991) Influence of SIMV plus
distributions during postoperative
A/Q
inspiratory pressure support on V
weaning. Intensive Care Medicine, 17, 136140.
Sassoon, C. S. H. (1991) Positive pressure ventilation. Chest, 100, 14211429.
Sassoon, C. S. H., Light, R. W., Lodia, R. et al. (1991) Pressure-time product
during continuous positive airway pressure, pressure support ventilation,
and T-piece during weaning from mechanical ventilation. American Review
of Respiratory Disease, 143, 469475.
Schettini, A., Stahurski, B. and Young, H. F. (1982) Osmotic and osmotic-loop
diuresis in brain surgery. Effects on plasma and CSF electrolytes and ion
secretion. Journal of Neurosurgery, 56, 679.
Schmidt, G. A. and Wood, L. D. H. (1993) Critical Care Medicine. Journal of the
Schreuder, W. O., Schneider, A. J., Groeneveld, A. B. J. et al. (1989) Effect of
dopamine vs norepinephrine on hemodynamics in septic shock. Chest, 95,
12821288.
Schumacker, P. T. and Cain, S. M. (1987) The concept of a critical oxygen
delivery. Intensive Care Medicine, 13, 223229.
Schumacker, P. T. and Samsel, R. W. (1989) Oxygen delivery and uptake by
peripheral tissues: Physiology and pathophysiology. Critical Care Clinics, 5,
255269.
Schumacker, P. T., Rhodes, G. R., Newell, J. C. et al. (1979) Ventilation
perfusion imbalance after head trauma. American Review of Respiratory
Disease, 119, 3343.
Shapiro, H. M. and Marshall, L. F. (1978) Intracranial pressure responses to
PEEP in head-injured patients. Journal of Trauma, 18, 254.
REFERENCES
Shelly, M. P., Lloyd, G. M. and Park, G. R. (1988) A review of the mechanisms
and methods of humidification of inspired gases. Intensive Care Medicine, 14,
19.
Shoemaker, W. C., Appel, P. L. and Kram, H. B. (1992) Role of oxygen debt in
the development of organ failure sepsis, and death in high-risk surgical
Shoemaker, W. C., Appel, P. L., Kram, H. B. et al. (1988) Prospective trial of
supranormal values of survivors as therapeutic goals in high-risk surgical
Silverman, H. J. (1991) Lack of a relationship between induced changes in
oxygen consumption and changes in lactate levels. Chest, 100, 10121015.
Silverman, H. J. and Tuma, P. (1992) Gastric tonometry in patient with sepsis.
Chest, 102, 184188.
Simard, J. M. and Bellefleur, M. (1989) Systemic arterial hypertension in head
trauma. American Journal of Cardiology, 63, 32C35C.
Simmons, R. L., Martin, A. M. Jr, Heisterkamp, C. A. III et al. (1969)
Respiratory insufficiency in combat casualties: II. Pulmonary edema
following head injury. Annals of Surgery, 170, 3944.
Simpson, D. A., Speed, I. E. and Blumbergs, P. C. (1991) Embolism of cerebral
tissue: a cause of coagulopathy and infarction? Surgical Neurology, 35,
159162.
Singbartl, G., Cunitz, G. and Hamrouni, H. (1982) Disturbed pulmonary gas
exchange in patients with cerebral trauma. Anaesthesist, 31, 228233.
Sohma, A., Brampton, W. J., Dunnill, M. S. et al. (1992) Effect of ventilation
with positive end expiratory pressure on the development of lung damage
in experimental acid aspiration pneumonia in the rabbit. Intensive Care
Medicine, 18, 112117.
Stone, J. L. (1989) Nonsurgical management of increased intracranial pressure.
Seminars in Neurology, 9, 218224.
Stone, D. J. and Bogdonoff, D. L. (1992) Airway considerations in the
management of patients requiring long-term endotracheal intubation.
Anesthesia and Analgesia, 74, 276287.
Strieter, R. M., Kunkel, S. L., Bone, R. C. (1993) Role of tumour necrosis factora in disease states and inflammation. Critical Care Medicine, 21,
S447S463.
Sudlow, M. F. and the Research Committee of the British Thoracic Society
(1992) Optimum duration of anticoagulation for deep-vein thrombosis and
pulmonary embolism. Lancet, 340, 873876.
Szekely, S. M., Webb, R. K., Williamson, J. A. et al. (1993) Problems related to
the endotracheal tube: an analysis of 2000 incident reports. Anaesthesia and
Takala, J., Keinanen, O., Vaisanen, P. et al. (1989) Measurement of gas exchange
in intensive care: laboratory and clinical validation of a new device. Critical
Talman, W. T. (1985) Cardiovascular regulation and lesions of the central
nervous system. Annals of Neurology, 18, 112.
361
Theodore, J. and Robin, E. D. (1976) Speculations on neurogenic pulmonary

edema (NPE). American Review of Respiratory Disease, 113, 405411.
Tomlinson, J. R., Miller, K. S., Lorch, D. G. et al. (1989) A prospective
comparison of IMV and T-piece weaning from mechanical ventilation.
Chest, 96, 348352.
Torres, A., Gonzalez, J. and Ferrer, M. (1991) Evaluation of the available
invasive and non-invasive techniques for diagnosing nosocomial pneumonias in mechanically ventilated patients. Intensive Care Medicine, 17,
439448.
Tracey, K. J. and Cerami, A. (1993) Tumour necrosis factor: An updated review
of its biology. Critical Care Medicine, 21, S415S422.
Trauma Committee, Royal Australasian College of Surgeons (1992) Early
Management of Severe Trauma (EMST), Royal Australasian College of
Surgeons, Melbourne, Victoria.
Trunkey, D. (1991) Initial treatment of patients with extensive trauma. New
Tsumo, K., Prato, P. and Kolobow, T. (1990) Acute lung injury from mechanical
ventilation at moderately high airway pressures. Journal of Applied
Physiology, 69, 956961.
Tsutsumi, H., Ide, K., Mizutani, T. et al. (1985) The relationship between
intracranial pressure, cerebral perfusion pressure and outcome in headinjured patients: the critical level of cerebral perfusion pressure, in
Intracranial Pressure VI, (eds J. D. Miller, G. M. Teasdale and J. O. Rowan),
Vincent, J. L. (1990) The relationship between oxygen demand, oxygen
uptake, and oxygen supply. Intensive Care Medicine, 16, S145S148.
Vincent, J. L. and Bihari, D. (1992) Sepsis, severe sepsis or sepsis syndrome:
need for clarification. Intensive Care Medicine, 18, 255256.
Walls, R. M. (1993) Rapid-sequence intubation in head trauma. Annals of
Emergency Medicine, 22, 10081013.
Walsh, J. M., Vanderwarf, C., Hoscheit, D. et al. (1989) Unsuspected
hemodynamic alterations during endotracheal suctioning. Chest, 95,
162165.
Weismann, S. J. (1939) Edema and congestion of the lungs resulting from
intracranial hemorrhage. Surgery, 6, 722729.
West, J. B. (1985) Respiratory Physiology The Essentials, Williams & Wilkins,
Baltimore, MD, pp. 1167.
Yang, K. L., and Tobin, M. J. (1991) A prospective study of indexes predicting
the outcome of trials of weaning from mechanical ventilation. New England
Zwissler, B., Schosser, R., Schwickert, C. et al. (1991) Perfusion of the
interventricular septum during ventilation with positive end-expiratory
pressure. Critical Care Medicine, 19, 14141424.
18
SEDATION AND ANESTHESIA

Guy L. Ludbrook
18.1
Introduction
The patient with a severe acute head injury, often with

other bone and soft-tissue injuries, presents a complex
management problem involving a large number of
medical and paramedical staff. A coordinated
approach is essential, with clear and frequent communication between the involved personnel on all aspects
of critical care, including initial resuscitation, anesthesia, radiological imaging, surgical intervention and
intensive care management.
The anesthetist is an active member of this team and
is involved in many stages of management. Indeed,
the anesthetist is often the one medical specialist in
constant attendance from hospital admission through
to the postoperative ward or intensive care unit.
Furthermore, therapy instituted by the anesthetist,
such as the choice of sedative agents, paralysis,
intubation and fluid management, will directly
impact upon other aspects of management, such as
neurosurgical and intensive care treatment.
The anesthetist must have a sound knowledge of
cerebral pathophysiology and pharmacology and of
the acute medical care of the multiply injured patient.
Many of these aspects are covered in other sections of
this book. The conduct of basic anesthetic techniques
is well covered in general anesthetic texts. This
chapter will therefore concentrate on the pharmacology of drugs in current use and aspects of anesthetic
management specifically related to the patient with a
severe head injury.
18.2
Cerebral pharmacology
There is a wide range of potent drugs available for

sedation, analgesia, anesthesia, paralysis, anticonvulsant therapy, cardiovascular manipulation and fluid
balance and most have actions that affect the injured
brain and its treatment.
Knowledge of drug effects on cerebral oxygen
consumption (CMR), cerebral blood flow (CBF), intracranial pressure (ICP), cerebral perfusion pressure
(CPP) and seizure activity are essential for planning

the optimal drug regimen for each patient.
Partly because these parameters are quite difficult to
measure both experimentally and in clinical practice,
there is no agreed optimal drug regimen for anesthesia
or sedation of the head-injured patient. Indeed, views
are constantly changing as new information and new
drugs come to hand. Examples of this include the
recently described effects of nitrous oxide on CBF and
CMR, the increases in CBF and ICP reported with the
synthetic opioids and the introduction into clinical
practice of the volatile agents desflurane and
sevoflurane.
More difficult to resolve is whether a particular
anesthetic technique or drug regimen with theoretical
advantages and perhaps significant financial cost
improves patient outcome. This was highlighted in a
recent study where no differences in short-term
outcome were found between three anesthetic drug
regimens used during elective neurosurgery, even
though each had very different pharmacological cerebral effects (Todd et al., 1993). These difficulties were
well summarized in an editorial on monitoring by
Eichhorn (1993), where he stated: There will never be
enough p < 0.05 scientific data on outcome of anesthesia to satisfy the staunch traditionalists. We will
continue, however, to have practices based on a
modern synthesis of the best information available at
the time.
18.3
Intravenous anesthetic agents
18.3.1
PROPOFOL
Propofol, a substituted phenol, is a relatively new

intravenous induction agent. It has achieved widespread popularity, largely because it has a short
effective half-life and hence a rapid onset and offset of
action with a short recovery time. Intravenous administration via bolus produces rapid onset of anesthesia
and a dose-dependent coupled fall in CBF and CMR,
effects similar to those well described following
364
barbiturate administration (Ludbrook et al., 1996;

Ramani et al., 1992; Eng et al., 1992; Figure 18.1).
The effects on CBF are probably secondary to a
reduction in CMR, but the effectiveness of propofol in
reducing raised ICP may be less if CMRCBF coupling
is impaired by the head injury itself. Hence cerebral
vasoconstriction and a decrease in ICP could be
demonstrated in patients receiving propofol during
elective surgery (Ravussin et al., 1988) but not in a
group of patients with severe head injuries, despite
evidence of a reduction in CMR (Stewart et al., 1994).
CO2 reactivity and autoregulation of CBF appear
reasonably well preserved during propofol admin-
istration (Fox et al., 1992; Eng et al., 1992) and

consequently, in the normal brain, CBF is not greatly
affected by changes in CPP between approximately 50
and 150 mmHg. This may not be so in the patient with
head injury if cerebral autoregulation is impaired
(Bouma et al., 1992).
In the head-injured patient with raised ICP, rapid
induction with propofol may reduce ICP and improve
cerebral oxygenation. This may prevent or attenuate
the marked, though relatively brief ICP increases that
may occur with endotracheal intubation.
Propofol may cause dose-dependent cardiovascular
depression and systemic hypotension especially when
Figure 18.1 Changes in CBF and CMR in a sheep following a bolus dose of propofol, 100 mg i.v. (Source: reproduced from
Ludbrook et al., 1996, with permission.)
INTRAVENOUS ANESTHETIC AGENTS
anesthesia is rapidly induced by bolus drug administration. Hypotension is partly due to systemic
vasodilatation (Pensado, Molins and Alvarez, 1993;
Petros, Bogle and Pearson, 1993) but may also be due
to direct myocardial depression and decreased cardiac
output (Lindgren, 1993). This may induce cerebral
ischemia by one of two mechanisms. If autoregulation
is intact, a reduction in mean arterial pressure will
produce reflex cerebral vasodilatation and risk further
increase in ICP and fall in CPP; if autoregulation is
impaired, hypotension may produce a critical decrease
in CPP and CBF. The risk of hypotension is greater in
the presence of hypovolemia due either to multiple
injury or to deliberate dehydration for ICP control.
Because of the cardiovascular effects and brief duration of action, propofol requires a careful choice of
dose regimen for induction in order to avoid hypotension or raised ICP at laryngoscopy. A slow rate of
administration may provide greater cardiovascular
stability (Stokes and Hutton, 1991) and allow titration of dose against effect, but this is usually not
possible in patients with head injury because of the
risk of regurgitation in the unfasted patient and the
dangers of hypoventilation and hypercarbia.
The potentially beneficial cerebral effects of propofol and its brief duration of action even after prolonged administration make it suitable for maintaining anesthesia or sedation by infusion. Thiopentone,
opioids and benzodiazepines can be used in this
manner, but they have a greater tendency to accumulate, which can delay recovery from anesthesia or
sedation. This is particularly relevant if early or
frequent postoperative neurological assessment is
required.
Infusion regimens which will provide stable planes
of anesthesia without prolonged recovery times have
been developed for propofol, either alone or in
combination with inhaled anesthetic agents (Valanne,
1992). For similar reasons, propofol infusion is becoming more popular for sedation of ventilated patients in
intensive care (Roekaerts, Huygens and de Lange,
1993; Farling, Johnstone and Coppel, 1989). Whether
used in anesthesia or for sedation, hypotension with
propofol needs to be carefully avoided by aggressive
fluid and/or pharmacological therapy.
There has been considerable debate about the effects
of propofol on seizure threshold. This is clearly relevant
in patients with head injuries where massive increases
in CBF have been demonstrated during seizures, even
during artificial ventilation (Meldrum and Nilsson,
1976). The anticonvulsant properties of propofol are
now well described. It reduces the duration of seizures
during electroconvulsive therapy (ECT) when compared to more traditional barbiturate-based anesthesia (Simpson et al., 1988; Dwyer et al., 1988;
although its effect on ECT efficacy is as yet unclear)
365
and there are clearly documented examples of its

efficacy in the treatment of status epilepticus (Melloni
et al., 1992; Mackenzie, Kapadia and Grant, 1990).
Excitatory movements and patient arousal have been
described during induction with propofol but, rather
than being true seizure activity, are probably related to
light planes of anesthesia and are more akin to the
excitation that frequently occurs during induction with
methohexitone (Reddy et al., 1993). For a more
complete discussion the reader is directed to recent
reviews of the current evidence regarding propofol and
convulsions (Bevan, 1993; Smith et al., 1994).
The clinical experience of this author is that
propofol, when administered at both sedative and
anesthetic doses by infusion, is effective in controlling
seizures after neurosurgery. This adds to its potential
benefits in the management of the head-injured
patient during and after surgery.
18.3.2
BARBITURATES
Thiopentone has been one of the most commonly used

induction agents in anesthesia for many years, and is
frequently used in neuroanesthesia and intensive care.
As with propofol, intravenous administration of
thiopentone produces rapid induction of anesthesia
accompanied by dose-dependent parallel falls in CBF
and CMR (Michenfelder, 1974). The preserved coupling
of CBF and CMR and the potential benefits on ICP led
to the extensive use of barbiturates in the head-injured
patient.
The cerebral vasoconstriction induced by barbiturates is effective in controlling raised ICP, even when
it is refractory to mannitol or hyperventilation. Barbiturates may also improve blood flow to underperfused
areas of the injured brain which have lost or reduced
vasoreactivity (reverse steal; Levin et al., 1979; Marshall et al., 1978; Frost et al., 1981; Rockoff, Marshall and
Shapiro, 1979). As with propofol, cerebral vasoconstriction is secondary to reduced CMR, and is dependent on
the presence of coupling of CBF and CMR after injury.
Messeter et al. (1986) found that beneficial effects on
CBF and ICP could be demonstrated in patients with
intact CO2 reactivity, but there was no change in CVR
and only a transient ICP reduction in patients in whom
CO2 reactivity was impaired.
The protective effects of barbiturates in experimental models of cerebral ischemia or hypoxia appear
greater than can be explained by decreases in CBF and
CMR alone, and other mechanisms including decreasing cellular calcium influx, inhibition of free radical
formation, membrane stabilization and actions as a
GABA agonist have been proposed (Blaustein and
Ector, 1975; Smith et al., 1980; Yoshida, Inoh and
Asano, 1983). However, despite the evidence of cerebral protection in experimental models, there is little
366
evidence that barbiturates improve outcome after

severe head injuries (Ward et al., 1985; Chapter 19).
Thiopentone may cause hypotension, due primarily
to myocardial depression (Azari and Cork, 1993).
Hypotension and a fall in CPP is a particular risk in
the hypovolemic patient and, as with propofol,
careful choice of dose and aggressive use of vasopressors should be used to avoid this during induction
of anesthesia. If barbiturate coma is used to control
raised ICP concomitant use of vasopressors is often
necessary.
To maintain anesthesia, thiopentone can be administered simply and successfully via continuous infusion (Crankshaw and Karasawa, 1989), avoiding
inhaled anesthetic agents, many of which produce
cerebral hyperemia and may increase ICP. However,
thiopentone has largely been replaced by propofol for
intravenous anesthesia because of reduced recovery
times.
Barbiturates are very effective anticonvulsants.
Thiopentone is probably the most effective agent
available for the rapid control of seizures. For prolonged control, however, thiopentone must be given
by continuous infusion. Hence if barbiturates are
necessary for more than 2448 hours, longer-acting
agents such as pentobarbitone are usually
substituted.
Methohexitone, a barbiturate, is widely used as an
induction agent because it has a short duration of
action allowing a more rapid recovery than thiopentone. Induction is frequently accompanied by
excitatory movements which may resemble seizures
(Dundee and Moore, 1961). For this reason it is not
commonly used in neuroanesthesia.
Longer-acting barbiturates such as pentobarbitone
(half-life 2030 h) and phenobarbitone (half-life
3090 h) are rarely used in anesthetic practice because
of their delayed onset and long duration of action, but
they may be considered for prolonged anticonvulsant
therapy.
Ketamine, a phencyclidine derivative, is a useful
induction agent, which may support or increase blood
pressure because of a drug-induced increase in sympathetic activity, and it may induce postoperative
confusion or nightmares. There is conflicting evidence
regarding the effects of ketamine on CBF. Early
studies found marked increases in CBF, CMR and ICP
and interference with autoregulation (Dawson,
Michenfelder and Theye, 1971; Shapiro, Whyte and
Harris, 1974; Whyte, Shapiro and Turner, 1972). In
addition, the postoperative confusion which occurs
particularly in the elderly, may complicate neurological assessment. For these reasons, ketamine was
not recommended for neuroanesthesia by most clinicians. Later studies have shown no change or falls in
CBF and ICP (Bjorkman et al., 1992; Pfenniger and
Reith, 1990; Houggard, Hansen and Broderson, 1974;

Takeshita and Michenfelder, 1972). Furthermore, in
some animal models a cerebral protective effect
following cerebral ischemia and head injury has been
reported (Church, Zeman and Lodge, 1988; Hoffman
et al., 1992; Shapira, Artru and Lam, 1992; Shapira et
al., 1994). Ketamine is an antagonist at N-methyl
D-aspartate (NMDA) receptors, and this is one proposed mechanism behind this protective action.
Ketamines popularity has recently increased in
general anesthesia, where it may be given either by
infusion alone or in combination with other hypnotic
agents (Royblat et al., 1992) and, for the reasons
outlined above, there is also some renewed interest in
its use in neurosurgery.
18.3.3
ETOMIDATE
Etomidate is a hypnotic agent that will rapidly

induces anesthesia after bolus intravenous administration. As with propofol and thiopentone, onset of
anesthesia is accompanied by parallel decreases in
CBF and CMR (Milde et al., 1985; Renou et al., 1978).
There is some evidence that etomidate provides
protection against cerebral ischemia (Sano et al., 1993;
Watson et al., 1992). CO2 reactivity is preserved
during etomidate administration (Renou et al., 1978).
Accompanying cardiovascular changes are usually
minimal (Criado et al., 1980) and it is therefore the
induction agent that is most likely to reduce ICP and
CMR with least compromise of CPP.
These effects are likely to be most beneficial when
inducing anesthesia in a head-injured patient, particularly in the presence of hypovolemia.
However, the onset of anesthesia is not uncommonly accompanied by myoclonic movements resembling seizures (Reddy et al., 1993). There have also
been reports of adrenal suppression following even
brief administration (Ellis et al., 1985; Fellows et al.,
1983) limiting its usefulness for maintenance of
anesthesia.
18.3.4
BENZODIAZEPINES
Benzodiazepines have sedative and anticonvulsant

properties, making them potentially very useful for
head-injured patients during anesthesia and intensive
care treatment. They bind to benzodiazepine receptors
linked to GABA-dependent chloride channels in the
cell membrane and, as the actions of both propofol
and the barbiturates are likely to be at least partially
GABA-mediated, it might be expected that their
cerebral effects would be similar.
Diazepam has been used extensively as an intravenous sedative/hypnotic agent but disadvantages
include a long elimination half-life (2050 h), irritant
MUSCLE RELAXANTS
effects on veins and incompatibility with a number of

other agents. The use of a lipid emulsion carrier
improves intravenous administration but its prolonged sedative and ventilatory depressive effects,
especially after repeated or infusion administration,
are a disadvantage.
Midazolam is a water-soluble short-acting benzodiazepine that has largely replaced diazepam for
parenteral use because it has a shorter half-life, is
miscible with most drugs and does not cause venous
irritation. Midazolam produces a dose-dependent fall
in CMR, CBF and ICP in both experimental animals
and humans (Nugent, Artru and Michenfelder, 1982;
Giffen, Cotrell and Schwirey, 1984; Fleischer et al.,
1988; Forster, Juge and Morel, 1982; Hoffman, Miletich
and Albrecht, 1986; Bjorkman et al., 1992). It is a
potent anticonvulsant.
In line with most drugs that reduce CBF and CMR,
there is some evidence of cerebral protective effects
with both midazolam and diazepam (Nugent, Artru
and Michenfelder, 1982). Midazolam is a safe and
effective agent for sedating patients with head injury
undergoing ventilation in intensive care, where it is
often combined in an infusion with opioids such as
fentanyl or morphine.
Midazolam is not as effective an anesthetic agent,
since the onset of anesthesia is slow, although without
excitation, and it has a less reliable doseresponse
relationship than many other hypnotic induction
agents. It has a reputation for cardiovascular stability
but the large doses required for induction of anesthesia (in the order of 0.2 mg/kg) may produce a
critical reduction in CPP, especially in the hypovolemic patient or in the presence of raised ICP
(Derbyshire et al., 1984; Forster, Juge and Morel, 1982;
Papazian et al., 1993). Low doses of midazolam can
effectively supplement anesthesia, allowing lower
doses of volatile or hypnotic agents without reducing
CPP.
Flumazenil is a specific antagonist to benzodiazepines that acts competitively at the benzodiazepine
receptor (Hunkeler et al., 1981). Its half-life is significantly shorter than either midazolam or diazepam
and repeated administration or infusions are necessary to avoid recurrent sedation. Given intravenously
it rapidly reverses the sedative and anticonvulsant
effects of benzodiazepines but not surprisingly it
increases CBF and ICP and can induce seizures
(Fleischer et al., 1988; Kumano et al., 1993). In two of
15 head-injured patients administered flumazenil,
ICP increased to over 40 mmHg (Chiolero et al., 1986).
It cannot be recommended in patients with severe
head injury. Rather, if a patient remains excessively
sedated after benzodiazepine administration, ventilation and airway protection should be undertaken
until benzodiazepines levels have fallen.
18.3.5
367
OPIOIDS
Opioids provide potent analgesia and are widely used

intraoperatively to permit lower doses of volatile or
hypnotic anesthetic agents and to provide cardiovascular stability. Fear that ICP would be increased due to
opioid-induced respiratory depression has frequently
limited their use in head-injured patients.
Morphine and pethidine (meperidine) were the
main opioids in anesthetic practice prior to the
introduction of the newer synthetic opioids. They
have similar durations of action and produce identical
respiratory depression in equipotent analgesic doses.
They may be therefore be associated with similar ICP
increases in self-ventilating patients. However, if
ventilation is controlled, they produce either a slight
decrease or little change in CBF and CMR (Takeshita,
Okuda and Sari, 1972; Jobes et al., 1977). They may be
used safely with controlled ventilation for analgesia
during anesthesia, or combined with hypnotic agents
in ventilated patients in intensive care. However their
use during anesthesia should be restricted if early
postoperative extubation is anticipated.
The newer synthetic opioids have relatively brief
durations of action and high dose rates can be used to
supplement anesthesia without causing excessive postoperative recovery times. Recent evidence suggests
that synthetic opioids may cause CO2-independent ICP
increases. This has led to doubts about their safety in
the presence of raised ICP. Indeed there have been
numerous recent studies of the effects of synthetic
opioids on CBF and ICP during ventilation. The
findings are somewhat inconsistent, but overall it
seems likely that any increases in CBF due to the
administration of synthetic opioids are insignificant.
In fact, From et al. (1990) found that there were no
detectable differences in intracranial operating conditions or in neurological outcome when fentanyl,
alfentanil or sufentanil were used.
However despite their reputation for cardiovascular
stability, they can induce hypotension and systemic
arterial pressure must be maintained. Sufentanil and
alfentanil can cause significant increases in ICP accompanied by marked decreases in blood pressure. It is
possible that some of the adverse intracranial effects
are due to cerebral vasodilatation induced by systemic
hypotension. For an excellent summary of this subject
the reader is directed to Mayberg and Lam, 1993.
18.4
Muscle relaxants
18.4.1
DEPOLARIZING RELAXANTS
Suxamethonium is the only depolarizing muscle

relaxant in clinical anesthetic practice. Because of its
short onset time, it is the relaxant most often used
when rapid intubation is required.
368
Suxamethonium 11.5 mg/kg in an adult (1.52 mg/

kg in a child) administered intravenously provides
good intubating conditions in 3060 seconds, thus
minimizing the time between loss of airway protection and airway control. It is sometimes avoided in
neuroanesthesia because of concern about increases in
ICP.
There are several suggested mechanisms for the
increase in ICP. The dominant view is that afferent
signals from muscle spindles activated during muscle
fasciculation produce cortical stimulation, a rise in
CMR and CBF, and hence an increase in ICP (Lanier,
Milde and Michenfelder, 1986). Other mechanisms of
action, perhaps of lesser importance, are an increase in
cerebral venous pressure due to fasciculations and a
fasciculation-induced small but measurable increase in
global metabolic rate leading to a slight rise in arterial
CO2 tension. These ICP changes are transient and
usually quite small, even in the presence of a spaceoccupying lesion, especially when suxamethonium is
given in combination with anesthetic induction agents
known to reduce ICP (Ducey, Deppe and Foley, 1989).
Indeed, Kovarik et al. (1994) were unable to demonstrate any significant changes in intracranial dynamics
when suxamethonium was administered to sedated
ventilated patients with neurological injuries.
Suxamethonium can provide excellent intubating
conditions rapidly with insignificant changes in
intracranial dynamics. It is the relaxant of choice for
rapid-sequence induction in the head-injured patient.
Pretreatment with non-depolarizing relaxants has
been advocated if ICP is raised, to ameliorate any
further ICP increases without increasing the risk of
aspiration (Stirt et al., 1987), but is probably not
necessary.
pressure and heart rate. While this may be an

advantage for circulatory support in the hypotensive
patient, it may produce prolonged increases in ICP if
cerebral autoregulation is impaired. The excretion of
pancuronium is dependent on renal function, which
may be impaired in the presence of major trauma.
(c)
This analog of pancuronium has a shorter half-life, is

cleared by the liver and has negligible direct effects on
the systemic circulation. It does not adversely effect
intracranial dynamics, but bradycardia and consequent hypotension due to its lack of vagolytic
activity may present a problem when it is given in
combination with drugs such as suxamethonium and
synthetic opioids and ICP is raised.
(d)
NON-DEPOLARIZING RELAXANTS
Non-depolarizing relaxants in general have little

direct action on the cerebral circulation but may affect
it indirectly through effects on the systemic circulation
and histamine release.
(a)
Curare
This drug was the earliest muscle relaxant but has now
largely been replaced in clinical practice. It can produce
systemic hypotension and significant histamine release
and has been reported to increase ICP (Tarkkanen,
Laitinen and Johannsen, 1974). Therefore it is not
generally recommended in the head-injured patient.
(b)
Pancuronium
This amino steroid is a long-acting neuromuscular

blocker that causes significant increases in blood
Atracurium
Atracurium is a relatively short-acting agent which is

quite unique in that its offset is partially dependent on
spontaneous degradation in blood. It is useful in
neuroanesthesia because stable neuromuscular blockade can be easily achieved by infusion and intracranial
dynamics are not adversely affected (Rosa et al., 1986).
Rapid bolus administration may produce significant
histamine release and hypotension, but this can be
ameliorated by slow i.v. injection. Although one of its
metabolites (laudanosine) can induce seizure activity
(Sheepstra et al., 1986), sufficient levels are not
achieved when atracurium is administered at clinically appropriate doses and so should not prevent its
use in neuroanesthesia.
(e)
18.4.2
Vecuronium
Mivacurium
This is a newer agent, characterized by a brief

duration of action of only 1015 minutes, and degradation by plasma cholinesterases (Ali, 1988). It does
not act as rapidly as suxamethonium and administration of larger doses in an attempt to reduce onset time
is limited by histamine-related hypotension (Mangat
et al., 1993; Silverman and Brull, 1993). Although
experience is still limited, infusion provides stable
relaxation without risk of accumulation and it appears
suitable for use in neuroanesthesia (Diefenbach,
1992).
(f)
Rocuronium
This steroidal relaxant, has a rapid onset of action. It is

a low potency drug and large doses can be given
safely to decrease onset time without adverse hemodynamic effects (Kopmann, 1992; Puhringer, 1992).
However prolonged duration of action will result
INHALED ANESTHETIC AGENTS
because of its relatively long half-life. Suxamethonium, because of its faster onset, remains the drug of
choice for rapid sequence induction in the headinjured patient.
18.5
Inhaled anesthetic agents
18.5.1
NITROUS OXIDE
Nitrous oxide (N2O) has been widely used for many

years to maintain anesthesia. It has been considered to
be an excellent adjunctive anesthetic agent with
minor side effects. There are significant advantages
using N2O as an adjunctive agent in the head-injured
patient. Its rapid onset and offset of action allows
rapid control of the depth of anesthesia, thus minimizing the response to intubation and surgical stimulus early in the course of anesthesia and allowing
rapid postanesthetic recovery for early assessment of
neurological function. Furthermore, it does not contribute to postoperative respiratory depression.
N2O is also useful in a patient with multiple injuries
because its effects on the cardiovascular system are
minimal. Maintaining CPP in these patients in the
presence of hypovolemia may be more difficult if high
concentrations of volatile agents or intravenous infusions of drugs such as propofol and barbiturates alone
are used.
However, with the advent of techniques such as
transcranial Doppler there is now good evidence that
N2O can cause increases in CBF and ICP and this has
resulted in a decline in its use (Lam and Mayberg,
1992; Sakabe et al., 1978; Jung et al., 1992; Henriksen
and Jorgensen, 1973). Near-maximal vasodilatation
occurs at relatively low inspired concentrations,
suggesting an all or none phenomenon at clinically
useful doses. There is little doubt that N2O can
adversely affect CPP but its advantages in cardiovascular control and in allowing rapid changes in the
depth of anesthesia must still be considered.
The exact mechanism behind the cerebrovascular
effects is not clear. N2O administration increased CBF
in awake humans but pial vessel diameter did not
change with direct exposure to N2O in vitro (Reinstrup
et al., 1994). It was suggested therefore that the
increase in CBF might be due to increase in CMR. This
was supported by a study in which addition of N2O to
propofol caused no change in CBF, suggesting that
CMR depression by propofol negated any CMR or
CBF increase due to N2O (Eng et al., 1992). Studies of
the effects of co-administration of isoflurane suggested an alternate explanation. In an elegant study
Lam et al. (1994) showed that reducing the dose of
isoflurane from 1.1 to 0.5 MAC and adding N2O
increased both CBF and CMR, whereas adding N2O to
1.1 MAC increased CBF but not CMR. They suggested
369
that the CBF changes following N2O may result

directly from vasodilatation and indirectly from a
decrease in CMR.
Co-administration of N2O and volatile agents is
clearly likely to increase cerebral blood volume and
ICP in patients with head injury and should be
avoided. Addition of N2O to an anesthetic based on
agents such as propofol or a benzodiazepine may
augment anesthesia with minimal increase in ICP, but
it should also be used cautiously in the head-injured
patient. As previously discussed, the ability of these
drugs to induce cerebral vasoconstriction will depend
on the status of autoregulation and Matta and Lam
(1995) have recently shown increases in CBF when
N2O is added to propofol anesthesia, which may be a
result of increases in CMR.
Nitrous oxide is much more soluble than nitrogen in
blood and will rapidly diffuse into gas-filled spaces.
Consequently, when N2O is administered during
anesthesia, there will be expansion of compliant
nitrogen-containing gas-filled spaces (such as the
bowel or a pneumothorax) or an increase in the
pressure within non-compliant spaces (such as the
middle ear or a pneumocephalus). Therefore, in the
head-injured patient with a pneumothorax or pneumocephalus, N2O should be avoided.
Currently there is divided opinion on whether N2O
has any place in neuroanesthesia. Although it should
not be used indiscriminately, it should also not be
forgotten that N2O has been used successfully in
neuroanesthesia for many years. It can be administered safely and, as with any other drug, its risks
and benefits should be carefully considered when
planning an anesthetic.
18.5.2
VOLATILE AGENTS
Volatile anesthetic agents remain the mainstay for

maintaining anesthesia despite a recent trend towards
continuous administration of intravenous agents.
Anesthetists are experienced with the dose regimens
of volatile agents appropriate to achieve required
depths of anesthesia and with the cardiovascular
changes that are likely to occur. The equipment for
their delivery is both readily available and familiar.
However, their place in neuroanesthesia has been
debated since halothane was first implicated in increases in ICP.
(a)
Halothane
Halothane, used in neuroanesthesia for many years, is

now rarely seen in the neurosurgery theater. It is the
volatile anesthetic against which the effects of newer
agents on intracranial dynamics are often compared;
therefore its neuropharmacology remains relevant.
370
Halothane uncouples CBF and CMR in a dosedependent manner, producing cerebral vasodilation
and a simultaneous fall in CMR. This relative CBF
excess (hyperemia) can readily be demonstrated by
measuring the increase in cerebral venous oxygen
content under halothane anesthesia.
At normocarbia the increase in CBV consistently
results in increased ICP with the potential to critically
reduce cerebral perfusion (McDowall, 1965, 1967).
These ICP effects can be prevented, or at least
ameliorated, by hyperventilation as the CBF response
to changes in arterial CO2 tension (CO2 reactivity) is not
abolished by halothane. However in order to prevent
an increase in CBF, PCO2 must be lowered prior to
halothane administration (Adams et al., 1972).
It is important also to recognize that both halothane
administration, particularly at doses greater than 1
MAC, and head injury itself can impair cerebral
autoregulation so that CBF is more directly dependent
on perfusion pressure (Miletich, Ivankovich and
Albrecht, 1976; Figure 18.2). Even minor degrees of
hypotension may then reduce CBF, while hypertension can result in increased CBF and ICP.
(b) Enflurane
Enflurane was the successor to halothane in general
anesthetic practice because of a more rapid onset and
offset of action and a lower risk of volatile-anestheticinduced hepatitis, a very rare but well publicized
condition. It is epileptogenic, however, and this has
led to a decline in its use in neuroanesthesia.
Its effects on intracranial dynamics represent some
improvement over halothane. CBF and CMR are still
uncoupled but the increases in CBF are probably

lower than with halothane at equivalent depths of
anesthesia (Drummond and Shapiro, 1981). Carbon
dioxide reactivity is well maintained; hence the CBF
increases can be minimized with hyperventilation. As
with halothane, at high doses autoregulation becomes
impaired and CBF becomes more pressure-dependent
(Miletich, Ivankovich and Albrecht, 1976).
(c)
Isoflurane
Isoflurane is generally considered to be the volatile

anesthetic agent of choice for neuroanesthesia. Of all
volatile agents it causes the greatest dose-dependent
fall in CMR (Todd, Drummond and Shapiro, 1982).
Complete suppression of neuronal electrical activity,
as demonstrated by an isoelectric EEG, is achievable
at the clinically useful concentrations of 23 MAC
(Newberg, Milde and Michenfelder, 1983). Like all
volatile anesthetics, isoflurane is a cerebral vasodilator. However CBF and ICP increases are less marked
with isoflurane than with other volatile agents at
equivalent depths of anesthesia one reason for its
popularity in neuroanesthesia.
CO2 reactivity is well preserved and CBF increases
can be prevented by hyperventilation (Adams et al.,
1981). Simultaneous hyperventilation is sufficient to
prevent increases in CBF and can therefore be used to
control ICP. Furthermore, autoregulation is reasonably
well preserved (McPherson and Traystman, 1987, 1988;
Adams et al., 1981). Isoflurane may be neuroprotective,
(a particular advantage over other volatile agents) and
flows of only 10 ml/100 g/min may be sufficient to
prevent ischemia under isoflurane anesthesia (Messick
Figure 18.2 The relationship between blood pressure and CBF with autoregulation intact (solid line) and impaired
(broken line).
OTHER AGENTS USED IN NEUROANESTHESIA
et al., 1987). Increased seizure activity does not occur

with isoflurane and indeed it may be used successfully
to treat recurrent seizures. Even marked hypocarbia,
which will potentiate seizures, does not induce seizures
during isoflurane administration.
(d)
Desflurane
Desflurane is a new volatile anesthetic agent of

particular interest because of its physicochemical
properties. It has very low partition coefficients, which
results in rapid induction and recovery from anesthesia (Jones, 1990). It also has a low boiling point
(23.5C) and requires expensive, specifically designed
heated and pressurized vaporizers (Graham, 1994).
It is structurally similar to isoflurane and it is not
surprising that it has similar effects on the cardiovascular and cerebrovascular systems. With increased
inspired concentrations it produces minimal myocardial depression and a fall in peripheral resistance
and blood pressure. Increases in CBF are similar to
those seen with isoflurane (Ornstein et al., 1993) and
ICP may increase in the presence of an intracranial
mass lesion (Muzzi et al., 1992). CO2 reactivity appears
to be well maintained (Lutz, Milde and Milde, 1991)
hence increases in CBF and ICP may be avoided by
hyperventilation.
Desflurane-induced hypotension can reduce CBF
(Milde and Milde, 1991) but this can be avoided by
pressor support. No EEG evidence of seizure activity
has been detected at therapeutic concentrations of
desflurane.
If inhaled anesthetic agents are to be used in
neuroanesthesia, desflurane would appear to have
many advantages. Intracranial dynamic changes are
similar to those of isoflurane but its physicochemical
properties allow more rapid control of depth of
anesthesia and more rapid recovery. However it is a
relatively new agent, and further clinical experience
and laboratory research will be valuable.
(e)
Sevoflurane
Sevoflurane is another new agent with physicochemical properties that, like desflurane, produce very
rapid onset and offset. Cardiovascular effects are
similar to those of isoflurane, producing a fall in
peripheral resistance and hypotension but with less
tachycardic response. Administration of either agent
in concentrations up to 1.52 MAC produces similar
changes in CBF and CMR (Scheller et al., 1990).
Preservation of CBF under sevoflurane anesthesia
during drug-induced hypotension has been demonstrated, suggesting that autoregulation remains intact
(Kitaguchi et al., 1992, 1993). CBF CO2 response is also
preserved (Kitaguchi et al., 1993). In a study of
371
hyperventilated dogs, increase in ICP was seen with

halothane and enflurane anesthesia but not with
sevoflurane. Sevoflurane, however, caused significant
hypotension (Takahashi, Murata and Ikeda, 1993).
There is no evidence of seizure activity on EEG at
therapeutic concentrations (Scheller et al., 1990).
There appear to date to be few differences in the
intracranial dynamic effects of sevoflurane, desflurane
and isoflurane but experience with the two newest
agents is still limited.
18.6 Other agents used in

neuroanesthesia
18.6.1
MANNITOL
Mannitol is an osmotically active agent which has

been used for many years as a diuretic and is now also
now well established for the treatment of ICP (Chapter
19). For many years it was assumed that it acted
primarily by increasing the osmotic gradient between
blood and brain, producing a net movement of water
out of the brain and a reduction in intracranial brain
volume and pressure, an action that may be compromised by injury-induced bloodbrain barrier defects
(Shapira et al., 1993). Recent evidence suggests that the
initial rapid fall in ICP may be due to plasma
expansion, with reduced blood viscosity. CBF increases and there is a compensatory vasoconstriction,
which reduces blood volume. The more delayed and
sustained fall in ICP may be due to the osmotic effects.
Both of these actions are greater with bolus administration than with continuous infusion.
However, bolus administration can produce transient
systemic vasodilatation and hypotension, particularly
in a hypovolemic patient, with an acute fall in CPP
(Cote, Greenhow and Marshall, 1979; Nissenson,
Weston and Kleeman, 1979), and plasma expansion
may precipitate cardiac failure in the patient with
compromised cardiac performance.
Dosages from 0.252 g/kg of mannitol have been
advocated. However the higher doses do not necessarily provide better ICP control and the side effects
are increased (Smith et al., 1986).
In the patient with a severe head injury and raised
ICP, 0.250.5 g/kg given over 1020 minutes will
usually reduce ICP and provide improved operating
conditions. This dose can be repeated 24-hourly as
necessary with regular monitoring of serum
osmolarity.
Co-administration of frusemide (furosemide) can
decrease the onset time and increase the duration of
ICP effects of mannitol (Wilkinson and Rosenfeld,
1983) but there is increased risk of hypovolemia and
this combination should be avoided in patients with
head injury.
372
18.6.2
CARDIOVASCULAR AGENTS
Numerous drugs are available for control of blood

pressure and heart rate peroperatively. The merits of
some of the more commonly used drugs will be
described briefly.
Glyceryl trinitrate (GTN) is a short-acting vasodilator that can be given by infusion to control hypertension intraoperatively. It is not an ideal agent for use
in the head-injured patient as it is also a cerebral
vasodilator. A recent study using SPECT scanning and
transcranial Doppler found that sublingual GTN
produced a fall in cerebral blood velocity without a
change in CBF (Dahl et al., 1989). Not surprisingly, this
cerebral vasodilatation with a resultant increase in
cerebral blood volume has been shown to increase ICP
(Hartmann et al., 1989; Lagerkranser, 1992). GTN
administered for systemic hypertension in the presence of decreased intracranial compliance risks further increases in ICP. Hyperventilation may be less
effective in reducing ICP than expected in these
circumstances because GTN-induced vasodilation
will reduce CO2 reactivity (Hartmann et al., 1989).
Sodium nitroprusside (SNP) has a very rapid onset
and brief duration of action, which allows very precise
blood pressure control. Hamaguchi et al. (1992) found
no change in CBF during SNP-induced hypotension
under enfluraneN2O anesthesia. This suggested a
cerebral vasodilatory action but could not differentiate
between direct SNP-induced vasodilatation and
impairment of normal autoregulation. It has been
associated with ICP increases in clinical practice,
presumably due to increase in cerebral blood volume.
SNP produced greater increases in ICP than GTN in a

baboon model at both normal and raised levels of ICP
and reduced CO2 (Hartmann et al., 1989).
Hydralazine is a direct-acting vasodilator which
acts directly on vascular smooth muscle producing
delayed onset of action; the peak hypotensive effects
occurring 1520 minutes after intravenous administration. As with GTN and SNP, it can increase ICP by
increasing cerebral blood volume (Herpin, 1989). This
cerebral vasodilatation may be prolonged and appears
to be a direct effect as it can occur in the absence of
significant change in blood pressure (Figure 18.3).
Because of its prolonged action, hydralazine should be
avoided in the presence of raised ICP.
Beta-blockers may be the antihypertensives of
choice for the treatment of perioperative hypertension
in the presence of raised ICP, since they lower blood
pressure by reducing cardiac output without reducing
systemic or cerebrovascular resistance or increasing
cerebral blood volume.
Esmolol is a relatively new selective beta-antagonist
which has an extremely short half-life (810 min) due
to rapid hydrolysis in blood. This allows rapid and
readily reversible changes of blood pressure and it is
therefore particularly useful for the control of brief
periods of hypotension, such as those occurring at
induction. It appears to have little effect on CBF when
autoregulation is intact (Bunegin, Albin and Gelineau,
1987), but esmolol-induced hypotension may decrease
CBF if autoregulation is impaired.
Inotropic or vasoconstrictor drugs under physiological circumstances can significantly increase blood
pressure without altering CBF. Adrenalin (epineph-
Figure 18.3 CBF in a sheep following low-dose bolus hydralazine administration, producing no changes in systemic
arterial pressure (Source: reproduced from Ludbrook et al., 1995, with permission.)
CONDUCT OF ANESTHESIA IN SEVERE HEAD INJURY
rine) has minimal effects on CBF in the normal brain

(Olesen, 1972; King, Sokoloff and Wechsler, 1952).
However, if the bloodbrain barrier is disrupted, as
may be the case after severe head injury, CBF may be
significantly altered (Abdul-Rahman et al., 1979). The
reported success of therapeutic strategies that aim to
maintain CPP with drugs such as phenylephrine,
metaraminol and noradrenalin (norepinephrine) suggests that they have a place in management of the
head-injured patient (Rosner et al., 1994; Chapter 19).
18.7 Conduct of anesthesia in severe

head injury
This section will concentrate on aspects of anesthetic
practice that relate to the patient with an severe acute
head injury, focusing particularly on preventing secondary injury.
18.7.1
PATIENT ASSESSMENT AND PREPARATION
Preoperative assessment is often brief because of the

urgent nature of this condition, and management will
usually have begun with the acute resuscitation team
outside hospital or in the emergency room. This
period of management is described in Chapter 15.
The anesthetist must often rely on limited information about the premorbid medical and anesthetic
history of the patient. Relatives are often the most
valuable source of this information. Details of the
nature of the trauma involved (such as speed and
direction of impact in the case of a motor vehicle
accident) should be sought from the retrieving medical or paramedical staff, and detailed records of all
treatment regimens and investigations performed
should be available to, and studied by, the anesthetist
at the outset.
Although a detailed examination should have been
performed by other medical staff, the anesthetist must
undertake a basic assessment of cardiorespiratory
systems at least, with particular emphasis on traumatic injuries. The significance to the anesthetist of
such basic findings as broken teeth may not be
appreciated by non-anesthetic personnel.
Depending on the stage at which the anesthetist is
involved in patient management, extensive investigations including radiological surveys may have already
been performed. It is advisable for the anesthetist to
check these results personally. The presence of pneumocephalus, broken ribs or a small untreated pneumothorax may not be emphasized by other attending
staff and the experience of medical personnel reporting on investigations such as cervical spine X-rays is
not always known.
Acute resuscitation may have commenced (Chapter
15). Adequate intravenous access should be ensured,
373
especially if blood loss from other injuries or for

surgery is anticipated, and the anesthetist should
check that blood is readily available for transfusion.
Adequate patient monitoring should be begin preoperatively. An arterial line should be sited, central
venous access obtained for monitoring as well as drug
administration, and a urinary catheter inserted.
18.7.2
INDUCTION OF ANESTHESIA
The aim of induction of anesthesia is to achieve an

adequate depth of anesthesia and relaxation in order
to rapidly and safely control airway and ventilation
without patient awareness and to simultaneously
improve intracranial dynamics. All this must be
achieved in a patient who may have a full stomach, be
hypovolemic and have other injuries or medical
problems. Many of these factors place conflicting
demands on the anesthetist for example, whether to
use a rapid large dose of sodium thiopentone in order
to reduce the ICP response to intubation but risk
hypotension. A diversity of techniques are used,
according to the clinical setting and the preference of
the anesthetist (Nakajama et al., 1992).
The advantage of having an anesthetic plan, prepared equipment, adequate assistance, and a second
anesthetist if particular difficulty is anticipated,
cannot be stressed enough.
If aspiration of gastric contents is a significant risk,
a rapid-sequence induction including the use of
suxamethonium is necessary. A priming dose of a
non-depolarizing relaxant such as vecuronium may be
considered to offset the ICP effects of suxamethonium
(Stirt et al., 1987) but is probably not necessary.
If hypovolemia or concomitant cardiovascular disease are not major concerns then an induction dose
of sodium thiopentone (36 mg/kg) or propofol
(23 mg/kg) will provide rapid onset of anesthesia
while at the same time reducing CBF and ICP.
However, depending on the dose regimen, these
agents alone may be insufficient to ablate the cardiovascular responses to laryngoscopy and intubation
(Giffen, Cotrell and Schwirey, 1984) and so adjunctive drugs may be necessary. Lignocaine (11.5 mg/
kg) given approximately 3 minutes prior to intubation may cause sufficient cerebral vasoconstriction to
further reduce the ICP response to intubation (Bedford et al., 1980). Fentanyl (5 g/kg) may be considered but to be effective should be given at least
35 minutes prior to intubation and unexpected loss
of airway protection and muscle rigidity which
compromises ventilation may occur in that time
period. Doses of all drugs should take into account
the preanesthetic conscious state, which may have
been affected by the head injury or both therapeutic
and recreational drugs.
374
If hypovolemia/hypotension and aspiration risk

coexist, the challenge to the anesthetist is great.
Hypovolemia should be corrected before, or if surgery
is urgent, during induction by fluid replacement
guided by central pressure monitoring. Etomidate
(0.10.5 mg/kg) may provide cardiovascular support
without reducing cerebral perfusion but can still
precipitate hypotension, and furthermore is not available in a number of countries. The combination of an
opioid such as fentanyl with small doses of midazolam
or sodium thiopentone may avoid profound hypotension. Ketamine can be a useful induction agent in the
presence of hypovolemia, but it cannot yet be recommended in a patient with a severe head injury because
marked increases in ICP have been recorded following
intravenous bolus administration.
The anesthetist must be prepared to treat hypotension rapidly and aggressively as episodes of hypotension following a severe head injury are associated
with a worse neurological outcome. Adequate intravenous access to allow rapid infusion of fluid should
be obtained, and vasopressors such as metaraminol
should be drawn up ready for use. Simple maneuvers
such as leg elevation (rather than the Trendelenburg
position, which may do little to improve CPP) should
not be forgotten in treating anesthesia-related
hypotension.
Oral endotracheal intubation is usually preferred,
especially as it allows the airway to be most rapidly
secured with less trauma than nasal intubation. Nasal
intubation, preferred in some centers if long-term
ventilation is anticipated, should be avoided in the
presence of a basal skull fracture or major maxillary
trauma as passage of the endotracheal tube out of the
airway and even intracranially is possible. The postoperative phase should be considered when selecting
the endotracheal tube. Depending on the surgical
approach and patient positioning, armored or preformed tubes may be most convenient intraoperatively
but if postoperative ventilation is planned a straight
plastic tube with a high-volume low-pressure cuff
should be substituted after completion of surgery.
Cervical spine injury occurs in 1020% of patients
with severe head injury (Hills and Deanse, 1993;
Heiden et al., 1987). There are differing opinions about
the optimal intubation technique for these patients.
However it is generally considered that oral intubation with head immobilization (rather than in-line
traction) is a safe technique and does not compromise
the spinal cord (Wood and Lawler, 1992; Suderman,
Crosby and Lui, 1992).
As for any surgery, particularly when it may be
prolonged, careful attention should be paid to patient
positioning. This should be discussed with the surgeon preoperatively, so that the position of the
anesthetic machine, endotracheal tube selection and
the position of monitoring lines and equipment can be

planned in advance. Particular care should be taken
with positioning of the patient before draping and the
surgeons prevent access. A head-up posture (1520)
has been shown to improve ICP and not adversely
affect CPP or CBF, and is therefore recommended
when practical (Durward et al., 1983; Wilson, 1992;
Feldman et al., 1992). There must be no obstruction to
venous drainage from the head by endotracheal tube
tapes or other equipment.
18.7.3
MAINTENANCE OF ANESTHESIA
There are some aspects of neuroanesthesia for the

severely head-injured patient that warrant specific
mention.
(a)
Anesthetic agents
The complex and variable pathophysiology of brain

and systemic injuries makes a simple anesthetic
recipe impossible. For example, head injury can be
associated with either high or low CBF and theoretically no one anesthetic technique is ideal for both
situations. While sophisticated techniques to differentiate between the two states exist, this is rarely
possible preoperatively, even in the few centers
capable of such measurements.
There is in fact little outcome-based evidence that
any particular anesthetic technique is superior to any
other in neuroanesthesia in general. This is in part a
result of the inherent difficulties of outcome-based
studies in anesthesia (Eichhorn, 1993) as well as of the
difficulties in performing well controlled trials in
head-injured patients. Indeed, as pointed out earlier,
even halothane, a drug with substantial theoretical
disadvantages for neuroanesthesia, has not been
proved to detrimentally affect patient outcome. The
choice of i.v. anesthetic technique will depend largely
on the personal experience of the anesthetist and an
assessment of the particular case. For example, a total
intravenous technique using propofol to reduce CBF
may be an excellent choice for a patient who is
normotensive or hypertensive and may have cerebral
hyperemia, while an opioidbenzodiazepine technique may be preferred when postoperative ventilation is anticipated. Because of the risk of adversely
affecting CPP, cerebral vasodilators such as the volatile agents should avoided.
(b)
CVS homeostasis
Maintaining adequate and sustained oxygen delivery

is fundamental to avoiding secondary hypoxic injury,
and cardiovascular homeostasis is a vital step in the
oxygen delivery chain.
Hypotension is detrimental whether or not autoregulation is compromised. In the presence of impaired

cerebral autoregulation, any degree of hypotension
will produce a decrease in CBF and, although ICP may
also decrease, cerebral ischemia may result.
This has been demonstrated in animal models
where either hypotension or acute head injury alone
did not impair CBF, but CBF fell significantly when
both conditions occurred simultaneously (DeWitt et
al., 1992). As described by Bouma et al. (1992) cerebral
ischemia may still occur following hypotension and
head injury if autoregulation is preserved. Cerebrovascular resistance will fall in response to a fall in
MAP, producing an increase in cerebral blood volume,
a rise in ICP and a net fall in CPP. Hypotension has
been associated with an increased morbidity and
mortality (Pietropaoli, Rogers and Zhuang, 1992).
Hypertension may or may not be detrimental. It may
increase CBF and CPP and reduce ischemia due, for
example, to cerebral vasospasm following traumatic
subarachnoid hemorrhage (Martin et al., 1992; Lang
and Chesnut, 1994). However, in the presence of
impaired autoregulation and hyperemia it may
increase CBF, ICP and cerebral edema and thus
ultimately increase cerebral ischemia.
It is apparent therefore that both hypo- and hyper
perfusion must be avoided, but the difficulty lies in
deciding what level of MAP is ideal.
Because of the risk of increasing ICP with fluid
overload (Hariri et al., 1993), central pressure monitoring is mandatory to optimize intravascular volume. CPP should not fall below 6070 mmHg and
pharmacological support should be used to maintain
this as necessary (Bendo, 1994).
Oxygen-carrying capacity is another vital link in
avoiding secondary hypoxic injury. Hematocrit
should be measured regularly and kept between 30%
and 35%, which probably gives the optimal balance
between reduced blood viscosity and adequate oxygen-carrying capacity.
(c)
Monitoring
Observation of the patient and equipment by an alert,

skilled anesthetist once constituted an acceptable level
of monitoring; however it is now generally recognized
that correctly used specialized equipment provides
more accurate measurements and an increased margin
of safety.
Pulse oximetry
Pulse oximetry has been the most important advance
in monitoring during anesthesia since the introduction of the sphygmomanometer. Not only can its
waveform provide an index of peripheral perfusion
375
but it also continuously displays the saturation of the

hemoglobin in arterial blood, thus providing valuable
information about the cardiovascular and respiratory
systems
Clinical acumen is notoriously unreliable in detecting cyanosis (Hanning, 1985) and pulse oximeters are
now mandatory in anesthetic practice in many centers.
They are reasonably accurate under most clinical
circumstances (Webb, Ralston and Runciman, 1991);
however, reduced tissue hypoperfusion in hypotensive, hypothermic or vasoconstricted patients
situations not uncommon in severe head injury may
give falsely low readings. If there is doubt about the
accuracy of the pulse oximeter reading, measurement
of arterial blood gas tensions is essential.
Capnography
Capnography to monitor endotracheal tube placement
and ventilation is essential. It also provides a guide to
changes in cardiac output. It must be recognized that
large alveolararterial gradients in carbon dioxide
tension can occur, especially in patients with traumatic
injuries, and end-expired carbon dioxide tensions
should be calibrated against blood gas measurements.
Electrocardiography
The electrocardiogram (ECG) may be used to measure
heart rate, detect and characterize arrhythmias and
conduction defects, monitor pacemaker function and
provide some indication of myocardial ischemia.
There are specific indications for its use intraoperatively (Ludbrook et al., 1993) but it is an essential
monitor for the head-injured patient because of the
increased risk of dysrhythmias and the high incidence
of ischemic changes in these patients (Kaufman et al.,
1993; section 17.3.3(b)).
Blood pressure measurement
Blood pressure measurement is essential for determining CPP and must be continuously and accurately
recorded during anesthesia. Indirect blood-pressure
measuring devices such as blood pressure cuffs or
oscillometric devices can be inaccurate by up to 30%
(Raftery and Ward, 1968; Bruner et al., 1981) and the
degree of error may change with the patients cardiovascular status (Rutten et al., 1986). This is particularly relevant in the head-injured patient, where
vascular tone and the accuracy of blood pressure
measurement may change dramatically during an
operation (Cohn, 1967). It is therefore essential that an
arterial line be placed preoperatively and connected to
a cannulatappressure-tubetaptransducer system,
which will provide reasonable accuracy.
376
Several systems in clinical use have surprisingly

large errors (Gardner, 1981). In practice, a fluid-filled
system comprising a 4 cm 20 gauge, non-tapered
arterial cannula connected to a low-displacement
pressure transducer by non-compliant pressure tubing
and a three-way stopcock is sufficiently accurate to
gauge mean and systolic pressures. Care must be taken
to exclude all air bubbles and the amplifierrecorder
system must have an adequate frequency response.
Central venous pressure measurement
This is a high priority. A catheter can usually be
inserted quickly and safely to the junction of the
superior vena cava and right atrium via the basilic
vein from the cubital fossa. If time permits, however, a
subclavian or internal jugular catheter inserted using a
Seldinger technique is preferable, being more suitable
for long-term use in intensive care. Pneumothorax
following subclavian line insertion is uncommon in
experienced hands (Eerola, Kaukinen and Kaukinen,
1985); nonetheless a chest X-ray should be taken prior
to anesthesia. Monitoring pulmonary capillary wedge
pressure and cardiac output may require a Swan
Ganz-type catheter. This is covered more extensively
in Chapter 17.
Cerebral perfusion measurement
More specialized monitoring of intracranial perfusion
developed for intensive care use has little application
in the operating room at present.
Transcranial Doppler can be used to differentiate
between low- and high-flow states and autoregulatory
break points (Chapter 14; Chan et al., 1992; Compton
and Teddy, 1987; Mayberg and Lam, 1992) but this is
frequently impractical intraoperatively.
Jugular bulb catheters incorporating oximetry to
measure the saturation of cerebral effluent blood has
proven useful in Intensive Care management (Sheinberg et al., 1992; Robertson et al., 1992; Matta, Lam and
Mayberg, 1994) and there is now some experience of
their use intraoperatively (Matta et al., 1994).
Near-infrared spectroscopy has the potential to
provide a continuous measurement of cerebral oxygenation intraoperatively but this promising technique requires further refinement before it can be
accepted as a reliable monitor of brain mitochondrial
oxygenation (Hazeki and Tamura, 1988; Brown, 1993;
Lewis et al., 1994).
Temperature
Core body temperature is usually tightly regulated by
peripheral and central thermosensors, a central control system in the hypothalamus and effector mecha-
nisms operating largely via the sympathetic nervous

system. The central control mechanisms can be disrupted through cerebral injury, or by anesthetic or
vasoactive drugs (Morley-Forster, 1986). Severe hypothermia has been intentionally induced in neurosurgery when cerebral perfusion is likely to be compromised or even deliberately arrested for a short period
(Silverberg, Reitz and Ream, 1981). The adverse
systemic effects of profound hypothermia have largely
limited its use in neurosurgery to operations such as
clipping of giant cerebral aneurysms under circulatory
standstill.
Mild hypothermia is almost inevitable during most
surgery and considerable effort can be expended in
trying to minimize this by using warmed intravenous
fluids, warming blankets and temperature monitoring. Many of the adverse effects of severe hypothermia
such as cardiac dysrhythmias, increased blood viscosity and coagulopathy do not occur with mild
hypothermia, but there is a marked increase in oxygen
consumption during rewarming.
There is renewed interest in the intentional use of
mild hypothermia (3335C) during neurosurgical
operations and in the intensive care management of
patients with head injury (Baker et al., 1994). The main
advantage of hypothermia was initially considered to
be the reduction in CBF and CMR (a reasonable rule of
thumb equates each 1C fall in temperature with an
8% fall in CMR and CBF), an effect similar to that
following administration of barbiturates and
propofol.
It is now apparent that even small reductions in
brain temperature provide a greater tolerance to
ischemia than can be explained by the predicted falls
in CMR (Busto, Dietrich and Globus, 1987; Kader et
al., 1992). A number of mechanisms such as reduced
levels of nitric oxide or excitatory amino acids, or
effects on ion transfer and lipid peroxidation, have
been proposed.
Recent clinical trials of moderate hypothermia have
demonstrated apparent improvement in the outcome
following head injury. Clifton and coworkers (1993)
found a significant reduction in seizure rate and some
improvement in cerebral function during recovery
when all patients with severe head injuries were
treated with mild hypothermia. In another study,
patients with large rises in ICP that did not respond
adequately to dehydration, hyperventilation or intravenous barbiturates were treated with mild hypothermia. There were significant increases in CPP and
reductions in ICP, CBF and CMR (Shiozaki et al., 1993).
Marion and coworkers (1993) also found reduction in
CBF, CMR and ICP in head-injured patients after
induction of mild hypothermia, but this reversed as
patients were rewarmed. This form of therapy does
not appear to significantly increase the incidence of
dysrhythmias and hypokalemia (Clifton and Christensen, 1992). Hypothermia (3335C) may therefore
be advantageous in the treatment of the severely headinjured patient.
A major clinical trial is currently being conducted to
test the ability of moderate hypothermia to improve
outcome after severe head injury.
Hypothermia can often be relatively easily achieved
as patients are frequently cold on arrival to hospital as
a result of such factors as exposure, fluid therapy and
alcohol (Luna et al., 1987), and exposure to the cold
theater environment and heat redistribution commonly leads to decreases in core body temperature
between induction of anesthesia and beginning of
surgery. Temperature can be reduced further by the
use of cooling/heating blankets (air-circulating types
are particularly effective) and cooled intravenous
fluids. Lower esophageal probes give a reasonable
estimation of core body temperature with a rapid
response time. External cooling measures carry the
risk of overshoot during the initiation of hypothermia, and therefore active cooling should be discontinued before the target temperature is reached.
Shivering is an early response to hypothermia and
can increase BP, ICP, respiratory effort and oxygen
consumption; it should be avoided in severe head
injury. Therefore, if hypothermia is induced intraoperatively, muscle relaxation should be used and
temperature corrected before relaxation is reversed.
Muscular relaxation may be continued postoperatively if ongoing artificial ventilation is planned.
While hypothermia may be beneficial, hyperthermia
after trauma induced by loss of cerebral temperature
autoregulatory mechanisms, drugs or infection may
worsen outcome. Hyperthermia has been shown
experimentally to increase neurological damage following cerebral ischemia. Hyperthermia should therefore be treated aggressively.
18.7.4
FLUID ADMINISTRATION
Concepts of fluid administration in the patient with

head injury have undergone significant change in
recent years. In particular, the greater appreciation of
the dangers of hypotension and cerebral hypoperfusion following head injuries have led to an emphasis
on maintaining normovolemia. Care must still be
taken to avoid fluid overload by aggressive fluid
therapy, since this may increase ICP and reduce
cerebral perfusion (Hariri et al., 1993).
Where possible, fluid resuscitation should be instituted before or during induction of anesthesia. The
delicate balance between the risks of under- and overresuscitation described above make monitoring of
central pressures, or pulmonary artery pressures, a
high priority.
377
The ideal fluids for resuscitation remains a subject

of debate but there are some reasonable rules of
thumb.
In general, hypotonic solutions should be avoided
since they will reduce plasma osmotic pressure and
this may worsen cerebral edema (Kaieda, Todd and
Cook, 1989; Shackford, Zhuang and Schmoker, 1992).
Glucose-containing solutions should be used cautiously. Hyperglycemia has been linked to poor
neurological outcome after traumatic and ischemic
brain injury (Lam et al., 1991; Araki et al., 1992) and
may be associated with a worsening of intracerebral
acidosis. Although less common, hypoglycemia may
compound cerebral damage. Hence if glucose-containing solutions are used, for example in the diabetic,
hypothermic or intoxicated patient, serum glucose
levels should be monitored closely.
The debate over colloids versus crystalloids for fluid
resuscitation continues in most fields of anesthesia,
including neuroanesthesia and in the management of
head injury. However there appears to be little
evidence that isotonic crystalloid solutions are harmful following head injury
Hypertonic saline solutions have recently been used
for the resuscitation of hypovolemic patients since
relatively smaller infusion volumes are required to
achieve blood volume expansion. Studies have shown
them to successfully increase CPP and perhaps reduce
cerebral edema (Fisher, Thomas and Peterson, 1992;
Prough et al., 1985; Battistella and Wisner, 1991;
Shackford, Zhuang and Schmoker, 1992).
Some reservations remain about the possibility of
adverse affects of large and prolonged rises in serum
osmolarity, particularly renal failure, and further
studies are required before they can be recommended
for routine use in head injury.
18.7.5
VENTILATION
It is not surprising that hyperventilation during

neuroanesthesia became widespread considering the
marked reductions and ICP and brain tension noted
intraoperatively with hypocarbia. It is now recognized
that persistent hypocarbia may have significant side
effects and there is a move towards using normocarbia
or only moderate hypocarbia in neuroanesthesia
(Chapters 17 and 19).
In the head-injured patient, hyperventilation may
be relatively ineffective as CO2 reactivity may be
impaired (Bouma et al., 1992). Indeed for this reason
hyperventilation may be the least effective in controlling raised ICP in the most severe injuries. Secondly,
hyperventilation may induce cerebral ischemia. Cerebral injury is inhomogeneous and, in some areas of the
injured brain, critical falls in CBF may be induced by
hyperventilation (Cold, 1989; Obrist et al., 1984).
378
Furthermore, hypocarbia may lower MAP and CPP,

particularly if there is concomitant hypovolemia, and
it may induce dysrhythmias. Hyperventilation has
been associated with increased morbidity after head
injury (Muizelaar et al., 1991).
Thus PaCO2 during anesthesia should be maintained
at 35 2 mmHg and monitored with intermittent
arterial blood gas analysis.
Despite the evidence against profound and sustained hyperventilation, temporary hyperventilation
may be vitally important for the rapid control of raised
ICP. If CPP is compromised by raised ICP, brief
periods of hyperventilation may allow time for other
measures such as surgical evacuation of a hematoma,
osmotic or barbiturate therapy, to be instituted.
Hyperventilation is only a temporary measure, however, as resetting of CBF occurs within hours. CSF pH
returns towards normal and rebound increases in CBF
and ICP may occur when normocarbia is resumed
(Muizelaar et al., 1988).
18.7.6
SEIZURE MANAGEMENT
Seizure management (Table 18.1) is an important aspect

in the care of the head-injured patient, and involves
treating and preventing seizures occurring early in the
course of the injury, and preventing late seizures. There
are recognized risk factors for the development of
seizures after head injury. Patient factors include age
(there being a higher incidence in the pediatric age
group), a history of alcoholism, previous seizures and a
family history of seizures (Yablon, 1993; Annegers,
1980). Injury factors include injury severity patients
with minor head injuries (PTA < 24 h) have a very low
seizure rate and thus the occurrence of seizures in this
group may signify a more serious cerebral injury than
was originally believed (Feuerman et al., 1988; Lee and
Lui, 1992), and the pattern of injury. Patterns associated
with an increased risk of seizures include intracranial
Table 18.1
hemorrhage, linear and depressed skull fractures, focal

or penetrating injuries and the presence of focal
neurological deficit or prolonged amnesia (i.e. PTA >
24 h; Jennett, 1975). The presence of intracerebral blood
or iron may be a trigger for late seizures (Lee and Lui,
1992).
Anticonvulsants are administered during the acute
phase of head injury management to prevent or treat
early post-traumatic seizures or to prevent the development of post-traumatic epilepsy. Seizures occurring
during the acute phase can increase ICP and interfere
with ventilation and airway control and the management of other injuries. Indeed, there is a positive
association between the occurrence of seizures and
morbidity and mortality, although whether this simply reflects the presence of a more severe injury or is a
result of the seizures is unclear (Yablon, 1993).
That acute seizures should be treated immediately is
clear. Prophylactic therapy has been more controversial and its effectiveness has been debated (North et
al., 1983; Temkin et al., 1990; Hauser, 1990; Young et al.,
1983). However, it appears likely that anticonvulsant
administration, carefully monitored to ensure therapeutic blood levels, is effective in reducing the
incidence of seizures occurring early after head injury.
They are most effective in the first week after an injury
and less so in preventing delayed seizures (Temkin et
al., 1990).
The rationale for using anticonvulsants to prevent
late seizures is based largely on the kindling model.
This assumes that recurrent stimulation by seizures
lowers the stimulus threshold for further seizures,
eventually reaching the stage where spontaneous
seizures may occur. Clinical trials, however, have not
shown a convincing benefit (Temkin et al., 1990; Young
et al., 1983) and the place of prophylactic therapy
beyond the first week is still unclear.
What guidelines should therefore be used for the
institution of anticonvulsant therapy?
Seizure management in acute head injury
Drug
Initial dose
Maintenance dose
Initial treatment of a seizure

Diazepam
Sodium thiopentone
Propofol
Clonazepam
0.050.15 mg/kg i.v.

14 mg/kg i.v.
0.52 mg/kg i.v.
0.0030.01 mg/kg i.v.
0.10.2 mg/kg i.v.

15 mg/kg/h i.v.
16 mg/kg/h i.v.
0.0020.004 mg/kg i.v.
Treatment of recurrent seizures

Phenytoin
15 mg/kg i.v. over 20 min
312 mg/kg/24 h (titrated to

therapeutic range of 1020 g/ml)
0.10.2 mg/kg/h i.v.
15 mg/kg/24 h (titrated to
therapeutic range of 45130 g/ml
0.0020.004 mg/kg i.v.
Diazepam
phenobarbitone
0.050.15 mg/kg i.v.
Clonazepam
0.0030.01 mg/kg i.v. or po
EMERGENCE AND RECOVERY

Any seizure occurring early in the course of the

injury should be rapidly treated, and therapy
should be continued because of the high risk of
recurrence.
Prophylactic anticonvulsant therapy should be
instituted to those patients considered at high risk
of seizures, but should be discontinued during the
second week after injury. Those at high risk include
patients with
intracranial hemorrhage;
linear and depressed skull fractures;
focal or penetrating injuries;
focal neurological deficit or prolonged amnesia
(post-traumatic amnesia > 24 h).
A seizure should not automatically be attributed to the

recent head injury. Other causes, which may be
treatable, should be excluded. A family or personal
history of epilepsy should be sought and reversible
conditions such as hypoglycemia, hyponatremia,
hypoxemia and extreme hypocarbia should be excluded. Alcohol, antidepressant medication and recreational drugs have all been reported to cause seizures
and a history of their use should be sought or a drug
screen performed.
In the ventilated and paralyzed patient seizure
activity may be disguised. Neuromuscular blockade
should therefore be avoided unless it is essential for
ICP control or because of severe lung pathology. In
these patients consideration should be given to prophylactic anticonvulsant therapy. Ideally, continuous
EEG monitoring should be instituted to detect seizures
and monitor the effectiveness of therapy.
(a)
Anticonvulsant therapy
Many of the hypnotic anesthetic agents in clinical use

have anticonvulsant actions. Intravenous bolus
administration of even small doses of sodium thiopentone or propofol provides rapid-onset, shortduration seizure control, but with the risks of loss of
airway reflexes, neurological and respiratory depression and hypotension. They should only be used if
personnel and equipment for airway management
and artificial ventilation are available. Because of the
cardiorespiratory depression, infusion administration
for prolonged anticonvulsant therapy is usually
restricted to the ventilated patient, where their sedative and cerebrovascular effects may be beneficial.
Phenobarbitone has a longer onset time and more
prolonged duration of action than thiopentone and is
therefore more appropriate for prolonged therapy.
Intravenous benzodiazepines (diazepam and midazolam) are also effective but have a variable dose
response relationship and a more prolonged duration
of action than the other sedative/hypnotic i.v. anesthetic agents.
379
Phenytoin, a diphenylhydantoin, is one of the most

commonly used anticonvulsants in head injury. It does
not cause significant CNS depression; it is effective
both intravenous and orally; and it has a long duration
of action. Its onset is considerably longer than that of
barbiturates and benzodiazepines; hence it is relatively ineffective in the immediate treatment of
seizures. Other agents should be used until therapeutic
levels have been achieved. Dilantin levels should be
monitored regularly to maintain concentrations in the
therapeutic range of 820 g/ml.
18.8
Emergence and recovery
A period of ventilation is often undertaken postoperatively and clear communication between anesthetist and intensive care staff is essential. Transportation of the critically ill patient must be undertaken
with great care. It requires continued adequate sedation and analgesia to prevent emergence during
transportation. Sedative infusions, such as propofol or
fentanyl/midazolam, administrated via a portable
syringe pump are one satisfactory solution. Ventilation and cardiovascular homeostasis must be continued, often with suboptimal monitoring.
If extubation is anticipated, there is a choice
between extubation under deep anesthesia or awake,
i.e. after reflexes have fully returned. In the non-fasted
patient the awake approach should be used to reduce
the risk of postoperative aspiration, even if attempts
have been made to empty the stomach with gastric
tubes intraoperatively. With this technique it is difficult to avoid coughing and postoperative ventilation
should be considered if ICP is still high. In the fasted
patient, extubation while still deeply anesthetized
reduces the risk of coughing or gagging, but is
inevitably associated with some ventilatory depression and potential ICP increase.
18.8.1
ANALGESIA
Head-injured patients frequently experience moderate

to severe pain due to associated soft tissue or bony
injury. This is a common cause of restlessness and
difficulty in managing and assessing. There may be
some reluctance to prescribe analgesic agents because
of the fear of drug-induced respiratory depression
which may increase ICP, and because sedation may
affect the neurological status.
If sedation, intubation and artificial ventilation are
necessary early, opioids are indicated for analgesia
and to assist in tolerating the endotracheal tube.
For the patient in whom control of ventilation or
airway protection is not considered necessary (either
pre- or postoperatively), but who has severe pain, the
decision is more difficult. The practice of prescribing
380
mild analgesia with non-opioid analgesics in the

presence of major injuries cannot be condoned. The
restlessness, straining and systemic hypertension
associated with severe pain increases ICP, makes
nursing more difficult and hazardous and causes
unnecessary suffering.
One solution is to administer short-acting opioids
such as fentanyl or alfentanil intravenously and to
titrate the dose according to analgesic effect. If there is
a significant deterioration in mental state due to
oversedation, the effect is likely to be short-lived or
can be readily reversed with opioid antagonists such
as naloxone. This may indicate a reduced intracranial
compliance and ventilation or the use of pharmacological measures to reduce ICP need to be considered.
If this technique is successful in producing adequate
analgesia without significant change in the neurological status or respiratory function, equivalent doses
of longer-acting opioids such as morphine, intravenously or subcutaneously, can then be considered.
Patient-controlled analgesia should be avoided in a
confused patient.
Codeine phosphate has often been used as an
alternative to drugs such as morphine but it should be
recognized that codeine is an opioid (and in fact it acts
mainly through metabolism to morphine) and will
produce as much respiratory depression for the same
degree of analgesia as any other opioid.
Regional anesthetic techniques in patients with
head injury should be used whenever possible. If used
appropriately, they have the advantages of excellent
analgesia with little risk of cerebral side effects. For
example, femoral nerve blockade can provide good
pain relief for a fractured femur without affecting
conscious state, and intercostal nerve or interpleural
blocks may aid in analgesia for fractured ribs
Non-steroidal anti-inflammatory drugs, have
gained popularity for perioperative analgesia, particularly now that parenteral preparations are available
(Gillies et al., 1987). For the patient with an acute head
injury, they have the advantages of minimal respiratory depression and sedation, but their side effects
include impairment of platelet aggregation and of
renal function and have tended to limit their use in
fields such as neuroanesthesia, when hemostasis is
critical (Merry, 1994; Walker, 1991).
18.9
References
Abdul-Rahman, A., Dahlgren, N., Johansson, B. B. and Siesjo,

B. K. (1979)
Increase in local cerebral blood flow induced by circulating adrenaline:
involvement of blood brain barrier dysfunction. Acta Physiologica Scandinavica, 107, 227.
Adams, R. W., Gronert, G. A., Sundt, T. M. and Michenfelder, J. D. (1972)
Halothane, hypocarbia and cerebrospinal fluid pressure in neurosurgery.
Adams, R. W, Cucchiara, R. F., Gronert, G. et al. (1981) Isoflurane and
cerebrospinal fluid pressure in neurosurgical patients. Anesthesiology, 54(2),
9799.
Ali, H. H. (1988) Clinical pharmacology of mivacurium chloride infusion.

Annegers, J. F. (1980) Seizures after head trauma: a population study.
Araki, N., Greenberg, J. H., Sladky, J. T. et al. (1992) The effect of
hyperglycemia in intracellular calcium in stroke. Journal of Cerebral Blood
Flow and Metabolism, 12, 469476.
Azari, D. M. and Cork, R. C. (1993) Comparative myocardial depressive effects
of propofol and thiopental. Anesthesia and Analgesia, 77(2), 324329.
Baker, K. Z., Young, W. L., Stone, J. G. et al. (1994) Deliberate mild
intraoperative hypothermia for craniotomy. Anesthesiology, 81, 361376.
Battistella, F. D. and Wisner, D. H. (1991) Combined haemorrhagic shock and
head injury: effects of hypertonic saline (7. 5%) resuscitation. Journal of
Trauma, 31, 182188.
Bedford, R. F., Persing, J. A., Pobereskin, L. and Butler, A. (1980) Lidocaine or
thiopental for rapid control of intracranial hypertension. Anesthesia and
Analgesia, 59, 435.
Bendo, A. A. (1994) Anesthetic management of the head injured patient. ASA
Annual Refresher Course Lectures, 272, 16.
Bevan, J. C. (1993) Propofol related convulsions. Canadian Journal of Anaesthetics, 40, 805809.
Bjorkman, S., Akeson, J., Nilsson, F. et al. (1992) Ketamine and midazolam
decrease cerebral blood flow and consequently their own rate of transport
to the brain: an application of mass balance pharmacokinetics with a
changing regional blood flow. Journal of Pharmacokinetics and Biopharmacy,
20(6), 637652.
Blaustein, M. P. and Ector, A. C. (1975) Barbiturate inhibition of calcium
uptake by depolarised nerve terminals. Molecular Pharmacology, 11,
369378.
Bouma, C. J., Muizelaar, J. P., Bandoh, K. and Marmarou, A. (1992) Blood
pressure and intracranial pressure-volume dynamics in severe head injury:
relationships with cerebral blood flow. Journal of Neurosurgery, 77, 1519.
Brown, R. W. (1993) Continuous monitoring of cerebral hemoglobin oxygen
saturation. International Anesthesiology Clinics, 31, 141158.
Bruner, J. M. R., Krenis, L. J., Kunsman, J. M. and Sherman, AP. (1981)
Comparison of direct and indirect methods of measuring arterial blood
pressure. Medical Instruments, 15, 182.
Bunegin, L., Albin, M. S. and Gelineau, M. F. (1987) Effect of esmolol on
cerebral blood flow during intracranial hypertension and hemorrhagic
hypovolaemia. Anesthesiology, 67, A424.
Busto, R., Dietrich, W. D. and Globus, M. Y. T. (1987) Differences in intra
ischemic brain temperature critically determine the extent of ischemic
neuronal injury. Journal of Cerebral Blood Flow and Metabolism, 7, 729.
Chan, K. H., Dearden, N. H., Miller, J. D. et al. (1992) Transcranial Doppler
waveform differences in hyperemic and nonhyperemic patients after severe
head injury. Surgical Neurology, 38, 433436.
Chiolero, R. L., Ravussin, P., Anderes, J. P. et al. (1986) The effects of
midazolam reversal by RO 151788 on cerebral perfusion pressure in
patients with severe head injury. Intensive Care Medicine, 14(3), 196200.
Church, J., Zeman, S. and Lodge, D. (1988) The neuroprotective action of
ketamine and MK 801 after transient cerebral ischaemia in rats. Anesthesiology, 69(5), A310.
Clifton, G. L. and Christensen, M. L. (1992) Use of moderate hypothermia
during elective craniotomy. Texas Medicine, 88(12), 6669.
Clifton, L., Allen, S., Barrodale, P. et al. (1993) A phase II study of moderate
hypothermia in severe brain injury. Journal of Neurotrauma, 10(3), 263271.
Cohn, J. N. (1967) Blood pressure measurement in shock. Journal of the
Cold, G. E. (1989) Does acute hyperventilation provoke cerebral oligaemia in
comatose patients after acute head injury. Acta Neurochirurgica (Vienna), 96,
100106.
Compton, J. S. and Teddy, P. J. (1987) Cerebral arterial vasospasm following
severe head injury: a transcranial Doppler study. British Journal of
Cote, C. J., Greenhow, E. and Marshall, B. E. (1979) The hypotensive response
to rapid intravenous administration of hypertonic solutions in man and
rabbit. Anesthesiology, 50, 3035.
Crankshaw, D. P. and Karasawa, F. (1989) A method for implementing
programmed infusion of thiopentone and methohexitone with a simple
infusion pump. Anaesthesia and Intensive Care, 17(4), 496499.
Criado, A., Maseda, J., Navarro, E. et al. (1980) Induction of anaesthesia with
etomidate: haemodynamic study of 36 patients. British Journal of Anaesthesia, 52, 803805.
Dahl, A., Russell, D., Nyberg-Hansen, R. and Rootwelt, K. (1989) Effect of
nitroglycerin on cerebral circulation measured by transcranial Doppler and
SPECT. Stroke, 20(12), 17331736.
Dawson, B., Michenfelder, J. D. and Theye, R. A. (1971) Effects of ketamine on
canine cerebral blood flow and metabolism: modification by prior administration of thiopental. Anesthesia and Analgesia, 50, 443.
Derbyshire, D. R., Hunt, P. C. W., Achola, K. and Smith, G. (1984) Midazolam
and thiopentone: catecholamine and arterial pressure responses to induction and endotracheal intubation. British Journal of Anaesthesia, 56, 429P.
DeWitt, D. S., Prough, D. S., Taylor, C. L. and Whitley, J. M. (1992) Reduced
cerebral blood flow, oxygen delivery and electroencephalographic activity
REFERENCES
after traumatic brain injury and mild hemorrhage in cats. Journal of
Diefenbach, C. (1992) Mivacurium: doseresponse relationship and administration by repeated injection or infusion. Anesthesia and Analgesia, 74, 420423.
Drummond, J. C. and Shapiro, H. M. (1981) Cerebral physiology, in Anesthesia,
(ed. R. D. Miller), Churchill Livingstone, New York.
Ducey, J. P., Deppe, A. S. and Foley, F. T. (1989) A comparison of the effects of
suxamethonium, atracurium and vecuronium in intracranial hemodynamics in swine. Anaesthesia and Intensive Care, 17, 448455.
Dundee, J. W. and Moore, J. (1961) Thiopentone and methohexital: a comparison
as main anaesthetic agents for standard operation. Anaesthesia, 35, 5060.
Durward, Q. J., Amacher, A. L., DelMaestro, R. F. and Sibbald, W. J. (1983)
Cerebral and cardiovascular responses to changes in head elevation with
Dwyer, R., McCaughey, W., Lavery, J. et al. (1988) Comparison of propofol and
methohexitone as anaesthetic agents for electroconvulsive therapy. Anaesthesia, 43(6), 459462.
Eerola, R., Kaukinen, L. and Kaukinen, S. (1985) Analysis of 13,800 subclavian
vein catheterizations. Acta Anaesthesiologica Scandinavica, 29, 193197.
Eichhorn, J. H. (1993) Pulse oximetry as a standard of practice in anesthesia.
Anesthesiology, 78(3), 423426.
Ellis, R., Berberich, J. J., Ettigi, P. A. and Kwentus, J. A. (1985) Adrenocortical
function following multiple anaesthesia inductions with etomidate. Anesthesiology, 63, 571.
Eng, C., Lam, A. M., Mayberg, T. S. et al. (1992) The influence of propofol with
and without nitrous oxide on cerebral. blood flow velocity and CO2
reactivity in humans. Anesthesiology, 77, 872879.
Farling, P. A., Johnstone, J. R. and Coppel, D. L. (1989) Propofol infusion for
sedation of patients with head injury in intensive care. A preliminary
report. Anaesthesia, 44(3), 222226.
Feldman, Z., Kanter, M. J., Robertson, C. S. et al. (1992) Effect of head elevation
on intracranial pressure, cerebral perfusion pressure and cerebral blood
flow in head injured patients. Journal of Neurosurgery, 76, 207211.
Fellows, I. W., Bastow, M. D., Byrne, A. J. and Allson, S. P. (1983)
Adrenocortical suppression in multiply injured patients: a complication of
etomidate treatment. British Medical Journal, 287, 18351837.
Feuerman, T., Wackym, P. A., Gade, G. F. and Becker, D. P. (1988) Value of skull
radiography, head computed tomographic scanning and admission for
observation in cases of minor head injury. Neurosurgery, 22, 449453.
Fisher, B., Thomas, D. and Peterson, B. (1992) Hypertonic saline lowers raised
intracranial pressure in children after head trauma. Journal of Neurosurgical
Fleischer, J. E., Milde, J. H., Moyer, T. P. and Michenfelder, J. D. (1988) Cerebral
effects of high dose midazolam and subsequent reversal with Ro 151788 in
dogs. Anesthesiology, 68, 234242.
Forster, A., Juge, O. and Morel, D. (1982) Effects of midazolam on cerebral
blood flow in human volunteers. Anesthesiology, 56, 453455.
Fox, J., Gelb, A. W., Enns, J. et al. (1992) The responsiveness of cerebral blood
flow to changes in arterial carbon dioxide is maintained during propofolnitrous oxide anesthesia in man. Anesthesiology, 77(3), 453456.
From, R. P., Warner, D. S., Todd, M. M. and Sokoll, M. D. (1990) Anesthesia for
craniotomy: a double blind comparison of alfentanil, fentanyl and
sufentanil. Anesthesiology, 73, 896904.
Frost, E. AM., Tabaddor, K., Kim, B. Y. and Chung, S. K. (1981) Efficacy of
induced barbiturate coma in control of intracranial hypertension. Anesthesiology Reviews, 8, 1419.
Gardner, R. M. (1981) Direct blood pressure measurement dynamic response
requirements. Anesthesiology, 54, 227236.
Giffen, J. P., Cotrell, J. E. and Schwirey, B. (1984) Intracranial pressure, mean
arterial pressure and heart rate following midazolam or thiopental in
humans with tumours. Anesthesiology, 60, 491494.
Gillies, G. W. A., Kenny, G. N. C., Bullingham, R. E. S. and McCardle, C. S.
(1987) The morphine sparing effect of ketorolac tromethamine. Anaesthesia,
42, 727731.
Graham, S. G. (1994) The desflurane Tec 6 vaporizer. British Journal of
Hamaguchi, M., Ishibashi, T., Katsumata, N. et al. (1992) Effects of sodium
nitroprusside (MR7S1) and nitroglycerin on the systemic, renal, cerebral,
and coronary circulation of dogs anesthetized with enflurane. Cardiovascular Drugs Therapy, 6(6), 611622.
Hanning, C. D. (1985) Monitoring oxygenation during anaesthesia. British
Journal of Anaesthesia 57, 359360.
Hariri, R. J., Firlich, A. D., Shepard, S. R. et al. (1993) Traumatic brain injury,
hemorrhagic shock, and fluid resuscitation: effects on intracranial pressure
and brain compliance. Journal of Neurosurgery, 79, 412427.
Hartmann, A., Buttinger, C., Rommel, T. et al. (1989) Alteration of intracranial
pressure, cerebral blood flow, autoregulation and carbon dioxide reactivity
by hypotensive agents in baboons with intracranial hypertension. Neurochirurgia (Stuttgart), 32(2), 3743.
Hauser, W. A. (1990) Prevention of post traumatic epilepsy. New England
Hazeki, O. and Tamura, M. (1988) Quantitative analysis of hemoglobin
oxygenation state of rat brain in situ by near infrared spectrophotometry.
Journal of Applied Physiology, 64, 799802.
381
Heiden, J. S., Weiss, M. H., Rosenberg, A. W. et al. (1987) Management of

cervical spinal cord trauma in Southern California. Journal of Neurosurgery,
43, 732.
Henriksen, H. T. and Jorgensen, P. B. (1973) The effect of nitrous oxide on
intracranial pressure in patients with intracranial disorders. British Journal
of Anaesthesia, 45, 486492.
Herpin, D. (1989) The brain and arterial hypertension. 2. Effects of
antihypertensive treatment on cerebral circulation. Annales de Cardiologie et
Angeiologie (Paris), 38(5), 273278.
Hills, M. W. and Deanse, S. A. (1993) Head injury and facial injury: is there an
increased risk of cerebral spine injury?. Journal of Trauma, 34, 549554.
Hoffman, W. E., Miletich, D. J. and Albrecht, R. A. (1986) The effects of
midazolam on cerebral blood flow and oxygen consumption and its
interaction with nitrous oxide. Anesthesia and Analgesia, 65, 729733.
Hoffman, W. E., Pelligrino, D., Werner, C. et al. (1992) Ketamine decreases
plasma catecholamines and improves outcome from incomplete cerebral
ischemia in rats. Anesthesiology, 76(5), 755762.
Houggard, K., Hansen, A. and Broderson, P. (1974) The effect of ketamine on
regional cerebral blood flow in man. Anesthesiology, 41, 562.
Hunkeler, W., Mohler, H., Pieri, L. et al. (1981) Selective antagonists of
benzodiazepines. Nature, 290, 514516.
Jennett, B. (1975) Epilepsy After Non-missile Head Injuries, Heinemann Medical
Books, London.
Jobes, D. R., Kennell, E. M., Bush, G. L. et al. (1977) Cerebral blood flow and
metabolism during morphine-nitrous oxide anesthesia in man. Anesthesiology, 47, 16.
Jones, R. M. (1990) Desflurane and sevoflurane: inhalation anaesthetics for
this decade? British Journal of Anaesthesia, 65, 527536.
Jung, R., Reisel, R., Marx, W. et al. (1992) Isoflurane and nitrous oxide:
comparative impact on cerebro spinal fluid pressure in patients with brain
tumours. Anesthesia and Analgesia, 75, 724728.
Kader, A., Brisman, M. H., Maraire, N. et al. (1992) The effect of mild
hypothermia on permanent focal ischemia in the rat. Neurosurgery, 31,
10561061.
Kaieda, R., Todd, M. M. and Cook, L. N. (1989) Acute effects of changing
plasma osmolality and colloid oncotic pressure on the formation of brain
edema after cryogenic injury. Neurosurgery, 24, 671.
Kaufman, H. H., Timberlake, G., Voelker, J. and Pait, T. G. (1993) Medical
complications of head injury. Medical Clinics of North America, 77, 4376.
King, B. D., Sokoloff, L. and Wechlser, R. L. (1952) The effects of epinephrine
and nor epinephrine upon cerebral circulation and metabolism in man.
Journal of Critical Investigation, 31, 273.
Kitaguchi, K., Nakajima, T., Takaki, O. et al. (1992) The change in cerebral
blood flow during hypotensive anesthesia induced by prostaglandin E1.
Masui, 41(5), 76671.
Kitaguchi, K., Ohsumi, H., Kuro, M. et al. (1993) Effects of sevoflurane on
cerebral circulation and metabolism in patients with ischemic cerebrovascular disease. Anesthesiology, 79(4), 704709.
Kopmann, A. F. (1992) Onset of action of neuromuscular blockade: contributing factors. Current Opinion in Anesthesiology, 5, 561567.
Kovarik, W. D., Mayberg, T. S., Lam, A. M. et al. (1994) Succinylcholine does
not change intracranial pressure, cerebral blood flow velocity or the
electroencephalogram in patients with neurological injury. Anesthesia and
Analgesia, 78(3), 469473.
Kumano, H., Shimomura, T., Furuya, H. et al. (1993) Effects of flumazenil
during administration of midazolam on pial vessel diameter and regional
cerebral blood flow in cats. Acta Anaesthesiologica Scandinavica, 37(6),
567570.
Lagerkranser, M. (1992) Effects of nitroglycerin on intracranial pressure and
cerebral blood flow. Acta Anaesthesiologica Scandinavica, 97, 3436.
Lam, A. M. and Mayberg, T. S. (1992) Use of nitrous oxide in neuroanaesthesia:
Why bother? Journal of Neurosurgical Anesthesiology, 4, 290294.
Lam, A. M., Winn, H. R., Cullen, B. F. and Sundling, N. (1991) Hyperglycaemia and neurological outcome in patients with head injury. Journal of
Lam, A. M., Mayberg, T. S., Eng, C. C. et al. (1994) Nitrous oxideisoflurane
anesthesia causes more cerebral vasodilation than an equipotent dose of
isoflurane in humans. Anesthesia and Analgesia, 78(3), 462468.
Lang, E. W. and Chesnut RM. (1994) Optimizing cerebral perfusion pressure
in targeted therapy of severely head injured patients Proceedings, 1994
International Conference on Recent Advances in Neurotrauma, 293294.
Lanier, W. L., Milde, J. H. and Michenfelder, J. D. (1986) Cerebral stimulation
following succinylcholine in dogs. Anesthesiology, 64, 551.
Lee, S. T. and Lui, T. N. (1992) Early seizures after mild closed head injury.
Levin, A. B., Duff, T. A., Javid, M. J. et al. (1979) Treatment of increased
intracranial pressure. Neurosurgery, 5(5) 570575.
Lewis, S. B., Myburgh, J. A., Thornton, E. L. and Reilly, P. L. (1994) Cerebral
oxygenation monitoring by jugular bulb monitoring (SjO2) and near
infrared spectroscopy (NIRS) in patients with severe head injury. Proceedings, 1994 International Conference on Recent Advances in Neurotraumatology, 1,
357359.
Lindgren, L. (1993) Cardiovascular effects of propofol. British Journal of
382
Ludbrook, G. L., Upton, R. N., Grant, C. and Gray, E. C. (1996) The cerebral
effects of propofol following bolus administration in sheep. Anaesthesia and
Ludbrook, G. L., Upton, R. N., Grant, C. and Gray, E. C. (1995) Cerebrovascular effects of hydralazine and nitroprusside. Anaesthesia and
Intensive Care, 23(5), 637.
Ludbrook, G. L., Russell, W. J. R., Webb, R. K. et al. (1993) The electrocardiograph: applications and limitations: an analysis of 2000 incident reports.
Luna, G. K., Maier, R. V., Pavlin, E. G. et al. (1987) Incidence and effect of
hypothermia in seriously ill patients. Journal of Trauma, 27, 10141018.
Lutz, L. J., Milde, J. H. and Milde, L. N. (1991) The response of the canine
cerebral circulation to hyperventilation during anesthesia with desflurane.
Anesthesiology, 74(3), 504507.
McDowall, D. G. (1965) Effects of anaesthetic agents on cerebral blood flow
and cerebral metabolism. British Journal of Anaesthesia, 37, 236245.
McDowall, D. G. (1967) The effects of clinical concentrations of halothane on
the blood flow and oxygen uptake of the cerebral cortex. British Journal of
Mackenzie, S. J., Kapadia, F. and Grant, I. S. (1990) Propofol infusion for
control of status epilepticus. Anaesthesia, 45(12), 10431045.
McPherson, R. W. and Traystman R. J. (1987) Effects of isoflurane on cerebral
autoregulation. Anesthesiology, 67(3A), A576.
McPherson, R. W. and Traystman, R. J. (1988) Effects of isoflurane on cerebral
autoregulation in dogs. Anesthesiology, 69(4), 493499.
Mangat, P. S., Evans, D. E. N., Harmer, M. and Lunn, J. N. (1993) A
comparison between mivacurium and suxamethonium in children. Anaesthesia, 48, 866869.
Marion, D. W., Obrist, W. D., Carlier, P. M. et al. (1993) The use of moderate
therapeutic hypothermia for patients with severe head injury: a preliminary
report. Journal of Neurosurgery, 79, 354362.
Marshall, L. F. et al. (1978) Pentobarbital therapy for intracranial hypertension
and metabolic coma. Critical Care Medicine, 6(1), 15.
Martin, N. A., Doberstein, C., Zane, C. et al. (1992) Postraumatic cerebral
arterial spasm: transcranial Doppler ultrasound, cerebral blood flow, and
angiographic findings. Journal of Neurosurgery, 77, 575583.
Matta, B. F. and Lam, A. M. (1995) Nitrous oxide increases cerebral blood flow
velocity during pharmacologically induced EEG silence in humans. Journal
of Neurosurgical Anesthesiology, 7(2), 8993.
Matta, B. F., Lam, A. M. and Mayberg, T. S. (1994) The influence of arterial
oxygenation on cerebral venous oxygen saturation during hyperventilation. Canadian Journal of Anaesthetics, 41, 10411046.
Matta, B. F., Lam, A. M., Mayberg, T. S. et al. (1994) A critique of the
intraoperative use of jugular venous bulb catheters during neurosurgical
procedures. Anesthesia and Analgesia, 79(4), 745750.
Mayberg, T. S. and Lam, A. M. (1992) Perioperative use of transcranial Doppler
in patients with head injury. Anesthesia and Analgesia, 74(Suppl.), 5197.
Mayberg, T. S. and Lam, A. M. (1993) Management of central nervous system
trauma. Current Opinion in Anesthesiology, 6, 764771.
Meldrum, B. S. and Nilsson, B. (1976) Cerebral blood flow and metabolic rate
early and late in prolonged epileptic seizures induced in rats by bicuculline.
Brain, 99(3), 523542.
Melloni, C., Fusari, M., Scesi, M. et al. (1992) The use of propofol in status
epilepticus. Minerva Anestesiologica, 58(9), 553556.
Merry, A. (1994) The place of NSAIDs in perioperative pain relief, in
Australasian Anaesthesia, (ed. J. Keneally), Australian and New Zealand
College of Anaesthetists, Melbourne, Victoria, pp. 149158.
Messeter, K., Nordstrom, C. H., Sundbarg, G. et al. (1986) Cerebral
Neurosurgery, 64(2), 231237.
Messick, J. M. Jr, Casement, B., Sharbrough, F. W. et al. (1987) Correlation of
regional cerebral blood flow (rCBF) with EEG changes during isoflurane
anesthesia for carotid endarterectomy: critical rCBF. Anesthesiology, 66,
344.
Michenfelder, J. D. (1974) The interdependency of cerebral function and
metabolic effects following massive doses of thiopental in the dog.
Milde, L. N. and Milde, J. H. (1991) The cerebral and systemic hemodynamic
and metabolic effects of desflurane-induced hypotension in dogs. Anesthesiology, 74(3), 513518.
Milde, L. N., Milde, J. H. and Michenfelder, J. D. (1985) Cerebral function,
metabolic and hemodynamic effects of etomidate in dogs. Anesthesiology, 63,
371.
Miletich, D. J., Ivankovich, A. D. and Albrecht, R. F. (1976) Absence of
autoregulation during halothane and enflurane anaesthesia. Anesthesia and
Analgesia, 55, 100.
Morley-Forster, P. K. (1986) Unintentional hypothermia in the operating room.
Canadian Anaesthetists Society Journal, 33, 515527.
Muizelaar, J. P., van der Poel, H. G., Li, Z. C. et al. (1988) Pial arteriolar vessel
rabbit. Journal of Neurosurgery, 69(6), 923927.
prolonged hyperventilation in patients with severe head injury: a randomised clinical trial. Journal of Neurosurgery, 75, 731739.
Muzzi, D. A., Lasasso, T. J., Dietz, N. M. et al. (1992) The effect of desflurane
and sevoflurane on cerebrospinal fluid pressure in humans with supratentorial lesions. Anesthesiology, 76, 720724.
Nakajama, D. K., Waggoner, T., Venkataraman, S. T. et al. (1992) The use of
drugs in emergency airway management in pediatric trauma. Annals of
Surgery, 216(2), 205211.
Newberg, L. A., Milde, J. H. and Michenfelder, J. D. (1983) The cerebral
metabolic effects of isoflurane at and above concentrations that suppress
cortical electrical activity. Anesthesiology, 59, 23.
Nissenson, A. R., Weston, R. E. and Kleeman, C. R. (1979) Mannitol. Western
Journal of Medicine, 131(4), 277284.
North, J. B., Penhall, R. K., Hanieh, A. et al. (1983) Phenytoin and post
operative epilepsy. Journal of Neurosurgery, 58, 672677.
Nugent, M., Artru, A. A. and Michenfelder, J. D. (1982) Cerebral metabolic,
vascular and protective effects of midazolam maleate. Anesthesiology, 56,
172176.
Obrist, W. D., Langfitt, T. W, Jaggi, J. L. et al. (1984) Cerebral blood flow and
Olesen, J. (1972) The effect of intracarotid epinephrine, norepinephrine and
angiotensin on the regional cerebral blood flow in man Neurology 22, 978.
Ornstein, E., Young, W. L., Fleischer, L. H. and Ostapkovich, N. (1993)
Desflurane and isoflurane have similar effects on cerebral blood flow in
patients with intracranial mass lesions. Anesthesiology, 79(3), 498502.
Papazian, L., Albanese, J., Thirion, X. et al. (1993) Effect of bolus doses of
midazolam on intracranial pressure and cerebral perfusion pressure in
patients with severe head injury. British Journal of Anaesthesia, 71(2),
267271.
Pensado, A., Molins, N. and Alvarez, J. (1993) Effects of propofol on mean
arterial pressure and systemic vascular resistance during cardiopulmonary
bypass. Acta Anaesthesiologica Scandinavica, 37(5), 498501.
Petros, A. J., Bogle, R. G. and Pearson, J. D. (1993) Propofol stimulates nitric
oxide release from cultured porcine aortic endothelial cells. British Journal of
Pharmacology, 109(1), 67.
Pfenniger, E. and Reith, A. (1990) Ketamine in intracranial pressure, in Status
of Ketamine in Anesthesiology, (ed. E. F. Domino), NPP Books, Ann Arbor, MI,
pp. 109118.
Pietropaoli, J. A., Rogers, F. B. and Zhuang, J. (1992) The deleterious effects of
intraoperative hypotension on outcome in patients with severe head
injuries. Journal of Trauma, 33, 403407.
Prough, D. S., Johnson, J. C., Poole, J. V. et al. (1985) Effects on intracranial
pressure of resuscitation from hemorrhagic shock with hypertonic saline
versus lactated Ringers solution. Critical Care Medicine, 13, 407411.
Puhringer, F. K. (1992) Evaluation of the endotracheal intubating conditions of
rocuronium and succinylcholine in outpatient surgery. Anesthesia and
Analgesia, 75, 3740.
Raftery, E. B. and Ward, A. P. (1968) The indirect method of recording blood
pressure. Cardiovascular Research, 2, 210218.
Ramani, R., Todd, M. M., Warner, D. S. et al. (1992) A dose related study of the
influence of propofol on cerebral blood flow. Journal of Neurosurgical
Ravussin, P., Guinard, J. P., Ralley, F. and Thorin, D. (1988) Effect of propofol
on cerebrospinal fluid pressure and cerebral perfusion pressure in patients
undergoing craniotomy. Anaesthesia, 43(Suppl.), 3741.
Reddy, R. V., Moorthy, S. S., Dierdorf, S. F. et al. (1993) Excitatory effects and
electroencephalographic correlation of etomidate, thiopental, methohexital
and propofol. Anesthesia and Analgesia, 77(5), 10081011.
Reinstrup, P., Ryding, E., Algotsson, L. et al. (1994) Effects of nitrous oxide on
human regional cerebral blood flow and isolated pial arteries. Anesthesiology, 81(2), 396402.
Renou, A. M., Vernhiet, J., Macrez, P. et al. (1978) Cerebral blood flow and
metabolism during etomidate anaesthesia in man. British Journal of
Robertson, C. S., Contantin, C. F., Gokaslan, Z. L. et al. (1992) Cerebral blood
flow, arteriovenous oxygen difference and outcome in head injured
patients. Journal of Neurology, Neurosurgery and Psychiatry, 55, 594603.
Rockoff, M. A., Marshall, L. F. and Shapiro, H. M. (1979) High dose
barbiturate therapy in humans: a clinical review of 60 patients. Annals of
Roekaerts, P. M., Huygen, F. J. and de Lange, S. (1993) Infusion of propofol
versus midazolam for sedation in the intensive care unit following coronary
artery surgery. Journal of Cardiothoracic and Vascular Anesthesia, 7(2),
142147.
Rosa, G., Orfei, P., Sanfilippo, M. et al. (1986) The effects of atracurium
besylate on intracranial pressure and cerebral perfusion pressure. Anesthesia
and Analgesia, 65, 381384.
Rosner, M. J., Rosner, S. D., Johnson, A. H. and Langford K . (1994) CPP
management: results. Proceedings, 1994 International Conference on Recent
Advances in Neurotraumatology, 1, 436437.
Royblat, L., Katz, J., Rozentsveig, V. et al. (1992) Anaesthetic interaction
between thiopentone and ketamine. European Journal of Anesthesiology, 9,
307312.
Rutten, A. J., Ilsley, A. H., Skowronski, G. A. and Runciman, W. B. (1986)
Measurement of mean arterial blood pressure by automatic oscillometers,
REFERENCES
arterial cannulation and auscultation. A comparative study. Anaesthesia and
Sakabe, T., Kuramoto, T., Inoue, S. et al. (1978) Cerebral effects of nitrous oxide
in dogs. Anesthesiology, 48, 195200.
Sano, T., Patel, P. M., Drummond, J. C. and Cole, D. J. (1993) A comparison of
the cerebral protective effects of etomidate, thiopental, and isoflurane in a
model of forebrain ischaemia in the rat. Anesthesia and Analgesia, 76(5),
990997.
Scheller, M. S., Nakakimura, K., Fleischer, J. E. and Zornow, M. H. (1990)
Cerebral effects of sevoflurane in the dog: comparison with isoflurane and
enflurane. British Journal of Anaesthesia, 65(3), 388392.
Shackford, S. R., Zhuang, J. and Schmoker, J. (1992) Intravenous fluid tonicity:
effect on intracranial pressure, cerebral blood flow, and cerebral oxygen
delivery in focal brain injury. Journal of Neurosurgery, 76, 9198.
Shapira, Y., Artru, A. A. and Lam, A. M. (1992) Ketamine decreases cerebral
infarct volume and improves neurological outcome following experimental
head injury in rats. Journal of Neurosurgical Anesthesiology, 4, 231240.
Shapira, Y., Setton, D., Artru, A. A. and Shomai, E. (1993) Bloodbrain barrier
permeability, cerebral edema and neurological function after closed head
injury in rats. Anesthesia and Analgesia, 77(1), 141148.
Shapira, Y., Lam, A. M., Eng, C. C. et al. (1994) Therapeutic time window and
dose response of the beneficial effects of ketamine in experimental head
injury. Stroke, 25(8), 16371643.
Shapiro, H. M., Whyte, S. R. and Harris, A. B. (1974) Ketamine anaesthesia in
patients with intracranial pathology. British Journal of Anaesthesia, 44,
12001204.
Sheepstra, G. L., Vree, T. B., Crul, J. F. et al. (1986) Convulsive effects and
pharmacokinetics of laudanosine in rats. European Journal of Anesthesiology,
3, 371383.
monitoring of jugular venous oxygen saturation in head injured patients.
Journal of Neurosurgery, 76; 212217.
Shiozaki, T., Sugimoto, H., Taneda, M. et al. (1993) Effect of mild hypothermia
on uncontrollable intracranial hypertension after severe head injury. Journal
of Neurosurgery, 79(3), 363368.
Silverberg, G. D., Reitz, B. A. and Ream, A. K. (1981) Hypothermia and
cardiac arrest in the treatment of giant aneurysms of the cerebral
circulation and hemangioblastoma of the medulla. Journal of Neurosurgery,
55(3), 337346.
Silverman, D. G. and Brull, S. J. (1993) Depth of block after divided doses of
mivacurium spaced 60 seconds apart. Anesthesia and Analgesia, 77,
164167.
Simpson, K. H., Halsall, P. J., Carr, C. M. and Stewart, K. G. (1988) Propofol
reduces seizure duration in patients having anaesthesia for electroconvulsive therapy British Journal of Anaesthesia, 61(3), 343344.
Smith, D. S. et al. (1980) Inhibitory effects of different barbiturates on lipid
peroxidation in brain tissue in vitro. Comparison with the effects of
promethazine and chlorpromazine. Anesthesiology, 53(3), 186194.
Smith, H. P., Kelly, D. L., McWhorter, J. M. et al. (1986) Comparison of
mannitol regimens in patients with severe head injury. Journal of Neurosurgery, 65, 820.
Smith, I., White, P. F., Nathanson, M. and Gouldson, R. (1994) Propofol. An
update on its clinical use. Anesthesiology, 81(4), 10051043.
Stewart, L., Bullock, R., Rafferty, C. et al. (1995) Propofol sedation in severe
head injury fails to control high ICP, but reduces brain metabolism. Acta
Neurochirurgica (Supplement) (Vienna), 60, 544546.
Stirt, J. A., Grosslight, K. R., Bedford, R. F. and Vollmer, D. (1987)
Defasciculation with metocurine prevents succinylcholine induced increases in intracranial pressure. Anesthesiology, 67, 5053.
383
Stokes, D. N. and Hutton, P. (1991) Rate-dependent induction phenomena

with propofol: implications for the relative potency of intravenous
anaesthetics. Anesthesia and Analgesia, 72(5), 578583.
Suderman, V. S., Crosby, E. T. and Lui, A. (1992) Elective oral intubation in
cervical spine injured patients. Canadian Journal of Anaesthetics, 39(5),
516517.
Takahashi, H., Murata, K. and Ikeda K (1993) Sevoflurane does not increase
intracranial pressure in hyperventilated dogs. British Journal of Anaesthesia,
71(4), 551555.
Takeshita, H. and Michenfelder, J. D. (1972) The effects of morphine and
N-allylnormorphine on canine cerebral metabolism and circulation. Anesthesiology, 37(6), 605612.
Takeshita, H., Okuda, Y. and Sari, A. (1972) The effects of ketamine on cerebral
circulation and metabolism in man. Anesthesiology, 36, 69.
Tarkkanen, L., Laitinen, L. and Johannsen, G. (1974) Effects of D-tubocurarine
on intracranial pressure and thalamic electrical impedance. Anesthesiology,
50, 247251.
Temkin, N. R., Dikmen, S. S, Wilensky, A. J. et al. (1990) A randomized double
blind study of phenytoin for the prevention of post traumatic seizures. New
Todd, M. M., Drummond, J. C. and Shapiro, H. M. (1982) Comparative
cerebrovascular and metabolic effects of halothane, enflurane and isoflurane. Anesthesiology, 57, A332.
Todd, M. M., Warner, D. S., Sokoll, M. D. et al. (1993). A prospective
comparative trial of three anesthetics for elective supratentorial craniotomy: propofolfentanyl, isofluranenitrous oxide and fentanylnitrous
oxide. Anesthesiology, 76, 743754.
Valanne, J. (1992) Recovery and discharge of patients after long propofol
infusion vs isoflurane anesthesia for ambulatory surgery. Acta Anaesthesiologica Scandinavica, 36(6), 530533.
Walker, R. J. (1991) Paracetamol, non steroidal antiinflammatory drugs and
nephrotoxicity. New Zealand Medical Journal, 104, 182183.
Ward, J., Becker, D. P., Miller, J. D. et al. (1985) Failure of prophylactic
Watson, J. C., Drummond, J. C., Patel, P. M. et al. (1992) An assessment of the
cerebral protective effects of etomidate in a model of incomplete forebrain
ischemia in the rat. Neurosurgery, 30(4), 540554.
Webb, R. K., Ralston, A. C. and Runciman, W. B. (1991) Potential errors in
pulse oximetry. II. Effects of changes in saturation and signal quality.
Anaesthesia, 46(3), 207212.
Whyte, S. R., Shapiro, H. M. and Turner, P. (1972) Ketamine induced
intracranial hypertension: two case reports. Anesthesiology, 35, 226.
Wilkinson, H. A. and Rosenfeld, S. (1983) Furosemide and mannitol in the
treatment of acute experimental intracranial hypertension. Neurosurgery, 12,
405.
Wilson, D. V. (1992) Anesthesia for patients with head trauma. Veterinary
Clinics of North America Small Animal Practice, 22(2), 486491.
Wood, P. R. and Lawler, P. G. (1992) Managing the airway in cervical spine
injury. A review of the Advanced Trauma Life Support Protocol. Anaesthesia, 47, 792797.
Yablon, S. A. (1993) Posttraumatic seizures. Archives of Physical Medicine and
Rehabilitation, 74, 9831001.
Yoshida, S., Inoh, S. and Asano, T. (1983) Brain free fatty acids, oedema and
mortality in gerbils subjected to transient bilateral ischaemia, and effect of
barbiturate anesthesia. Journal of Neurochemistry, 40, 1278.
Young, B., Rapp, R. P., Norton, J. A. et al. (1983) Failure of prophylactically
administered phenytoin to prevent late posttraumatic seizures. Journal of
19
MANAGEMENT OF
AND CEREBRAL PERFUSION
Peter Reilly
19.1
Introduction
The treatment of traumatic brain injury is based on the

conceptual distinction between primary and secondary injury. Primary injury comprises the immediate
effects of impact. Secondary injury is due to factors
other than the direct impact and may begin almost
immediately or some time later. The distinction is of
course, not so sharp. It is increasingly clear that the
impact initiates a series of events, including an
evolving axonal injury and release of mediator substances, which cause injury over minutes to hours
after impact, suggesting that it may become possible
to intervene therapeutically in the evolution of these
processes initiated by the primary injury.
Secondary injury, on the other hand, may begin at or
very close to the impact. Asphyxia or massive blood
loss may cause overwhelming ischemichypoxic secondary damage, irrespective of the severity of the
primary injury.
However the model of primary and secondary
injury has clinical value and this chapter will concentrate on the two of the main end points in managing
the secondary effects of injury, cerebral perfusion
pressure (CPP) and intracranial pressure (ICP).
The major cause of raised ICP, apart from hematomas, is brain swelling. The factors responsible for
post-traumatic brain swelling are not clear. There is
convincing evidence for both cytotoxic and ischemic
hypoxic influences, and specific means of combating
them are now emerging from the laboratory into
clinical trial. However, until brain swelling can be
prevented the primary goals in preventing secondary
brain injury will be to provide adequate cerebral
perfusion and oxygenation and to prevent brain shift
(herniation).
Ideally, treatment should be based on direct knowledge of regional CBF, metabolism and function.
Although this may soon be possible, at present ICP
and CPP, which can be measured easily and continuously, provide the best indirect measurements of
tissue perfusion. ICP and CPP are however, global
measurements and this has implications in defining
treatment goals for a heterogeneous injury. This will
be discussed later. Indirect indices of CBF and metabolism, such as jugular venous oxygen saturation (PjvO2),
transcutaneous Doppler (TCD; Chapter 14), electrical
function (Chapter 13) and near-infrared spectroscopy
(Chapter 11) may assist in defining adequate cerebral
perfusion and oxygen delivery and in selecting the
most appropriate treatment.
19.2
Intracranial pressure
Some aspects of ICP pathology relevant to the management of head injury will be reviewed briefly. (This
subject is covered in greater detail in Chapter 6.)
Our understanding of the relationship between
intracranial volume and pressure has its origin in the
appreciation of Monro (1783) and Kellie (1824) that
cranial volume remains constant after sutural fusion
and that the intracranial contents are incompressible.
Burrows (1846) subsequently made the important
additional observation that, while total volume
remains constant, CSF volume could interchange with
blood volume. Hence provided that a change in
volume in one compartment is balanced by near-equal
change in another, ICP will remain constant.
The major intracranial volumes are brain parenchyma (12001600 ml), blood (100150 ml) and CSF
(100150 ml; Figure 19.1). These latter two constitute
about 20% of total intracranial volume and part of
each is capable of rapid extracranial displacement.
Therefore they are of prime importance in the volume
equilibration of an expanding mass.
The extracellular fluid component of brain parenchyma is also capable of change, both pathologically and in response to treatment. Therefore
386
MANAGEMENT OF INTRACRANIAL PRESSURE AND CEREBRAL PERFUSION

Figure 19.1
The major intracranial contents by volume.
manipulation of these volumes underlies the treatment of raised ICP after head injury.
Since the landmark study of Lundberg (1960) on the
clinical value of ICP monitoring, many studies have
drawn attention to the importance of raised ICP after
severe head injury.
In one major trauma center, ICP was found to be
elevated (to greater than 10 mmHg) in 82% of patients
with severe head injury. Of those with raised ICP,
one-third (16% of the entire series) developed uncontrolled intracranial hypertension and died. Conversely, in about 50% of those who died, raised ICP
was the major cause. In survivors, moderately raised
ICP was a major cause of morbidity (Miller et al.,
1977; 1981).
In the multicenter Traumatic Coma Data Bank
(TCDB) study from the USA, ICP was found to be an
important independent variable, significantly related
to outcome (Marmarou et al., 1991). Although in some
patients intracranial hypertension may simply be a
sign of an overwhelming and irrecoverable brain
injury, the evidence for actively treating raised ICP
derives mainly from the observation that mortality is
lower in those patients in whom ICP can be controlled
(Eisenberg et al., 1988).
19.2.1 INTRACRANIAL PRESSURE AND CEREBRAL
PERFUSION PRESSURE
The relationships between ICP, CPP and brain function

are complex. However certain points can be made.
It is doubtful whether raised ICP itself directly

alters neuronal function. Although it has been
suggested that the effects of raised ICP on cellular
metabolism and membrane integrity may not be
entirely related to CPP (Chesnut and Marshall,
1993), it seems likely that the absolute level of ICP is
not as important as its effect on CPP and its
relationship to brain herniation.
Reduced CPP and brain herniation are principal

mechanisms of secondary brain damage following
severe head injury (Johnston, Johnston and Jennett,
1970; Rosner and Daughton, 1990). In support of
this, raised ICP, even over 60 mmHg, without
displacement and with adequate CPP may produce
no neurological deficit. This is seen quite dramatically in benign intracranial hypertension.
In the absence of herniation, ICP measured in one
brain region will reflect the pressure of the entire
cranial cavity (Langfitt, Weinstein and Kassell,
1964a, b). Even so, its effects on regional perfusion
may be heterogeneous. Perfusion in and around
areas of injury may be reduced further by an
increase in regional tissue pressure creating regional
pressure gradients, by vessel distortion or by
regional loss of autoregulation. ICP and therefore
global CPP may thus overestimate regional
perfusion.
The viscoelastic properties of the brain as measured
by compliance tests are altered by brain injury
(Figure 19.2(a)) and by the standard methods of
treating raised ICP (Figure 19.2(b)). Mannitol compared with hyperventilation increases compliance
at the same pressure level (Leech and Miller, 1974;
Rowed et al., 1975).
19.2.2
INTRACRANIAL PRESSURE GRADIENTS
Interhemispheric pressure gradients are not observed

in patients with diffuse injury but can occur in patients
with focal lesions and midline shift and may last
several hours (Sahuquillo et al., 1994). In this latter
group, ICP might best be measured from the side of
the focal lesion.
The relationship between ICP and cerebral herniation is variable. Herniation is the result of a pressure
gradient between anatomical compartments generated by an expanding mass, but the speed of development depends on anatomical factors and determines the functional significance of the herniation.
Temporal lobe masses can only expand medially and
posteriorly, being constrained anteriorly, laterally and
inferiorly by bone. They will therefore result in more
tentorial herniation at lower ICP than will occipital
or frontal masses. The lateral displacement of deep
brain structures may itself be an important basis for
loss of consciousness through the shearing effect on
small perforating vessels to deep diencephalic structures (Andrews et al., 1988; Ropper, 1986, 1989;
Marshall et al., 1983). Herniation is a form of volume
compensation for high ICP and once herniation and
brain-stem compression have occurred, the intercompartmental pressure gradients will disperse and
ICP may even be low for a short time (Johnston and
Jennett, 1973).
387
Figure 19.2 The influence of viscoelastic changes on the

volume pressure curve. (a) Edema and hemorrhage reduce
compliance, as well as take up volume. Hence pressure
begins to rise sooner and more steeply. (b) Mannitol reduces
brain volume and shifts the curve to the right. The brain is
more compliant, hence the initial curve is flatter. Hypocarbia
reduces blood volume and also shifts the curve to the right
but does not alter compliance (the slope of the curve).
(a)
(b)
19.2.3 THE GENESIS OF RAISED INTRACRANIAL

PRESSURE
Marmarou, Shulman and LaMorgese (1975) developed a theoretical model of ICP based on a constant
production of CSF independent of pressure, circulation through a compliant storage space and absorption
via pressure-dependent outflow pathways at the
arachnoid villi. Dural sinus pressure was defined as
the exit pressure of the system. The steady state
relationship may be expressed as:
ICP = If Ro + PD
where If = formation rate; Ro = outflow resistance; PD
= dural sinus pressure.
In this model, steady state ICP is proportional to:

rate of CSF formation;

resistance to CSF absorption;
dural sinus pressure.
The factors leading to raised ICP can be considered as

being of CSF or vascular origin.
The CSF component of ICP derives from the product
of CSF formation and outflow resistance. Pathological
events that increase the CSF component are associated
with an increase in brain water. This may take the form
of ventricular dilatation or brain edema.
The vascular component of ICP under normal
physiological conditions is equivalent to dural sinus
pressure. Pathological events that affect the vascular
component are associated with dilatation of the
vascular bed and congestive or hyperemic brain
swelling. The contribution of these two components to
raised ICP can be determined by analysis of pressure
volume relationships. In head-injured patients, vascular factors are twice as common as non-vascular
factors and make the major contribution to raised ICP,
probably from compression of compliant veins (Marmarou et al., 1987). Hyperemic swelling is particularly
common in children (Muizelaar et al., 1989)
388
19.3
Cerebral perfusion pressure
Perfusion pressure across any vascular bed is the

mean systemic arterial pressure minus venous outflow pressure. Intracranial venous outflow pressure at
the dural sinuses (the pressure in the cerebral veins
within the subarachnoid space) remains slightly
above and is linearly related to CSF pressure over a
wide range, so in practice cerebral perfusion pressure
(CPP) = mean BP mean ICP (Rowan et al., 1972).
Normally, within a CPP range of 60160 mmHg, CBF
remains constant. However, in brain injury this
autoregulatory relationship may be altered by a
number of factors.
CPP is equivalent to the transmural pressure across
the cerebral vessel walls. CPP at the arteriolar level is
the stimulus for the autoregulatory response and, at
the capillary level, is the driving force for fluid
exchange.
19.4
Intracranial volumes
19.4.1
CEREBRAL BLOOD VOLUME
Some 70% of cerebral blood volume (CBV) resides in

the capacitance vessels. As CPP falls, arteriolar dilatation occurs in an attempt to maintain constant CBF.
CBV will vary with these autoregulatory adjustments
in vessel diameter. Within the autoregulatory range, as
CPP rises, arteriolar constriction leads to a fall in CBV
and hence in ICP. Below the autoregulatory range,
maximal vasodilatation has occurred but, if CPP falls
further, CBV will fall paradoxically as a result of
vasocollapse (Kontos et al., 1978)
Hence, although CBV and ICP are directly related,
the relationship between CBF and CBV depends on the
status of autoregulation and the level of CPP.
Hypoxia causes cerebral vasodilation at PaO2 levels
below 50 mmHg. The stimulus for vasodilatation is
reduced oxygen delivery and therefore depends on
Hb, CBF and O2 saturation.
CO2 is also a cerebral vasodilator probably acting
through H+ concentration in smooth muscle cells of
the cerebral vessels (Gotoh, Meyer and Takagi, 1965).
The relationship between PaCO2 and CBF is sigmoid
(Harper and Glass, 1965). Within the range of
3050 mmHg CBF bears a linear relation to PaCO2
(Hayes and Tindall, 1969; Paul et al., 1972). The CO2
response is very robust and is usually at least partly
preserved after severe head injury, even when pressure
autoregulation is lost.
Indeed the absence or severe reduction of the CO2
response has serious prognostic significance (Shalen,
Messeter and Nordstrom,
1991; Cold, Jensen and

Malmros, 1977; Enevoldsen and Jensen, 1978; Overgaard and Tweed, 1974; Yoshihara, Bandoh and
Marmarou, 1995). It may also predict failure to

respond to barbiturates (Nordstrom et al., 1988) since
the ability of these drugs to reduce ICP depends
upon the capacity of the cerebral vessels to respond
by vasoconstriction to reduced metabolic demand.
19.4.2
CEREBROSPINAL FLUID VOLUME
Some 70% of CSF is produced by the choroid plexus

and 30% enters the ventricles across the ependyma
from brain parenchyma as excess extracellular fluid
(brain lymphatic fluid). CSF production involves
active transport, particularly of Na+, and may be
reduced by the carbonic anhydrase inhibitor acetozolamide (Rubin et al., 1966; Sahar and Tsipstein,
1978), by frusemide (furosemide) and steroids (Sato
et al., 1973). The rate of production is constant at
about 0.35 ml/min (20 ml/h or 500 ml/d) giving a
fivefold turnover per day. Production is not influenced by ICP except at very high levels (Brgesen
and Gjerris, 1989).
Total CSF volume in an adult is approximately
140 ml, of which 30 ml is spinal and 20 ml in the
ventricles (Davson, Welch and Segal, 1987).
A total of 80% absorption occurs via the arachnoid
granulations, particularly those related to the dural
sinuses, and at normal pressures is determined both
by the pressure gradient across the granulations and
by active transport. The other 20% is absorbed via
spinal nerve root sheaths and the VirchowRobin
perivascular spaces. Hence CSF absorption increases
with ICP. It can be inhibited by obstruction in any
part of the CSF pathways or at the arachnoid villi.
19.4.3
BRAIN VOLUME
Brain volume contributes about 70% of intracranial

volume and is normally constant. Traumatic brain
swelling may be due to an increase in one or more
of the three brain fluid compartments blood,
intracellular or extracellular fluid and is the major
cause of raised ICP after head injury. (See Chapter 7
for a detailed discussion of traumatic brain swelling.) Brain is the least displaceable of the intracranial
volumes. Displacement or herniation of brain is a
limited and final form of accommodation of an
intracranial mass.
19.5
Volumepressure relations
The initial, nearly horizontal part of the exponential

volumepressure curve represents the phase when
intracranial (venous) blood and CSF can be readily
displaced to accommodate an increase in another
VOLUMEPRESSURE RELATIONS
389
(Rosner and Becker, 1984a) has attributed them to a

reduced CPP, which leads to a vasodilatory cascade.
Providing CPP is within the autoregulatory range
and autoregulation is preserved, a fall in CPP will
induce vasodilatation, an increase in CBV and a
parallel increase in ICP. This will further reduce CPP
and tend to perpetuate the circle.
The process may be terminated by a rise in
BP (Cushing response) the increase in CPP will
then lead to vasoconstriction and a fall in CBV
or by hyperventilation, which reduces ICP by
vasoconstriction.
19.5.2 MEASURING INTRACRANIAL VOLUME
RESERVE
Figure 19.3 The contribution of CSF, venous and arterial
blood volume displacement to the volume pressure curve.
Pressure is defined here as the force necessary to displace
contents. The volume pressure curve is seen as the resultant
of curves derived from the pressure necessary to displace
CSF, venous and arterial blood. (Source: reproduced from
Rosner, 1993, with permission)
volume or the addition of a new volume, without rise

in pressure. Further volume additions (V) cause
increasingly steep rises in ICP as the force needed to
displace compensating fluid volumes increases The
relative contribution of the two chief displaceable
volumes, blood and CSF, are represented in Figure
19.3 (Rosner, 1993).
CSF can be displaced readily into the spinal subarachnoid space. A total of 70% of the CSF buffering
capacity is contained in the intracranial compartment
and 30% in the spinal compartment, where dural
expansion can occur by compression of epidural veins
(Marmarou, Shulman and LaMorgese, 1975). Blockage
at the foramen magnum by tonsillar herniation prevents this compensatory displacement and accelerates
further rapid rises in ICP.
CBV within the venous (capacitance vessels) can be
readily and passively displaced to the extracranial
venous system. A smaller volume of blood is contained on the arterial side but the reactivity of the
arterioles (autoregulation) has a major influence on
CBV and hence on ICP as already outlined.
19.5.1
INTRACRANIAL PRESSURE WAVES
Lundberg (1960) first recognized the significance of

pressure waves occurring under conditions of reduced
intracranial compliance. Of the three main waves he
described, A waves (Plateau) and the shorter duration
B waves are of greatest clinical relevance. Rosner
Theoretically, measurement of intracranial volume

reserve would allow treatment of brain swelling
before ICP rises (Marmarou, 1986). However compliance measurements (the pressure rise associated
with a given volume increase) have not gained a place
in routine clinical practice and it is not known whether
such a pre-emptive approach would make the
treatment of brain swelling more effective. However,
compliance measurements have given some insights
into the effects of head injury and various treatments.
Following head injury, compliance increases, altering
the shape of the volume pressure curve so that it rises
more steeply, and the breakpoint, the point at which
ICP begins to rise, may be moved to the left, indicating
a reduced buffering capacity (Figure 19.2(a)).
Treatments used to reduce ICP may also affect
compliance. Hyperventilation moves the breakpoint
to the right, increasing the intracranial buffering
capacity, but does not change its slope (Rowed et al.,
1975). Yoshihara, Bandoh and Marmarou (1995)
found that the PVI, and hence the slope of the
volumepressure curve did not change with moderate changes of PaCO2. Mannitol appears to shift the
breakpoint and also to change the slope of the curve,
so having the additional advantage of rendering the
brain more compliant (Leech and Miller, 1974; Figure
19.2(b)).
19.5.3
VISCOELASTIC PROPERTIES
As indicated by the changing slope of the volume

pressure curve under different circumstances, the
viscoelastic properties of the brain, the skull and the
dural sac affect the equilibrium relationship between
ICP and intracranial volume. If the brain is tight
(high elastance, low compliance), a small volume
addition will produce a marked rise in ICP. Hence
pathological processes, such as brain edema, will
affect volumepressure relationships both by the
increase in volume of tissue water and by altering the
390
Figure 19.4 The relationship between CSF inflow and outflow and viscoelasticity. The viscoelastic properties of brain affect
the equilibrium relationship between CSF pressure and volume (the starting point of the volume pressure curve) and also the
dynamic relationship, i.e. the effect of a small volume addition. (Source: reproduced from Miller, 1985, with permission.)
viscoelastic properties of the brain The relationship

between CSF inflow and outflow and viscoelasticity
are represented in Figure 19.4 (Miller, 1985).
19.5.4 IDENTIFYING THE CAUSES OF RAISED ICP
AFTER HEAD INJURY
After mass lesions (e.g. hematomas ) have been

excluded and extracerebral factors such as raised
intrathoracic pressure or hyperthermia have been
rectified, persistently raised ICP is likely to be due to
brain swelling.
Brain swelling may be due to hyperemia (vascular
engorgement, mainly venous) or edema (an increase in
brain water content). Brain edema has been further
classified as cytotoxic, vasogenic or hydrostatic
(Chapter 7).
Ways of identifying the vascular and non-vascular
factors contributing to raised ICP by a combination of
ICP wave form analysis, cerebral electrical function
monitoring and continuous monitoring of jugular
venous oxygen saturation have been suggested (Piper,
Dearden and Miller, 1989; Dearden and Miller, 1989).
Jugular venous oxygen saturation combined with
arterial oxygen saturation measurements will indicate
cerebral oxygen consumption and so identify ischemic
and hyperemic states.
It must always be remembered that the pathology
of head injury is both heterogeneous and dynamic, and
that the dominant cause of raised ICP may vary with
time. Brain injuries do not produce simple or pure
forms of brain swelling.
Indeed one form of brain swelling may inevitably
lead to the development of another. For example, an
increase in ICP due to extravascular fluid increase
(brain edema) may lead to ischemia and vasodilatation, a vascular cause for increased ICP. Conversely,
ischemic cytotoxic swelling may later damage the
bloodbrain barrier and cell membranes, leading to
vasogenic brain swelling.
Swelling that is predominantly due to increase in

blood volume may respond to vasoconstriction and
hence to barbiturates or to hyperventilation (Yoshihara, Bandoh and Marmarou, 1995). Indeed, the initial
effect of the osmotic diuretic mannitol may also be
through vasoconstriction induced by lowering blood
viscosity, and therefore dependent on intact vascular
reactivity.
Mannitol also increases brain tissue specific gravity,
most probably by reducing brain tissue water; hence
non-vascular swelling may also respond to osmotic
therapy (Nath and Galbraith, 1986).
19.6
Principles of management
19.6.1
EVIDENCE-BASED GUIDELINES
The many and often conflicting reports and reviews

on aspects of head injury pathophysiology and treatment have supported a range of different plans of
management (Gahjar et al., 1995; Fearnside et al., 1993).
There is now an emphasis on forming guidelines for
treatment based on a rigorous assessment of the
available evidence. It is inherent to such a process that
the guidelines be constantly updated in the light of
new evidence. Guidelines for the Management of Severe
Head Injury was published in 1995 by the Brain
Trauma Foundation and the American Association of
Neurological Surgeons and forms an excellent precis
of present knowledge.
Because of the clear relationship between ICP and
outcome, most treatment protocols have focused on
avoiding high ICP. Others have argued that the
primary objective of treatment should be to maintain
an adequate CPP, i.e. above the lower limit of
autoregulation, as far as it can be judged in the headinjured patient. Rosner suggested that a falling CPP
may set in train a vasodilatory cascade. As CPP falls,
vasodilation occurs to maintain CBF. The consequent
increase in CBV raises ICP and further reduces CPP.
PRINCIPLES OF MANAGEMENT
The cascade is halted by increasing CPP. The consequent vasoconstriction will reduce ICP (Rosner and
Coley, 1986).
Initial reports of the use of CPP-based protocols
suggest improved outcome compared with the TCDB
(Rosner, Rosner and Johnston, 1995). However, the
principal benefit of focusing on CPP may be to reduce
the incidence of hypotension, one of the commonest
causes of secondary insult after head injury (Chesnut
et al., 1993a). No particular management protocol has
yet had the benefit of a controlled trial.
It seems clearer now that, although ICP and CPP are
interrelated, both need to be taken into account.
In treating ICP and CPP the following questions
need to be addressed.

What is the likely cause of increased ICP or reduced

CPP?
What levels of ICP and CPP should be maintained
in this patient?
What methods of treatment should be applied and
in what order?
What effects might the treatment of ICP or CPP
have on cerebral oxygenation and metabolism?
Since ICP and CPP are indirect indices of brain

oxygenation, treatment goals must allow for the
anticipated but usually unmeasurable variability in
CBF and metabolism resulting from the head injury.
For example, pericontusional CBF may be borderline
ischemic (18 ml/100 g/min), while whole-brain CBF is
high. CBF is generally reduced in the first 24 hours
after injury but in the first 46 hours both very low
flows and hyperemic flows (i.e. greater than metabolic
requirements) are found (Salvant and Muizelaar, 1993;
Bouma et al., 1991; Chapter 5).
19.6.2
CLINICAL SIGNS OF RAISED ICP
The traditional clinical signs of raised ICP and of a

mass lesion are in fact quite unreliable (Chapter 8)
Furthermore Chesnut et al. (1994) reported that pupillary asymmetry was a poor indicator of either the
presence or the side of an intracranial lesion. They
found pupil asymmetry of 1 mm or more in only 40%
of patients with a mass lesion of 25 ml or more and
only 33% of patients with pupil inequality of 3 mm or
more had an ipsilateral mass lesion. This study
emphasizes the importance of early CT scanning of all
patients with severe head injury.
Nonetheless, it is important to watch for clinical
signs of brain-stem compression, in so far as this is
possible. Signs of herniation demand treatment whatever the level of ICP, and indeed may not be reliably
related to the level of ICP or to the degree of midline
shift (Marshall, Toole and Bowers, 1983).
19.6.3
391
WHO IS MONITORED?
Some 80% of patients with severe closed head injury

(GCS < 8) will have some degree of raised ICP when
first monitored. In about 40% ICP is greater than
20 mmHg at some time after injury, despite therapy
(Miller et al., 1979).
CT features will identify most patients who have or
are likely to have raised ICP (Chapter 9).
The most important features are:

midline shift;
obliteration of CSF cisterns around the brain stem
(Eisenberg et al., 1990; Marshall et al., 1992).
Even in a patient who is able to talk, a midline shift of

greater than 1.5 cm due to an acute traumatic lesion
will invariably predict later deterioration (Marshall,
Toole and Bowers, 1983).
Patients with severe head injury and a normal initial
CT scan have a 1015% chance of developing raised
ICP later (Eisenberg et al., 1990; Narayan et al., 1982).
Risk factors for later onset of high ICP in this group
were:

age over 40 years;

BP less than 90 mmHg;
unilateral or bilateral motor posturing.
If none or one of these factors was present, the

incidence of raised ICP was only 4%, whereas if two or
more were present it rose to 60%.
However, those with a normal initial scan who were
not monitored should have a repeat CT scan at 1224
hours (Narayan et al., 1982).
In another study, seven of eight patients with severe
head injury (GCS < 8) whose initial CT scan was
normal developed intracranial hypertension of more
than 20 mmHg for 5 minutes, and five of these
developed ICP greater than 30 mmHg. CPP was
reduced below 60 mmHg in five patients (OSullivan
et al., 1994).
Hence, in severely head-injured patients with a
normal initial CT scan, there is an incidence of raised
ICP and critically reduced CPP, which will only be
apparent if they undergo close and continuous monitoring of ICP and CPP or, if they have low risk factors,
close clinical monitoring and repeated CT.
It is our practice for all patients in coma, defined as
a GCS of 8 or less after resuscitation, to be ventilated
for CT scanning. If the CT scan is normal and there are
no other adverse features, then sedation will cease
within 24 hours and their neurological status is
reviewed. Patients who have been transferred from
another hospital by retrieval team are usually intubated and ventilated for stability during transfer and
their initial neurological status may be in doubt
392
(Chapter 8). A decision about continuing monitoring

will also depend on the CT and, if that is normal, on
neurological status after reversal of sedation.
The criteria for ventilation and ICP monitoring
are:

GCS 8 after resuscitation with an abnormal scan,

or with a normal scan and at least two adverse
features i.e. age > 40 years; BP < 90 mmHg after
resuscitation; unilateral or bilateral motor
posturing;
GCS 10 and abnormal CT scan;
a tight brain at operation following removal of a
clot;
ventilation necessary for other injuries, especially
chest injuries.
19.6.4
(a)
(b)
CPP should be maintained above the expected lower

limit of autoregulation, that is above the level at
which CBF will be expected to fall.
In a particular patient, this level will depend upon
a number of factors.
TREATMENT GOALS
ICP levels
A number of studies have considered the level of ICP

that should be treated.
Patients were found to have a higher morbidity
when ICP exceeded 20 mmHg (Miller et al., 1981) and
improved outcome was reported when ICP was
maintained below 15 mmHg (Marshall, 1980; Saul and
Ducker, 1982). Saul and Ducker (1982) suggested that
early and aggressive treatment of modest rises of ICP
might avoid more significant rises late. Contant et al.
(1993) found that critical values independently related
to outcome were ICP more than 25 mmHg, BP less
than 80 mmHg and CPP less than 60 mmHg.
The safe level of ICP in a particular patient may
depend on the accompanying CPP and on the
pathology.
Miller (1992) recommended treatment when ICP
exceeded 25 mmHg in the first 2 days and 30 mmHg
thereafter, providing there were no signs of brain
herniation.
As stated earlier, signs of herniation may occur at
only moderately elevated ICP, e.g. less than 20 mmHg,
particularly with bifrontal or temporal contusions
(Andrews et al., 1988), and these signs demand
immediate treatment whatever the pressure level
(Johnston and Jennett, 1973, Marshall et al., 1983). For
this reason Chesnut, Marshall and Marshall recommended always treating ICP greater than 20 mmHg
lasting more than 20 minutes but, with lesions in the
temporal fossa or deep inferior frontal lobes, treating
ICP greater than 15 mmHg (Chesnut, Marshall and
Marshall, 1993).
It is our present practice to aim for ICP below
20 mmHg but on occasion to accept higher pressures if
CPP can be held over 70 mmHg and there are no signs
of brain-stem herniation.
CPP levels
Pre-existing hypertension. In chronically hypertensive patients the upper and lower limits of
autoregulation are raised, so they tolerate hypotension poorly; i.e. CBF will begin to fall at a
higher level of CPP than in normotensive patients
(Strandgaard and Paulson, 1992). Conversely they
will tolerate higher levels of systemic hypertension without increase in CBF.
Susceptibility of injured brain to additional
injury. Prior injury may reduce brain tolerance to
hypotension and hypoxia (Jenkins et al., 1989)
Hence, more severe injuries may be more susceptible to the effects of secondary injury.
Mechanism of reduction of CPP. CBF will begin
to fall at a higher level of CPP when CPP is
reduced by hemorrhage hypotension rather than by
raised ICP or hypotensive medication (Figure
19.5).
The pattern of brain injury. CPP is a global
measurement and brain injury is typically inhomogeneous. There may be brain regions where
perfusion is marginal even though global CPP and
CBF are adequate. CBF may be reduced around
contusions or beneath subdural hematomas (Salvant and Muizelaar, 1993), by compression of
vessels due to brain shift or by vasospasm. In
areas of ischemia, tissue acidosis may cause maximal vasodilatation so that any fall in global
perfusion will result in a fall in regional blood
flow.
Without direct measurement of CBF, the lower

limit of autoregulation can be estimated from changes in middle cerebral artery flow velocity measured
by transcutaneous Doppler (TCD). As described in
Chapter 13, below a CPP of 70 mmHg, pulsatility
index may rise and jugular venous oxygen saturation
fall, indicating the onset of hypoperfusion (Chan et
al., 1992). Increasing CPP above 70 mmHg did not
improve cerebral oxygenation further, suggesting
that a CPP of 70 mmHg was adequate.
Hence although ICP and CPP are inter-related,
both must be taken into consideration in setting
treatment goals. Brain swelling may progress despite
an adequate CPP and in this situation steps must be
taken to reduce ICP. However, providing there is no
midline shift or evidence of other brain herniation,
and CPP is maintained, moderate increases in ICP
can probably be accepted (Miller, 1992).
TREATMENT
393
Figure 19.5 The lower limit of autoregulation under different conditions. (Source: reproduced from Miller, 1985, with
permission)
19.7
Treatment
19.7.1 CONTROLLING ICP AND CPP DURING

TRANSFER
(a)
From the accident site
Apart from intracranial hemorrhage, the major early

risks to the patient with a head injury are hypoxia
(PaO2 < 60 mmHg) and hypotension (BP < 90 mmHg;
Fearnside et al., 1993; Chesnut et al., 1993b). Cardiopulmonary resuscitation is always the first priority, as
set out in the ATLS program in Chapter 15. The patient
with a severe closed head injury should be presumed to
have raised ICP and factors that may further increase
ICP should be avoided. Urgent attention to airway,
breathing and circulation will reduce cerebral hypoxia
and hypoperfusion (Chapter 15). Normoxia and normocarbia or moderate hypocarbia (PaCO2 = 28
32 mmHg) should be achieved as soon as possible. ICP
can be dangerously increased by overperfusion, hypertension, hyperpyrexia, inadequate sedation and inappropriate anesthetic agents. Intubation and endotracheal suction cause sharp although usually
transitory increases in ICP. In patients with reduced
volume reserve, however, these stimuli may initiate a
prolonged rise in ICP.
Pressure waves may be avoided by adequate fluid
resuscitation to maintain CPP. Sedation should be
sufficient to prevent coughing or gagging during
transport (Chapter 17).
(b)
During transfer between hospitals
Patients who have been intubated and ventilated for

transfer are usually moderately hyperventilated
(PaCO2 = 2832 mmHg). If lateralizing signs or pupillary asymmetry develop, signs that are highly suggestive of a mass, then mannitol may be given. At the
same time, the neurosurgeon at the receiving hospital
should be alerted. However, mannitol should not be
given routinely as it may interfere with volume
resuscitation and furthermore mask the signs of a
developing mass lesion.
19.7.2
DEFINITIVE TREATMENT
Following the initial assessment, CT scanning and

treatment of immediate causes of raised ICP such as
intracranial clots and enlarged ventricles, a decision
must be made as to whether to institute continuous
ventilation in order to treat brain swelling. Basic
monitoring consists of ICP, CPP and central venous
pressures. In some institutions it will also include
jugular venous oxygen monitoring and continuous
transcranial ultrasonic Doppler examination.
Raised ICP is treated by removing mass lesions
and/or increasing the volume available for expansion
of injured tissue. This may be achieved by:
reducing one of the other available intracranial fluid

volumes:
394

CSF by ventricular drainage;

cerebral blood volume by hyperventilation or
mannitol;
brain tissue water content by mannitol;
removing swollen and irreversibly injured brain;
increasing cranial volume by decompression.
(a)
CSF drainage
A ventricular catheter is the most accurate method of

measuring ICP and also permits CSF drainage. Since
reliable, safe and robust pressure-tip catheters that can
be placed easily in the subdural space or within the
brain parenchyma have become available, ventricular
catheters are being used less often. Certainly, when
ventricles are compressed or displaced, they may be
impossible to catheterize, even with ultrasound guidance. However, outcome was reported to be better in
patients in whom ICP was controlled by ventricular
drainage alone (Gahjar, Hariri and Patterson, 1993).
Ventricular catheterization should be used whenever possible and particularly if the ventricles are
enlarged. Drainage of even small amounts of CSF
produces an immediate fall in ICP and a rise in CPP.
However if systolic BP has been increased by a
Cushing response, then sudden release of ICP may
precipitate a fall in BP and CPP. This can be minimized
by i.v. fluid preloading.
CSF drainage pressure should be set at about
20 cmH2O since continuous drainage at a low pressure
may lead to collapse of the ventricles and loss of this
avenue for ICP control.
(b)
Hyperventilation
Hyperventilation will reduce ICP by vasoconstriction

induced by alkalosis. The fall in ICP parallels the fall
in CBV.
The relationship between PaCO2 and ICP is exponential and follows the volumepressure curve. By
measuring the change in ICP induced by change in
PaCO2, vascular reactivity to CO2 can be calculated. In
one study (Yoshihara, Bandoh and Marmarou, 1995),
patients with severe closed head injury (GCS < 8) were
studied within 24 hours of injury. CO2 reactivity to
hypocapnia was reduced by 50% compared with
hypercapnia, indicating that at this phase after injury
resistance vessels were in a state of persistent vasoconstriction, possibly due to vasospasm or compression.
The volume of blood needed to increase ICP was
estimated from the position on the volumepressure
curve, i.e. on the baseline ICP and the compliance
measured as the pressure volume index (PVI). The PVI
is derived from the slope of the semilogarithmic
transformation of the exponential pressurevolume
curve and is the theoretical volume necessary to
increase ICP tenfold. At a baseline ICP of 15 mmHg,

only 2.5 ml of blood were needed to increase ICP to
20 mmHg. This was equivalent to a PaCO2 increase of
about 3 mmHg. When PVI was less than 15 ml, only
4 ml of blood were needed to raise ICP by 10 mmHg,
whereas if the PVI was greater than 15 ml then twice
that amount was necessary. As the brain became
swollen, so the amount of intracranial blood necessary
to increase ICP also fell, as predicted by the exponential volume pressure curve. This study also confirmed
that impaired blood vessel reactivity was associated
with a worse outcome.
Apart from reducing ICP rapidly, hyperventilation
may have other beneficial effects.
It may correct brain and CSF lactoacidosis which is

commonly present and associated with poor outcome from head injury (DeSalles, Muizelaar and
Young, 1987).
It may restore pressure autoregulation (Paulson,
Olesen and Christensen, 1972) and increase global
O2 metabolism (Obrist et al., 1984; Gennarelli et al.,
1979).
It may reverse post-traumatic hyperemia, reduce
the associated raised ICP and normalize cerebral
glucose uptake (Cruz, 1995).
Cerebral resistance vessels are most sensitive to PaCO2

around the normal level of 40 mmHg. Even in severe
injury where pressure autoregulation is impaired or
absent, cerebral vessels usually retain some responsiveness to PaCO2, although it may be less than normal
(Enevoldsen and Jensen, 1978; Muizelaar et al., 1991;
Newell et al., 1996). Hyperventilation is the most rapid
method of reducing ICP other than ventricular drainage. Yoshihara, Bandoh and Marmarou (1995) noted a
15-second delay between initiating respiratory change
and fall in ICP. The peak effect occurs in about 30
minutes.
Since hyperventilation has its maximal vasoconstrictor effect on normal vessels, it has been suggested that it may cause a favorable redistribution of
blood from less injured to more injured or ischemic
areas where the vessels are already fully dilated and
unresponsive to hypocapnia (reverse steal). Any
redistribution, however occurs within the context of
an overall reduction in CBF and indeed it was
suggested some time ago that prolonged hyperventilation might be harmful (Jennett et al., 1977).
This has been now been supported by a randomized
trial, which found that prophylactic hyperventilation
applied to a consecutive group of patients with
severe head injury was associated with a worse
outcome. Prophylactic hyperventilation (PaCO2
25 2 mmHg) was applied for 5 days to patients with
a GCS of 8 or less, 14% of whom had an ICP of more
than 20 mmHg on admission. Hyperventilated
TREATMENT
patients with a GCS of 4 and 5 had worse outcomes

at 3 and 6 months. The bad outcome did not occur
in a similar group treated by hyperventilation and
the base tromethamine (THAM). Ischemic levels of
CBF were not recorded in any group either by xenon
CT or by AVDO2 measurements. However, fluctuations in ICP were most marked in the hyperventilation-alone group, perhaps because of vessel
hypersensitivity to CO2. It was speculated that these
fluctuations and the time spent at ICP greater than
20 mmHg were the determining factors in the
increased mortality in hyperventilated patients, but
that the deleterious effects of hyperventilation might
be avoided by using THAM (Muizelaar et al.,
1991).
Cruz (1995) recommended profound hyperventilation (PaCO2 < 25 mmHg), with continuous monitoring
of cerebral oxygen uptake, in patients with raised ICP
and diffuse swelling on CT, and found that in most
patients oxygen and glucose uptake was optimal
under these conditions.
Other clinical studies have, however, reported a fall
in global AVDO2 to potentially ischemic levels in some
patients during hyperventilation to PCO2 = 25 mmHg
(Sheinberg et al., 1992; Lewis, Myburgh and Reilly,
1995).
Once hyperventilation is initiated, CSF begins to
compensate for the systemic alkalosis immediately. In
a study of non-injured rabbits, vessel diameters and
arterial and CSF pH returned to baseline by 24 hours,
despite continued hypocapnia (Muizelaar et al.,
1988).
CBF measured within 6 hours after head injury by
xenon CT was below the threshold of infarction
(CBF 18 ml/100 g/min) in one-third of patients
(Bouma et al., 1991) and is generally below normal in
more than 50% of patients. This early ischemia may be
worsened by hyperventilation or by a fall in BP.
Hyperventilation might also increase the potential for
ischemia in patients with vasospasm, which is an
increasingly recognized complication of head injury
(Cold, 1989; Yoshida and Marmarou, 1991; Turner et
al., 1984).
Hence, hyperventilation to PaCO2 less than
28 mmHg should be avoided, especially in the first
day after injury, unless all other measures to control
ICP have failed. If profound hyperventilation is
undertaken, the return to normocarbia needs to be
slow because the increase in PaCO2 will now act as a
vasodilatory stimulus and cause a rise in ICP (Havill,
1984). ICP must be monitored carefully during this
phase.
The present view is that PaCO2 levels below
28 mmHg are acceptable only as an emergency, for
about 30 minutes, or when all other means of ICP
control have failed.
395
If prolonged hyperventilation is necessary because

other ICP control measures have failed, then its
potentially harmful effects may be overcome by the
use of THAM. However, the effects of THAM on renal
function must be monitored carefully.
Marked hyperventilation and vasoconstriction, particularly in young patients, may reduce venous return
and cardiac output (Muizelaar et al., 1991) and it is
important to ensure that CPP is maintained, particularly if the patient is nursed head-up (section 19.9.1).
In summary, hyperventilation is an effective method
of ICP control. However, it is best reserved as specific
treatment, e.g. following initial resuscitation until CT
scanning, and as short-term treatment for acute ICP
rises, rather than prophylactically. If profound hyperventilation is used (PaCO2 < 28 mmHg), CPP and PjvO2
should be monitored carefully and ventilation titrated
to keep the PjvO2 above 6065 mmHg, or AVDO2
below 6 ml. However, it is safer to keep the PaCO2 at
3035 mmHg, thus avoiding the potential for
ischemia.
Ventilation parameters
Hyperventilation for rapid control of ICP can be
achieved by increasing stroke rate and tidal volume.
However high ventilatory rates and tidal volumes
may also reduce systemic arterial pressure and hence
CPP. This may also occur with vigorous hand ventilation or bagging, often a very effective way of rapidly
reducing ICP. However, this may be dangerous
because PaCO2 is not controlled and may fall to
1520 mmHg.
Positive end-expiratory pressure (PEEP) improves
oxygenation without the need for a high inspired
oxygen fraction (FiO2). It has been suggested that the
higher end-expiratory pressure of PEEP might raise
venous pressure and hence ICP. However, studies in
patients have suggested that ICP is not affected, even
with intrathoracic pressures of up to 40 cmH2O sustained for up to 18 hours (Frost, 1977; Shapiro and
Marshall, 1978; Cooper, Boswell and Choi, 1985).
Others have reported that PEEP is associated with
increased ICP in patients with reduced compliance
(Apuzzo et al., 1977; Burchiel, Steege and Wyler, 1981).
Hence the effects of PEEP are hard to predict and, if it
is used, ICP and CPP should be careful monitored. For
a more extensive discussion on ventilation in head
injury, refer to Chapter 17.
(c)
Osmotic agents and diuretics
Osmotic agents have been one of the principal means

of treating raised ICP for many years (Weed and
McKibben, 1919), yet their principal modes of action
are still debated
396
Mannitol
Mannitol, a six-carbon hexhydric alcohol of the sugar
mannose, molecular weight 182, is a powerful osmotic
diuretic and draws fluid into the vascular compartment, thereby increasing circulating blood volume
and reducing blood viscosity. It is not metabolized and
the intact bloodbrain barrier is relatively impermeable to it, although in high concentrations or after
prolonged use the normal bloodbrain barrier may
open, allowing it or other large molecules to enter the
extracellular space.
It was originally proposed that the principal action of
mannitol was to reverse the bloodbrain osmotic
gradient, thereby drawing water from the brain
extracellular space. On this premise the aim of
treatment has been to maintain an osmotic gradient of
1020 mosmol. This requires repeated doses of mannitol with consequent problems of progressive dehydration, hypotension and prerenal failure. CVP and
urine output need to be carefully watched, particularly
if mannitol is combined with the renal loop diuretic
frusemide. Serum osmolarity needs to be checked and
hypokalemia corrected with potassium supplements.
Normovolemia should be maintained (Chapter 17).
Evidence of a cerebral dehydrating action has come
from clinical studies. MRI studies have suggested that
mannitol reduces tissue water in edematous but not in
normal brain (Bell et al., 1987). In a group of patients
with traumatic intracerebral hematomas and contusions, mannitol increased the specific gravity of white
matter, probably by reducing brain water content
(Nath and Galbraith, 1986). On the other hand, in noninjured cats there was no change in brain water
content after mannitol and it was suggested that
mannitol acted by reducing CSF volume (Takagi et al.,
1983). Also in non-injured cats, pial artery constriction
was observed and taken to indicate autoregulation to
change in viscosity (Muizelaar et al., 1983).
In patients with severe head injury, when autoregulation was intact, mannitol reduced ICP by 27.2%
without changing CBF. However, when autoregulation was defective ICP fell by only 4.7% and CBF
increased. This strongly suggests that vasoconstriction, as well as cerebral dehydration, is a major action
of mannitol (Muizelaar, Lutz and Becker, 1984). In
another study of patients with severe head injury,
mannitol consistently reduced ICP and increased CBF
and CPP, the greatest increases in CBF after mannitol
being seen in patients with low initial CPP and very
high ICP, i.e. those most likely to have impaired
autoregulation (Mendelow et al., 1985).
It is probable that the initial rapid reduction in ICP
is due to plasma expansion and decreased blood
viscosity which leads to constriction of the cerebral
vasculature to maintain constant CBF. This rapid
response is more likely with bolus injection rather

than infusion. The more prolonged reduction in ICP
may be due to the osmotic effect.
Mannitol may have other possible beneficial
actions.

It may reduce red blood cell rigidity, thus allowing

them to penetrate small vessels and areas of
marginal perfusion more easily (Burke et al., 1981).
It may scavenge free radicals, which have been
implicated in ischemic brain damage.
It may reduce CSF production (Sahar and Tsipstein,
1978) and a reduction in CSF volume has been
reported experimentally (Tagaki et al., 1983). However, as argued by Nath and Galbraith (1986), this is
unlikely to be an important mechanism for reducing
raised ICP when CSF has already been displaced.
Treatment based on the dehydration theory aims at

providing a constant osmotic gradient. However,
treatment based on the theory of viscosity autoregulation employs the opposite strategy i.e. fluid losses
are replaced accurately to maintain normal osmolality
and normovolemia (Rosner, 1993).
Mannitol dose
Mannitol is usually given in a 20% solution in bolus
doses, rather than as a continuous infusion. ICP falls
within 510 minutes. The maximum effect occurs in
about 60 minutes and the total effect may last 34
hours (James et al., 1977).
Bolus administration minimizes hemoconcentration
and prolongs the effects. Boluses of 0.250.5 g/kg
(given over 1020 min) may be used and repeated
depending on the response. A dose of 0.25 g/kg seems
as effective as 1 g/kg in reducing ICP but may not act
as long (Marshall et al., 1978). A Foley catheter should
be inserted and fluid losses should be replaced.
Central pressures should be carefully watched and the
patient kept euvolemic. Even so, mannitol may
become less effective with repeated doses because:
replacing the diuresis becomes increasingly difficult

and hemoconcentration develops; at serum osmolality of greater than 320 mosmol/l there is a risk of
uremia and prerenal renal failure (Stuart et al.,
1970);
as osmolality rises, the increase in blood viscosity
causes a fall in CPP, which will lead to cerebral
vasodilatation and hence to a rise in ICP;
mannitol may enter the extracellular space by
opening the normal bloodbrain barrier or where it
is disrupted, taking fluid with it and cause a
rebound rise in ICP (Kaufmann and Cardoso,
1992). In practice this was seen more often with urea
and is uncommon with mannitol (James et al., 1977).
TREATMENT
However the gradual increase in extracellular fluid

osmolality reduces the osmotic gradient and renders mannitol less effective. This change will occur
more slowly if mannitol is given in boluses rather
than as an infusion.
It is our practice to avoid osmolality above 320 mosmol/l. Calculated osmolality may be significantly less
than measured osmolality in the presence of mannitol.
Complications of mannitol

Hyperosmotic prerenal renal failure (Cottrell et al.,

1977; Feig and McCurdy, 1977).
Electrolyte disorders. Hypokalemia usually occurs
after a few days of treatment and requires potassium supplements.
Dehydration and hypotension. Reduced vascular
volume leads to hypotension and the added risk of
cerebral ischemia due to a fall in CPP. This is a
particular risk in those with multiple injuries, the
aged and those with pre-existing cardiac disease.
These patients need careful monitoring of central
pressures (Chapter 16).
Expansion of an intracranial hemorrhage. By
shrinking the brain, osmotic diuretics may assist the
continued expansion of an intracranial clot. When
the effects of mannitol wear off, there may be an
acute deterioration. This risk is not proven and the
acute deterioration under these circumstances may
simply represent the inevitable enlargement of the
clot temporarily masked by the effects of the
mannitol. In a patient who may harbor a clot,
mannitol is very effective in allowing time for
transfer, CT scanning and surgical treatment.
Renal diuretics
Frusemide (furosemide) and other renal loop diuretics
such as ethacrynic acid increase plasma osmolarity by
diuresis (Cottrell et al., 1977; Cottrell and Marlin,
1981). They may reduce CSF formation directly (Sahar
and Tsipstein, 1978). In combination with mannitol,
frusemide produces a slightly greater reduction in ICP
and increases the duration of its effect (Pollay et al.,
1983; Wilkinson and Rosenfeld, 1983). Given alone,
frusemide and other renal diuretics do not reliably
reduce ICP.
The combination of frusemide and mannitol increases the risk of hypovolemia and renal failure and
should be reserved for patients with actual or incipient cardiac failure or pulmonary edema. When used it
is important to avoid excessive dehydration and
sodium loss.
Dose: frusemide 2040 mg i.v.; monitoring: serum
and urinary electrolytes and fluid balance.
(d)
397
Barbiturates
High doses of barbiturates will frequently lower ICP

even when osmolar therapy and hyperventilation
have failed. Barbiturates and the other hypnotic drugs
(such as etomidate and propofol), act by reducing
cerebral metabolic rate for oxygen (CMRO2) and
producing a coupled reduction in CBF. The consequent reduction in CBV leads to a parallel fall in ICP
(Michenfelder, 1974). Some degree of retained vascular
reactivity to CO2 (Nordstrom et al., 1988; Messeter et
al., 1986) and brain electrical activity must be present
for barbiturates to act. The maximum effect has been
correlated with EEG burst suppression, again underscoring the relationship between reduced metabolic
rate and fall in ICP.
Anesthetic equivalent doses of hypnotics are necessary and hence ventilation and ICP monitoring are
mandatory. Systemic hypotension and pulmonary
failure may occur. Cardiovascular status, including
central venous and pulmonary artery wedge pressures,
must be carefully monitored.
If BP falls and cannot be maintained by volume
transfusion then CPP must be maintained by inotrope
infusion. It is safest to set up an inotrope infusion in
anticipation whenever barbiturates are given.
Barbiturate therapy is generally reserved for the
1215% of patients in whom other forms of intensive
medical therapy for raised ICP have failed. Two
uncontrolled trials reported an improved outcome
with barbiturates (Marshall, Smith and Shapiro, 1979;
Rea and Rockswold, 1983). In one trial about half those
patients who failed to respond to other ICP therapy
responded to barbiturates. The mortality in the barbiturate responders was 33% compared with 75% in the
non-responders: 69% of survivors were classed as good
recovery/moderate disability at follow-up (Rea and
Rockswold, 1983). Similar results were reported in the
earlier study (Marshall, Smith and Shapiro, 1979). In a
controlled trial, patients were randomized to barbiturate therapy after other medical management for raised
ICP had failed. ICP was controlled in about 30% of the
barbiturate group compared to 16% of those treated
with conventional therapy (Eisenberg et al., 1988).
Thus, in each trial, barbiturates were able to control ICP
in some patients, even when all other medical means
had failed and failure to respond to barbiturates carried
a high mortality and morbidity. In contrast to these
studies, when barbiturates were given prophylactically, i.e. to all patients, no benefit was found (Ward et
al., 1985; Schwartz et al., 1984).
Thus there appears to be a small subgroup of
patients with severe head injury in whom barbiturate
treatment is beneficial.
At present barbiturates are generally used in the
following situations:
398
as part of a stepwise protocol for ICP control, once

other medical management has failed (see Figure
19.6).
for acute intraoperative brain swelling (Shapiro et
al., 1973; Bricolo and Glick, 1981).
It has been suggested that hypnotic therapy, using

barbiturates and other anesthetic drugs, may be
specifically indicated in patients in whom ICP is
elevated by vascular or hyperemic brain swelling, as
indicated by decreased arteriovenous oxygen difference (SjvO2 > 75%) and preserved electrical activity
(Dearden and McDowall, 1985; Miller, Piper and
Dearden, 1993). Miller et al. have recommended that in
this group of patients, hypnotic therapy should be
given early, since barbiturate and other hypnotic
agents will only work if vascular reactivity to arterial
CO2 is still present (Nordstrom et al., 1988) and brain
electrical activity is preserved (Bingham et al., 1985).
Hyperemic diffuse brain swelling is found most
often in young patients, who usually show an ICP
trace in which the amplitude of the arterial pressure
oscillation is more than twice that of the respiratory
oscillation (Miller, 1992; Dearden and Miller, 1989;
Aldrich et al., 1992). Chesnut, Marshall and Marshall
(1993) have also suggested that young patients with
markedly increased ICP and CT grade 3 or 4 swelling
(Chapter 9) may sometimes benefit from early barbiturate treatment. However, they emphasized that the
evidence for such a specific benefit is anecdotal and
early aggressive barbiturate treatment cannot be generally recommended.

Hypothermia. Moderate hypothermia may be beneficial. However, below a core temperature of 32C the
risk of cardiac complications increases markedly.
Complications of prolonged coma such as decubitus ulcers and venous thrombosis.
Gastric stasis. This usually necessitates parenteral
feeding.
Inability to diagnose brain death by clinical
criteria for 24 hours or more after treatment has
stopped.
Monitoring
Barbiturate coma should only undertaken in an
experienced Intensive Care Unit. Monitoring should
include ICP, BP via an arterial line, CPP, central venous
and pulmonary artery catheters and core temperature.
Pupillary responses
Moderate levels of barbiturates cause small pupils
that are sluggish or non-reacting. They may still
dilate in response to brain-stem pressure. Larger doses
may produce mid-position to 5 mm non-reacting
pupils (Chapter 8), which may produce a state
indistinguishable from brain death (Chesnut, Marshall and Marshall, 1993). Widely dilated pupils, particularly if unequal, should always suggest the possibility of a mass lesion. During reduction of barbiturate
dosage, pupil reactions are the first neurological
function to recover and motor responses are the last
(Chesnut and Marshall, 1993).
Actions of barbiturates
Endpoints
The actions of barbiturates include:

a decrease in metabolic rate with coupled reduction

in CBF;
free radical scavenging and reduction in lipid
peroxidation;
a direct increase in vasomotor tone.
Complications and contraindications

Hypotension. This is the main complication of

barbiturate therapy and is due to reduced systemic
vascular resistance and myocardial depression.
Existing cardiovascular instability, hypotension and
hypovolemia increase the risks of barbiturate therapy, lowering the safe dose that can be given and
therefore reducing the likelihood of ICP control.
Consequently patients must be euvolemic before
barbiturates are started.
Anergy, so that infection may occur without leukocytosis or fever.
ICP control: this is the principal endpoint;

EEG: EEG burst suppression is a useful guide to
optimal barbiturate dosage; there is no benefit in
achieving higher barbiturate levels;
Serum barbiturate levels: most authors advocate
serum levels of about 3040 mg/dl for optimal ICP
control; however it is doubtful whether routine
serum levels are really necessary except to determine when clinical tests of brain function are likely
to be valid.
Dose and duration

Thiopentone. An i.v. bolus of thiopentone (e.g.

250 mg) will reduce ICP within seconds, but the
effect will last for only 1520 minutes. This may be
a useful test to see whether a patient is likely to
respond to prolonged barbiturate therapy. Once
tissue levels have built up after continuous or
repeated doses, the half-life is similar to that of the
longer-acting barbiturates more commonly used.
TREATMENT
Pentobarbitone.
Loading dose 10 mg/kg over 30 min or 5 mg/kg
hourly for 3 hours;
Maintenance 1 mg/kg/h or adjusted to serum
level of 3040 mg/dl or to burst suppression
(Eisenberg et al., 1988).
399
(tirilazad), has shown promise in experimental head

injury and is presently undergoing phase 3 clinical
trials (Hall et al., 1988). The present status of this
and other neuroprotective agents is discussed in
Chapter 21.
Gamma hydroxybutyrate
Reducing barbiturate therapy

Once ICP control has been achieved for 2436 hours
barbiturates may be reduced. Despite their long halflife it is preferable to reduce them slowly in order to
prevent ICP overshoot.
(e)
Other agents
Hypertonic saline
A 15% solution of sodium chloride was the first agent
used experimentally to reduce raised ICP. It was
presumed to act by osmosis but the effect was found
to be short-lived and it gained no place in clinical
practice (Weed and McKibben, 1919). However hypertonic saline has more recently been found to be useful
in patients who no longer respond to mannitol and are
developing uremia. Intravenous hypertonic saline
(30% or 5 mmol/ml) over 10 minutes may produce a
prolonged reduction in ICP without diuresis and
improve renal function in dehydrated patients (Worthley, Cooper and Jones, 1988). Small amounts of
hypertonic saline may help to prevent hypovolemia in
patients requiring large amounts of mannitol (Ropper,
1993). In animal models of focal brain injury and
hemorrhagic shock, resuscitation with hypertonic
saline increased CBF at a lower ICP than did Ringers
solution (Walsh, Zhuang and Shackford, 1991).
Dose: hypertonic saline (5 mmol/ml) up to 50 ml
over 1015 min; monitoring; serum Na+; urinary
output.
Steroids
Because of their effectiveness in treating the brain
swelling associated with brain tumors and abscesses,
steroids were at one time widely used in traumatic
brain swelling. However, several trials, some using
very high-dose steroid regimes, have failed to identify
a benefit (Saul et al., 1981; Dearden et al., 1986;
Gudeman, Miller and Becker, 1979; Cooper et al.,
1979).
The was a single exception (Gobiet, 1977). However
when this study was reanalyzed the changes after
steroids were found not to be statistically significant
(Cooper et al., 1979). Steroids probably have no place
in the routine management of closed head injury.
However a non-glucosteroid, the 17 amino steroid
Gamma hydroxybutyrate is an analog of the inhibitory neurotransmitter gamma-amino butyrate acid

(GABA) and is used as a short-acting anesthetic agent.
It is a potent vasoconstrictor and reduces cerebral
metabolic rate. When given as a bolus it has been
shown to be as effective as sodium thiopentone in
reducing ICP and with a longer duration of action. As
with sodium thiopentone it will reduce systemic blood
pressure and CPP must be carefully monitored (Leggate, Dearden and Miller, 1986).
Dose: 60 mg/kg body weight; monitoring: cerebral
perfusion pressure, since it may reduce systemic blood
pressure.
Indomethacin
Indomethacin, a prostaglandin inhibitor, may reduce
raised ICP even though other medical treatments,
including barbiturate therapy, hyperventilation and
mannitol have failed (Jensen et al., 1991; Biestro et al.,
1995). Indomethacin may act by restoring the capacity
for autoregulation and hence the response to hyperventilation and to mannitol. If indomethacin infusion
is withdrawn suddenly there may be a rebound rise in
ICP (Biestro et al., 1995).
Dose: bolus 50 mg in 20 min; infusion 1030 mg/h.
Lidocaine
Lidocaine is a local anesthetic which may be given as
an endotracheal aerosol to prevent ICP rises during
intubation (Yano et al., 1986). It may also be given
intravenously for the same purpose (Donegan and
Bedford, 1980). This may cause myocardial depressant
and requires ECG monitoring. Lidocaine has been
used to treat refractory status epilepticus, but high
doses may lower the seizure threshold. There is some
evidence that it will lower raised ICP refractory to
other methods but it has not been widely used because
of the potential complications. At the doses used
preintubation, such complications are unlikely.
Dose: 11.5 mg/kg i.v. within 6090 s of intubation,
or as a 4% intratracheal aerosol; monitoring: ECG.
Tromethamine (THAM)
The base tromethamine has been proposed as a means
of correcting the lactic acidosis frequently found in
400
patients after head injury (Rosner and Becker, 1984b).

In a randomized trial of the effects of prophylactic
hyperventilation, Muizelaar et al. (1991) found that the
potentially adverse effects of hyperventilation were
prevented and that ICP was more stable. In a further
prospective, controlled study, the addition of THAM
to conventional treatment resulted in better ICP
control. However there was no improvement in
outcome (Wolf et al., 1993). Hence, THAM may be
useful in controlling raised ICP but its place needs to
be more clearly defined in further studies. THAM may
impair renal function and this needs to be carefully
monitored.
Dose: 1 mg/kg/h of 0.3 mmol/l solution; monitoring: renal function.
(f)
Hypothermia
Moderate hypothermia (core body temperature

3034C) was recommended many years ago in the
treatment of head injury (Sedzimir, 1959) but after
some initial interest it was little used. However,
recent experimental evidence suggests that moderate
hypothermia is cytoprotective after severe global
ischemic insults. Hypothermia reduces the effects of
global ischemia in head injury. It reduces glutamate
release and prevents the depletion of high-energy
phosphate compounds (Chopp et al., 1991; Busto et al.,
1989). A significant improvement in recovery was
found following fluid percussion injury in rats when
hypothermia was initiated up to 15 minutes after the
injury (Clifton et al., 1991). Similarly, cerebral damage
following epidural compression was reduced in a
canine model (Pomeranz et al., 1993).
Recent clinical studies have suggested a benefit
from moderate hypothermia used prophylactically for
2448 hours following severe head injury (Marion et
al., 1993), or when ICP is not responding to medical
therapy, including high-dose barbiturates (Shoizaki et
al., 1993). In both studies there was a reduction in ICP,
CBF and CMRO2 and a trend towards improved
outcome in the hypothermic groups, more evident in a
later report (Marion and Carlier, 1995). There has been
no greater incidence of significant complications with
moderate hypothermia in the head injury studies
reported so far. Hypothermia is well known to cause
coagulopathy, but Resnick, Marion and Darby (1994)
found no greater incidence of measurable coagulopathy or of delayed intracerebral hemorrhage in a small
study. Metz et al. (1996) found that platelet count fell
during 24 hours of hypothermia and for up to 24
hours after rewarming, but there was no significant
change in coagulation tests. Four of ten patients
developed signs of pancreatitis, which reversed with
rewarming. Thus hypothermia may be a safe additional treatment in patients with high ICP. Currently, a
large multicenter placebo-controlled trial of prophylactic moderate hypothermia (32 2C) is under way
in neurosurgical centers in the USA to test its effects
upon outcome and ICP control.
(g)
Hyperbaric oxygen
Hyperbaric oxygen reduces CBF and ICP by vasoconstriction while preserving a high oxygen delivery to
the tissues. Those who respond to hyperbaric oxygen
are also likely to respond to hyperventilation, which
has fewer logistical problems. Hence, even where
pressure chambers are available, hyperbaric oxygen is
rarely used in the management of patients with head
injury. In one prospective study, hyperbaric oxygen
reduced mortality after severe head injury by nearly
50%, but did not increase the favorable outcome rate
among survivors (Rockswold et al., 1992).
19.8
Surgical treatment
19.8.1
INDICATIONS (Chapter 20)
When there is clear evidence of neurological deterioration and a CT scan shows an intracranial mass lesion,
the need for urgent removal is clear. The results of
surgery for a mass lesion are directly proportional to
the neurological state at the time of evacuation (Becker
et al., 1977; Mendelow et al., 1983). Unnecessary delay
in achieving surgical evacuation and lowering ICP
must be avoided under any circumstances. Rarely, if
deterioration is rapid and there are clear localizing
features, then the CT scanner may be bypassed and
the patient taken directly to surgery (Chesnut et al.,
1994).
In comatose patients with hemispheric or diffuse
swelling and a small surface clot the need for surgical
removal may be less clear. However the removal of
even a small surface clot 35 mm thick, extending over
several cuts, may improve compliance and make ICP
management easier. Moreover, duraplasty and craniotomy itself may reduce ICP dramatically in these
circumstances. These potential advantages must be
balanced by the risk of brain herniation, venous
compression and increased swelling, even with duraplasty. When the patient is stable, there is only
moderate brain shift, i.e. 5 mm or less, and ICP is less
than 20 mmHg it may be reasonable not to treat small
surface clots by surgery initially. However, such
patients need careful ICP monitoring and CT scans
should be repeated routinely. In general we believe
that it is better to err towards removing too many clots
rather than to tolerate high ICP. Knuckey, Gelbard and
Epstein (1989) noted that patients with small extradural hematomas were more prone to late deterioration if the hematoma was diagnosed within 6 hours of
MANAGEMENT OF RAISED ICP AND REDUCED CPP
injury than if it was diagnosed later. Deterioration in

the latter group was therefore most often due to other
factors.
Delayed intracerebral hemorrhage
Delayed intracerebral hemorrhages have been recognized for many years. However, the CT scan has
enabled more precise definition and distinction from
swelling around or enlargement of a previous hemorrhage (Gentleman and Macpherson, 1989). A delayed
intracerebral hemorrhage is defined on CT as a lesion
of increased attenuation occurring in a previously
normal area. They are uncommon, being only 2.6% of
traumatic intracranial hematomas in the series of
Gentleman and Macpherson, and developed within
hours to days. Some 80% were detected within 48
hours of injury. Less than half needed surgical
evacuation. Their potential occurrence makes repeat
CT scan mandatory whenever there is an unexplained
clinical deterioration, failure to improve or a rise in
ICP, even if the initial scan was normal. It was
suggested many years ago that delayed hemorrhage
occurred in areas of focal injury where vasoparalysis,
congestion and increased capillary permeability resulted in multiple areas of diapedetic hemorrhage (Evans
and Scheinker, 1946). Delayed hemorrhage has been
associated with clotting abnormalities, in particular
disseminated intravascular coagulation (Kaufman et
al., 1980), hypotension and hypoxia (Ninchoji et al.,
1984).
19.9 Plan of management of raised ICP

and reduced CPP
Although the evidence for certain components of
head injury management is sufficient to provide
management guidelines, there is no controlled study
of any particular combination of treatments that
would constitute a recommended protocol or critical
pathway. The treatment plans outlined here are
those of two separate institutions and represent
somewhat different approaches to the same
problem.
It must be recognized that head injury is a
dynamic process that varies in its pathophysiology
between patients, with time, and in its regional
distribution. The usual practice of using a single
protocol of management for all patients is not ideal.
However at present it is difficult to measure cerebral
blood flow and metabolism directly and treatment
tends to be based on probabilities rather than on
specific information. Furthermore there are still few
effective treatments available for brain swelling and
increased ICP.
401
In order to tailor the treatments that are available

to the specific and evolving pattern of injury in a
particular patient, continuous bedside monitoring of
brain metabolism and substrate delivery are
required. Methods of providing this information are
becoming available but until it is possible to distinguish specific causes for raised ICP and reduced
CPP in individual patients, treatment regimes will
remain sequential.
Despite these limitations it is important that a
generally consistent plan of approach be adopted. The
goals of management should be:

to identify the dominant pathology;

to target treatment to the cause;
to avoid secondary insults.
There are three levels of observation upon which

treatment is based:

clinical;
continuous multimodality monitoring;
imaging.
19.9.1
NURSING POSITION
Patients are usually nursed 30 head up after intracranial surgery and head injury in order to reduce
venous congestion and brain swelling. In most
patients ICP is lower in this position.
Rosner and Coley (1986) found that in many
patients, although ICP fell with head elevation, there
was also a significant fall in CPP. A fall in CPP was less
likely when circulating volume was adequate. However this variable response to positional change
highlights the need for adequate fluid replacement
and for CPP monitoring (Chapter 17).
19.9.2 EXTRACRANIAL CAUSES OF RAISED ICP
(Table 19.1)
ICP may be increased by a number of extracranial

factors, including nursing procedures such as endotracheal suction and physiotherapy. Before concluding
Table 19.1 Factors that affect intracranial pressure

Increase
Hypercarbia
Hypoxia (PaO2 < 50 mmHg)
Epilepsy
Hyperthermia
Decrease
Hypocapnia
Hyperoxia
Hypothermia
Barbiturates
Venous congestion due to intrathoracic pressure; fluid

overload
402
that raised ICP is due to intracranial factors, a

checklist should be considered.

Airway and head position. Ensure that there is no

venous obstruction.
Arterial blood gases. Ensure that PaCO2 is
32 2 mmHg.
Core temperature. A rise of 1C can increase
metabolic rate by 10%, thus increasing ICP by
several millimeters of mercury. Core body temperature must be kept below 38C.
Occult seizures. Anticonvulsants have been recommended in all ventilated patients with severe closed
head injury because seizures occurring in heavily
sedated and paralyzed patients are difficult to
detect without EEG monitoring, yet they can lead to
marked increases in CBF and ICP (Marienne, Robert
and Bagnet, 1979)
Sedation level. Consider whether the patient is
becoming more conscious and fighting the
ventilator.
19.9.3
TREATING RAISED ICP AND REDUCED CPP
Treatment protocols should begin with simple nursing

procedures and proceed stepwise to more complex
methods, depending on the response at each stage.
Recent studies from Edinburgh and other centers have
helped to define what constitutes a significant secondary insult and hence to set treatment goals (Jones et al.,
1994). When methods of reliably monitoring cerebral
oxygenation and metabolism become available, these
parameters can be taken into account.
Treatment endpoints should take into account the
time since injury and nature of the pathology. There
also needs to be a balance between treating every
change in ICP or CPP from normal and using up the
available and limited methods of treatment
prematurely.
Treatment should be on the background of:

cardiopulmonary homeostasis and euvolemia;

adequate sedation.
These aspects are covered in Chapters 16, 17 and 18.

Table 19.2 indicates the changes in the emphasis of
treatment that have occurred in recent years.
Table 19.2 Changing emphasis in management of acute

head injury: SjvO2 = jugular venous oxygen saturation;
PEEP = positive end-expiratory pressure; ICP =
intracranial pressure (Source: adapted from Myburgh and
Lewis, 1996.)
Previous strategies
Current strategies
Reduce intracranial
pressure
Maintain cerebral perfusion

pressure and reduce
Elective dehydration
Euvolemia
Routine osmotherapy
Selective osmotherapy
(< 300 mosmol/l)
Routine hyperventilation:
(PaCO2 < 30 mmHg)
Normocapnia: acute
hyperventilation to control
rises in ICP prior to
imaging, maintain SjvO2
above 55%
Routine barbiturates
Limited barbiturates
Routine corticosteroids
No corticosteroids
Avoid sedation, use muscle

relaxants
Avoid muscle relaxants,

ensure sedation and
analgesia
Avoid PEEP
Use PEEP to maintain PaO2
ICP monitoring
intraventricular or subdural
fluid-filled catheters
ICP monitoring
intraventricular or
intraparenchymal solidstate systems
Refractory intracranial hypertension

If ICP rises above the desired treatment level and CPP
falls despite first-tier therapy, the second-tier therapies
to be considered are:

induced hypertension;
barbiturate therapy;
decompressive craniotomy;
hypothermia;
marked hyperventilation (Guidelines for the Management of Severe Head Injury, 1995).
Before proceeding to more complex treatments it is

always important to recheck for easily reversible
causes of raised ICP, as set out in section 19.9.2, and to
consider repeating the CT scan.
Treatment algorithms
First-tier therapy
If ICP rises the first steps are:

check blood gases to ensure normoxia and

normocarbia;
drain CSF;
administer mannitol;
institute moderate hypothermia.
The staircase protocol used at the Medical College of

Virginia is shown in Figure 19.6. The indications for
decompressive craniotomy within this protocol are
summarized in Chapter 20. The protocol currently
used at the Royal Adelaide Hospital (Figure 19.7) is
based on maintaining ICP below 20 mmHg and CPP
above 70 mmHg. However, it also takes into account
SjvO2 and TCD data in distinguishing hyperemia and
MANAGEMENT OF RAISED ICP AND REDUCED CPP
403
Figure 19.6 The staircase approach to management of ICP and CPP at the Medical College of Virginia. Treatment is
initiated when ICP > 20 mmHg or CPP < 70 mmHg for more than 10 minutes. Inotropes are used at any time, aiming to keep
CPP above 7080 mmHg.
determining the lower limit of autoregulation in order

to define the optimal level of CPP (Lewis, Reilly and
Myburgh, 1996). If jugular bulb oximetry is used, a
careful protocol must be observed to minimize false
readings.
Duration of monitoring
ICP and CPP monitoring should continue until the
patient is stable without treatment and the risk of a
later deterioration is low. The period of vulnerability
to secondary insults may be up to two weeks after
injury. Studies from Edinburgh found two peaks of
reduced CPP. The first, within 24 hours, was most
often due to hypotension. The second occurred 56
days after injury, usually as a result of increased ICP
(Cortbus et al., 1995). Unterberg et al. (1993) reported
the peak incidence of high ICP to be 13 days after
injury. Late rises occurred after 310 days, particularly
in patients with multiple contusions, and coincided
with a marked leukocytosis. These reports suggest
that patients with severe head injury and abnormal
scans should be monitored for at least 6 days, and that
weaning from the ventilator should occur slowly.
If the ICP is low from the outset or has been low
without treatment for 24 hours, and the CT scan does
not show a mass lesion or diffuse swelling, then it is
reasonable to cease sedation. If ICP and CPP remain
satisfactory, then monitoring may be withdrawn. As

noted earlier, it is important to withdraw barbiturate
therapy slowly.
Intractable ICP: who manages the patient?
One of the most important advances in the management of head injury in recent years has been the
formation of integrated multidisciplinary services that
extend from the accident site to the trauma hospital
and indeed beyond to the phase of rehabilitation. The
intensive care specialist has played a key role in all
stages of acute care. With the emphasis on intensive
care management for patients, the intensivist needs, in
addition to expertise in cardiorespiratory care, a
detailed knowledge of the pathophysiology of head
injury and the effects of drugs knowledge that needs
to be shared by the neurosurgeon. The neurosurgical
role extends beyond removing hematomas. The neurosurgeon should be involved in determining treatment protocols, assessing and deciding the treatment
plan for each patient and continuing care after the
acute phase. During the acute phase neurosurgeons
and intensivists should work in cooperation in determining treatment steps and in counseling relatives.
This is particularly important when the question of
withdrawing treatment arises, which is discussed in
more detail in Chapter 21.
404
Figure 19.7 Algorithm for the management of acute head injury based on CPP and ICP and taking into account SjO2 and
TCD measurements. ICP = intracranial pressure; CPP = cerebral perfusion pressure; CSF = cerebrospinal fluid; RAP = right
atrial pressure; MAP = mean atrial pressure; SjO2 = jugular venous oxygen saturation; TCD = transcranial Doppler. (Source:
reproduced from Myburgh and Lewis, 1996, with permission)
REFERENCES
19.10 References
Aldrich, E. F., Eisenberg, H. M., Saydjari, C. et al. (1992) Diffuse brain swelling
in severely head-injured children. A report from the NIH Traumatic Coma
Andrews, B. T., Chiles, B. W., Olsen, W. L. and Pitts, L. H. (1988) The effect of
intracerebral hematoma location on the risk of brain-stem compression and
on clinical outcome. Journal of Neurosurgery, 69, 518522.
Apuzzo, J. L., Wiess, M. H., Petersons, V. et al. (1977) Effect of positive end
expiratory pressure ventilation on intracranial pressure in man. Journal of
Bell, B. A., Smith, M. A., Kean, D. M. et al. (1987) Brain water measured by
magnetic resonance imaging. Correlation with direct estimation and
changes after mannitol and dexamethasone. Lancet, i, 6669.
Biestro, A. A., Alberti, R. A., Soca, M. C. et al. (1995) Use of indomethacin in
brain-injured patients with cerebral perfusion pressure impairment: preliminary report. Journal of Neurosurgery, 83, 627630.
Bingham, R. M., Procaccio, F., Prior, P. F. and Hinds, C. J. (1985) Cerebral
electrical activity influences the effects of etomidate on cerebral perfusion
pressure in traumatic coma. British Journal of Anaesthesia, 57, 843848.
Brgesen, S. E. and Gjerris F. (1989) CSF production in patients with increased
intracranial pressure, in Intracranial Pressure VII, (eds J. T. Hoff and A. L.
Betz), Springer-Verlag, Berlin, pp. 332335.
metabolism after severe traumatic brain injury: the elusive role of ischemia.
Bricolo, A. P. and Glick, R. P. (1981) Barbiturate effects on acute experimental
Burchiel, K. J., Steege, T. D. and Wyler, A. R. (1981) Intracranial pressure
changes in brain-injured patients requiring positive end-expiratory pressure ventilation. Neurosurgery, 8, 443449.
Burke, A. M., Quest, D. O., Chien, S. and Cerri, C. (1981) The effects of
mannitol on blood viscosity. Journal of Neurosurgery, 55, 550553.
Burrows, G. (1846) Disorders of the Cerebral Circulation, London.
Busto, R., Globus, M. Y. T., Dietrich, W. D. et al. (1989) Effect of mild
hypothermia on ischemia-induced release of neurotransmitters and free
fatty acids in rat brain. Stroke, 20, 904910.
Chan, K. H., Miller, J. D., Dearden, N. M. et al. (1992) The effect of changes in
and jugular bulb venous oxygen saturation after severe brain injury. Journal
Chesnut, R. M. and Marshall, L. F. (1993) Management of severe head injury,
in Neurological and Neurosurgical Intensive Care, 3rd edn, (ed. A. H. Ropper),
Raven Press, New York, pp. 203246.
Chesnut, R. M., Marshall, L. F. and Marshall S. B. (1993) Medical management
of intracranial pressure, in Head Injury, 3rd edn, (ed. P. R. Cooper), Williams
& Wilkins, Baltimore, MD, p. 225246.
Chesnut, R. M., Marshall, L. F., Piek, J. et al. (1993a) Early and late systemic
hypotension as a frequent and fundamental source of cerebral ischaemia
following sever brain injury in the Traumatic Coma Data Bank. Acta
Chesnut, R. M., Marshall, L. F., Klauber, M. R. et al. (1993b) The role of the
secondary brain injury in determining outcome from severe head injury.
Journal of Trauma, 34: 216222.
Chesnut, R. M., Gautille, T., Blunt, B. A. et al. (1994) The localizing value of
asymmetry in pupil size in severe head injury: relation to lesion type and
location. Neurosurgery, 34, 840846.
Chopp, M., Chen, H., Dereski, M. O. and Garcia, J. H. (1991) Mild
hypothermic intervention after graded ischemic stress in rats. Stroke, 22,
3743.
Clifton, G. L., Jiang, J. L., Lyeth, B. G. et al. (1991), Marked protection by
moderate hypothermia after experimental traumatic brain injury. Journal of
Cold, G. E. (1989) Does acute hyperventilation provoke cerebral oligaemia in
comatose patients after acute head injury? Acta Neurochirurgica (Vienna), 96,
100106.
Cold, G. E., Jensen, F. T. and Malmros, R. (1977) The cerebrovascular CO2
reactivity during the acute phase of brain injury. Acta Anaesthesiologica
Scandinavica, 21, 222231.
Contant, C. F., Robertson, C. S., Gopinath, S. P. et al. (1993) Determination of
clinically important thresholds in continuously monitored patients with
head injury (abstract). Journal of Neurotrauma, 10, S57.
Cooper, K. R., Boswell, P. A. and Choi, S. C. (1985) Safe use of PEEP in patients
with severe head injury. Journal of Neurosurgery, 63, 552555.
Cooper, P. R., Moody, S., Clark, W. K. et al. (1979) Dexamethasone and severe
head injury. A prospective double-blind study. Journal of Neurosurgery, 51,
307316.
Cortbus, F., Jones, P. A., Miller, J. D. et al. (1995) Cause distribution and
significance of episodes of reduced cerebral perfusion pressure following
head injury (abstract). Journal of Neurotrauma, 12, 368.
405
Cottrell, J. E. and Marlin, A. E. (1981) Furosemide and human head injury.

Cottrell, J. E., Robustelli, A., Post, K. and Turndorf, H. (1977) Furosemide- and
mannitol-induced changes in intracranial pressure and serum osmolality
and electrolytes. Anesthesiology, 47, 2830.
Cruz, J. (1995) An additional therapeutic effect of adequate hyperventilation
in severe acute brain trauma: normalization of cerebral glucose uptake.
Davson, H., Welch, K. and Segal, M. B. (1987) The Physiology and Pathophysiology of the Cerebrospinal Fluid, Churchill Livingstone, Edinburgh, pp.
23.
Dearden, N. M. and McDowall, D. G. (1985) Comparison of etomidate and
althesin in the reduction of increased intracranial pressure after head injury.
Dearden, N. M. and Miller, J. D. (1989) Paired comparison of hypnotic and
osmotic therapy in the reduction of intracranial hypertension after severe
head injury, in Intracranial Pressure VII, (eds J. T. Hoff and A. L. Betz),
Dearden, N. M., Gibson, J. S., McDowall, D. G. et al. (1986) Effect of high-dose
dexamethasone on outcome from severe head injury. Journal of Neurosurgery,
64, 8188.
DeSalles, A. F., Muizelaar, J. P. and Young, H. F. (1987) Hyperglycemia,
cerebrospinal fluid lactic acidosis and cerebral blood flow in severely headinjured patients. Neurosurgery, 21, 4550.
Donegan, M. F. and Bedford, R. F. (1980) Intravenously administered lidocaine
prevents intracranial hypertension during endotracheal suctioning. Anesthesiology, 52, 516518.
Eisenberg, H. M., Frankowski, R. F., Contant, C. F. et al. (1988) High-dose
Eisenberg, H. M., Gary, H. E. Jr, Aldrich, E. F. et al. (1990) Initial CT findings
in 753 patients with severe head injury. A report from the NIH Traumatic
Coma Data Bank. Journal of Neurosurgery, 73, 688698.
Enevoldsen, E. M. and Jensen, F. T. (1978) Autoregulation and CO2 responses
of cerebral blood flow in patients with acute severe head injury. Journal of
Evans, J. P. and Scheinker, I. M. (1946) Histological studies of the brain
following head trauma. II. Post traumatic petechial and massive intracerebral hemorrhage. Journal of Neurosurgery, 3, 101113.
Feig, P. U. and McCurdy, D. K. (1977) The hypertonic state. New England
Fearnside, M. R., Cook, R. J., McDougall, P. and McNeil, R. J. (1993) The
Westmead Head Injury Project outcome in severe head injury. A comparative analysis of pre-hospital, clinical and CT variables. British Journal of
Frost, E. A. (1977) Effects of positive end-expiratory pressure on intracranial
pressure and compliance in brain-injured patients. Journal of Neurosurgery,
47, 195200.
Gahjar, J. B., Hariri, R. J. and Patterson, R. H. (1993) Improved outcome from
traumatic coma using only ventricular CSF drainage for ICP control.
Advances in Neurosurgery, 21, 173177.
Gahjar, J. B., Hariri, R. J., Narayan, R. K. et al. (1995) Survey of critical care
management of comatose head-injured patients in the United States. Critical
Gennarelli, T. A., Obrist, W. D., Langfitt, T. W. et al. (1979) Vascular and
metabolic reactivity to changes in pCO2 in head injured patients, in Neural
Trauma, (eds A. J. Popp, B. S. Bourke, L. R. Nelson et al.), Raven Press, New
York, pp. 18.
Gentleman, D. and Macpherson. P. (1989) Diagnosis and management of
delayed traumatic haematomas. British Journal of Neurosurgery, 3, 367372.
Gobiet, W. (1977) Monitoring of intracranial pressure in patients with severe
head injury. Neurochirurgia (Stuttgart), 20, 3547.
Gopinath, S. P., Robertson, C. S., Contant, C. F. et al. (1994) Jugular venous
Gotoh, F., Meyer, J. S. and Takagi, Y. (1965) Cerebral effects of hyperventilation
in man. Archives of Neurology, 12, 410423.
Gudeman, S. K., Miller, J. D. and Becker, D. P. (1979) Failure of high-dose
steroid therapy to influence intracranial pressure in patients with severe
Guidelines for the Management of Severe Head Injury (1995) Guidelines for the
Management of Severe Head Injury, a joint initiative of the Brain Trauma
Foundation, the American Association of Neurological Surgeons, the Joint
Section on Neurotrauma and Critical Care, Brain Trauma Foundation, New
York.
Hall, E. D., Yonkers, P. A., McCall, J. M. and Braughler, J. M. (1988) Effects of
the 21-aminosteroid U74006F on experimental head injury in mice. Journal
Harper, A. M. and Glass, H. I. (1965) Effect of alterations in the arterial carbon
dioxide tension on the blood flow through the cerebral cortex at normal and
low arterial blood pressures. Journal of Neurology, Neurosurgery and
Havill, J. H. (1984) Prolonged hyperventilation and intracranial pressure.
406
Hayes, T. M. and Tindall, G. T. (1969) Effects of altering arterial carbon dioxide

pressure on internal carotid blood flow and cerebrospinal fluid pressure in
man. Surgical Forum, 20, 421423.
James, H. E., Langfitt, T. W., Kumar, V. S. and Ghostine, S. Y. (1977) Treatment
of intracranial hypertension. Analysis of 105 consecutive, continuous
recordings of intracranial pressure. Acta Neurochirurgica (Vienna), 36,
189200.
Jenkins, L. W., Moszynski, K., Lyeth, B. G. et al. (1989) Increased vulnerability
of the mildly traumatised rat brain to cerebral ischaemia: the use of
controlled secondary ischaemia as a research tool to identify common or
different mechanisms contributing to mechanical and ischaemic brain
injury. Brain Research, 477, 211224.
Jennett, B., Teasdale, G., Galbraith, S. et al. (1977) Severe head injuries in
three countries. Journal of Neurology, Neurosurgery and Psychiatry, 40,
291298.
Jensen, K., Ohrstrom, J., Cold, G. E. and Astrup, J. (1991) The effects of
indomethacin on intracranial pressure, cerebral blood flow and cerebral
metabolism in patients with severe head injury and intracranial hypertension. Acta Neurochirurgica (Vienna), 108, 116121.
Johnston, I. H. and Jennett, B. (1973) The place of continuous intracranial
pressure monitoring in neurosurgical practice. Acta Neurochirurgica
(Vienna), 29, 5363.
Johnston, I. H., Johnston, J. A. and Jennett, B. (1970) Intracranial-pressure
changes following head injury. Lancet, ii, 433436.
Kaufmann, A. M. and Cardoso, E. R. (1992) Aggravation of vasogenic cerebral
edema by multiple-dose mannitol. Journal of Neurosurgery, 77, 584589.
Kaufman, H. H., Moake. J. L., Olson, J. D. et al. (1980) Delayed and recurrent
intracranial hematomas related to disseminated intravascular clotting and
Kellie, G. (1824) An account of the appearances observed in the dissection of
two of three individuals presumed to have perished in the storm of the 3rd
and whose bodies were discovered on the morning of the 4th November,
1821 with some reflections on the pathology of the brain. Transactions of the
Edinburgh Medico-Chirurgical Society, 1, 84122.
Knuckey, N. W., Gelbard, S. and Epstein, M. H. (1989) The management of
asymptomatic epidural hematomas. A prospective study. Journal of
Kontos, H. A., Wei, E. P., Raper, A. J. et al. (1978) Role of tissue hypoxia in local
regulation of cerebral microcirculation. American Journal of Physiology, 234,
H582H591.
Langfitt, T. W., Weinstein, J. D. and Kassell, N. F. (1964a) Transmission of
increased intracranial pressure. II. Within the supratentorial space. Journal
Langfitt, T. W., Weinstein, J. D. and Kassell, N. F. (1964b) Transmission of
increased intracranial pressure. I. Within the craniospinal axis. Journal of
Leech, P. and Miller, J. D. (1974) Intracranial volumepressure relationships
during experimental brain compression in primates. 3. Effect of mannitol
and hyperventilation. Journal of Neurology, Neurosurgery and Psychiatry, 37,
11051111.
Leggate, J. R. S., Dearden, N. M. and Miller, J. D. (1986) The effects of
gammahydroxybutyrate and sodium thiopentone on intracranial pressure
in severe head injury, in Intracranial Pressure VI, (eds J. D. Miller, G. M.
Teasdale and J. O. Rowan), Springer-Verlag, Berlin, pp. 754757.
Lewis, S. B., Myburgh, J. A. and Reilly, P. L. (1995) Detection of cerebral
venous desaturation by continuous jugular bulb oximetry following acute
neurotrauma. Anaesthesia and Intensive Care, 23, 307314.
Lewis, S. B., Reilly, P. L. and Myburgh, J. A. (1996). Determining cerebral
perfusion pressure thresholds in severe head trauma. Journal of
Neurotrauma.
pressure in neurosurgical practice. Acta Psychiatrica et Neurologica Scandinavica (Supplement), 36 (Suppl. 149), 1193.
Marienne, J. P., Robert, G. and Bagnet, E. (1979) Post traumatic acute rise in
ICP related to subclinical epileptic seizures. Acta Neurochirurgica (Supplement) (Vienna), 28, 8992.
Marion, D. W. and Carlier, P. (1995) Moderate therapeutic hypothermia
improves outcome following severe traumatic brain injury (abstract).
therapeutic hypothermia for patients with severe head injuries: a preliminary report. Journal of Neurosurgery, 79, 354362.
Marmarou, A. (1986) Progress in the analysis of intracranial pressure
dynamics, in Intracranial Pressure VI, (eds J. D. Miller, G. M. Teasdale and J.
O. Rowan), Springer-Verlag, Berlin, pp. 781788.
Marmarou, A., Shulman, K. and LaMorgese, J. (1975) Compartmental analysis
of compliance and outflow resistance of the cerebrospinal fluid system.
Marmarou, A., Maset, A. L., Ward, J. D. et al. (1987) Contribution of CSF and
vascular factors to elevation of ICP in severely head-injured patients.
Journal of Neurosurgery, 75, S59S66.
Marmarou, A., Anderson, R. L., Ward, J. D. et al. (1991) Impact of ICP stability
and hypotension on outcome in patients with severe head trauma. Journal
Marshall, L. F. (1980) Treatment of brain swelling and brain edema in man.
Advances in Neurology, 28, 459469.
aggressive treatment in severe head injuries. Part II: Acute and chronic
barbiturate administration in the management of head injury. Journal of
Marshall, L. F., Toole, B. M. and Bowers, S. A. (1983) The National Coma Data
Bank. Part 2: Patients who talk and deteriorate: implications for treatment.
Journal of Neurosurgery,. 59, 285288.
Marshall, L. F., Smith, R. W., Rauscher, L. A. and Shapiro, H. M. (1978)
Mannitol dose requirements in brain-injured patients. Journal of Neurosurgery, 48, 169172.
Marshall, L. F., Barba, D., Toole, B. M. and Bowers, S. A. (1983) The oval pupil:
clinical significance and relationship to intracranial hypertension. Journal of
of head injury based on computerized tomography. Journal of Neurosurgery,
75(Suppl.), S14S20.
Mendelow, A. D., Teasdale, G. M., Jennett, B. et al. (1983) Risks of intracranial
haematoma in head injured patients. British Medical Journal, 287,
11731176.
Mendelow, A. D., Teasdale, G. M., Russell, T. et al. (1985) Effect of mannitol on
cerebral blood flow and cerebral perfusion pressure in human head injury.
Messeter, K., Nordstrom, C. H., Sundbarg, G. et al. (1986) Cerebral
Metz, C., Holzschuh, M., Bein, T. et al. (1996) Moderate hypothermia in
patients with severe head injury: cerebral and extracerebral effects. Journal
Michenfelder, J. D. (1974) The interdependency of cerebral functional and
metabolic effects following massive doses of thiopental in the dog.
Miller, J. D. (1985) Raised intracranial pressure and its effects on brain
function, in Neurosurgery: The Scientific Basis of Clinical Practice, (eds A.
Crockard, R. Hayward and J. T. Hoff), Blackwell Scientific Publications,
Oxford, pp. 266267.
Miller, J. D. (1992) Evaluation and treatment of head injury in adults.
Neurosurgery Quarterly, 2, 2843.
Miller, J. D., Becker, D. P., Ward, J. D. et al. (1977) Significance of intracranial
hypertension in severe head injury. Journal of Neurosurgery, 47, 503516.
Miller, J. D., Gudeman, S. K., Kishore, P. R. and Becker, D. P. (1979) CT scan,
ICP and early neurological evaluation in the prognosis of severe head
injury. Acta Neurochirurgica (Supplement) (Vienna), 28, 8688.
289299.
Monro, A. (1783) Observations on the Structure and Function of the Nervous
System, Creech & Johnson, Edinburgh.
Moss, E., Powell, D., Gibson, R. M. and McDowall, D. G. (1979) Effect of
etomidate on intracranial pressure and cerebral perfusion pressure. British
Muizelaar, J. P., van der Poel, H. G., Li, Z. C. et al. (1988) Pial arteriolar vessel
rabbit. Journal of Neurosurgery, 69, 923927.
Muizelaar, J. P., Marmarou, A., DeSalles, A. A. et al. (1989) Cerebral blood flow
and metabolism in severely head-injured children. Part 1: Relationship with
GCS score, outcome, ICP, and PVI. Journal of Neurosurgery, 71, 6371.
Myburgh. J. A. and Lewis, S. B. (1996) Goal directed therapy in neurotrauma,
in 1996 Yearbook of Intensive Care and Emergency Medicine, (ed. J. L. Vincent),
Springer-Verlag, New York, pp. 716730.
Narayan, R. K., Kishore, P. R., Becker, D. P. et al. (1982) Intracranial pressure:
to monitor or not to monitor? A review of our experience with severe head
Nath, F. and Galbraith, S. (1986) The effect of mannitol on cerebral white
matter water content. Journal of Neurosurgery, 65, 4143.
Newell, D. W., Weber, J. P., Watson, R. et al. (1996) Effect of transient moderate
hyperventilation on dynamic cerebral autoregulation after severe head
injury. Neurosurgery, 39, 3544.
REFERENCES
Ninchoji, T., Uemura, K., Shimoyama, I. et al. (1984) Traumatic intracerebral
haematomas of delayed onset. Acta Neurochirurgica (Vienna), 71, 6990.
Nordstrom, C. H., Messeter, K., Sundbarg, G. et al. (1988) Cerebral blood flow,
OSullivan, M. G., Statham, P. F., Jones, P. A. et al. (1994) Role of intracranial
pressure monitoring in severely head-injured patients without signs of
intracranial hypertension on initial computerized tomography. Journal of
Overgaard, J. and Tweed, W. A. (1974) Cerebral circulation after head injury.
1. Cerebral blood flow and its regulation after closed head injury with
emphasis on clinical correlations. Journal of Neurosurgery, 41, 531541.
Paul, R. L., Polanco, O., Turney, S. Z. et al. (1972) Intracranial pressure
responses to alterations in arterial carbon dioxide pressure in patients with
head injuries. Journal of Neurosurgery, 36, 714720.
Paulson, O. B., Olesen, J. and Christensen, M. S. (1972) Restoration of
autoregulation of cerebral blood flow by hypocapnia. Neurology, 22, 286293.
Piper, I. P., Dearden, N. M. and Miller, J. D. (1989) Can waveform analysis of
ICP separate vascular from non-vascular causes of intracranial hypertension?, in Intracranial Pressure VII, (eds J. T. Hoff and A. L. Betz), SpringerVerlag, Berlin, pp. 157163.
Pollay, M., Fullenwider, C., Roberts, P. A. and Stevens, F. A. (1983) Effect of
mannitol and furosemide on bloodbrain osmotic gradient and intracranial
pressure. Journal of Neurosurgery, 59, 945950.
Pomeranz, S., Safar, P., Radovsky, A. et al. (1993) The effect of resuscitative
moderate hypothermia following epidural brain compression on cerebral
damage in a canine outcome model. Journal of Neurosurgery, 79, 241251.
Prior, J. G., Hinds, C. J., Williams, J. and Prior, P. F. (1983) The use of etomidate in
the management of severe head injury. Intensive Care Medicine, 9, 313320.
Rea, G. L. and Rockswold, G. L. (1983) Barbiturate therapy in uncontrolled
intracranial hypertension. Neurosurgery, 12, 401404.
Resnick, D. K., Marion, D. W. and Darby, J. M. (1994) The effect of
hypothermia on the incidence of delayed traumatic intracerebral hemorrhage. Neurosurgery, 34, 252256.
Rockswold, G. L., Ford, R. N., Anderson, M. D. et al. (1992) Results of a
prospective randomized trial for treatment of severely brain-injured
patients with hyperbaric oxygen. Journal of Neurosurgery, 76, 929934.
Ropper, A. H. (1986) Lateral displacement of the brain and level of
consciousness in patients with an acute hemispheral mass. New England
Ropper, A. H. (1989) A preliminary MRI study of the geometry of brain
displacement and level of consciousness with acute intracranial masses.
Ropper, A. H., ORourke, D. and Kennedy, S. K. (1982) Head position,
intracranial pressure, and compliance. Neurology, 32, 12881291.
Rosner, M. J. (1993) Pathophysiology and management of increased intracranial pressure, in Neurosurgical Intensive Care, (ed. B. T. Andrews),
McGraw-Hill, New York, pp. 57112.
Rosner, M. J. and Becker, D. P. (1984a) Origin and evolution of plateau waves.
Experimental observations and a theoretical model. Journal of Neurosurgery,
60, 312324.
Rosner, M. J. and Becker, D. P. (1984b) Experimental brain injury: successful
therapy with the weak base, tromethamine. With an overview of CNS
acidosis. Journal of Neurosurgery, 60, 961971.
Rosner, M. J. and Coley, I. B. (1986) Cerebral perfusion pressure, intracranial
pressure, and head elevation. Journal of Neurosurgery, 65, 636641.
Rosner, M. J. and Daughton, S. (1990) Cerebral perfusion pressure management in head injury. Journal of Trauma, 30, 933940.
Rosner, M. J., Rosner, S. D. and Johnson, A. H. (1995) Cerebral perfusion
pressure: management protocol and clinical results. Journal of Neurosurgery,
83, 949962.
Rowan, J. O., Johnston, I. H., Harper, A. M and Jennett, W. B. (1972) Perfusion
pressure in intracranial hypertension, in Intracranial Pressure, (eds M. Brock
and H. Dietz), Springer-Verlag, Berlin, pp. 165170.
Rowed, D. W., Leech, P. J., Reilly, P. L. and Miller, J. D. (1975) Hypocapnia and
intracranial volume-pressure relationship. A clinical and experimental
study. Archives of Neurology, 32, 369373.
Rubin, R. C., Henderson, E. S., Ommaya, A. K. et al. (1966) The production of
cerebrospinal fluid in man and its modification by acetozolamide. Journal of
Sahar, A. and Tsipstein, E. (1978) Effects of mannitol and furosemide on the rate
of formation of cerebrospinal fluid. Experimental Neurology, 60, 584591.
407
Sahuquillo, J., Poca, M. A., Monforte. L. et al. (1994) Interhemispheric

supratentorial ICP gradients in head injury patients: are they clinically
important?, in Intracranial Pressure IX, (eds H. Nagai, K. Kamiya and S.
Ishii), Springer-Verlag, Berlin, pp. 4851.
Salvant, J. B. Jr and Muizelaar, J. P. (1993) Changes in cerebral blood flow and
metabolism related to the presence of subdural hematoma. Neurosurgery 33,
387393.
Sato, O., Hara, M., Asai, T. et al. (1973) The effects of dexamethasone
phosphate on the production rate of cerebrospinal fluid in the spinal
subarachnoid space of dogs. Journal of Neurosurgery, 39, 480484.
Saul, T. G., Ducker, T. B., Salcman, M. and Carro, E. (1981) Steroids in severe
head injury: a prospective randomized clinical trial. Journal of Neurosurgery,
54, 596600.
Schalen, W., Messeter, K. and Nordstrom,
C. H. (1991) Cerebral vasoreactivity
and the prediction of outcome in severe traumatic brain lesions. Acta
Anaesthesiologica Scandinavica, 35, 113122.
Sedzimir, C. B. (1959) Therapeutic hypothermia in cases of head injury. Journal
Shackford, S. R., Shmoker, J. D. and Zhuang, J. (1995) The effect of hypertonic
resuscitation on pial arteriolar tone after brain injury and shock. Journal of
Trauma, 37: 899904.
Shapiro, H. M. and Marshall, L. F. (1978) Intracranial pressure responses to
PEEP in head-injured patients. Journal of Trauma, 18, 254256.
Shapiro, H. M., Galindo, A., Wyte, S. R. and Harris, A. B. (1973) Rapid
intraoperative reduction of intracranial pressure with thiopentone. British
on uncontrolled intracranial hypertension after severe head injury. Journal of
Strandgaard, S. and Paulson, O. B. (1992) Regulation of cerebral blood flow in
health and disease. Journal of Cardiovascular Pharmacology, 19, S89S93.
Stuart, F. P., Torres, E., Fletcher, R. et al. (1970) Effects of single, repeated and
massive mannitol infusion in the dog: structural and functional changes in
kidney and brain. Annals of Surgery, 172, 190204.
Tagaki, H., Saitoh, T., Kitihara, T. et al. (1983) The mechanism of ICP reducing
effect of mannitol, in Intracranial Pressure V, (eds S. Ishii, H. Nagai and M.
Brock), Springer-Verlag, Berlin, pp. 729733.
Turner, E., Hilfiker, O., Braun, U. et al. (1984) Metabolic and hemodynamic
response to hyperventilation in patients with head injuries. Intensive Care
Medicine, 10, 127132.
Unterberg, A., Kiening, K., Schmiedek, P. and Lanksch, W. (1993) Long term
observations of intracranial pressure after severe head injury. The phenomenon of secondary rise of intracranial pressure. Neurosurgery, 32,
1724.
Walsh, J. C., Zhuang, J. and Shackford, S. R. (1991) A comparison of
hypertonic to isotonic fluid in the resuscitation of brain injury and
hemorrhagic shock. Journal of Surgical Research, 50, 284292.
Weed, L. H. and McKibben, P. S. (1919) Pressure changes in cerebrospinal
Wilkinson, H. A. and Rosenfeld, S. R. (1983) Furosemide and mannitol in the
treatment of acute experimental intracranial hypertension. Neurosurgery, 12,
405410.
Wolf, A. L., Levi, L, Marmarou, A. et al. (1993) Effect of THAM in severe head
injury: a randomized prospective clinical trial. Journal of Neurosurgery, 78:
5459.
Worthley, L. I., Cooper, D. J. and Jones, N. (1988) Treatment of resistant
intracranial hypertension with hypertonic saline. Report of two cases.
Yano, M., Nishiyama, H., Yokota, H. et al. (1986) Effect of lidocaine on the ICP
response to endotracheal suctioning. Anesthesiology, 64, 651653.
Yoshida, K. and Marmarou, A. (1991) Effects of tromethamine and hyperventilation on brain injury in the cat. Journal of Neurosurgery, 74, 8796.
Yoshihara, M., Bandoh, K. and Marmarou, A. (1995) Cerebrovascular carbon
dioxide reactivity assessed by intracranial pressure dynamics in severely
head injured patients. Journal of Neurosurgery, 82, 386393.
20
THE ROLE OF SURGERY FOR

INTRACRANIAL MASS LESIONS
AFTER HEAD INJURY
Nigel Jones, Ross Bullock and Peter Reilly
20.1
Introduction
Many uninformed neurosurgeons have regarded surgery for post-traumatic intracranial hematomas as
unrewarding. This pessimism is based on the belief that
outcome is determined principally by the magnitude of
the initial injury and, therefore, frequently remains
poor despite optimal surgery. In fact, management of
post-traumatic epidural hematoma is one of the most
cost-effective of neurosurgical procedures in terms of
quality of life and years preserved (Pickard et al., 1990).
It is particularly in those with moderate head injury
that hematoma management may make the difference
between survival with permanent disability and a good
outcome. Intracranial hematoma is by far the most
common cause of secondary deterioration after all head
injuries and constitutes over 70% of the causes of death
in patients who talk and die (Reilly et al., 1975; Rose,
Valtonen and Jennett, 1976). Emergency surgery for
post-traumatic intracranial hematomas may be among
the most difficult procedures performed by neurosurgeons because of the frequency of complications such
as heavy bleeding and brain swelling yet, because these
operations frequently occur at night, it is too often the
less experienced surgeons who are delegated to do
them. Similarly, decisions regarding removal of hematomas, particularly contusions and intracerebral hematomas, may be extremely difficult, especially so when
surgery is prophylactic and intended to prevent
deterioration. There have been major changes in the
patterns of management of intracranial hematoma in
recent years, and this chapter has in mind neurosurgeons in training and aims to provide some guidelines
for surgical management.
20.2
Post-traumatic lesions on CT
CT scanning may reveal a number of patterns of

intracranial hematoma often associated with cerebral
swelling:

small, scattered hemorrhages associated with diffuse injury;

isolated, single or multiple deep cerebral
hematomas;
intraventricular hemorrhage;
subdural or extradural extra-axial hematomas.
Each of these, particularly intraventricular hemorrhage, may occur with diffuse axonal injury, ischemic
cell damage, contusion or any of the other types of
hemorrhage.
Large hematomas occur most frequently at the
frontal and temporal poles, often associated with
extensive contusions and subdural hematoma (the
exploded pole). Swelling around these severe parenchymal injuries may increase in the days following
injury and ICP may become more difficult to
control.
Marshall has proposed a reproducible and quantifiable CT classification for head injury, which is now
widely accepted as a method of stratifying head injury
in addition to the Glasgow Coma Scale (Marshall et al.,
1991; Table 20.1).
Although brain swelling may occur in relation to
hematomas, the degree is quite variable and need not
relate to the size of the hematoma. Indeed, swelling of
a hemisphere or of the whole brain may also occur
without CT evidence of bleeding.
Temporal lobe hematomas are prone to causing
brain-stem compression at low ICP and with little
midline shift.
20.3 Indications for evacuation of

intracranial hematomas
Craniotomy is not required in all patients with mass
lesions due to severe closed head injury. In the
Traumatic Coma Data Bank series only 37% of
comatose patients underwent surgery for removal of
intracranial hematomas (Eisenberg et al., 1991).
410
THE ROLE OF SURGERY FOR INTRACRANIAL MASS LESIONS
Table 20.1
Diagnostic categories of types of abnormality visualized on CT scanning (Source: from Marshall et al., 1991)
Category
Definition
Diffuse injury I (no visible pathology)
No visible intracranial pathology seen on CT scan
Diffuse injury II
Cisterns are present with midline shift 05 mm and/or

Lesion densities present
No high- or mixed-density lesion > 25 cm3
May include bone fragments and foreign bodies
Diffuse injury III (swelling)
Cisterns compressed or absent with midline shift 05 mm; no high- or

mixed-density lesion > 25 cm3
Evacuated mass lesion
Any lesion surgically evacuated
Non-evacuated mass lesion
High- or mixed-density lesion > 25 cm3 not surgically evacuated
CT is being used earlier and more often in patients

who sustain head injuries, even in those with mild
injuries, partly because of the medicolegal implications
of failure to detect an intracranial hematoma. In
consequence, patients with minimal or no neurological
signs or symptoms are often found to have intracranial
mass lesions. Neurosurgeons must then weigh up the
risks and benefits of conservative treatment, knowing
that the majority of such lesions will resolve spontaneously, yet a relatively small number will develop
raised ICP and secondary brain damage which could
have been prevented by prophylactic surgery
At least 25% of patients with intracranial mass
lesions will show clinical or radiological worsening in
the first 23 days after the injury. This is particularly
likely in patients with cerebral contusions, where
swelling is the norm rather than the exception.
Recently, several studies have provided guidelines
for the management of extradural and subdural
hematomas in asymptomatic patients (Bullock, Smith
and Van Dellen, 1985; Knuckey, Gelbard and Epstein,
1989; Hamilton and Wallace, 1992; Pozzati and Tognetti, 1986; Mathew et al., 1993). Management decisions
in individual patients must take into account a number
of factors, such as extracranial injuries, the age of the
patient and the presence of associated intracerebral
contusions or hemisphere swelling. Unfortunately,
such management decisions are even more difficult in
patients with intraparenchymal lesions such as contusions and intracerebral hematomas.
20.3.1 GUIDELINES FOR INITIAL CONSERVATIVE
THERAPY
Non-operative therapy should only be considered:

in patients who are fully conscious;

when the extra-axial mass lesion is the single
dominant lesion, i.e. there should not be multiple
contusions or potentially significant contralateral
mass lesions (which may be preventing midline
shift);
when there are no features of mass effect such as

midline shift greater than 3 mm, or basal cistern
effacement (Bullock and Teasdale, 1991).
In the conscious patient with an acute subdural

hematoma where points 13 above are fulfilled and
the lesion is less than 10 mm at its thickest point,
conservative therapy has been shown to be successful
in most cases (Mathew et al., 1993). The subdural
hematoma will usually resorb within 1 month,
although there are occasional instances of chronic
subdural hematoma formation.
Similarly, deep-seated interhemispheric or tentorial
subdural hematomas and small extradural hematomas in a stable, conscious patient may not need
surgical evacuation.
Indications for surgical removal are:

presence of a mass lesion >40 ml;

in the conscious, communicating, non-ventilated
patient:
decline in conscious state;
development of focal signs;
severe and especially worsening headache, nausea or vomiting;
in the unconscious, non-communicating, ventilated
patient:
decline in neurological state; (this may only be
indicated by the development of brain-stem
signs);
increase in ICP, e.g. >25 mmHg;
Either of these developments should lead to an

urgent CT scan.
increase in hematoma size on CT scan (Galbraith

and Teasdale, 1981).
20.3.2 INTRACEREBRAL HEMATOMAS AND FOCAL

DEFICIT
There has been debate about the benefit of removing

an intracerebral hematoma causing a focal deficit but
TECHNIQUES FOR CRANIOTOMY
without clinical effects of raised pressure or significant

midline shift. Experimental intracerebral hemorrhage
produces ischemia in the surrounding brain, apparently due to blood constituents as well as compression
(Jenkins et al., 1990). However, clot removal does not
relieve the ischemic neuronal damage, since this has
already occurred as a result of the clot. Removal
should therefore be based on evidence of harmful
mass effect. When mass effect is present, removal has
been reported to lead to a more rapid resolution of
neurological deficits (Levinthal and Stern, 1977).
20.4
Techniques for craniotomy
20.4.1
PREOPERATIVE PREPARATION
In general, cranial surgery should not be performed

until a stable blood pressure and adequate lung
function, confirmed by blood gas analysis, have been
achieved (Chapter 17).
However, in a patient who is comatose, particularly
if there has been a documented deterioration in
conscious level or development of focal signs, removal
of the hematoma is a matter of great urgency. The
patient should be intubated and hyperventilated, if
this has not already been done, and mannitol 1 g/kg
should be given immediately as the patient is taken to
the operating theater.
Prior to, or during the preparation for craniotomy,
the following checklist should be completed:
1. blood to laboratory for:
(a) cross matching (2 units of whole blood);
(b) coagulation studies
(i) prothrombin index
(ii) partial thromboplastin time
(iii) platelet count;
(c) blood gas analysis;
(d) routine full blood count and electrolytes;
2. X-rays of chest and cervical spine (or keep cervical
spine in collar);
3. consent for surgery;
4. Foley catheter in bladder;
5. two large bore peripheral i.v. lines, or one peripheral and one central line (maintaining
CVP > 5 cmH2O);
6. arterial catheter;
7. protection of both eyes from fluids and pressure;
8. adequately secured cuffed endotracheal tube.
The head should be placed on a horseshoe or
doughnut headrest, turned to place the operative side
uppermost and slightly elevated above the level of the
heart. A sandbag placed beneath the ipsilateral shoulder makes turning the head easier. Pressure points
should be carefully padded. Unless deterioration is
rapid, the scalp should be shaved and prepared with
411
povidone-iodine as for any other intracranial procedure. The drapes can be stapled into place to prevent
them becoming dislodged if the head has to be turned
or moved during the procedure. Antibiotics, anticonvulsants and mannitol are used as required. Strict
attention to anesthetic techniques is vital to avoid
hypercarbia and further elevation of intracranial
pressure.
20.4.2
EXPLORATORY BURRHOLES
With the wider availability of CT, the necessity for

exploratory burrholes is declining. They will be
required only very rarely in hospitals with a CT
scanner; however, they may be life-saving in rural
locations where transfer to a CT-equipped facility may
involve long delays.
The use of exploratory burrholes implies that
confirmation of the position or even the presence of a
hematoma is lacking. It is therefore important to have
access to the whole head and to explore all likely sites
bilaterally before discarding the diagnosis of intracranial hematoma. The extent of this exploration will
be determined at least in part by the experience and
skills of the operator, given that this will frequently be
a non-neurosurgeon working in less than ideal circumstances. It should always be possible to obtain
neurosurgical advice by telephone and this should be
encouraged. Specific guidelines for the management
of head injury in remote locations in Australia have
been established by the Neurosurgical Society of
Australasia.
(a)
Technique
The patient is placed supine on a horseshoe or

doughnut headrest. The whole head is shaved, prepared and draped to allow access to both frontal,
parietal and temporal areas. The site for the initial
burrhole is determined according to the most likely
site of the suspected hematoma. If a dilated pupil is
present, it will usually be ipsilateral to the hematoma.
The next most valuable localizing feature is a hemiparesis, which will usually be contralateral to a hematoma. If a fracture is present it is most likely to overlie
an extradural hematoma. It must be stressed that none
of these signs is absolute and if no hematoma is found
on the suspected side, the other side should be
explored in all cases.
Unless a fracture is present in a different location,
the first burrhole should be temporal, as this is the site
of most extradural hematomas. It is frequently written
that a burrhole placed as little as 2 mm away from the
edge of an extradural hematoma will fail to identify it,
and this tends to discourage the inexperienced doctor
in a difficult situation. The temporal burrhole is not
412
intended to be placed in any position to that degree of

accuracy. It should ideally be just above the zygoma
and approximately one finger breadth in front of the
tragus. The head is turned away from the side of
exploration if the cervical spine is normal or the
patient is placed in the lateral position in a suitable
collar if there may be a cervical spine fracture.
A vertical scalp incision is made as shown in Figure
20.1. The scalpel is used to cut directly through the
temporalis muscle down to the skull. To prevent
injury to the facial nerve the inferior limit of the
incision is the zygomatic arch. Bleeding will be
encountered from the superficial temporal artery but
this can be easily controlled with diathermy or ligation
after a retractor is placed. The periosteum and
temporalis muscle are scraped off the temporal bone
with a sharp periosteal elevator and a self-retaining
retractor is inserted to hold back the muscle. A
burrhole is drilled, using the perforator until an
irregular wobbling is felt and there is a perforation
through the central part of the hole that can be
palpated with forceps (Figure 20.2). Through this
small hole, the surgeon may see either dura or
extradural blood, but for a confident diagnosis the
hole needs to then be enlarged with the burr to
complete the hole, but only if the perforator has
breached the inner table. The temporal bone is usually
quite thin (23 mm), whereas the bone in the frontal
and parietal areas may be more than 10 mm thick. If an
extradural hematoma is present, the hole will need to
be enlarged to allow removal of the clot. For an
experienced operator in a well-equipped site this will
usually involve a craniotomy; for an inexperienced
operator or when facilities are limited a craniectomy is
Figure 20.1 The inital burr hole is made through a vertical

incision just above the zygoma and one fingerbreadth in
front of the tragus.
Figure 20.2 The preforator makes a small hole in the inner

table of the skull through which extradural blood may be
seen.
often more appropriate. If no abnormality is found at

the temporal burrhole, further burrholes are performed in the ipsilateral frontal and posterior parietal
areas. The technique is the same but the bone will be
thicker. The scalp incisions should be made in such a
way that they can easily be converted into a questionmark-type flap (Figure 20.3).
If all three burrholes on one side are negative, the
process should be repeated on the opposite side.
Opinions regarding opening of the dura vary. It is very
difficult to determine whether a significant intradural
hematoma exists without first opening the dura. The
presence of blue, bulging dura seen through a burrhole is a very poor diagnostic feature. Following
Figure 20.3 The scalp incision for exploratory burr holes

should be made in a direction suitable for later conversion to
a flap (as shown by the dotted line).
trauma, the brain is often swollen and frequently there

is a thin rim of acute subdural blood which is of little
significance. Inexperienced operators performing
exploratory burrholes for the first time will often
mistake this for a significant acute subdural hematoma and may be dissuaded from looking elsewhere,
thereby missing a more important hematoma. We
believe it is better to complete the standard six
burrholes to exclude an extradural hematoma before
opening the dura in search of an acute subdural
hematoma. The dura is opened by gently incising the
outer layers with a number 15 blade, picking these up
with a sharp hook to lift the dura off the brain before
completing the incision. The dural edges can then be
picked up with forceps and a cruciate incision can be
completed. The dural edges are diathermied, taking
care to avoid any underlying cerebral vessels. A thin
layer (23 mm) of blood is common and should not be
mistaken for a significant acute subdural hematoma. If
a thicker layer of blood is encountered, it should be
exposed by performing a craniotomy or craniectomy,
depending on the circumstances.
20.4.3
(a)
EXTRADURAL HEMATOMA
Craniectomy
Although craniotomy is preferable when skills and

conditions permit, craniectomy is often used as a
means of extending the exposure after exploratory
burrholes. Initial evacuation should be performed
rapidly through the craniectomy in order to relieve
brain-stem pressure, then either a formal craniotomy
or a more extensive craniectomy may be performed,
the hematoma evacuated and the bleeding point
secured.
The technique is simple. A rongeur is used to nibble
bone away and expose the entire hematoma. When the
bone is very thick, it is sometimes easier to make
several burrholes and nibble the bone away between
them. The dura will have already been separated from
the inner table of the skull by the hematoma and is not
prone to injury. At the margins of the hematoma the
dura will be adherent to the bone and should not be
stripped any further. If there is a bleeding point on the
dura it should be coagulated with bipolar diathermy
or, if this is not available, unipolar diathermy may be
used on a low setting. Bleeding from the edges of the
wound may be troublesome and difficult to stop with
diathermy. It is best to pack a little Surgicell or Avitene
(postage-stamp size) under the bone edge and then
hitch up the dura to surrounding pericranium with
fine sutures (Figure 20.4). It is however, important not
to traumatize the underlying brain with these sutures,
and this may be achieved by first lifting the dura off
the brain with forceps, or opening the dura to allow
413
Figure 20.4 Hitching the dura to the pericranium is an

effective means of controlling venous bleeding from beneath
the bone edges.
air or saline to lift the dura from the brain. The bone
fragments should be left out, and the wound closed in
layers with a suction drain in the subgaleal (extradural) space for 12 hours. One should not be deterred
by the amount of bone that is removed since an
elective cranioplasty is easily performed when the
patient has recovered.
(b)
Craniotomy
When the site of the hematoma has been determined

preoperatively by CT scan, the operation of choice will
usually be a craniotomy. It is helpful to transpose the
position of the hematoma on to the CT scout film, as
unusual gantry angles may be confusing (Figure 20.5).
Most extradural hematomas are temporal and can be
adequately removed through the standard question
mark trauma scalp flap (Figure 20.3). More posterior
hematomas may be approached through an inverted
horseshoe flap (Figure 20.6). Frontal and subfrontal
hematomas should be exposed through a bicoronal
flap (Figure 20.7) to avoid an unsightly midline
forehead scar.
The patient is positioned with the head in a
horseshoe headrest and turned to bring the operative
site uppermost. (A three-pin headrest may be used if
the location of the hematoma is established.) The scalp
is shaved and prepared with povidone-iodine.
Depending on the urgency of the situation, the whole
scalp flap can be turned down initially, or a small
incision can be made and a burrhole craniectomy
performed to allow rapid decompression, before
completing the craniotomy. In most cases it is more
efficient to quickly turn down a myocutaneous scalp
flap followed by a free bone flap. The use of a high-
414
(a)
Figure 20.5 The extradural hematoma seen in the axial

scans in (a) has been measured and transposed on to the
corresponding scout film (b). This may be of considerable
assitance in planning the site for craniotomy.
(b)
speed drill system permits this to be done just as

quickly in skilled hands as hand-drilling a single
burrhole. A small hole is drilled in the thin temporal
bone or elsewhere depending on the location of the
hematoma. A craniotome cutting attachment is then
used to complete the craniotomy. The risk of dural
damage is small, especially in the presence of an
extradural hematoma. The medial edge of the bone
flap is taken no closer than 2 cm from the midline to

avoid injury to the superior sagittal sinus and its
draining veins. Unless the CT dictates otherwise, the
inferior edge of the bone flap is taken well down into
the temporal region to allow clear access to the middle
fossa.
After removing the bone flap, the hematoma is sucked
and irrigated away. Bleeding dural arteries and veins
415
also helps to control this bleeding and prevent

recurrent hematoma formation.
The bone flap is replaced and fixed with mini-plates
or stainless-steel wire if the brain is not swollen, and
suction drains are used in the extradural and subgaleal spaces for 12 hours. The muscle and galea are
closed with Vicryl and the skin with staples. In
patients who have sustained a significant primary
injury (that is all who were in coma), an ICP monitor
is inserted at the end of the procedure. When brain
swelling occurs the dura becomes tense and it should
always be opened to exclude an underlying subdural
hematoma.
20.4.4
Figure 20.6 An inverted horseshoe flap may be more

appropriate than a question mark flap for hematomas that
are positioned more posteriorly.
are then coagulated with bipolar diathermy. Bleeding

bone edges are controlled with bone wax. Hemostatic
agents such as Surgicell or Avitene can be used around
the margins of the craniotomy to control venous
bleeding, which often seeps out from just beneath the
bone edge. The addition of dural hitching sutures
Figure 20.7 A bicoronal or aneurysm flap for a frontal

hematoma avoids a visible forehead scar.
ACUTE SUBDURAL HEMATOMA
The treatment of acute subdural hematoma follows

the same principles as extradural hematoma, although
the prognosis is much worse (60% die or survive with
a bad outcome Chapter 18). The affected area is
exposed by craniectomy or preferably craniotomy as
described for extradural hematoma, taking care to use
all means available to reduce ICP first (see Chapter
18). The dura is opened using a sharp hook and
scalpel, revealing dark, clotted blood under pressure.
For the inexperienced, it is preferable to make several
slits in the dura without joining them together (Figure
20.8) as this will minimize herniation of the brain,
which can be dramatic and may even prevent closure
of the scalp. If a dural flap is used, it should be based
medially to avoid damage to the superior sagittal
sinus and the draining veins running into it (Figure
20.9). The hematoma is removed by suction and
irrigation, taking care not to use suction beyond the
limits of the craniotomy in areas that are not directly
visible and accessible. Often there will be no obvious
bleeding point. Bleeding from a contused cortical
Figure 20.8 To minimize brain herniation an acute subdural hematoma may be evacuated through several small
dural incisions rather than one large dural flap.
416
outweighed by the disadvantages of cerebral herniation through the wound with consequent venous and
even arterial compromise. The bone may be secured
with mini-plates or stainless-steel wire unless brain
swelling is present, in which case the bone is best left
free and secured only by a few pericranial sutures to
allow expansion. A suction drain is placed in the
subgaleal space with or without another in the
extradural space (if the dura has been closed). The
scalp is closed in two layers and an ICP monitor is
used routinely.
20.4.5
Figure 20.9 Basing a dural flap medially reduces the risk of
damage to the superior sagittal sinus and the draining veins
running into it.
surface can be controlled with bipolar diathermy or

Gelfoam. This may not be easy, especially if the brain
is swelling rapidly.
An attempt should always be made to identify a
parasagittal venous bleeder, if no cortical source of
bleeding is found. The medial cortex should be very
gently depressed, while irrigating parasagittally. Raising the venous pressure may reveal the bleeder. This is
best done by asking the anesthetist to increase
inflation pressure for three or four breaths to
40 cmH2O. If a bleeding parasagittal vein is found, it
should be coagulated with bipolar diathermy, If the
vein is avulsed from the sagittal sinus, bleeding may
be very difficult to arrest, requiring application of
ligaclips or a muscle patch for tamponade. Failure to
seek a parasagittal bleeding source (present in 4050%
of subdural hematomas) may account for the high
incidence of recurrent hematomas in some series
(Richards and Hoff, 1974).
Contused and necrotic brain is best removed with
suction and gentle irrigation, with consideration of
partial frontal or temporal lobectomy when the damage is severe. This should not be undertaken by nonneurosurgeons, who should only remove damaged
brain.
After removal of the hematoma the dura is closed.
This may require a duroplasty, using an insert of
temporalis fascia, pericranium or other dural substitute such as lyophilized bovine dura or woven
collagen. Commercial forms of human lyophilized
dura are no longer available in many countries
because of the known cases of transmitted CreutzfeldtJakob disease (Will and Matthews, 1982; Simpson et al., 1996).
It is generally preferable to replace the bone flap, as
the advantages of decompression in this situation are
POSTERIOR FOSSA HEMATOMAS
Suspicion of a posterior fossa hematoma should

mandate rapid transfer to a neurosurgical facility, or
a neurosurgical retrieval team (Chapter 15). The
major concerns are rapid onset of CSF obstruction
and brain-stem compression, and air embolism during surgery due to venous sinus injury. The patient
is positioned prone with the head fixed in a three-pin
headrest. The head should be slightly elevated to
reduce venous bleeding but not so much as to
promote air embolism. A midline incision and standard suboccipital craniectomy are performed. It is
important to continue the decompression through
the rim of the foramen magnum, as postoperative
swelling in the posterior fossa may cause death by
brain-stem compression. If the trauma has already
involved a sinus, bleeding should be controlled by
tamponade with muscle, Gelfoam, Surgicel or similar
substances while carefully elevating the head. An
esophageal stethoscope and end-tidal CO2 measurements should be used to detect any air embolism.
Many extradural hematomas extend both above and
below the level of the tentorium, overlying the
transverse sinus. In these cases, a bridge of bone
should be left over the sinus to allow the dura to be
tented up and tamponade bleeding sites.
20.5
Specific surgical problems
20.5.1
INTRAVENTRICULAR HEMORRHAGE
Intraventricular hemorrhage usually occurs in the

setting of a severe diffuse injury with other CT
evidence of parenchymal injury. In about 40% of cases
there is no evidence of an intraparenchymal source of
the bleeding, although there will often be an injury to
the corpus callosum or septum pellucidum (Sato et al.,
1987). Hydrocephalus is uncommon (Christie, Marks
and Liddington, 1988; Jayakumar et al., 1990) but
when it is present, raised ICP should be treated by
ventricular drainage. Mortality is high in most series
but tends to reflect the GCS on admission and hence
the severity of the primary injury (Chapter 10).
SPECIFIC SURGICAL PROBLEMS
20.5.2
INTRACEREBRAL HEMATOMAS
Some 88% of patients with intracranial hematomas

after trauma were found to have raised ICP postoperatively (Miller et al., 1981). When control of ICP by
all available medical means fails then removal of
contusions may be a highly effective form of ICP
control, especially when accompanied by a large
craniotomy and duroplasty. A swollen and contused
temporal pole may cause early tentorial herniation
and brain-stem compression, even at surprisingly low
ICP in the sedated, ventilated patient.
20.5.3 SURGICAL DECOMPRESSION FOR BRAIN
SWELLING
Decompression by large craniotomy has been advocated in several older studies when other methods of
controlling ICP have failed (Kjellberg and Prieto, 1971;
Ransohoff et al., 1971, Britt and Hamilton, 1978).
Although all these older clinical (Cooper, Rovit and
Ransohoff, 1976; Venes and Collins, 1975), and experimental studies (Cooper et al., 1979) suggested no
benefit, and indeed worsening of brain edema following craniectomy, there are many more recent anecdotal
experiences in many neurosurgical centers where
decompressive craniectomy has been successful. Gaab
et al. (1990) proposed indications for large unilateral or
bilateral fronto-temporo-parietal decompression in
selected trauma patients. This was considered when
medical treatment had failed in patients under 40
years, with initial GCS of 7 or more and evidence of
clinical deterioration, but without the signs of brainstem failure. In their report, 30 of their 37 carefully
selected patients made good recoveries or had only
moderate disability. Alexander, Ball and Laster reported on the application of the more limited surgical
approach used by Cushing. In patients who did not
respond to medical treatment, they performed a
subtemporal decompression, leaving dura open and
resecting the anterior temporal lobe if it was damaged.
(Alexander, Ball and Laster, 1987).
One of the well known hazards of craniotomy for
brain swelling is herniation through the craniotomy.
Hence, large craniotomies have generally been advocated. Bifrontal craniotomies should be made so that
the inferior margin is cut low across the supraorbital
margin, and they should extend well posterior so
that the brain can expand forward and not herniate
against the bony margin. For effective decompression
the dural envelope must be enlarged. The dura
should be opened widely, hinged medially and the
falx and sagittal sinus divided as low as possible
taking care to preserve draining veins whenever
possible. An augmentation duroplasty is used to
cover the brain.
417
At the Medical College of Virginia, a wide decompressive craniotomy with duraplasty and bone-flap
removal is performed for ICP control under the
following circumstances:

ICP control not achieved with maximal therapy

using mannitol, ventricular drainage, pressors,
moderate hyperventilation (PaCO2 = 32 2 mmHg)
and moderate hypothermia (32 2C);
pupils not fixed and dilated;
preserved brain-stem responses and central conduction time on evoked potential testing.
Decompression is used before barbiturate therapy, in

view of the propensity for barbiturate therapy to
lower CPP and worsen outcome (Figure 20.10).
20.5.4
LOBECTOMY FOR BRAIN SWELLING
Surgery for removal of swollen and contused brain is

more controversial. It frequently does not control ICP
and more than 80% of such patients may still have
raised ICP postoperatively (Miller et al., 1981). It is
therefore better to control ICP by medical means
initially, removing contusions only when this treatment has failed (Miller, 1992). However, as noted
earlier, a swollen and contused temporal pole may
cause tentorial herniation and brain-stem compression
at quite low ICP and, in the sedated, ventilated
patient, the only (late) sign of this may be pupillary
dilatation.
Removal of contused and swollen brain to control
intracranial pressure may be beneficial when other
means of control have failed. Litofsky et al. (1994) found
this to be so in patients under 40 years with higher
initial GCS scores. In this small retrospective report
55% of patients had a good or moderately disabled
outcome and there was no increase in vegetative
survivors. Lobectomy aimed to remove damaged brain
and to avoid eloquent areas (Figure 20.11).
20.5.5
INTRAOPERATIVE BRAIN SWELLING
As noted above, removal of acute hematomas may be

followed by severe brain herniation. The brain herniating through the craniotomy may occlude surface
veins and even rupture against the bonydural
margins. This is most commonly seen with acute
subdural hematomas and may be anticipated when
the preoperative CT scan shows brain swelling in
addition to the hematoma (Bullock and Teasdale,
1991). There are several potentially remediable causes
that must be considered urgently before concluding
that the swelling is simply a reflection of a nonsurvivable injury. The surgeon should quickly establish with the anesthetist that ventilation has not
become impaired (e.g. by pneumothorax, endotra-
418
(a)
(b)
Figure 20.10 A 22-year-old male was car surfing when he

fell from the top of the car, which was travelling at about
25 km/h. GCS on admission was 10. CT scan showed
bifrontal contusions, obliteration of the basal cisterns and an
occipital fracture. He was intubated, ventilated and monitored. ICP stayed above 30 mmHg, despite medical treatment, excluding barbiturate therapy. CT scans showed
increase in the size of the frontal contusions. A bifrontal
decompressive craniotomy without lobectomy was performed. He recovered with a mild short-term memory
impairment. (a) Preoperative scan showing bifrontal contusions and a tight brain. (b) Scan 2 weeks after decompressive craniotomy. (c) Following replacement of the bone
flap 1 month later.
(c)
cheal tube dislodgment or occlusion) and that the

patient is hyperventilated. Secondly, a hidden hematoma, either intracerebral or beyond the craniotomy
margin, may have developed. An intracerebral hematoma is possible if there is a cerebral contusion or
exploded pole. Acute hydrocephalus is uncommon
but should be excluded by attempting to cannulate the
ipsilateral frontal horn. Intraoperative ultrasound is
valuable in seeking and excluding these events.
Contralateral hematomas occasionally develop or
enlarge after removal of the major hematoma. This
possibility should be considered particularly if there is
a contralateral fracture or small hematoma on the
preoperative CT, but indeed has been reported if
neither of these features is present.
20.5.6 MANAGEMENT OF DIFFUSE INTRAOPERATIVE

BLEEDING
(a)
Prevention
No matter how urgent the need to evacuate a

hematoma, it is imperative that a fixed surgical
routine be followed in order to avoid excessive
bleeding. After marking out the large craniotomy flap
required to evacuate an acute subdural hematoma or
intracerebral contusions, a burr hole should be made
and the hematoma should be decompressed acutely.
This allows the surgeon to open the remainder of the
craniotomy in a controlled fashion, securing the
bleeding points in the scalp with diathermy, scalp
clips or hemostats. Bleeding points in the temporalis
SPECIFIC SURGICAL PROBLEMS
(a)
419
(b)
Figure 20.11 A 19-year-old pedestrian was struck by a car travelling at high speed. GCS on admission was 3. CT scan
showed a basal ganglionic hematoma with 1 cm midline shift. ICP was 26 mmHg. The clot was partially removed by
craniotomy but ICP rose to 3040 mmHg over the next 24 h, despite medical therapy, excluding barbiturates. A frontal
lobectomy was performed and ICP and CPP were well controlled thereafter. (a) CT on admission showing a deep
intracerebral hematoma with 1 cm midline shift. (b) Following partial frontal lobectomy.
muscle are coagulated. Several units of blood may be

lost from badly secured scalp edges.
(b)
Management
Profuse and diffuse bleeding from brain and meningeal surfaces almost always indicates a significant
coagulopathy. Blood should be sent intraoperatively
for coagulation studies, and a transfusion of fresh
frozen plasma and/or whole blood should be given
while awaiting the results of these studies. Diffuse
hemorrhage may be accompanied by massive brain
swelling and herniation through the craniotomy. The
surgeon is then faced with the prospect of worsening
the bleeding by performing a temporal or frontal
lobectomy in order to control the brain swelling. A
management checklist should be followed before
proceeding with lobectomy.
1. Check the coagulation studies.
2. Transfuse blood and fresh frozen plasma as appropriate, usually 610 units.
3. Consider intraoperative ultrasound to detect the
development of an intraoperative hematoma, either
intracerebral, epidural or subdural on the same or
opposite side.
4. Optimize cerebral perfusion pressure. Sometimes
increasing CPP intraoperatively to a mean of 80 or
90 mmHg will help to control brain swelling by
promoting vasoconstriction if the patient has intact
autoregulation. Clearly, however, it is equally possible that this maneuver will worsen the brain
swelling if the patient does not autoregulate. (These
patients almost always die.)
5. Give thiopental 250 mg1 g in incremental doses, to
suppress any cerebral electrical activity (Chapter
19). CPP must be maintained at about 70 mmHg,
using vasopressor agents if necessary.
6. Give more mannitol to a dose of 1 g/kg and lower
the PaCO2 to approximately 28 mmHg by increasing
ventilation.
7. Occasionally, massive brain swelling may be
caused by an intraoperative pneumothorax, which
raises inflation and ventilatory pressures and hence
intracranial venous pressure. The chest should,
therefore be auscultated and, if necessary, X-rayed,
during the surgery.
It is usually advisable to wait 15 minutes to assess the
effect of these measures before proceeding. If there is no
improvement, or if bleeding is ongoing, it is usually
necessary to perform a lobectomy. It is important to
emphasize that, for any chance of success in this
desperate situation, the surgeon and a highly skilled
anesthesiologist must work closely together throughout and necessary assistants should be called in.
The best course of action is to perform a swift and
generous lobectomy where the damage is greatest,
avoiding where possible the speech areas in the
420
dominant hemisphere. It may be necessary to remove

bone very rapidly by extending the craniotomy with a
craniotome or rongeur in order to obtain sufficient
access for this. Hemostatic adjuncts may include
thrombin-soaked gel foam, Surgicel, and peroxidesoaked cotton wool patties.
20.5.7
CRANIOBASAL SKULL FRACTURE
Penetrating open wounds of the skull are not within

the provenance of this book, but craniobasal skull
fractures may be internally compound and will be
briefly considered. The presence of a dural fistula is
indicated by intracranial air, cerebrospinal fluid rhinorrhea or otorrhea or later by meningitis. An intracranial aerocele may expand if the dural tear acts as a
one-way valve (North, 1971). Cerebrospinal fluid
leakage may occur early, or after some days. It may
cease spontaneously, but this may not in itself indicate
satisfactory dural healing and the risk of late meningitis may still exist.
Recommendations for treatment have ranged from
dural repair in all cases, irrespective of whether the
rhinorrhea ceases (Lewin, 1954), to a more selective
repair when there is continued CSF leakage, anosmia
or wide fractures (Loew et al., 1984).
We recommend operative repair when:

CSF leakage persists for 710 days;

CSF leakage ceases, but recurs after 710 days;
there is clinical evidence of a large dural defect
indicated by a large aerocele, or by escape of brain
tissue through the nostrils;
meningitis or brain abscess develop after trauma at
any time;
there is radiological evidence suggesting that natural dural repair is unlikely, e.g. a basal skull fracture
involving the paranasal sinuses, wide separation of
bone edges, a spike of bone or a possible cerebral
hernia into the nasal cavity (Reilly and Simpson,
1995).
The incidence of meningitis after cerebrospinal fluid

rhinorrhea associated with the Le Fort III fractures
appears to be lower than that associated with a vault
fracture (OBrien and Reade, 1984). Indeed, CSF
leakage often ceases once the midface fracture is
reduced, but the indications for dural repair remain as
outlined above. Whether prophylactic antibiotics are
protective against meningitis continues to be debated
(Brown, 1993; Working Party Report, 1994). Although
there is no strong evidence favoring prophylactic
antibiotics it is our policy to use intravenous metronidazole, co-trimoxazole and ampicillin if there is
evidence of a fistula. These are continued until CSF
leakage ceases or for several days after surgical repair
is undertaken.
(a)
Timing of transcranial repair
The transcranial repair of a dural fistula is usually

undertaken electively and therefore does not affect the
early acute management of the patient with severe
head injury. Acute transcranial repair is often very
difficult, because the brain is tight and resists
retraction and the repair may therefore be inadequate
or impossible. If there are facial fractures, a one-stage
combined transcranial repair and internal fixation of
the facial fractures may be undertaken with the plastic
surgery team, usually 510 days after injury. Recently,
well-localized fractures involving the sphenoid and/
or ethmoid sinuses have been approached subcranially via an ethmoidectomy or transnasally aided by
endoscopy. This has the merit of not requiring brain
retraction and may be performed within days of
injury. It is also less likely to cause anosmia when
olfaction is still present (Raveh, Redli and Markwalder, 1984).
20.5.8 COAGULOPATHY IN PATIENTS WITH HEAD
INJURY
Coagulopathy develops in one quarter to one third of

patients who undergo surgery for removal of an
intracranial hematoma (Bullock et al., 1990; Stein et al.,
1992). In many, this may be mild and self-limiting, but
in others it may be sufficiently severe to cause death
on the operating table. It is consequently mandatory to
determine the coagulation status in every patient who
is considered for surgical evacuation of a traumatic
intracranial hematoma. Frequently, this will need to be
performed as an emergency. Full evaluation should
include a prothrombin index, partial thromboplastin
time and a platelet count. A template bleeding time, if
available, can be used to assess bleeding time. The
increasing frequency of coagulation disorders in
patients with intracranial hematoma has many causes,
but the most important are:

prior medication with warfarin, heparin or

aspirin;
chronic alcoholism: excess bleeding in alcoholics
may be related to reduced levels of coagulation
factors; an ethanol-induced decrease in platelet
function and nutritional factors may contribute;
disseminated intravascular coagulation (DIC) due
to thromboplastin activator substances derived
from brain or from sites of multiple trauma (Simpson, Speed and Blumbergs, 1991);
coagulopathy associated with massive blood transfusion to replace blood loss from abdominal or limb
injuries;
a pre-existing coagulation problem, e.g. von Willebrands disease, thrombocytopenia or hemophilia.
REFERENCES
These patients may be predisposed to develop large

intracranial hematomas with relatively minor cranial
injuries. It is imperative, therefore, to obtain a history
from relatives as early as possible, so that relevant preexisting diseases are known prior to surgery.
A marked hyperemic response is frequently
encountered in the injured brain, scalp and meninges,
and relatively minor abnormalities of coagulation,
which might be acceptable for abdominal or orthopedic surgery, can make emergency neurosurgery
extremely hazardous. The management goal in such
patients should be complete normalization of the
hematological parameters by replacing the appropriate factors (see below). Hematological and blood
transfusion services should be aware that platelets or
fresh frozen plasma may be needed in patients with
persistent severe intracranial bleeding, even when
coagulation parameters are only mildly abnormal or
indeed normal, in patients with qualitative platelet
function disorders.
(a)
Qualitative platelet disorders
Chronic aspirin therapy, which is increasingly prevalent in the general population, and chronic alcoholism
are the two most common causes of coagulopathy
after neurotrauma. When the platelet count is marginally depressed, as in chronic alcoholism, and platelet
function is also significantly compromised, the coagulation problem can be particularly severe. Unfortunately, qualitative platelet disorders may be undetectable by standard tests such as bleeding time. More
complex measures such as glass bead platelet adhesion tests may be needed and these are not routinely
available in the emergency situation that often confronts neurotrauma patients.
20.6
References
Alexander, E. Jr, Ball, M. and Laster, D. W. (1987) Subtemporal decompression:

radiological observations and recurrent surgical experience. British Journal
Britt, R. H. and Hamilton, R. D. (1978) Large decompressive craniotomy in the
treatment of acute subdural hematoma. Neurosurgery, 2, 195200.
Brown, E. M. (1993) Antimicrobial prophylaxis in neurosurgery. Journal of
Antimicrobial Chemotherapy, 31(Suppl. B), 4963
Bullock, R., Smith, R. M. and Van Dellen, J. R. (1985) Nonoperative treatment
of extradural haematoma. Neurosurgery, 16, 602606.
Bullock, R. and Teasdale, G. (1990) Head injuries surgical management of
traumatic intracerebral haematomas, in Handbook of Clinical Neurology, vol.
57 Head Injury, (ed. R. Braakman), Elsevier, Amsterdam, ch. 10, pp.
249298.
Bullock R., Hanneman, C. O., Murray, L. and Teasdale, G. M. (1990) Recurrent
haematomas following craniotomy for traumatic intracranial mass. Journal
Christie, M., Marks, P. and Liddington, M. (1988) Post-traumatic intraventricular haemorrhage: a reappraisal. British Journal of Neurosurgery, 2,
343349.
Cooper, P. R., Rovit, R. L. and Ransohoff, J. (1976) Hemicraniectomy in the
treatment of acute subdural hematoma: a re-appraisal. Surgical Neurology, 5,
2528.
Cooper, P. R., Hagler, H., Clark, W. K. and Barnett, P. (1979) Enhancement of
experimental cerebral edema after decompressive craniectomy: implications for the management of severe head injuries. Neurosurgery, 4,
296300.
421
Eisenberg, H. M., Jane J. A., Marmarou, A. et al. (1991) The Traumatic Coma
Data Bank: design methods and baseline characteristics. Journal of Neurosurgery, 75, S8S13.
Gaab, M. R., Rittierodt, M., Lorenz, M. and Heissler, H. E. (1990) Traumatic
brain swelling and operative decompression: a prospective investigation.
Acta Neurochirurgica (Supplement) (Vienna), 51, 326328.
Galbraith, S. and Teasdale, G. (1981) Predicting the need for operation in the
patient with an occult traumatic intracranial hematoma. Journal of Neurosurgery, 55, 7581.
Hamilton, M. and Wallace, C. (1992) Nonoperative management of acute
epidural haematoma diagnosed by CT: the neuroradiologists role. American Journal of Neuroradiology, 13, 853859.
Jayakumar, P. N., Kolluri, V. R., Basavakumar, D. G. et al. (1990) Prognosis in
traumatic intraventricular haemorrhage. Acta Neurochirurgica (Vienna), 106,
4851.
Jenkins, A., Mendelow, A. D., Graham, D. I. et al. (1990) Experimental
intracerebral haematoma: the role of blood constituents in early ischaemia.
British Journal of Neurosurgery, 4, 4552.
Kjellberg, R. N. and Prieto, A. J. (1971) Bifrontal decompressive craniotomy
for massive cerebral edema. Journal of Neurosurgery, 34, 488493.
Knuckey, N. W., Gelbard, S. and Epstein, M. H. (1989) The management of
asymptomatic epidural hematomas. A prospective study. Journal of
Levinthal, R. and Stern, W. E. (1977) Traumatic intracerebral hematoma with
stable neurological deficit. Surgical Neurology, 7, 269273.
Lewin, W. (1954) Cerebrospinal rhinorrhoea in closed head injuries. British
Journal Surgery, 42, 118.
Litofsky, N. S., Chin, L. S., Tang, G. et al. (1994) The use of lobectomy in the
management of severe closed head trauma. Neurosurgery, 34, 628633.
Loew, F., Pertuiset, B., Chaumier, E. E. and Jaksche, H. (1984) Traumatic,
spontaneous and postoperative CSF rhinorrhoea. Advances and Technical
Standards in Neurosurgery, 11, 169207.
Marshall, L. F., Bowers-Marshall, S., Klauber, M. R. et al. (1991) A new
classification of head injury based on computerised tomography. Journal of
Mathew, P., Oluoch-Olunya, D. L., Condon, B. R. and Bullock, R. (1993) Acute
subdural haematoma in the conscious patient: outcome with initial nonoperative management. Acta Neurochirurgica (Vienna), 121, 100108.
Miller, J. D. (1992) Head injury in adults. Neurosurgery Quarterly, 2, 2843.
289299.
North, J. B. (1971) On the importance of intracranial air. British Journal of
Surgery, 58, 826829.
OBrien, M. D. and Reade, P. C. (1984) The management of dural tear resulting
from mid-facial fracture. Head and Neck Surgery, 6, 810818.
Pickard. J. D., Bailey, S., Sanderson, H. et al. (1990). Steps towards cost- benefit
analysis of regional neurosurgical care. British Medical Journal, 301,
629635.
Pozzatti, E. and Tognetti, F. (1986) Spontaneous healing of acute extradural
haematoma: study of 22 cases. Neurosurgery, 18, 696700.
Ransohoff, J., Benjamin, M. V., Gage, E. L. Jr and Epstein, F. (1971)
Hemicraniectomy in the management of acute subdural hematoma. Journal
Raveh, J., Redli, M. and Markwalder, T. M. (1984) Operative management of 194
cases of combined maxillofacial frontobasal fractures: principles and surgical
modifications. Journal of Oral and Maxillofacial Surgery, 42, 555564.
Reilly, P. L. and Simpson, D. A. (1995) Craniocerebral injuries, in Craniomaxillofacial Trauma, (ed. D. J. David and D. A. Simpson), Churchill Livingstone,
Edinburgh, pp. 367396.
Reilly, P. L., Graham, D. I., Adams, J. H. and Jennett, B. (1975) Patients with
head injury who talk and die. Lancet, i, 375377.
Richards, T. and Hoff, J. (1974) Factors affecting survival from subdural
hematoma. Surgery, 75, 253258.
Rose, J., Valtonen, S. and Jennett, B. (1977) Avoidable factors contributing to
death after head injury. British Medical Journal, 2, 615618.
Sato, M., Tanaka, S., Kohama, A. and Fujii, C. (1987) Traumatic intraventricular haemorrhage. Acta Neurochirurgica (Vienna), 88, 95103.
Simpson, D. A., Speed, I. E. and Blumbergs, P. C. (1991) Embolism of cerebral
tissue: a cause of coagulopathy and cerebral infarction? Report of a case.
Simpson, D. A., Masters, C. L., Ohlrich, G. et al. (1996) Iatrogenic Creutzfeldt
Jacob disease and its neurosurgical implications. Journal of the Clinical
Neurosciences, 3: 118123.
Stein, S. C., Young, G. S., Talucci, R. C. et al. (1992) Delayed brain injury after
head trauma: significance of coagulopathy. Neurosurgery, 30, 160165.
Venes, J. L. and Collins, W. F. (1975) Bifrontal decompressive craniectomy in
the management of head trauma. Journal of Neurosurgery, 42, 429433.
Will, R. G. and Matthews, W. B. (1982) Evidence for case-to-case transmission
of Creutzfeldt-Jacob disease. Journal of Neurology, Neurosurgery and Psychiatry, 45, 235238.
Working Party Report (1994) Antimicrobial prophylaxis in neurosurgery and
after head injury. Infection in Neurosurgery Working Party of the British
Society for Antimicrobial Chemotherapy. Lancet, 344, 15471551.
21
NEUROPROTECTION
IN HEAD INJURY
Graham M. Teasdale and
Paul E. Bannan
21.1
Introduction
The concept of neuroprotection can be traced to the

use of cold by ancient Greek physicians to treat
injuries and stroke which they believed were related
to excess body heat (Hoff, 1986). Modern application
was allied to the development of major cardiac and
vascular surgery in the 1950s (Cooley, Mahaffey and
De Bakey, 1955) and in the 1960s neurosurgeons
cooled patients during intracranial vascular surgery
(Drake et al., 1964). Profound hypothermia for head
injury was described by Lazorthes and Campan in
1958; systemic complications limited its benefit and it
lost favor, but interest in moderate hypothermia has
returned recently.
In the last decade, the impressive reductions in the
extent of brain damage that can be obtained experimentally from a variety of methods of neuroprotective
treatment have raised high expectations of clinical
benefit in stroke and head injury. Nevertheless, at the
time of writing these hopes remain unfulfilled and a
succession of trials have failed to show consistent
significant improvement in outcomes. Although this
has led to some skepticism, the results so far may
reflect only the inadequacies of the agents concerned
at least in the dosages and regimens employed and in
the patient populations treated. The prospects of
benefit from neuroprotection remain under intensive
investigation.
The original concept of neuroprotection depended
upon the initiation of treatment before the onset of
the event leading to brain damage and the methods
employed aimed to minimize the intensity of insult
or its immediate effects upon the brain. This reflected
the experience of investigators, who found that the
methods then available often failed to work, and
could even be harmful, when started after the insult.
More recently, the concept of neuroprotection has
been extended to include treatment started after the
onset of an insult (Figure 21.1). This change reflects
new understanding of the mechanisms of brain
damage and the development of an immense range
of specifically targeted, powerful, neuropharmacological treatments (Table 21.1).

Much research has been focused on the treatment of
cerebral ischemia but the findings, and interventions
employed, are likely also to be relevant in the
management of head injury. This reflects three
factors:

the increasing clinical awareness of the importance

of ischemic damage, due to secondary insults, in
worsening outcome after head injury;
the evidence that trauma heightens the vulnerability of the brain to ischemia;
the convergence of the concepts of the cellular and
molecular mechanisms of progressive ischemic and
traumatic brain damage.
21.1.1
SECONDARY INSULTS
It has become clear that hardly ever is it appropriate in

clinical practice to consider any severe head injury as
a single event. Instead, it is common for patients to
survive the initial injury but then suffer additional
damage as a result of delayed complications. These
have been highlighted clinically as secondary insults
(Miller and Becker, 1982) and neuropathological evidence of their occurrence, in the form of ischemic
damage, is still common in fatal cases (Graham et al.,
1989). The abundant evidence that these insults
significantly worsen outcome has been dealt with in
depth in previous chapters (e.g. Chapters 4, 5, 7 and 8).
There is therefore ample opportunity for treatment
started after the head injury to act as a pretreatment for subsequent secondary insults.
21.1.2 THE ADDITIVE EFFECTS OF ISCHEMIA AND
TRAUMA
The experimental work of Jenkins and colleagues

(1989) established the heightened vulnerability to
ischemic brain damage following an experimental
head injury. They found that a brief ischemic insult,
424
NEUROPROTECTION IN HEAD INJURY
The concept of neuroprotection.
Figure 21.1
too short by itself to cause damage, produced structural changes when preceded by a mild traumatic
injury. The mechanisms involved may include a loss of
cerebrovascular reactivity (Lewelt, Jenkins and Miller,
1980) and disturbances of cerebral function, metabolism, and ionic and neurotransmitter homeostasis
(Katayama et al., 1990).
21.1.3
PROGRESSION OF DAMAGE: CASCADES
Events thought previously to be completed and

irreversible within moments of injury are now under-
Table 21.1
Main targets and methods of neuroprotection
Mechanism
Neuroprotective method
Energy failure
Hypothermia
Barbiturates
Cell swelling
Diuretics
Acidosis
THAM
Free radicals
Superoxide dismutase
Lipid peroxidation
Steroids, amino steroids

Indomethacin
Calcium damage
Calcium antagonists
Neurotransmitter
Antagonists to glutamate, etc.
stood to lead to persisting damage only as a consequence of processes of maturation, with timescales ranging from minutes to hours and even days
(Siesjo,
1992a, b; Povlishock, 1992). Although complex,
these processes are open to amelioration by a wide
range of interventions.
In this chapter we consider the main classes of
proposed neuroprotective treatments, their actions
and the clinical experience of their use. However, this
is a rapidly developing and changing field and the
evidence, and our conclusions, are likely to be
superseded even by the time of publication.
21.2 From preclinical research to clinical

benefit
The intense focus on the discovery and development
of neuroprotective agents is reflected in descriptions
of a host of potential agents in a vast literature.
Readers are referred to articles by Faden and Salzman
(1992), McIntosh (1993) and Cohadon (1994) for
reviews. It is clear that only a few of the agents
effective experimentally can ever be investigated
under clinical circumstances. It is important that
these are well chosen, on the basis of as complete as
possible a portfolio of preclinical investigations
(Table 21.2). These should include studies in a relevant
range of models, under carefully defined experimental
TREATMENTS UNDERGOING CLINICAL EVALUATION
Table 21.2 Translation of neuroprotection to patients:

information needed from preclinical investigations

Mechanistic rationale for benefit and S/E

Evidence of efficacy: robustness > magnitude
Knowledge of limits to potency and S/E time, dosage
Regimens pharmacokinetics
Surrogate effect at desired dose
enhancement of cerebral perfusion, avoidance of

hyperglycemia, hypothermia and the use of hyperventilation, drainage of CSF and diuretics to counteract or prevent raised intracranial pressure.
21.3.1
(a)
conditions, using rigorous measures of outcome and
establishing clear dosing parameters. The initial clinical studies need to be equally thorough to establish
the appropriate doses in patients (Table 21.3). Failure
to meet these requirements may account for much of
the lack of success in phase 3 studies so far. It is
becoming clear that the challenge is not so much to
achieve benefit experimentally but to combine efficacy
with an acceptable margin of safety from adverse side
effects in patients.
21.3 Treatments undergoing clinical

evaluation
Two main avenues can be discerned. In the first the
focus is on improving metabolism and microenvironment; methods include hypothermia and barbiturates
to minimize the effects of energy failure, THAM to
correct acidosis and mannitol to counteract brain
swelling and edema. In the second the approaches are
pharmacological: glucocorticoids to suppress inflammation, superoxide dismutase and new amino steroids to antagonize free radicals and lipid peroxidation, indomethacin to minimize lipid degradation,
calcium antagonists to limit either the movement of
calcium or the intracellular effects of increased calcium, and antagonists to glutamate excitotoxicity and
other forms of neurotransmitter-induced damage.
Some agents may have several potentially beneficial
effects, and in some approaches a package or cocktail
of agents is employed in the hope of an additive effect.
Neuroprotective treatments are used along with strategies designed to optimize the milieu interieur, e.g.
Table 21.3 Translation of neuroprotection to patients:

aims of initial clinical program

Dose related tolerability

Volunteers, patients
Loading, maintenance
Pharmacokinetics, pharmacodynamics
Factors in variability
Experience of regimen chosen for efficacy
Knowledge of S/E: nature, incidence, duration,
management
Effect upon some marker?
Interactions with conventional treatments
425
HYPOTHERMIA
Rationale and activity
Hypothermia decreases cerebral blood flow by

approximately 5.2% per degree of reduction in body
temperature. Cerebral metabolic rate for oxygen
(CMRO2) and the arterio-jugular-venous oxygen difference (AVDO2) fall after the institution of moderate
hypothermia. This reflects a reduction in energy
requirements and hence less energy loss in injured
brain; stabilization of cell membranes (Ginsberg et al.,
1992) and reduction of neurotransmitter turnover may
also contribute to the benefit seen in models of
ischemia (Busto et al., 1989). Clifton et al. (1993)
showed, in an acute percussion head model in the rat,
improvement in behavioral outcome and tissue preservation when hypothermia was used during and
after an insult, but the effect was lost when cooling
was withheld for 30 minutes or longer (Clifton et al.,
1991). Hypothermia reduced brain damage in a model
of focal injury from compression (Pomeranz et al.,
1993).
Hypothermia has been associated with several
complications, including cardiovascular instability
(mainly arrhythmias) coagulopathies, hypokalemia
and an increased risk of infection. Prolonged treatment has resulted in coagulopathy and necrosis. These
complications have been particularly associated with
profound hypothermia (below 30C) and are considered to be less frequent at moderate hypothermia
(3233C).
(b)
Clinical status
Several series have reported on relatively small

numbers of patients treated with hypothermia and
compared with control subjects (Table 21.4). Shiozaki
and colleagues studied subjects selected because they
had persisting raised intracranial pressure, despite all
conventional therapy, and aimed for a temperature of
34C sustained for 2 days (Shiozaki et al., 1993).
Compared with controls, after initiating hypothermia
the patients had a lower average intracranial pressure,
a higher cerebral perfusion pressure and a higher rate
of independent outcome. They drew attention to
complications during rewarming from hypothermia,
including shock and raised intracranial pressure.
Marion et al. (1993) and Clifton et al. (1993) used
moderate hypothermia of 3233C, sustained for 24 or
48 hours. Clifton et al. noted no difference in intra-
426
Table 21.4
Randomized comparisons of normothermia and hypothermia in the management of severe head injury
Shiozaki et al., 1993
Marion et al., 1993
Clifton et al., 1993
Control
Hypothermia
Control
Hypothermia
Control
Hypothermia
20
20
17
16
22
23
Dead
Vegetative
Severe disability
2
3
7
0
2
6
14
1
1
8
1
1
14
11
Moderate disability
Good recovery
4
4
7
5
0
1
1
5
12
Total patients
cranial pressure in hypothermic and control groups

but a tendency for blood pressure and cerebral
perfusion to decline in the former. Marion et al., in
contrast, observed a reduction in intracranial pressure
but did not provide data on cerebral perfusion
pressure. Each of these studies also showed a trend to
an increased rate of independent outcome in patients
treated with hypothermia compared with the controls.
The foregoing results pointed to the need for a
definitive, randomized study with a sufficient large
number of patients and this trial is under way in
several North American centers.
21.3.2
(a)
THAM
Increased lactate and acidosis in the CSF have been

recognized in clinical investigations for many years
(Kurze, Tranquada and Benneact, 1966; Gordon, 1971),
and in experimental head injury models (Unterberg et
al., 1988). They have now been confirmed in brain
tissue in animals by neurochemical methods (Prasad et
al., 1994) and by magnetic resonance techniques in
animals (McIntosh et al., 1987) and in man (OluochOlunya et al., 1996). This has stimulated studies of
alkalizing agents such as THAM; (tromethamine;
(hydroxy-methyl) amino methane). Studies in cats
treated with THAM after fluid percussion injury
showed an improved energy state and less edema at 8
hours (Yoshida, Corwin and Marmarou, 1990), a lower
ICP and increased survival (Rosner and Becker, 1984).
Benefit was also seen in models of local brain injury
(Akioka et al., 1976; Gaab et al., 1980). Side effects of
THAM result from the production of bicarbonate,

causing an osmotic diuresis but also requiring
hyperventilation.
(b)
Clinical studies
Wolf et al. (1993) studied 49 patients with a severe

head injury. Three experimental groups were required
because of the need to combine THAM treatment with
hyperventilation to eliminate CO2. Tromethamine was
given for 5 days. Although intracranial pressure was
better controlled in the patients who received THAM,
there was no evidence of benefit on outcome (Table
21.5). An important result of the study was the poorer
outcome in the patients treated by hyperventilation
alone (Muizelaar et al., 1993). This provided strong
additional evidence for the dangers of hyperventilation. Indeed, the authors believed that an effect of
THAM was to counteract the ischemia produced by
marked hypocapnia in the group treated by hyperventilation alone. Although there is some continuing
support for using THAM and hyperventilation selectively, to assist control of intracranial pressure, it
cannot be recommended for routine use.
21.3.3
(a)
MANNITOL
Mannitol is widely used in neurosurgery to treat raised

intracranial pressure, to decrease brain bulk during
intracranial operations and to treat cerebral ischemia. It
has been shown to improve cerebral blood flow in a
primate model of head injury (Johnston and Harper,
Table 21.5 Randomized comparison of outcome of treatment of severe head injury with tromethamine (THAM),
hyperventilation and control cases (Source: from Muizelaar et al., 1993)
Control
Total
41
Dead/Veg/Severe
26
Hyperventilation
Total
36
Dead/Veg/Severe
28
THAM + hyperventilation
Total
36
Dead/Veg/Severe
23
1973). A 20% mannitol solution, given as a bolus dose,

has been shown to reduce raised ICP for up to 60
minutes in a rabbit model of cryogenic brain injury
(Zornow, Oh and Schelle, 1990). In an animal model of
brain retraction injury, mannitol was shown to increase
cerebral blood flow under the area of the brain retractor
but did not increase electrical activity. Mannitol was
thought to be more neuroprotective when given in
small, frequent doses, rather than by continuous
infusion (Andrews, Bringas and Muto, 1993). In the
injured brain or spinal cord the effect of mannitol on
cerebral blood flow has been related to compensatory
cerebral vasoconstriction in response to blood viscosity
(Muizelaar et al., 1983). Other neuroprotective effects of
mannitol have been related to improvement in the
microcirculation and the amelioration of cerebral
edema (Tanaka and Tomonaga, 1987).
Mannitol given in high doses induces hypernatremia, decreases hematocrit and increases osmolality
(Andrews, Bringas and Muto, 1993). It may also cause
acute renal failure. The pathogenesis of this is not yet
established but may be associated with renal vasoconstriction produced by high concentrations of mannitol
(Dorman, Sondheimer and Cadnapaphornchai, 1990).
Other detrimental effects of mannitol include hypotension, acidosis and hyperkalemia (Chapters 17 and 18).
(b)
Clinical status
Schwartz and colleagues (1984) showed that the use of

mannitol to treat raised ICP in severely head-injured
patients led to a better CPP than the use of barbiturates. Smith and colleagues (1986) randomized 80
patients to treatment with mannitol empirically
(0.25 g/kg 2-hourly and 0.75 g/kg after any neurological deterioration) or as guided by an ICP of more
than 25 mmHg. Favorable outcomes occurred in 48%
and 54% (i.e. not significantly different). Attempts are
now being made to identify the cause of intracranial
hypertension and to match the treatment with the
cause. Patients in whom raised ICP is thought to be
caused by increased brain water due to focal edema
related to a focal space-occupying lesion, contusion or
hematoma may best respond to osmotherapy (Miller,
Piper and Dearden, 1993). This is in accord with the
finding of Mendelow et al. (1985) that mannitol
increases blood flow in the most damaged hemisphere
in patients with a focal injury.
21.3.4
(a)
BARBITURATES
The main mechanism by which barbiturates are

neuroprotective has not been established (Moskopp et
al., 1991). Effects include a decrease in cerebral
metabolic rate, to a more profound extent than the
427
accompanying reductions in cerebral flow, as a reflection of a decrease in functional activity of the brain
(Nordstrom and Siesjo,
1978). Cerebral blood volume
falls, intracranial pressure is reduced and edema is
less. Other possibilities include a scavenging of oxygen free radicals (Flamm et al., 1977) and a stabilizing
of cell membranes (Demopoulos et al., 1977). Some
reports indicate that, when given before production of
either focal or diffuse cerebral ischemia or hypoxia,
barbiturates reduce subsequent damage (Michenfelder, Milde and Sundt, 1976), but later reports
disagree (Steen, Milde and Michenfelder, 1989). When
treatment is delayed until after the insult, the effects
can be deleterious (Spetzler and Hadley, 1989). Clinically, barbiturate administration has beneficial effects
in cerebrovascular surgery (Sundt, Piepgras and Ebersold, 1982) but not in cardiac surgery (Scheller, 1992),
nor after cardiac arrest.
The main complication of the use of barbiturates is
arterial hypotension which occurs in up to 58% of
patients (Schalen, Messeter and Nordstrom, 1992). The
decline in blood pressure may be greater than the
reduction in ICP, so that CPP actually falls, especially in patients with hypovolemia or cardiac disease. Other delayed complications include hypokalemia, hypernatremia and increased risk of infection.
Liver and renal dysfunction and cardiac failure have
also been reported (Oda et al., 1992; Schalen, Messeter
and Nordstrom, 1992).
(b)
Clinical status
Several studies show that the administration of barbiturates in doses sufficient to cause burst suppression of
the EEG reduces raised intracranial pressure in headinjured patients (Marshall, Smith and Shapiro, 1979;
Bricolo and Glick, 1981; Rea and Rockswold, 1983).
This effect is seen even in patients with severe rises
refractory to other treatments (Eisenberg et al., 1988)
and is best seen in patients with preserved cerebral
metabolic activity and cerebrovascular reactivity to
carbon dioxide (Nordstrom et al., 1988). Unfortunately,
this immediate effect on intracranial pressure has not
translated into consistent evidence of improved outcome, even though a variety of trial designs have been
employed.
Schwartz and colleagues (1984) allocated 59 patients
with severe head injury and raised intracranial pressure to treatment first either with barbiturate or with
mannitol and allowed crossover when the first
method was judged unsuccessful. Barbiturate treatment was not more successful than mannitol as a first
intervention. The group treated initially with barbiturate had a lower cerebral perfusion pressure and a
trend to a poorer outcome. Eisenberg et al. (1988), in a
multicenter study, randomized patients with refrac-
428
tory intracranial pressure to barbiturates or conventional treatment. Although barbiturates were significantly more successful in reducing intracranial
pressure, this was not reflected in an overall improvement in outcome. One possibility raised by these
studies was that treatment had begun too late, after
severe brain damage had been sustained. Therefore, it
was thought more fruitful to commence early treatment as a routine, in an effort to prevent the
development or progression of damage. Ward et al.
(1985) studied patients with a severe head injury and
compared the outcome of patients prophylactically
treated with barbiturates with randomized controls
and found no benefit. An overview of data from major
randomized studies does not show evidence to support a marked benefit of barbiturates (Table 21.6).
Barbiturate treatment, therefore, does not have a
place in the routine management of severe head injury,
although it can improve physiological values, especially ICP, in the acute stage. Similar effects can be
obtained by other new hypnotic agents, for example
propofol, that have shorter durations of action and
are more convenient to use.
21.3.5
CORTICOSTEROIDS
The publication of the second National Acute Spinal

Cord Injury Study (NASCIS II) renewed interest in the
neuroprotective pharmacology of glucocorticoids. The
trial showed a significant improvement in neurological recovery in those patients treated with highdose methylprednisolone within 8 hours of their
spinal cord injury (SCI; Bracken et al., 1990). An
interesting observation was the need both for high
intravenous dosage (30 mg/kg), and for early treatment, i.e. less than 8 hours postinjury.
(a)
Rationale and actions
acid release; Hall, 1992). It is also thought to have a

number of other effects, including the maintenance of
normal tissue blood flow, maintenance of aerobic
energy metabolism and reversal of intracellular calcium accumulation. The neuroprotective activity of
methylprednisolone is independent of its glucocorticoid activity. A single, large intravenous dose
enhanced early neurological recovery after head injury
in a rodent model (Hall, 1985). Prednisolone when
given in equal dose with methylprednisolone was
equally efficacious but half as potent in the same rodent
model. It is thought that both these glucocorticoids are
equally efficacious as lipid antioxidants
(b)
A number of studies have demonstrated the failure of

high-dose dexamethasone to influence ICP or outcome
in patients with severe TBI. Thus, a study from
Richmond, Virginia, showed that high-dose methylprednisolone (40 mg 6-hourly) did not decrease ICP in
patients with severe head injury. There was a high
incidence of gastric hemorrhage (50%) and hyperglycemia (85%; Gudeman, Miller and Becker, 1979). An
overview of major published studies does not suggest
efficacy, even with the use of dexamethasone treatment
up to 96 mg/d (Todd and Teasdale, 1989; Table 21.7).
Nevertheless steroids are still used in many neurosurgical centers in the USA and Germany (Ghajar et
al., 1995) and a few in the UK (Malta and Menon, 1996),
despite the lack of proven efficacy. A recent overview
has suggested that the question is still not definitively
answered (Roberts, Anderson and Rowan, 1996).
21.3.6 FREE RADICAL SCAVENGERS AND
INHIBITORS OF LIPID PEROXIDATION PEG-SOD
(a)
Methylprednisolone is thought to be an inhibitor of

lipid peroxidation and of lipid hydrolysis (arachidonic
Table 21.6 Overview of outcome in major randomized

trials of barbiturates in severe head injury
Clinical status
Oxygen-radical-mediated lipid peroxidation is

thought to be an important factor in post-traumatic
neuronal degeneration (Hall, 1989; Siesjo,
Agardh and
Bergtsson, 1989) The arachidonic acid cascade metabolites contribute to ischemia of the gray matter and
Protocol
Conventional first
Schwartz et al., 1984

Ward, 1985
Yano, 1986
Eisenberg et al., 1988
Totals
Mortality
Barbiturate first
Total
Dead
Total
Dead
31
26
60
36
153
14
13
37
19
83
28
27
68
37
160
170
14
40
23
94
54%
59%
Table 21.7 Overview of main randomized high-dose

steroids in head injury (Source: data from Faupel, 1986;
Cooper, 1978; Saul, 1981; Braakman et al., 1983; Dearden,
1986; Gianotta, 1987; Gaab, 1992)
Total (n)
Dead/vegetative (%)
Severe disability (%)
Moderate/good recovery (%)
Controls
Treated
399
40
10
51
482
40
12
48
stimulate production of oxygen free radicals. Oxygen

free radicals have been implicated in the pathogenesis
of microvascular damage (Kontos and Povlishock,
1986) brain edema, cerebral ischemia and spinal cord
trauma (Kontos and Povlishock, 1986). These highly
reactive free radicals cause peroxidation of membrane
phospholipids and oxidation of cellular protein and
nucleic acid, and can attack both neuronal membranes
as well as the cerebral vasculature (McIntosh, 1993).
Free radical scavengers such as super oxide dismutase
(SOD) reduce mechanically induced injury of brain
cells (McKinney et al., 1996) and reduce infarction in
experimental focal reperfusion ischemia, but have a
narrow therapeutic dose range (Liu, Beckman and
Freeman, 1989). Studies in experimental brain injury
show reduced microvascular damage (Wei et al., 1981)
and reduced bloodbrain barrier opening in cerebral
edema (Chan, Longar and Fishman, 1987; Schettini,
Lippman and Walsh, 1989). SOD has a very short
biological half-life, which limits its clinical utility, but
conjugation with polyethylene glycol (PEG) extends
the half-life of SOD to approximately 5 days.
(b)
Clinical status
A phase 2 trial using polyethylene-glycol-conjugated

super oxide dismutase (PEG-SOD) was reported to
show both safety and improved outcome in headinjured patients (Muizelaar et al., 1993). No complications were noted in the group using PEG-SOD and
unfavorable outcome (dead, vegetative or severe
disability) was reduced by one-fifth (Table 21.8). This
initial pilot study prompted the conduct of further
large trials to seek definitive evidence. Young et al.
(1996) reported on patients allocated to control or a
single administration of either 10 000 U/kg or
20 000 U/kg of PEG-SOD. A total of 463 patients were
studied and the groups were well balanced for major
prognostic factors. The primary comparison was the
429
outcome of patients at 3 months after injury, separating favorable (moderate/good) versus unfavorable
(dead/vegetative/severe). Patients treated with PEGSOD had a higher proportion of favorable outcomes
than controls and the difference was most marked in
the 10 000 U/kg group. However, outcome was not
significantly different at 6 months (Table 21.8). This
study indicated that no additional benefit was likely to
be obtained from the higher dosage of PEG-SOD and
that a further study was needed, powered to detect a
more realistic difference than had been suggested by
the pilot study. This further study, involving more
than 1000 patients, has been completed and did not
show a significant effect upon outcome (Muizelaar,
1996, personal communication).
21.3.7
(a)
INHIBITORS OF LIPID PEROXIDATION
The 21 amino steroids were introduced as a new class

of lipid peroxidation inhibitors, lacking glucocorticoid
effects. Of these, tirilazad mesylate (U-74006F) is the
most studied. It is a potent inhibitor of iron-catalyzed
lipid peroxidation in brain homogenates, decreases
iron-induced damage to cultured cortical neurons and
inhibits peroxidation in systems that contain neither
membranes or iron (Hall et al., 1989). Unfortunately,
these effects are of uncertain relevance because tirilazad does not cross the cerebral endothelium to enter
the brain. Any beneficial effects are, therefore, likely
to be related to a reduction of microvascular damage.
The effects of tirilazad in experimental ischemia have
been variable: both benefit (Park and Hall, 1994;
Perkins et al., 1991; Xue, Silvka and Buchan, 1992) and
lack of efficacy (Hellstrom et al., 1994; Karlsson et al.,
1994) have been reported and the inconsistencies are
not simply a reflection of discrepancies in protocol or
the choice of focal or global insult or transient rather
Table 21.8 Overview of comparison of effect of PEG-SOD 10 000 U/kg and placebo on outcome at 3 months after
severe head injury (Source: Muizelaar, 1996, personal communication)
% Favorable outcome*
Treatment difference (%)
Protocol No.
Placebo
10 000 U/kg
Absolute
Relative
p value
007
Phase 2
45.8
(n = 24)
60.0
(n = 25)
14.2
31.0
0.32
005
Phase 3
46.3
(n = 162)
55.7
(n = 149)
9.4
20.3
0.11
006
Phase 3
47.0
(n = 487)
48.9
(n = 483)
1.8
3.9
0.61
007/005/006
Integrated summary
46.8
(n = 673)
50.8
(n = 657)
4.0
8.6
0.15
* Good recovery and moderate disability outcomes combined
430
than permanent ischemia. Tirilazad was reported to

improve early neurological recovery and survival
after experimental head injury (Hall et al., 1988;
McIntosh et al., 1992; Dumlich et al., 1990), to reduce
post-traumatic hydroxyl radical formation and to
attenuate traumatically induced opening of the BBB
(Mathew et al., 1996).
of 30 mg followed by 30 mg an hour has been shown to

reduce ICP below 20 mmHg for several hours (Biestro
et al., 1995). Cerebral blood flow decreased and there
was a fall in rectal temperature (Jensen et al., 1991;
Cold et al., 1990). However, evidence that this
improves outcome is lacking.
21.3.9
(b)
Two large Phase 2 studies utilizing tirilazad mesylate

in patients with head injuries have been performed,
one in the United States and Canada, involving 1198
patients, the other in Europe and Australia, involving
1132 patients (Marshall and Marshall, 1996). The
results of the latter are awaited, but the North
American study was stopped prematurely because of
concern by the monitoring committee about an excess
mortality in patients treated with tirilazad. This may
reflect imbalance in the severity of the groups. It is,
nevertheless, unlikely that this compound will enter
routine clinical use in head injuries. The second
generation of 21 amino steroids couple the function of
the tirilazad with the 2-methyl amino chroman ring
structure of Vitamin E. They have a greater antioxidant and free radical scavenging activity and exert
a more potent protective action than tirilazad (Hall,
1993).
21.3.8 MODULATORS OF ARACHIDONIC ACID
METABOLISM INDOMETHACIN
(a)
Rationale and actions
Calcium-activated proteases and lipases attack cell

membranes, degrading phospholipids and arachidonic acid into thromboxanes, prostaglandins and
leukotrienes (Leslie and Watkins, 1985). The arachidonic acid cascade has been implicated as a pathophysiological factor in models of experimental cerebral
edema, cerebral ischemia, cerebral vasospasm and
spinal cord trauma (Awad et al., 1983; Fukumori et al.,
1983; Hallenbeck, Jacobs and Faden, 1983). Cyclooxygenase inhibitors, e.g. ibuprofen and indomethacin, improved cerebral metabolism and blood flow in
rats following a cortical cryogenic lesion and
improved neurological function after a weight-drop
brain injury in mice (Pappius and Wolfe, 1983; Hall,
1988).
(a)
Clinical status
There has been recent interest in the effects of

indomethacin in human head injury and its effects on
intracranial pressure, cerebral blood flow and cerebral
metabolism. Indomethacin, given as a bolus injection
The initial agents available for clinical use were

introduced with the aim of reducing calcium overload
from all sources but were particularly active on
voltage-operated L channels. Typified by the dihydropyridine nimodipine, their vascular effects were
considered to be preferential for cerebral vessels. Their
actions included increases in cerebral blood flow, and
prevention or reversal of experimental vasospasm but
also hypotension at higher doses. The balance
between such vascular effects and benefits due to
reduction in neuronal calcium rises remains unclear.
Nevertheless, nimodipine treatment significantly and
reasonably consistently reduced damage in both focal
(Mohammed et al., 1985; Hadley et al., 1989) and global
ischemia (Steen et al., 1985), including reperfusion
models. Calcium antagonists have improved outcome
in models of traumatic and hemorrhagic damage
(Sinar et al., 1988). Clinically, nimodipine has an
undoubted benefit in the pretreatment of delayed
ischemia after spontaneous subarachnoid hemorrhage
(Pickard et al., 1989; Robinson and Teasdale, 1990). It
does not have a major effect on ischemic stroke and
given intravenously can cause hypotension in such
patients (Wahlgren et al., 1994), but a meta-analysis
has suggested a possible benefit if treatment is
commenced within 12 hours of onset of stroke (Mohr
et al., 1994). Compton et al. (1990) observed reductions
in cerebral blood flow velocity in head-injured
patients treated with nimodipine and considered that
this reflected a relief of vasospasm.
A new range of N calcium-channel blockers,
derived from spider venoms, block neurotransmission
and are effective in ischemia, even when treatment is
started some hours after an insult. Hypotension is a
side effect but one compound (SNX111; Omega Conotoxin) is being evaluated for safety and tolerability in
severe head injuries.
(b)
(b)
CALCIUM CHANNEL ANTAGONISTS
Clinical status
Clinical status
In the first two substantial randomized trials of

nimodipine in head injury, severely injured patients,
selected by their inability to obey commands, received
12 mg/h intravenously for 7 days (Table 21.9). In the
first, HIT I trial (Bailey et al., 1991), 351 patients in
431
Table 21.9
Effect of nimodipine on outcome of severe head injury

HIT II
HIT I
Total patients
Dead
Vegetative
Severe disability
Moderate disability
Good recovery
Missing
Unfavorable (D/V/S)
Favorable
Placebo
Nimodipine
Placebo
Nimodipine
75
50
2
37
33
53
176
49
8
25
31
62
1
82 (47%)
93 (53%)
429
98
19
51
98
148
15
168 (41%)
246 (59%)
423
90
16
57
101
144
18
160 (40%)
245 (60%)
89 (51%)
86 (49%)
whom treatment was commenced within 24 hours of

injury were analyzed. Unfavorable outcome (dead,
vegetative, severe disability), occurred in 51% of
control patients and 47% of treated patients. This trend
to an 8% relative decrease in unfavorable outcome was
not statistically significant but prompted a second
study, HIT II (European Study Group on Nimodipine
and Severe Head Injury, 1994). A total of 819 patients
were treated within 12 hours of becoming unable to
obey commands and within 24 hours of injury.
Overall, the difference in outcome was small: 40.6%
unfavorable in the placebo group and 39.5% unfavorable in nimodipine treated patients. Various subgroups were analyzed and in the group considered by
the investigators to show CT scan evidence of
traumatic subarachnoid hemorrhage, unfavorable
outcome was reduced from 56% in placebo patients to
45% in treated subjects. The CT interpretation by the
investigators was not always borne out by a Review
Committee, but where the latter considered subarachnoid to be present there was a similar reduction
in unfavorable outcome, from 60% to 52%. The
importance of traumatic subarachnoid hemorrhage
had not been appreciated when the HIT I study was
started (Kakarieka, Braakman and Schakel, 1994).
Since then correlations have been established between
Table 21.10
a large amount of blood on the CT scan, a high bloodflow velocity assessed by ultrasound, focal reductions
in blood flow and CT evidence of focal infarction.
Therefore, the data in the HIT I study were analyzed
retrospectively, after a review of the CT scans to select
those with subarachnoid hemorrhage. This showed
that 69% of patients with subarachnoid hemorrhage
given placebo had a poor outcome and 77% of patients
given nimodipine had a poor outcome (Murray,
Teasdale and Schmitz, 1996).
These findings prompted a new trial, carried out in
21 German neurosurgical centers (Harders et al., 1996).
The study included severe, moderate and even mildly
injured patients and subjects were selected primarily
on the basis of the investigators opinion that the CT
scan showed a traumatic subarachnoid hemorrhage
which was not supported by the Review Committee in
21% of cases. Nimodipine was given intravenously for
710 days and thereafter orally for 21 days. Unfavorable outcome occurred in 46% of the placebo group
and 25% of the nimodipine group. High transcranial
Doppler velocity was less frequently seen in nimodipine-treated patients. Hypotension was more frequent
in the nimodipine-treated patients, but this did not
adversely effect their outcome (Table 21.10). Thus there
is a plausible rationale for the use of nimodipine in
Outcome for patients classified by presence or absence of SAH on CT scan

HIT I*
HIT II*
German Study Group
Treatment group
Placebo
Nimodipine
Placebo
Nimodipine
Placebo
Nimodipine
Total
Not SAH
Poor outcome
133
97
40 (41%)
124
89
34 (38%)
414
269
81 (30%)
405
282
96 (34%)
61
60
SAH
Poor outcome
36
25 (69%)
35
26 (74%)
145
87 (60%)
123
64 (52%)
28 (46%)
15 (25%)
* Review group assessment on CT scan

Investigators assessment of CT scan
432
traumatic subarachnoid hemorrhage, even when it is

borne in mind that its benefits in spontaneous
subarachnoid hemorrhage do not appear to be as a
result of reduction of angiographic vasospasm (Pickard et al., 1989).
The debate at the moment is whether the existing
clinical evidence is sufficiently definitive for nimodipine to be adopted as a routine in traumatic SAH or
whether, given the disparities in evidence from the
three studies and the lack of complete consistency in
the findings, there should be a further large prospective randomized study. The extensive experience
already gained with nimodipine in neurosurgery
indicates that, although its efficacy is not overwhelming, this is balanced by a considerably better
safety and tolerability profile than several other
pharmacological approaches. It should be possible,
therefore, to conduct a mega trial of nimodipine
much more easily (and cheaply) than for certain other
potentially more powerful proposed interventions.
21.3.10
(a)
NEUROTRANSMITTER ANTAGONISTS
Introduction
The recognition of the role of agonist-gated channels

in allowing calcium entry has focused intense research
on the role of neurotransmitters in the progression of
ischemic and traumatic brain damage. The principal
interest has focused on the role of glutamate. Currently, the clinical value of glutamate-receptor
antagonists is the most important issue in neuroprotection research.
Glutamate is localized presynaptically and released
in a calcium-dependent fashion on electrical stimulation. Glutamate receptors are the main mediators of
synaptic excitation of the central nervous system.
They are involved in many important physiological
processes, including sensory information handling,
learning and memory. At least four distinct glutamate
receptor subtypes are recognized. Most interest is
focused on the N-methyl-D-aspartate NMDA receptor, which is coupled to a sodiumcalcium channel.
The NMDA receptor is modulated by glycine and
polyamine coantagonists, which facilitate opening of
the ion channel. This also requires membrane depolarization, in order to overcome a voltage-dependent
block by magnesium ions. Glutamate levels are
normally tightly controlled and there are important
uptake mechanisms into presynaptic terminals and
into glia. Other ionophore-coupled glutamate receptors, such as 2-amino-3 hydroxy-5 methyl isoxazole-4
propionic acid (AMPA) and kianate (KA), are more
important in sodium and potassium exchange but are
also important targets for neuroprotective drugs. The
metabotropic glutamate receptor is linked to phospho-
lipase C and activation leads to an increase in

intracellular calcium.
High concentrations of glutamate have been known
for some years to be neurotoxic (Rothman and Olney,
1986; Olney et al., 1978, 1989). In experimental models
of cerebral ischemia and trauma, and in human
trauma and stroke there are immediate increases in
extracellular concentrations of glutamate and aspartate at levels of blood flow that are associated with
membrane failure and the development of neuronal
damage (Zauner and Bullock 1995; Bullock et al.,
1996). In models of primary diffuse head injury the
elevation in glutamate is transient (Katayama et al.,
1990) but it is more long-lasting in models of subdural
hematoma in which there is associated focal cortical
ischemic damage underlying the hematoma (Miller et
al., 1990; Chen et al., 1992). These observations in
animals are now being replicated in studies in patients
with severe head injury (Bullock 1995) and stroke,
using microdialysis (Bullock et al., 1996). The results
show that high levels of glutamate occur in regions of
focal cerebral damage, in association with cerebral
blood flow below the threshold value of 20 ml/
100 g/min, in particular in association with raised
intracranial pressure.
Much of the evidence for the importance of glutamate in producing brain damage, and the stimulus to
clinical applications, has come from the wealth of
evidence that administration of glutamate antagonists,
in particular against the NMDA receptor, are markedly efficient in reducing the amount of damage
produced experimentally (McCulloch, 1995; McCulloch et al., 1991; Myseros and Bullock, 1995). This is
most clearly seen for NMDA antagonists in models of
permanent focal ischemia (Bullock 1990) where the
evidence is overwhelmingly consistent. There is an
emerging view that NMDA antagonists are not so
consistently effective in models of diffuse ischemia
(Buchan and Pulsinelli, 1991; Pulsinelli, Sarokin and
Buchan, 1993), where antagonists of AMPA and other
receptors may have more potential benefits. In focal
ischemia, benefit has been consistently found with
pretreatment and, in many studies, benefit has been
also obtained from post-treatment.
Pretreatment with an NMDA antagonist has
reduced behavioral effects in a model of diffuse injury.
In the model of subdural hematoma, pretreatment
reduces the extent of ischemic damage in the cortex
(Chen et al., 1991) and, perhaps as a reflection of this,
lowers intracranial pressure (Kuroda et al., 1991). The
lack of efficacy of NMDA antagonists in conditions of
extremely dense ischemia may reflect the over-riding
prime importance of non-receptor-mediated calcium
overload when there is severe energy failure. In these
circumstances, other antagonists, e.g. AMPA, may be
more beneficial.
Blockade of an important excitatory neurotransmission mechanism can be expected to have side effects to
offset against potential benefit and the effects of the
blockade of NMDA receptors were foreseen by the
effects of phencyclidine (PCP). This non-competitive
NMDA-channel antagonist was used in psychiatric
research as a model of schizophrenia but rapidly
achieved street notoriety as a drug of abuse (angel
dust). There are also similarities with the effects of
ketamine, a low-affinity NMDA-channel blocker
widely used for dissociative anesthesia. Psychoactive
effects of more recent NMDA receptors antagonists in
conscious subjects occur in a dose-dependent way
(Muir and Lees, 1995) and include light-headedness,
dizziness, paresthesia, disinhibition, nystagmus, paranoid ideation, hallucinations and catatonia. NMDA
antagonists also have beneficial sedative, analgesic
and possibly anticonvulsant effects.
The behavioral effects of NMDA antagonists can be
controlled by administration of benzodiazepines but
other studies have raised concern about the possibility
that, in themselves, NMDA antagonists may be
neurotoxic. The area at risk is the limbic system, which
is maximally activated by their administration. This is
disclosed histologically by the appearance of vacuoles
in neurones in the lingulate cortex, reported by Olney,
Labruyere and Price, 1989. The precise importance
and relevance of this Olney effect is still under
review. The vacuoles are transient and more prominent in rodents than in primates but do occur at
dosages that may be similar to those likely to be
required for clinical benefit.
Cardiovascular effects are of particular importance
in subjects with brain damage. Experimentally, the
Figure 21.2
433
effects of NMDA antagonists are influenced by anesthesia, which changes a dose-dependent elevation to a
hypotensive effect. There is some evidence for a
similar pattern with high doses in human subjects.
The complexity of glutamate receptors, in particular
the NMDA receptor, is reflected in the large number of
modulatory sites amenable to pharmacological intervention (Choi, 1990; Figure 21.2). The possibilities
include reduction of presynaptic release of glutamate
by sodium channel blockers, postsynaptic antagonism at the NMDA receptor by either a high- or lowaffinity non-competitive antagonist, competitive
receptor-site antagonists, and glycine- and polyaminesite modulators. The differing properties of the different classes may influence their utility. For example the
more rapid brain penetration of non-competitive
NMDA-receptor antagonists offers a better pharmacokinetic profile than the much less lipid-soluble
competitive agents. Examples of each class of antagonist have been applied to subjects with severe head
injury to assess safety and tolerability, and in some
cases these have been followed by efficacy trials (Table
21.11). At the time of writing, information is available
only from abstracts and presentations at meetings and
definitive reports are awaited.
The safety and tolerability studies have employed
increasing dosages and duration of administration in
order to seek evidence of safety and tolerability in
regimens appropriate to achieve neuroprotective
effects. The first study of an NMDA site competitive
antagonist concerned the high-affinity, non-competitive antagonist GCS 19755 (Selfotel). The agent was
given in increasing doses, 16 mg/kg on 82 successive
days. There were no important adverse experiences
Schematic of excitatory amino acid pathways and potential sites for pharmacological intervention.
434
Table 21.11
Glutamate antagonists potentially clinically applied to head injury
Class
Drugs
Company
Comments
Presynaptic release
inhibition
619C89
Wellcome/Glaxo
Phase 2 study halted because of

company strategy
Na+ channel block
Enadoline
Parke Davis
Phase 2 in progress
Postsynaptic competitive
Riluzole
Rhone-Poulenc
Rorer
Reported to benefit ALS

No experience in head injury
CGS 19755 Selfotel
CIBA-GEIGY
Phase 2 2 completed
Phase 3 halted by safety committee
d-CPP-ene EAA 49
Sandoz
Phase 2 completed
Phase 3 in progress in Europe
Non-competitive highaffinity
Aptiganel H (Cerestat)
Cambridge
Neuroscience
Boehringer
Phase 2 completed
Phase 3 in progress in North America
and Europe
Polyamine-site
SL820715 Eliprodil
Synthelabo
Phase 2/3 study completed
CP101-606
Pfizer
Phase 2 completed
Glycine-site
ACEA 1021
CIBA-GEIGY/Cocensys
Phase 2 in Progress
Lubeluzole
Janssen
Results of Phase 3 studies in stroke

awaited. Phase 2 in head injury planned
Dexanabinol HV-211
Pharmos
Phase 2 in progress in Israel
that might have been related to the drug in dosages up

to 3 mg/kg (Stewart et al., 1997a). At high doses,
reductions in blood pressure were observed and
intracranial pressure also declined in some patients
but this was not significantly related to drug dosage
(Stewart et al., 1997b) and in some cases cerebral
perfusion pressure fell. In a further phase 2 study
dosages of 3 mg/kg or 5 mg/kg were compared to
placebo in a randomized blind manner. The preliminary findings supported the initiation of two randomized prospective studies aimed to determine the
efficacy of 5 days treatment with 5 mg/kg daily for
severe head injuries. These were conducted in North
America and in a cooperative group of European/
Australasian centers. The studies were stopped in
December 1995 because of a concern in the safety
monitoring committee about an excess of adverse
events in patients in these trials and in parallel trials of
patients with stroke. The data that led to this action
are not available at the time of writing. The safety of a
second high-affinity non-competitive antagonist
(CPP-ene, SZEA494) was studied in a double-blind
dose escalation study with administration of up to 7
days after injury. Dosages were titrated to achieve a
final target serum concentration known to be neuroprotective in experimental circumstances. Preliminary observations in this study indicated an acceptable safety profile and a Phase 3 efficacy study has
been initiated in Europe.
Although the non-competitive ion-channel antagonist originally responsible for much of the enthusiasm
for glutamate blockade, Dizocilpine MK801, was not
pursued clinically, a similar non-competitive antagonist, aptiganel HCL, CNS1102 Cerestat (McBurney et
al., 1992), which is effective in focal ischemia (Minematso et al., 1993), has been studied in dose escalation
protocols in head injury. In the first open study,
administration of increasing doses showed reductions
of intracranial pressure, increases in blood pressure
and a net increase in cerebral perfusion pressure
(Wagstaff et al., 1995). In a subsequent study the
duration of administration was increased to 3 days,
with an acceptable safety and tolerability at a plasma
concentration well above the level known to be
neuroprotective in animals. A study in a combined
group of North American and European centers, aimed
at determining efficacy, is currently in progress.
Eliprodil (SL820715) is a synthetic antagonist of the
polyamine site on the NMDA receptor and has
preclinical neuroprotective efficacy (Gotti et al., 1988).
Eliprodil in normal volunteers is associated with dosedependent prolongations of the QT interval but
appears to have less psychological effects, at least in
the dosages employed, than either competitive or noncompetitive antagonists. Trials of eliprodil have been
conducted in selected groups of head-injured patients
in Europe and the results are awaited at the time of
writing.
CONCLUSION
The binding of glycine to the glutamate receptors is

required for NMDA-channel activation and endogenous concentrations of glycine are sufficiently high for
the site to be fully saturated under normal conditions.
Partial agonists such as ACEA 10201 reduce damage
in ischemic models and are under study for safety and
tolerability in head-injured patients. Two agents initially considered to reduce presynaptic release of
glutamate commenced safety and tolerability studies
in severely injured patients. Studies with 619C89 were
halted as a result of policy changes following the
merger of Wellcome and Glaxo. Studies with enadoline continue and this agent is now considered to have
effects in blocking abnormal sodium currents, representing a distinctly different approach to neuroprotective treatment. Riluzole, which has shown benefit
in patients with amyotrophic lateral sclerosis, is
considered to be a sodium-channel blocker and to
inhibit glutamate release, and is under consideration
for studies in head injury. Lubeluzole inhibits glutamate-induced brain damage but is thought to act at an
intracellular level through a nitric oxide mechanism
and application to head injury will depend upon the
results of efficacy studies in patients with stroke.
21.3.11
FUTURE
Other strategies of neuroprotection under consideration for clinical evaluation, on the basis of experimental studies, include an antagonist of bradykinin to
reduce bloodbrain-barrier damage, agents that limit
white cell adhesion to endothelial receptors and hence
reduce inflammation, agents that inhibit neuronal
nitric oxide production without blocking vascular
dilatation, adenosine analogs that reduce transmitter
release, inhibitors of calcium-activated enzymes,
agonists of the GABA receptor (e.g. chloromethiazole),
as well as neurotrophins and growth factors. On the
horizon are approaches that interfere with gene
expression after injury, for example by encouraging
production of protective proteins or blocking so-called
death genes that are involved in either necrotic or
apoptotic cell death.
21.4
Conclusion
The high incidence of ischemic damage in fatal head

injuries, a third of whom have spoken after injury, and
the frequent occurrence of secondary insults despite
the best medical and surgical treatment point to the
great potential for pharmacological treatment to
improve outcome after head injury. The agents that are
now entering early clinical studies are the fruits of
intensive experimental and clinical investigation over
the last two decades. Irrespective of the results of the
current studies, much remains to be learned and new
435
approaches can already be foreseen. Realization of

these opportunities will involve answering a number
of questions and facing a number of challenges.
Many mechanisms have been identified that may
contribute to the initiation or progression of brain
damage after head injury and ischemia (Siesjo et al.,
1992). Which of these are truly important in clinical
circumstances? Is it appropriate to envisage a
sequence initial events being dominated by ionic and
neurotransmitter events, followed by membrane degradation through free radical mechanisms, and thereafter balances between inflammatory and reparative
processes, such as necrotic and apoptotic cell death,
determined by gene responses and growth and survival factor production? This view would indicate that
a sequential program of neuroprotective treatment is
appropriate. The alternative view is that many processes are occurring concurrently, in which case
cocktails may be most efficacious.
More attention may need to be paid to distinctive
mechanisms in different types of injury, dictating the
appropriate choice of treatment. For example, it is
likely that it is only in patients with traumatic
subarachnoid hemorrhage that calcium-antagonist
treatment may be appropriate at present. Patients with
a focal injury such as contusion or hematoma, with
marked space-occupying effect, increased intracranial
pressure and reduced perfusion, may be most appropriate candidates for glutamate-antagonist treatment,
whereas this is unlikely to benefit those with uncomplicated diffuse primary injury. The unraveling of the
potentially complex series of combinations will
depend upon imaginative interactions between clinical and experimental investigators, aimed at identifying indices of the different mechanisms of damage in
patients, along with therapeutic trials of sufficient size
to determine differential effects in preselected subgroups of injuries.
There are major conceptual and practical issues in
applying neuroprotective drug treatment. Ideally,
treatment is best commenced as soon as possible after
the injury but this is increasingly difficult to achieve
in investigative trials. In part this reflects uncertainties about procedures for consent for research in
emergency circumstances relating to differing regulatory attitudes rather than fundamental ethical
problems. Is there a time window beyond which it is
not appropriate either to initiate or to continue
treatment? There is an indication from the study of
methyl prednisolone in spinal-cord injury that treatment initiated after 8 hours had an adverse effect, but
this is not necessarily relevant to head injury, in which
there is much evidence that secondary insults can
occur for days after injury.
The translation of the exciting preclinical findings
into benefit in patients will require careful, rigorous
436
clinical trials. Whereas in the past, such studies were

often developed within the pharmaceutical industry,
and their protocols to some extent imposed upon
clinical investigators, it is now being recognized that
clinicians must take a prominent role in the planning
and execution of clinical trials.
Two new clinical organizations have developed in
response to the issues raised by the clinical application
of the agents derived from the years of research and
development by industry. These are the American
Brain Injury Consortium and the European Brain
Injury Consortium. Each consists of a large number of
investigators working in centers committed to highquality clinical research aimed at improving outcome
of head injuries. Each has an administrative structure,
coordinating the groups activities through a central
hub in America in Richmond, Virginia, in Europe in
Glasgow, UK. Head injury is a new area for many
pharmaceutical companies. The Consortia provide a
way of harnessing the accumulated wisdom of experienced clinical investigators, backed by the influence of
the large number of cooperating centers, with the aim
of ensuring that trials are carried out in the most
rigorous, relevant and clinically influential manner.
This is achieved through developing studies in partnership with industry, discussing and, if need be,
negotiating over the specifics of the protocol, patient
population, data to be collected, method of analysis
and eventual reporting.
This wealth of clinical experience and potential
patient population in the American and European
consortia, combined with the major resources being
allocated to neuroprotection of head injury by industry, should guide the clinical application of neuroprotection. It will be crucial that this leads to results
that are relevant and comprehensible to individual
clinicians, and most importantly that it maximizes the
likelihood of translating the effects seen preclinically
from neuroprotection treatment to benefits for patients
with severe head injury.
21.5
References
Akioka, T., Ota, K., Matsumoto, A. et al. (1976) The effect of THAM on acute
intracranial hypertension. An experimental and clinical study, in Intracranial Pressure III, (eds J. W. F. Beks, D. A. Bosch and M. Brock), SpringerVerlag, Berlin, pp. 219223.
Andrews, R. J., Bringas, J. R. and Muto, R. P. (1995) Effects of mannitol on
cerebral blood flow, blood pressure, blood viscosity, hematocrit, sodium
and potassium. Surgical Neurology, 39, 218222.
Awad, I., Little, J. R., Lucas, F. et al. (1983) Modification of focal cerebral
ischaemia by prostacyclin and indomethacin. Journal of Neurosurgery, 58,
714719.
Bailey, I., Bell, A., Gray, J. et al. (1991) A randomised trial of the effect of
nimodipine on outcome after head injury. Acta Neurochirurgica, 110,
97105.
Biestro, A. D., Alberti, R. A., Soca, A. E. et al. (1995) Use of indomethacin in
brain injured patients with cerebral perfusion pressure impairment:
preliminary report. Journal of Neurosurgery, 83, 627630.
Braakman, R., Schouten, J. H., Dischoeck, M. B. and Minderhoud, J. M. (1983)
Megadose steroids in severe head injury: results of a prospective doubleblind clinical trial. Journal of Neurosurgery, 58, 326330.
Bracken, M. C., Shepard, M. J., Collins, W. F. et al. (1990) A randomised,

controlled trial of methylprednisolone or naloxone in the treatment of acute
spinal-cord injury. Results of the Second National Acute Spinal Cord Injury
Study. New England Journal of Medicine, 322, 14051411.
Bricolo, A. P. and Glick, R. P. (1981) Barbiturate effects on acute experimental
Buchan, A., Li, H. and Pulsinelli, W. A. (1991) The N methyl D aspartate
antagonist MK801 fails to protect against neuronal damage caused by
transient severe forebrain ischaemia in adult rats. Journal of Neuroscience, 11,
10491056.
Bullock, R., Graham, D. I., Chen, M.-H. et al. (1990) Focal cerebral ischaemia
in the cat: Pretreatment with a competitive NMDA receptor antagonist,
D-CPP-ene. Journal of Cerebral Blood Flow and Metabolism, 10, 668674.
Bullock, R., Zauner, A., Tsuji, O. et al. (1995) Patterns of excitatory amino acid
release and ionic flux after severe head trauma, in Neurochemical Monitoring
in the Intensive Care Unit, (eds T. Tsubokawa, A. Marmarou, C. Robertson
and G. Teasdale), Springer-Verlag, Tokyo, pp. 6467.
Busto, R., Globus, M. Y. T., Dietrich, W. D. et al. (1989) Effect of mild
hypothermia on ischaemia-induced release of neurotransmitters and free
fatty acids in rat brain. Stroke, 22, 3743.
Chan, P., Longar, S. and Fishman, R. (1987) Protective effects of liposomeentrapped super oxide dismutase on post traumatic brain oedema. Annals
of Neurology, 21, 540547.
Chen, M.-H., Bullock, R., Graham, D. I. et al. (1991) Ischaemic neuronal
damage after acute subdural hematoma in the rat: effects of pretreatment
with a glutamate antagonist. Journal of Neurosurgery, 74, 944950.
Choi, D. W. (1990) Methods for antagonising glutamate neurotoxicity.
Cerebrovascular and Brain Metabolism Reviews, 2, 105147.
Clifton, G. L., Allen, S., Barrodale, P. et al. (1993) A phase II study of moderate
hypothermia in severe brain injury. Journal of Neurotrauma, 10(3), 263271.
Clifton, G. L., Jiang, J. Y., Lyeth, B. G. et al., (1991) Marked cerebral protection
by moderate hypothermia after experimental brain injury. Journal of Cerebral
Blood Flow and Metabolism, 1, 114121.
Cohadon, F. (1994) Brain protection. Advances and Technical Standards in
Cold, G. E., Jensen, K., Astrup, J. and Ordstrom, J. (1990) Indomethacin in
severe head injury, in Correspondence in Neurosurgery, 27(4), 660661.
Compton, J. S., Lee, T., Jones, N. R. et al. (1990) A double-blind placebo
controlled trial of the calcium entry blocking drug, nicardipine, in the
treatment of vasospasm following severe head injury. British Journal of
Cooper, P. R., Moody, S., Chard, W. K. et al. (1979) Dexamethasone and severe
Dearden, N. M., Gibson, J. S., McDowall, D. G. et al. (1986) Effects of high-dose
dexamethasone on outcome from severe head injury. Journal of Neurosurgery,
64, 8188.
Demopoulos, H. B., Flamm, E. S., Seligman, M. L. et al. (1977) Anti-oxidant
effects of barbiturates in model membranes undergoing free radical
damage. Acta Neurologica Scandinavica, 56(Suppl. 64), 152153.
Dorman, H. R., Sondheimer, J. H. and Cadnapaphornchai, P. (1990) Mannitolinduced acute renal failure. Medicine (Baltimore), 69, 153159.
Drake, C. G., Barr, H. W. K., Coles, J. C. and Gergely, N. F. (1964) The use of
extra-corporeal circulation and profound hypothermia in the treatment of
ruptured intracranial aneurysm. Journal of Neurosurgery, 21, 575581.
Dumlich, R. V. W., Tornheim, P., Kindel, R. M. et al. (1990) Effects of an amino
steroid (U74006F) on cerebral metabolites and oedema after severe
experimental head trauma. Advances in Neurology, 52, 365375.
Eisenberg, H. M., Frankowski, R. F., Contant, C. F. et al. (1988) High dose
European Study Group on Nimodipine in Severe Head Injury (1994) A
multicenter trial of the efficacy of nimodipine on outcome after severe head
Faden, A. I. and Salzman, S. (1992) Pharmacological strategies in CNS trauma.
Trends in pharmacology. Science, 13, 2935.
Faupel, G., Reulen, J. H. and Muller, D. (1976) Double blind study on the
effects of steroids on severe closed hea injury, in Dynamics of Brain Edema,
(eds H. M. Pappius and W. Feindel), Springer-Verlag, Berlin, pp. 337434.
Flamm, E. S., Demopoulos, H. B., Seligman, M. L. et al. (1977) Possible
molecular mechanisms of barbiturate-mediated protection in regional
cerebral ischaemia. Acta Neurologica Scandinavica, 56(Suppl. 64), 150151.
Fukumori, R., Tani, E., Maeda, Y. and Sukenaga, A. (1983) Effects of
prostacyclin and indomethacin on experimental delayed cerebral vasospasm. Journal of Neurosurgery, 59, 829834.
Gaab, M. R., Knoblich, O. E., Sophr, A. et al. (1980) Effects of THAM on ICP,
EEG and tissue oedema parameters in experimental and clinical brain
oedema, in Intracranial Pressure IV, (eds K. Shulman, A. Marmarou, J. D.
Miller et al.), Springer-Verlag, Berlin, pp. 664668.
Gaab, M. R., Trost, H. A., Alcantara, A. et al. (1994) Ultra high dexamethasone
in acute brain injury. Results from a prospective randomised double blind
multicentre trial (GUD HIS). Zentralblatt fur
Neurochirurgie, S5, 135143.
Ghajar, J., Hariri, R. J., Narayan, R. K. et al. (1995) Survey of critical care
management of comatose, head injured patients in the United States.
REFERENCES
Gianotta, S. L., Weiss, M. H., Apuzzo, M. L. J. et al. (1984) High dose
glucocorticoids in the management of severe head injury. Neurosurgery, 15,
497501.
Ginsberg, M. D., Sternau, L. L., Globus, M. Y. T. et al. (1992) Therapeutic
modulation of brain temperature: relevance to ischaemic brain injury.
Cerebrovascular and Brain Metabolism Reviews, 4, 189225.
Gordon, E. (1971) Some correlations between the clinical outcome and the
acidbase status of blood and cerebrospinal fluid in patients with traumatic
brain injury. Acta Anaesthesiologica Scandinavica, 15, 209228.
Gotti, B., Duverger, D., Bertin, J. et al. (1988) Ifenprodil and SL 829715 as
cerebral anti-ischaemic agents. I. Evidence for efficacy in models of focal
cerebral ischaemia. Journal of Pharmacology and Experimental Therapeutics,
247, 12111221.
Graham, D. I., Tora, I., Adams, J. H. et al. (1989) Ischaemic damage is still
common in fatal non-missile head injury. Journal of Neurology, Neurosurgery
Gudeman, S. K., Miller, J. D. and Becker, D. P. (1979) Failure of high-dose
steroid therapy to influence intracranial pressure in patients with severe
Hadley, M. N., Zabranski, J. M., Spetzler, R. F. et al. (1989) The efficacy of
intravenous nimodipine in the treatment of focal cerebral ischaemia in a
primate model. Neurosurgery, 25, 6370.
Hall, E. D. (1985) High-dose glucocorticoid treatment improves neurological
recovery in head-injured mice. Journal of Neurosurgery, 62, 882887.
Hall, E. D. (1988) Beneficial effects of acute intravenous ibuprofen on
neurologic recovery of head injured mice. Comparison of cyclo-oxygenase
inhibition with inhibition of thromboxane A2 synthetase of 5-lipoxygenase.
Hall, E. D. (1989) Free radicals and CNS injury. Critical Care Clinics, 5,
793805.
Hall, E. D. (1992) The neuroprotective pharmacology of methylprednisolone.
Hall, E. D. (1993) Cerebral ischaemia, free radicals and anti-oxidant
protection. Biochemical Society Transactions, 21, 334339.
Hall, E. D. and Braughler, J. M. (1989) Central nervous system trauma in
stroke. II. Physiological and pharmacological evidence for involvement of
oxygen radicals and lipid peroxidation. Free Radicals and BioMedicine, 6,
303313.
Hall, E. D., Yonkers, P. A., McCall, J. M. and Braughler, J. M. (1988) Effect of
21 amino steroid U-74006F on experimental head injury in mice. Journal of
Hallenbeck, J. M., Jacobs, M. A. and Faden, A. I. (1983) Combined PGI2,
indomethacin and heparin improves neurological recovery after spinal
trauma in cats. Journal of Neurosurgery, 58, 749754.
Harders, A., Kakarieka, A., Braakman, R. and the German TSAH Study Group
(1996) Traumatic subarachnoid haemorrhage and its treatment with
nimodipine. Journal of Neurosurgery, 85, 8289.
Hellstrom, H. O., Waynhainen, A., Valtysson, J. et al. (1994) Effect of Tirilazad
mesylate given after permanent middle cerebral artery occlusion in the rat.
Acta Neurochirurgica (Vienna), 129, 188192.
Hoff, J. T. (1986) Cerebral protection. Journal of Neurosurgery, 65, 579591.
Jenkins, L. W., Moszynksi, K., Lyeth, B. G. et al. (1989) Increased vulnerability
of the mildly traumatized rat brain to cerebral ischaemia: the use of
controlled secondary ischaemia as a research tool to identify common or
different mechanisms contributing to mechanical and ischaemic brain
injury. Brain Research, 477, 211224.
Jensen, K., Ohrstrom, J., Cold, G. E. and Astrup, J. (1991) The effects of
indomethacin on intracranial pressure, cerebral blood flow and cerebral
metabolism in patients with severe head injury and intracranial hypertension. Acta Neurochirurgica (Vienna), 108, 116121.
Johnston, I. H. and Harper, A. M. (1973) The effect of mannitol on cerebral
blood flow. Journal of Neurosurgery, 38, 461471.
Kakarieka, A., Braakman, R. and Schakel, E. H. (1994) Clinical significance of
the finding of subarachnoid blood on CT scan after head injury. Acta
Neurochirurgica, 129, 15.
Karlsson, B. R., Loberg, E. M., Grogaard, B. and Stern, P. A. (1994) The
antioxidant tirilazad does not affect cerebral blood flow or histopathologic
changes following severe ischaemia in rats. Acta Neurologica Scandinavica,
90, 256262.
Katayama, Y., Becker, D. P., Tamura, T. et al. (1990) Increase in extracellular
glutamate and associated massive ionic fluxes following concussive brain
Kontos, H. A. and Povlishock, J. (1986) Oxygen radicals in brain injury. Journal
Kuroda, Y., Strebel, S., McCulloch. J. et al. (1991) An evaluation of the effects
of NMDA antagonists on intracranial pressure in a model of acute subdural
haematoma in the rat, in Intracranial Pressure VIII, (eds C. J. J. Avezaat, J. H.
M. van Eindhoven, A. I. R. Maas and J. T. J. Tans), Springer-Verlag,
Berlin.
Kurze, T., Tranquada, R. E. and Benneact, K. (1966) Spinal fluid lactic acid
levels in acute cerebral injury, in Head Injury, (eds W. F. Caveness and A. E.
Walker) J. B. Lippincott, Philadelphia, PA, pp. 254259.
Lazorthes, G. and Campan, L. (1958) Hypothermia in the treatment of
craniocerebral traumatism. Journal of Neurosurgery, 15, 162167.
437
Leslie, J. B. and Watkins, W. D. Eicosanoids in the central nervous system.

Lewelt, W., Jenkins, L. W. and Miller, J. D. (1980) Auto-regulation of cerebral
blood flow after experimental fluid percussion injury of the brain. Journal of
Liu, T. H., Beckman, J. S. and Freeman, B. A. (1989) Polyethylene glycolconjugated superoxide dismutase and catalase reduce ischaemic brain
injury. American Journal of Physiology, 56, H589H593.
McBurney, R. N., Daly, D., Fischer, J. B. et al. (1992) New CNS-specific calcium
antagonists. Journal of Neurotrauma, 9, 531543.
McCulloch, J. (1995) Excitatory amino-acids antagonists and their potential
for the treatment of ischaemia brain damage in man. British Journal of
Clinical Pharmacology, 34, 106114.
McCulloch, J., Bullock, R. and Teasdale, G. M. (1991) Excitatory amino acids
antagonists: opportunities for the treatment of ischaemic brain damage in
man, in Excitatory Amino Acids, (ed. B. S. Meldrum), Blackwell, Oxford, pp.
287325.
McIntosh, T. K. (1993) Novel pharmacologic therapies in the treatment of
experimental traumatic brain injury: a review. Journal of Neurotrauma, 10,
215261.
McIntosh, T. K., Faden, A. L., Bendall, M. R. and Vink, R. (1987) Traumatic
brain injury in the rat: alterations in brain lactate and pH as characterised
by IH and 31P NMR. Journal of Neurochemistry, 49, 15301540.
McIntosh, T. K., Thomas, M., Smith, D. et al. (1992) The novel 21 amino
steroids U74006F attenuates oedema and improves survival after brain
injury in the rat. Journal of Neurotrauma, 9, 3346.
McKinney, J. S., Willoughby, K. A., Liang, S. and Ellis, E. F. (1996) Stretch
induced injury of cultured neuronal, glial and endothelial cells. Effect of
polyethylene glycol conjugated superoxide dismutase. Stroke, 22, 934940.
Malta, B. and Menon, D. (1996) Severe head injury in the United Kingdom and
Ireland: a survey of practice and implications for management. Critical Care
Medicine, 24, 17431748.
therapeutic hypothermia for patients with severe head injuries: a preliminary report. Journal of Neurosurgery, 79, 354362.
Marshall, L. F. and Marshall, S. B. (1986) Pitfalls and advances from the
international tirilazad trial in moderate and severe head injury. Journal of
aggressive treatment in severe head injuries. Journal of Neurosurgery, 50,
2630.
Mathew, P., Bullock, R., Teasdale, G. et al. (1996) Changes in local
microvascular permeability and the effect of intervention with 21-amino
steroid (Tirilazad) in a new experimental model of focal cortical injury in
the rat. Journal of Neurotrauma, 13, 465472.
Mendelow, A. P., Teasdale, G. M., Russell, T. et al. (1985) Effect of mannitol on
cerebral blood flow in human head injury. Journal of Neurosurgery, 63, 4348.
Michenfelder, J. D., Milde, J. H. and Sundt, T. M. Jr (1976) Cerebral protection
by barbiturate anaesthesia. Use after cerebral artery occlusion in Java
monkeys. Archives of Neurology, 33, 345350.
Miller, J. D. and Becker, D. (1982) Secondary insults to the injured brain.
Journal of the Royal College of Surgeons of Edinburgh, 27, 292.
Miller, J. D., Bullock, R., Graham, D. I. et al. (1990) Ischaemic brain damage in
a model of acute subdural haematoma. Neurosurgery, 27, 433439.
Minematsu, K., Fisher, M., Li, L. et al. (1993) Effects of a novel NMDA
antagonist on experimental stroke rapidly and quantitatively assessed by
diffusion weighted MRI. Neurology, 43, 397403.
Mohammed, A. A., Goldh, O. l., Graham, D. I. et al. (1985) Effect of pretreatment with the calcium antagonist nimodipine on local cerebral blood
flow and histopathology after middle cerebral activity occlusion. Annals of
Mohr, J. D., Orgogozo, J. M., Harrison, M. et al. (1994) Meta-analysis of
nimodipine trials in acute ischaemic stroke. Cerebrovascular Diseases, 4,
197203.
Moskopp, D., Ries, F., Wassmann, H. and Nadstawek, J. (1991) Barbiturates in
severe head injuries? Neurosurgery Reviews, 14, 195202.
Muir, K. W. and Lees, K. R. (1995) Clinical experience with excitatory amino
acid antagonist drugs. Stroke, 26, 503513.
Muizelaar, J. P., Marmarou, A., Young, H. F. et al. (1993) Improving the
outcome of severe head injury with the oxygen radical scavenger
polyethylene glycol-conjugated superoxide dismutase: a Phase II trial.
Murray, G. D., Teasdale, G. M. and Schmitz, H. (1996) Nimodipine in
traumatic subarachnoid haemorrhage. A re-analysis of the HIT I and HIT II
trials. Acta Neurochirurgica, p. 180.
Myseros, J. S. and Bullock, M. R. (1995) The rationale for glutamate
antagonists in the treatment of traumatic brain injury. Annals of the New York
Academy of Science, 765, 262272.
438
Nordstrom, C.-H. and Siesjo,

B. K. (1978) Effects of phenobarbital in cerebral
ischaemia. Part I: Cerebral energy metabolism during pronounced incomplete ischaemia. Stroke, 9, 327335.
Nordstom, C.-H., Messeter, K., Sundbarg, G. et al. (1988) Cerebral blood flow,
severe traumatic brain lesion. Journal of Neurosurgery, 68, 424431.
Oda, S., Shimoda, M., Yamada, S. et al. (1992) Problems in general
management during barbiturate therapy. No Shinkei Geka, 20, 12411246.
Olney, J. W. (1978) Neurotoxicity of excitatory amino acids, in Kainic Acid as a
Tool in Neurobiology, Raven Press, New York, pp. 95121.
Olney, J. W. (1989) Excito-toxicity and N-methyl-D-aspartate receptors. Drug
Development Research, 17, 299319.
Olney, J. W., Labruyere, J. and Price, M. T. (1989) Pathological changes
induced in cerebrocortical neurons by phencyclidones and related drugs.
Science, 244, 13601362.
Oluoch Oluyna, D., Condon, B., Teasdale, G. et al. (1996) IH magnetic
resonance spectroscopy of acute head injury. Journal of Neurology, Neurosurgery and Psychiatry.
Pappius, H. M. and Wolfe, L. S. (1983) Effects of indomethacin and ibuprofen
on cerebral metabolism and blood flow in traumatised brain. Journal of
Park, C. K. and Hall, E. D. (1994) Dose response analysis of the effect of 21
amino steroid tirilazad mesylate (U74006F) upon neurological outcome and
ischaemic brain damage in permanent focal cerebral ischaemia. Brain
Research, 645, 157163.
Perkins, W. J., Milde, L. N., Milde, J. H. et al. (1991) Pre-treatment with
U744006F improves neurological outcome following complete ischaemia in
dogs. Stroke, 22, 902909.
Pickard, J. D., Murray, G. D., Illingworth, R. et al. (1989) The effect of oral
nimodipine on cerebral infarction and outcome after subarachnoid haemorrhage. British aneurysm nimodipine trial. British Medical Journal, 298,
636642.
Pomeranz, S., Safar, P., Radovsky, A. et al. (1993) The effect of resuscitative
moderate hypothermia following epidural brain compression on cerebral
damage in a canine outcome model. Neurosurgery, 79, 241251.
Prasad, M., Ramaiah, C., McIntosh, T. et al. (1994) Regional levels of lactate
and norepinephrine after experimental brain injury. Journal of Neurochemistry, 63, 10861094.
Pulsinelli, W., Sarokin, A. and Buchan, A. (1993) Antagonism of the NMDA
and non-NMDA receptors in global versus focal brain ischaemia. Progress in
Brain Research, 96, 125135.
Rea, G. L. and Rockswold, G. L. (1983) Barbiturate therapy in uncontrolled
intracranial hypertension. Neurosurgery, 12, 401404.
Roberts, I., Anderson, P. and Rowan, K. (1996) Intensive care of severely head
injured patients: guidelines should be based on systematic review of the
evidence. British Medical Journal, 313, 297.
Robinson, M. and Teasdale, G. (1990) Calcium antagonists in the management
of subarachnoid haemorrhage. Cerebrovascular and Brain Metabolism Reviews,
2, 205224.
Rosner, M. J. and Becker, D. P. (1984) Experimental brain injury: successful
therapy with the weak base, tromethamine with an overview of CNS
acidosis. Journal of Neurosurgery, 60, 961971.
Rothman, S. M. and Olney, J. W. (1986) Glutamate and the pathophysiology of
hypoxic-ischaemic brain damage. Annals of Neurology, 19, 105111.
Saul, T. G., Ducker, T. B., Salcman, M. and Cadozo, E. (1981) Steroids in severe
head injury: prospective randomised clinical trial. Journal of Neurosurgery,
54, 596600.
Schalen, W., Messeter, K. and Nordstrom, C. H. (1992) Complications and side
effects during thiopentone therapy in patients with severe head injuries.
Acta Anaesthesiologica Scandinavica, 36, 369377.
Scheller, M. S. (1992) Routine barbiturate protection during cardio-pulmonary
bypass cannot be recommended. Journal of Neurosurgery, 4, 6063.
Schettini, A., Lippman, C. H. and Walsh, E. K. (1989) Attenuation of
decompressive hypoperfusion and cerebral oedema of superoxide dismutase. Journal of Neurosurgery, 71, 578.
Schwartz, M. L., Tator, C. H., Rowed, D. W. et al. (1984) The University of
Toronto Head Injury Treatment Study: a prospective randomised comparison of pentobarbital and mannitol. Canadian Journal of Neurological Sciences,
11, 434440.
on uncontrollable intracranial hypertension after severe head injury. Journal
Siesjo,
B. K. (1992a) Pathophysiology and treatment of focal cerebral
ischaemia. Part I: Pathophysiology. Journal of Neurosurgery, 77, 169184.
Siesjo,
B. K. (1992b) Pathophysiology and treatment of focal cerebral
ischaemia. Part II: Mechanisms of damage and treatment. Journal of
Siesjo,
B. K., Agardh, C. D. and Bergtsson, F. (1989) Free radicals and brain
damage. Cerebrovascular and Brain Metabolism Reviews, 1, 165211.
Siesjo,
B. K., Katsura, K., Pahlmark, K. and Smith, M.-L. (1992) The multiple
causes of ischaemic brain damage: a speculative synthesis, in Pharmacology
of Cerebral Ischaemia 1992, (eds J. Krieglstein and Oberpichler-Schwenk, H.),
Philipps-Universitat, Marburg.
Sinar, E. J., Mendelow, A. D., Graham, D. I. et al. (1988) Intracerebral
haemorrhage: the effect of pre-treatment with nimodipine. Journal of
Smith, H. P., Kelly, D. L., McWhorter, J. M. et al. (1986) Comparison of
mannitol regimens in patients with severe head injury undergoing
intracranial monitoring. Journal of Neurosurgery, 65, 820824.
Spetzler, R. F. and Hadley, M. N. (1989) Protection against cerebral ischaemia.
The role of barbiturates. Cerebrovascular and Brain Metabolism Reviews, 1,
212219.
Steen, P. A., Gisvold, S. E., Milde, J. U. et al. (1985) Nimodipine improves
outcome when given after complete ischaemia in primates. Anaesthesiology,
62, 406414.
Steen, P. A., Milde, J. H. and Michenfelder, J. D. (1989) No barbiturate
protection in acting model of complete cerebral ischaemia. Annals of
Stewart, L., Bullock, R., Teasdale, G. M. and Wagstaff, A. (1996a) First
observations of safety and tolerability of a competitive antagonist to the
glutamate NMDA receptor (CGS 19755) in severely head injured patients.
(Submitted).
Stewart, L., Teasdale, G. M., Wagstaff, A. and Murray, G. D. (1996b) The
haemodynamic effects of increasing dosages of a competitive antagonist to
the glutamate NMDA receptor (CGS 19755) in severely head injured
patients. (Submitted).
Sundt, T. M. Jr, Piepgras, D. G. and Ebersold, M. J. (1982) Monitoring and
protecting cerebral function in neurovascular surgery: rationale and
techniques. Clinical Neurosurgery, 29, 390416.
Tanaka, A. and Tomonaga, M. (1987) Effect of mannitol on cerebral blood flow
and microcirculation during experimental middle cerebral artery occlusion.
Todd, N. V. and Teasdale, G. M. (1989) Steroids in human head injury: clinical
studies, in Steroids in Diseases of the Central Nervous System, (ed. R.
Capildeo), John Wiley & Sons, New York, pp. 151161.
Unterberg, A., Anderson, B., Clark, G. et al., (1988) Cerebral energy
metabolism following fluid percussion brain injury in rats. Journal of
Neurosurgery., 68, 594560.
Wagstaff, A., Teasdale, G. M., Clifton, G. and Stewart, L. (1995) The cerebral
haemodynamic and metabolic effects of the non-competitive NMDA
antagonist CNS 1102 in humans with severe head injury. Annals of the New
York Academy of Sciences, 765, 332333.
Wahlgren, N. G., MacMahon, D. G., DeKeyser, J. et al. (1994) Intravenous
nimodipine West European Stroke Trial (INWEST) of nimodipine in the
treatment of acute ischaemic stroke. Cerebrovascular Diseases, 4,
204210.
Wei, E. P., Kontos, H., Dietrich, W. D. et al. (1981) Inhibition by free radical
scavengers and by cyclo-oxygenase inhibitors of pial arteriolar abnormalities from concussive brain injury in cats. Circulation Research, 48,
95103.
Wolf, A. C., Levi, I., Marmarou, A. et al. (1993) Effect of THAM upon outcome
in severe head injuries: a randomised prospective clinical trial. Journal of
Xue, D., Silvka, A. and Buchan, A. M. (1992) Tirilazad reduces cortical
infarction after transient but not permanent focal cerebral ischaemia in
rates. Stroke, 23, 894899.
Yano, M., Ikeda, S., Kobayashi, S. et al. (1986) The outbome with barbiturate
therapy in severe head injuries, in Intracranial Pressure VI, (eds J. D. Miller,
G. M. Teasdale and J. O. Rowan), Springer-Verlag, Berlin, pp. 769773.
Yoshida, K., Corwin, F. and Marmarou, A. (1990) Effect of THAM on brain
oedema in experimental brain injury. Acta Neurochirurgica (Supplement), 51,
317319.
Young, B., Runge, J. W., Waxman, K. S. et al. (1996) Effects of pergorgolein on
neurologic outcome of patients with severe head injury. A multicenter,
randomised controlled trial. Journal of the American Medical Association, 276,
538543.
Zauner, A. and Bullock, R. (1995) The role of excitatory amino acids in severe
brain trauma opportunities for therapy: a review. Journal of Neurotrauma, 12,
547554.
Zornow, M. H., Oh, Y. S. and Schelle, M. S. (1990) A comparison of the cerebral
and haemodynamic effects of mannitol and hypertonic saline in an animal
model of brain injury. Acta Neurochirurgica (Supplement) (Vienna), 51,
324325.
22
OUTCOME AFTER SEVERE

HEAD INJURY
Bryan Jennett
22.1
Much more attention tends to be paid to describing

and classifying the initial diagnosis and severity of
injury than to defining the outcome after various types
of trauma. Yet when it is the brain that is injured the
need for accurate assessment of outcome is all the
more pressing, because many survivors are left with a
combination of mental and neurological deficits that
markedly affect the quality of life. Judging the medical
efficacy and the cost-effectiveness of interventions
both in the acute and rehabilitation phases depends
on measuring and valuing the ultimate outcome.
Apart from this the patient and his/her family are
clearly concerned to be kept informed about the
prospects of recovery as time passes and to be advised
about how to plan for the likely, and then to deal with
the actual, outcome.
The combined effect of mental and physical disabilities results in a global disability that is often
greater than the sum of these parts. This is because the
mental impairments limit the capacity to cope with the
physical disabilities, while for many patients the
mental changes comprise the major and often the only
persisting disability. For these reasons it is important
not only to list the various components of medical and
physical disability but also to make a global assessment of the patients state in terms of overall social
consequences of his/her brain damage.
Various terms previously used to describe outcome
tended to reflect the viewpoint of different observers,
some more optimistic than others. Optimistic assessments resulted from overemphasis on physical recovery and minimizing the mental impairments, and were
apt to be taken by those who had dealt with the patient
in the acute stage. For them the contrast with the
previously comatose state is striking, while they
naturally wish to justify their early therapeutic efforts
by claiming a reasonable recovery. Terms such as
practical, useful, reasonable, acceptable and worthwhile recovery usually prove to be euphemisms for
severe disability. On the other hand, assessments that

list every detectable neurological deficit, many of which
do not constitute any disability or handicap, can appear
too pessimistic. Nor can social measures such as return
to home or to work be relied on to indicate the degree of
recovery. Exceptional family efforts may enable some
very disabled patients to return home. Return to work is
an unsatisfactory guide for the many patients who were
not previously in work, while failure to return to work
may indicate a decision to retire early or, in times of high
unemployment, the patient being laid off for economic
rather than health reasons. Occasionally, return to work
is to a much lower level of job, provided by a
sympathetic employer. For these various reasons it is
much better to rely on some standard scale for assessing
outcome. Many of these exist for patients with stroke,
but they tend to focus in great detail on aspects of
physical capacity and the ability to undertake activities
of daily living. Moreover, most apply to elderly patients
whose quality of life has different dimensions and
expectations from that of the predominantly young
patients who suffer head injury. It was for these reasons
that we developed the Glasgow Outcome Scale.
22.2
Glasgow Outcome Scale
This scale is based on the overall social capability (or

dependence) of the patient, which takes account of the
combined effect of specific mental and neurological
deficits, but without listing these as part of the
definition (Jennett and Bond, 1975). It was devised for
brain damage in general because it was required for
studies both of head injury and of non-traumatic
coma. Its successful use in collaborative international
investigations has established that it can be reliably
and readily used by different observers. When 150
Glasgow survivors after severe head injury were
classified independently by a neurologist and by a
neurosurgeon there was over 90% agreement, both for
assessments at 6 months and at 12 months after injury
(Jennett et al., 1981).
440
OUTCOME AFTER SEVERE HEAD INJURY
The Glasgow Outcome Scale has five categories:

1.
2.
3.
4.
5.
Good recovery
Moderate disability
Severe disability
Vegetative state
Dead.
Four categories of survival are recognized.
disinhibited, irresponsible behavior. The least affected

patients in the category of severe disability are those
who are communicative and sensible (though often
with impaired cognitive and memory tests) but who
are dependent for certain physical activities on others
perhaps dressing, feeding, or cooking their meals.
Because such a person could not be left alone, even for
a weekend, he/she is not independent and is classified
as severely disabled.
1. Good recovery
The patient is able to participate in normal social life
and could return to work (although he or she may not
have done so). Quite a number of those with good
recovery do not, for a variety of reasons, return to
work, while some moderately disabled patients do.
Good recovery need not imply absence of sequelae, as
many patients have minor cranial nerve deficits,
residual mild limb weakness or some impairment on
cognitive testing or personality change.
2. Moderate disability (independent but disabled)
These patients look after themselves, can travel by
public transport, and some are capable of work. This
may be of a sheltered kind but certain marked
disabilities are compatible with a return to some
occupations. Moderately disabled patients may have
memory deficits or personality changes, varying
degrees of hemiparesis, dysphasia or ataxia, posttraumatic epilepsy, or major cranial nerve deficits. The
degree of independence required to reach this category is much higher than that commonly described by
geriatricians as independent for activities of daily
living, which may indicate no more than the ability to
attend to personal needs in their own room but
without independent mobility or the capacity to
organize their living without assistance; such patients
would be judged severely disabled on the Glasgow
scale.
3. Severe disability (conscious but dependent)
Patients in this category are dependent on some other
person for some activities during every 24 hours. The
worst affected are physically severely disabled, often
with spastic paralysis of three or four limbs. Marked
dysphasia, which limits communication, is a major
handicap in some, and in others dysarthria is a
problem. Marked physical deficits of this kind are
always associated with markedly restricted mental
activity. However, some patients who have little or no
persisting neurological disability are so seriously
affected mentally that they require permanent supervision by family or in residential care. Their mental
problems vary from severe organic dementia to
4. Vegetative state
The characteristic feature of this condition of nonsentient survival, defined by Jennett and Plum in 1972,
is that there is no evidence of psychologically meaningful activity, as judged behaviorally. The criteria for
the definition of this state should be strict and patients
who obey even simple commands or who utter even
occasional words should be assigned to the category
of very severe disability. Relatives or inexperienced
nurses sometimes interpret reflex grasping, groping or
withdrawal as evidence that commands are obeyed or
that purposeful movements are carried out. Likewise
groans, cries and chewing and pouting are sometimes
hopefully interpreted as no or Mum and claimed as
evidence of returning speech (see below).
22.2.1 NUMBER OF CATEGORIES, RELIABILITY AND
VALIDITY
According to the purpose for which an outcome

classification is required, more or fewer categories
than those of the original Glasgow Outcome Scale
(GOS) may be needed. When analyzing the rate and
degree of recovery it can be useful to recognize
improvement within one category by subdividing
each of the three categories of conscious survivors into
a better and a worse grade (Jennett et al., 1981). On the
other hand, when one is seeking statistical relationships between early features and outcome the fewer
categories the better. For example, dead and vegetative patients may be combined (bad outcomes),
or vegetative and severely disabled (dependent),
and moderate disability and good recovery
(independent).
In the study of 150 Glasgow survivors after severe
head injury classified independently by a neurologist
and by a neurosurgeon, referred to earlier, there was
over 90% agreement on the basis of written reports,
both for assessments at 6 months and at 12 months
after injury (Jennett et al., 1981). A more detailed
analysis of 57 cases confirmed good inter-rater reliability, this time between a psychiatrist and a neurosurgeon (Brooks, Hosie and Bond, 1986). This applied to
both the three-point and six-point scales for conscious
survivors, most disagreements being between good
GLASGOW OUTCOME SCALE
recovery and moderate disability. Cognitive tests

correlated much more closely with the three-point
scale in this study at 3 months than later after injury,
but did so at both 3 and 6 months in a later study
(Clifton et al., 1993).
22.2.2 ALTERNATIVES TO THE GLASGOW OUTCOME
SCALE
The main value of the GOS is in comparing the

outcome of groups of head-injured patients who have
been treated by different therapeutic regimes or in
different places. An obvious criticism is that its
simplicity makes it relatively insensitive to improvements occurring late in recovery that are not large
enough to make a change in category and yet are
significant. It has been claimed that the Disability
Rating Scale (Table 22.1) is more useful for this
purpose (Rappaport et al., 1982; Hall, Cope and
Rappaport, 1985).
A consensus conference on outcome measures for
clinical trials in head injury, held in Houston in 1992,
concluded that for severe injuries either the Glasgow
Outcome Scale or the Disability Rating Scale at 6
months was appropriate, but that for moderate
injuries (Glasgow Coma Score 912) the Disability
Rating Scale at 3 months was better (Clifton et al.,
1992).
Outcomes research in medicine in general is making
increasing use of state of health measures which
indicate the patients own perception of himself.
Table 22.1
Disability Rating Scale (Rappaport et al., 1982)
Scoring in eight categories with high indicating maximum

disability.
Eye opening
Verbal
Motor response
Toileting
Feeding
Grooming
Dependence
Employability
03
04
04
03
03
03
05
03
Maximum disability
30
Cognitive ability for
This gives ten degrees of disability.

0
1
23
46
711
1216
1721
2224
2529
30
Nil
Mild disability
Partial disability
Moderate disability
Moderate/severe disability
Severe disability
Extremely severe disability
Vegetative state
Extreme vegetative state
Death
441
Table 22.2 Nottingham Health Profile (Hunt, McKenna

and McEwan, 1981)
This consists of a questionnaire for patient completion,
consisting of a series of questions requiring yes/no
answers.
Part I has questions about six problems with health
physical mobility, pain, sleep, energy, social isolation and
emotional reactions.
Part II deals with seven areas of daily life affected by
health paid employment, looking after the house, social
life, love life, sex life, hobbies and holidays.
There are various weights for the yes answers to
different questions and these sum to 100 for each category
of disability for comparison with figures for abnormal
population or with the patient at some previous stage of
their recovery.
Widely used measures are the Nottingham Health

Profile (Table 22.2) in the UK (McEwan, 1983) and the
Sickness Impact Profile in the USA (Bergner et al.,
1976), but both countries are now recommending the
Short Form 36. This scores physical and social functioning, physical and emotional limitations, mental
health, energy/vitality, pain and general health perceptions (Ware, 1993; Jenkinson, Coulter and Wright,
1993).
22.2.3 COMPONENT DISABILITIES IN DIFFERENT
OUTCOME CATEGORIES
The mental component of the disability was judged to

be more important in more than half the survivors of
severe head injury in Glasgow; in only a quarter was
the physical disability more prominent than the
mental. This predominance of the mental disability
was found in each of the three categories of conscious
survivors. In a more detailed study of a subset of these
patients Bond (1976) found that mental disability
correlated more significantly than did physical deficits
with social handicap. Almost half the patients who
had made a good recovery had mild changes in
personality. In those with moderate disability, personality change and physical disability were recorded in
about equal proportions. Common neurophysical
sequelae in good recoveries were mild hemiparesis,
cranial nerve palsies or infrequent epilepsy. In the
moderately disabled, hemiparesis was again prominent and was sometimes severe, while dysphasia was
much more frequent than in the good recoveries.
Cranial nerve palsies were common in the moderately
disabled and ataxia was not uncommon. In both these
upper grades of recovery personality changes were
common; as expected these were more frequent and
more severe in the moderately disabled than in the
good recoveries.
442
All the severely disabled patients had personality

changes. Because of the small number of severely
disabled patients in whom formal cognitive testing
was feasible no marked difference could be shown
between measurable aspects of cognitive deficit in the
moderate and the severely disabled. This emphasizes
the value of making an overall clinical judgment
rather than basing a calculation of overall disability on
an aggregation of individual disabilities. Moreover
this can be reliably done at a brief outpatient
interview by medical, nursing or remedial staff from
various disciplines once they have been adequately
instructed in the use of a structured questionnaire.
However, according to one study, the GOS underestimates the continuing limitation of psychosocial
reintegration of less severely injured patients.
Although 52% of 87 patients made a good recovery,
only 24% made a good level of reintegration. There
was a substantial limitation of reintegration in more
than half of the good recoveries (Tate et al., 1989).
22.3
When to assess outcome
Reports on outcome sometimes fail to specify how

long after injury the assessments were made, or give
only the minimum or mean periods of follow-up. The
process of recovery in survivors can continue for
months, and claims are often made that substantial
recovery has occurred over a period of years. What
interval is chosen for assessment depends to some
extent on how detailed a categorization is used.
Deaths and survivors can reasonably be identified
within a few weeks, apart from the few late deaths
most of which are in vegetative or severely disabled
survivors. Attempts to record the ultimate degree of
disability at this early stage usually prove to be
predictions rather than assessments. Prognosis
becomes more confident as time passes but it should
always be made clear whether the expected or the
actual state is what is being described, and to what
interval after injury each applies.
Anecdotes about unexpected late improvement
many months or years after severe head injury lead
some clinicians to encourage patients, relatives and
therapists to look forward to continued recovery over
a long period of time. This is often unrealistic, and
what is required for practical purposes is to determine
an interval after injury by which most patients will
have achieved most of their recovery.
When more than 500 survivors in the Glasgowbased three countries study were assessed at 3, 6 and
12 months after injury (Jennett and Teasdale, 1981)
there were more good recoveries and fewer severely
disabled patients by the end of the year. Not only had
some improved but others with severe disability had
died during the period of follow-up. Of those who by
12 months had made a good or moderate recovery,

two-thirds had already reached this level of recovery
on the scale within 3 months of injury, and 90% had
done so by 6 months. Only 5% of 82 patients followed
for more than 18 months showed sufficient improvement after 12 months to change their outcome
category. In the 150 Glasgow patients it was confirmed that 10% of patients who were severe or
moderate at 6 months had improved to become
moderate or good respectively by 1 year.
This is not to deny that some degree of recovery
may continue after 6 months in many patients;
however, the degree of improvement seldom justifies
a change of category on the three-point Glasgow scale
for conscious survivors. With the six-point outcome
scale 20% of the 150 Glasgow survivors changed by
one category between 6 and 12 months; half of these
moved into the next higher category on the threepoint scale. That most recovery occurs in most
patients within the first 6 months is supported by
other studies, for example the 20-year follow-up by
Roberts (1979). Six months after injury is a useful time
to assess outcome, because it is practical to maintain
contact with the majority of patients for this long.
Six-month outcome was recently recommended as the
appropriate prime outcome measure for clinical trials
of therapies for the acute phase of injury (Clifton et
al., 1992).
Long-term follow-up reports of severe injuries are
now available from several centers. For patients aged
1564 years with at least 2 days PTA the Glasgow
Group made assessments at 3, 6 and 12 months and at
2, 5 and 7 years; the psychosocial and neurophysical
status of patients were assessed as well as the
psychiatric and social consequences for the main
caring relatives (Livingston, Brooks and Bond,
1985a,b; Brooks, Campsie and Symington, et al., 1986,
1987). Although neurophysical deficits (including dysphasia) tended to improve over the first few years no
consistent reduction in the symptom level was found
at 7 years compared with 2 years. Relatives were more
distressed at a year than at 3 months, and they
commonly complained of increasing burden over
subsequent years. More psychiatric symptoms and
negative social consequences became manifest as time
passed with the patient improving so little.
Brooks et al. (1987) emphasized that their Glasgow
patients did not have access to a well-structured
scheme of rehabilitation, coming as they did from a
wide area; they postulated that they might therefore
have fared less well than patients reported from
elsewhere.
However, a 7-year follow up of a series culled from
one of the few dedicated neurological rehabilitation
centers in Britain (Oddy et al., 1985) showed similar
results to those of Brooks. These patients were all
ETHICAL ISSUES
under 40 years of age and had a PTA of 7 days or

more. They were assessed 2 years after injury and
again at 7 years, when no significant change was
found in neurophysical status or in performance of
cognitive tests. Personality changes had persisted and
some patients had developed major psychiatric problems. There had been no major changes in employment status; no patient without a job at 2 years having
subsequently obtained work. Limited improvements
in social adjustments had occurred in a few, but these
were patients who were already well recovered by 2
years most of them from the minority who had by
then already returned to some kind of work. Reporting some late improvement in their somewhat less
severely injured series Brooks et al. (1987) emphasized
that they could find no reliable predictors of who
might have delayed improvement. Similar results
have been reported from Denmark, where Thomsen
(1984) studied 40 patients under 45 years of age with
PTA of a month or more. Although there is a small
added risk of accidental death, sometimes related to
physical disability, to epilepsy or to suicide in
disabled conscious survivors of severe head injury,
this makes for very little reduction in the normal
expectation of life (Roberts, 1979). Therefore most
patients left with severe disability will have to endure
this for 3040 years: the average age of such patients
at injury is 27 years.
The conclusion must be that it is unrealistic to expect
significant improvement in physical deficits, cognitive
functions, behavior or personality problems after 2
years. The improvements in social functioning after
that time (and indeed most such improvements after
the first year) probably reflect gradual acceptance of,
and adaptation to, a relatively fixed disability, on the
part of both the patient and the family. These studies
have all emphasized that the patients complaints are
fewer and less marked than the deficits perceived by
their carers, who are more aware of the relatively static
disability and of the contrast with the patients
443
previous state or their expectations for this future.

This may account for carers reporting increasing
burden as the years pass, and for their adverse
reactions to the situation.
The prognosis for patients in a vegetative state is
dealt with later.
(a)
Distribution of outcomes
There have been numerous reports of outcome after

severe head injuries in the last decade or so, with
varying definitions of initial severity and of the time
when outcome was assessed. Three large series indicate a similar distribution of recovered patients among
categories (Table 22.3). The lower mortality of the two
more recent series is partly explained by the exclusion
of patients who deteriorated because of intracranial
hematomas without having been in persisting coma
initially.
22.4
Ethical issues
The main ethical issue arises when active treatment

seems so unlikely to benefit the patient that a decision
to limit treatment has to be considered (Jennett,
1992a). This is a response to the four principles of
medical ethics to maximize benefit but to minimize
harm for the patient, to respect his/her autonomy and
to have regard to justice in the use and distribution of
health care resources. Two circumstances are now
widely agreed to justify the withdrawal of treatment,
namely when a patient is brain-dead and when he/she
has been vegetative for so long that no prospect of
recovery remains. These two situations are dealt with
in detail below. However, there are two less clear-cut
situations when a decision to limit treatment may be
considered. One is when, soon after admission, it is
judged that the patient has sustained an irrecoverable
injury whether only to the head or from multiple
injuries. This decision will usually be reached only
Table 22.3 Outcome after severe head injury (TCDB = Trauma Coma Data Bank)
Study
Ref
Years
Three countries
Jennett et al., 1979 197077
Trauma Coma Data Bank, USA Marshall et al., 1991 198487
Four UK centers
Murray et al., 1993
n
1000
746
198688 A: 1067
B: 1353
Interval*
6 months
Discharge
Last contact
6 months
6 months
Severe
Moderate/
Dead Vegetative disability Moderate Good
Good
(%)
(%)
(%)
(%)
(%)
(%)
49
33
36
45
37
2
14
5
1.6
1.3
10
28
16
20
18
17
19
16
18
18
22
7
27
15
23
Entry criteria: Three countries coma for 6 h; Trauma Coma Data Bank GCS 8 on admission or deteriorated to this in 48 h;
UK centers A, coma 6 h; B, including coma 6 h
*Time of outcome in TCDB: at discharge from Neurosurgery 46% in 30 days (mostly deaths), 79% in 60 days; last contact
two-thirds > 1 year, one-third > 2 years
39
26
43
36
45
444
after attempts at resuscitation, sometimes when a CT

scan of the head has also been done. Even if the latter
shows a large intracranial hematoma it may be judged
irrecoverable as the patient has been deeply comatose
with fixed pupils for some time, particularly if he/she
is elderly. Confidential inquiry of perioperative deaths
in the UK has identified a number of patients whom it
was considered should not have been subjected to
surgery because they were moribund, and this
included some with severe head injuries. In an audit
of deaths in a neurosurgical unit we found that
decisions to limit treatment were quite frequently
made, including a number where surgery was withheld after initial assessment (Barlow and Jennett,
1991). More often a decision to limit treatment is
made when the patient shows no sign of recovery after
a trial of treatment for example ventilation and
perhaps evacuation of an intracranial hematoma.
There are now formal computer programs for estimating the probability of survival and recovery, given a
limited number of variables that include the depth
and duration of coma, pupil reactions and the
patients age. We found that providing such data to
neurosurgeons led to less use of certain active treatments for patients with poor prognosis and greater
use in those with a better outlook (Murray et al., 1993).
There was no change in the proportion of patients for
whom a written decision to limit treatment was made
when predictions were available. The commonest
treatments to be withheld are cardiopulmonary resuscitation and antibiotics for new infections. Surgery
may be withheld either initially, as described above, or
when a secondary complication develops. Sometimes
a patient on a ventilator who is not brain-dead
remains deeply comatose with no sign of recovery. It
may then be decided to withdraw the ventilation,
accepting that spontaneous ventilation will probably
be inadequate. When the time comes for decisions
such as these it may be wise to consider discharge
from the intensive care unit (or the neurosurgical unit)
to a less intensive and less expensive setting.
In head-injured patients there is no prospect of
respecting the patients autonomy because he/she
cannot express his/her preferences. The family (or
friends) then become important and they should be
kept fully informed of the expectations of the doctor
regarding survival or useful recovery. They will then
often volunteer that they believe that this patient
would not want his/her life prolonged in such circumstances. Doctors should make every effort to be sure
that the relatives are genuinely attempting to put
forward the patients viewpoint rather than their
own. Only occasionally, but hopefully more often in
the future, patients may have an advance directive,
which can make such decision-making easier. What is
clear is that, if the patient has made such a declaration,
the doctor should respect it and act accordingly. So

should he/she act in accordance with the family
wishes if he/she is satisfied that these are in the
patients best interests. Under such circumstances
there should be no risk of civil or criminal liability for
the subsequent death of the patient.
There is increasing evidence that people do not
want life at any cost, and in particular are anxious to
avoid the risk of survival with severe brain damage. In
a survey of 500 Americans about their wish to refuse
life-saving or life-sustaining treatments given four
scenarios, more than 80% would refuse such treatments if vegetative while more than half would refuse
them if in coma with a small chance of complete
recovery (Emanuel et al., 1991). When a group of 59
neurosurgeons from several countries were asked at
what level of probability of a poor outcome after
severe head injury (dead, vegetative or severely
disabled) they would withhold ventilation or surgery,
most wanted more than 95% certainty (Barlow and
Teasdale, 1986). However, when then asked to imagine
that they themselves were injured, many wanted
treatment to be limited at a much lower probability of
a poor outcome. They were not prepared to accept the
risk of survival with severe brain damage that they
would recommend for their patients. Whatever the
generality of views on such matters, however, what
matters is to try to determine what a particular
patients previous attitudes were and thus to make a
decision that is both in his/her best interest and likely
to reflect what he/she would have wanted.
To assist in making such decisions it is helpful to
have guidelines, agreed by medical and nursing staff,
about the kinds of situation that would lead to a
treatment-limiting decision. These can be useful in
signaling that such decisions are part of good practice,
as well as reminding decision-makers of the factors to
consider when faced with such a situation. They are
therefore supportive of the doctor who makes such a
decision by indicating that this is in accordance with
agreed practice.
22.5
Brain death
Essential to the concept of brain death is the recognition that death is a process rather than an event.
Organs and tissues cease to function and later necrose
at different stages in the process of death, and when
death is declared is to some extent arbitrary. The
World Medical Association declaration of 1968 in
Sydney proposed that death is when the body as an
integrated whole has irreversibly ceased to function,
rather than when all organs and tissues are dead.
There are three common sequences that lead to
death. Most often, cardiac arrest is the initial event and
soon the cerebral cortex ceases to function; later the
BRAIN DEATH
brain stem also fails and respiration then stops. Less

often, respiratory arrest begins the sequence, leading
to anoxic cortical and then brain-stem failure, while
cardiac arrest may not occur for 1530 minutes.
Sometimes it is the brain stem that fails first, followed
by respiratory arrest with anoxic cardiac arrest occurring later. If artificial ventilation restores oxygenation
after the brain stem is dead but before the heart stops,
then ultimate cardiac arrest may be delayed for many
days. It is, however, a mistake to consider that there
are two kinds of death that evidenced by cardiorespiratory arrest and that by lack of brain-stem
function. This is because cardiorespiratory arrest is
considered to indicate death only when it has lasted
long enough to produce brain-stem death. When
patients are successfully resuscitated from cardiac
arrest or have been subjected to therapeutically
controlled stoppage of the heart during surgery, we do
not claim that they have been dead. It is therefore
brain-stem death that is the central feature of all
sequences of death, while the state of continued
cardiac function after this has occurred is an artifact
of nature resulting from technological intervention.
A distinction is sometimes drawn between brainstem death and whole brain death. However, the
function of the cerebral cortex is dependent on
upward impulses from the reticular formation in the
brain stem, and therefore when the brain stem is dead
the brain as a whole cannot function. This is not to
deny that some cells in the cerebral cortex and basal
ganglia may not continue to survive for a time, but
they are not able to maintain the function of the brain
as a whole. The logic of the situation is therefore that
if the brain stem is dead the brain is dead. It is also
now accepted that if the brain is dead the person is
dead. This last concept is explicitly stated in the UK
Royal Colleges memorandum of 1979, which states
that the time of death is when brain-stem death is
confirmed and not some later time when the heart
stops (Conference of Medical Royal Colleges, 1979). It
is important to explain this to those involved in
procedures associated with organ donation, when
there may be a delay of several hours before the
ventilator is withdrawn and the heart stops. It is easy
to refer carelessly to withdrawing life support or
letting the patient die, when in fact ventilation is
being stopped in a patient who is already dead.
22.5.1 THE DIAGNOSIS OF BRAIN DEATH THE UK
CRITERIA (Table 22.4)
These criteria were published by the UK Conference of

Medical Royal Colleges in 1976, and a further memorandum in 1979 confirmed these and indicated that
death could be declared once the criteria were satisfied. A feature of the diagnostic criteria is the
445
emphasis on satisfying the preconditions before considering the tests to confirm that the brain stem is
dead. There are four preconditions. The patient must
be in deep coma, must be apneic (and therefore on a
ventilator), must have irrecoverable structural brain
damage and reversible causes of brain-stem depression must have been excluded. Common causes of
brain damage leading to brain death are severe head
injury and spontaneous intracranial hemorrhage, but
a few result from brain tumor or intracranial infection.
Some cases follow delayed resuscitation after cardiac
arrest from various causes, including anoxia and drug
overdose. Reversible causes of brain-stem depression
include depressant drugs, neuromuscular blocking
agents used for intubation during resuscitation or as
an adjunct to mechanical ventilation, hypothermia
and gross metabolic abnormalities. These various
factors may not be the sole cause of brain-stem
depression but can aggravate the effect of structural
lesions. Screening for drugs will not normally be
necessary when there is a clear-cut story of sudden
coma from injury or hemorrhage. As for establishing
the irrecoverability of the brain damage, enough time
should elapse to correct temporary causes of brainstem depression such as hypotension, hypoxia, raised
intracranial pressure and barbiturate therapy. Normally the diagnosis would not be considered in less
than 6 hours but when the cause is anoxic damage or
when drugs are suspected the diagnosis should be
delayed for at least 24 hours.
The tests to confirm that there is no residual brainstem function are simple to perform and to interpret.
There should be no pupillary or corneal reflexes, no
movement of the facial muscles to pain or of the throat
muscles to movement of the endotracheal tube. The
caloric vestibulo-ocular reflex should be absent (no
eye movements following irrigation of the external
auditory meatus with at least 20 ml of ice cold water
on each side). Only when these reflexes are found all
to be absent is the final crucial test applied, to confirm
apnea. There should be no respiratory movements
when disconnection of the ventilator allows the PaCO2
to rise. The UK criteria require PaCO2 to reach
50 mmHg (6.65 kPa) but American codes recommend
60 mmHg (8.0 kPa). The rate of rise of PaCO2 in braindead patients can be slow (Benzel et al., 1989), and to
attain this level in 10 minutes requires that the PaCO2
be greater than 40 mmHg (5.3 kPa) before disconnection. This can be achieved by reducing the tidal
volume or by ventilating with 95% oxygen and 5%
CO2 for 5 minutes. To ensure that damaging hypoxia
does not occur during disconnection, preoxygenation
with 100% oxygen for 10 minutes before disconnection
is recommended and the maintenance during disconnection of 6 l/min of oxygen delivered down a
catheter in the trachea. Advice from experts is
446
Table 22.4
UK brain death criteria
Three preconditions
Patient on a ventilator
Coma due to irremedial structural brain damage
Exclusion of reversible factors
depressant or neuromuscular blocking drugs
primary hypothermia
metabolic or endocrine abnormalities
Five tests
No pupillary response to light
No tracheal, gag or cough reflex
No response to facial and peripheral pain
No cold caloric responses
No respiratory effort after achieving a PaCO2 of
50 mmHg for 10 min or more
required for patients with pre-existing chronic respiratory insufficiency who normally depend on a hypoxic
drive for respiration and may be unresponsive to
raised PaCO2.
The UK criteria specify that two doctors should be
involved in testing, one of them a consultant and the
other a senior registrar or consultant, and that the
tests be done on two separate occasions. Notice that
these criteria require no confirmatory laboratory
tests.
Provided the preconditions have been met before
the first test the interval between the two assessments
need be no more than half an hour. The Ad Hoc
Committee of the Harvard Medical School (1968)
recommended demonstrating absence of cerebral
activity on EEG, but this was declared optional by that
institution a year later (Beecher, 1969). This is still
frequently used in the US and in other countries, and
sometimes in the UK. In practice it is less useful than
might be expected, partly because it reflects activity in
the cerebral hemispheres rather than the brain stem
(and some residual activity may persist after unequivocal brain-stem death), and because securing an
isoelectric recording can be technically difficult in the
electronically active environment of an intensive care
unit. Those who use it sometimes say that they do so
to impress the family rather than to make a diagnosis.
Another confirmatory test is to demonstrate absence
of cerebral circulation over a period of time, either
visually by angiography or by showing no entry of
radioactive agents injected systemically. Both require
technical equipment and expertise and neither is
wholly reliable; they are virtually never used in the
UK and rarely in the USA.
Definitive guidance on the diagnosis of brain death
in children has been given by a US Task Force (Task
Force for the Determination of Brain Death in Children,
1987). This counseled that the diagnosis should not be
made in the first 7 days of life and the UK Conference of
Medical Royal Colleges (1988) subsequently endorsed

this for organ donation in the UK. From 7 days till 2
months of age the Task Force recommended two
isoelectric EEG records 48 hours apart, but for 212
months old the interval need only be 24 hours. In
children a year or more old the diagnosis by adult
criteria with up to 12 hours observation was considered adequate, without EEG confirmation.
22.5.2
VALIDITY OF THE CRITERIA
These criteria have now been applied to many

thousands of patients, many of whom were ventilated
until asystole before clinicians gained the confidence
to discontinue ventilation once brain death had been
diagnosed. Not one case is on record as having
recovered after the UK criteria were satisfied, according to Pallis (1990), who listed over 1900 published
cases. Nonetheless, sporadic press reports of patients
allegedly recovering after supposedly having been
brain-dead appeared during the 1970s and these
culminated in a challenge on BBC TV in 1980 about
whether organ donors were in fact always definitely
dead before organs were removed. The critics were
mostly from other countries (particularly the US), and
seemed mainly concerned that the UK criteria did not
require an EEG. In the event the original criteria did
not need to be modified (Robson, 1981), while subsequent guidelines in the US stressed that the use of
EEG was optional (Medical Consultants to the Presidents Commission, 1981).
It is, however, wise to be sensitive to misunderstandings that can arise in this sensitive area of
medicine. Most center on the issue of organ donation
and the suspicion that eagerness to secure organs
might tempt doctors to make a premature diagnosis of
brain death. An allegation may be made that a patient
recovered after they nearly removed his kidneys. A
likely explanation is that the family was told soon
after an acute episode of brain damage that the
outlook was almost hopeless, but that the results of
resuscitation were still awaited. When such a prognosis is given families now sometimes immediately
themselves raise the possibility of organ donation. The
response to this should be to indicate that it is too soon
to be sure that there will be no recovery (and certainly
too soon to diagnose brain death). However, doctors
may be tempted to accept this offer even though
noting that diagnostic tests will be needed later. As
occasionally such patients do recover or at least
survive for a time, it is easy to see how in retrospect it
may seem as though organs were nearly taken. The
same may happen when a reversible cause of brainstem depression such as drug overdose is discovered.
Clearly these are not examples of recovery after the
formal diagnosis of brain death.
THE VEGETATIVE STATE
Misunderstandings can also arise in the intensive

care unit when bystanders observe the activity of
spinal reflexes in patients declared brain-dead. In fact
these become more active the longer ventilation is
continued after brain death, and they may be precipitated by the removal of organs. The best safeguard
against such embarrassing allegations is never to
consider the diagnosis of brain death until the
preconditions have been met, to use widely accepted
diagnostic criteria, to have two doctors involved and
always to allow enough time to be certain that the
situation is not reversible. These conditions have been
established in the UK by the Health Departments Code
of Practice published in 1979, and revised in 1983
(Health Departments of Great Britain and Northern
Ireland, 1983), which reproduces the UK Colleges
criteria and memorandum as well as giving detailed
guidelines for the removal of organs for transplantation. This has been widely distributed and the
diagnostic criteria have recently been reproduced yet
again in the form of a checklist for inclusion in the
patients notes (OBrien, 1990). In practice there is now
little continuing controversy in the UK about brain
death, a concept that the public at large and the families
actually involved seem able to accept. A recent review
of the UK criteria has emphasized certain features, but
broadly endorsed the original recommendations
(Royal College of Physicians, 1995).
22.5.3
ORGAN DONATION
It is important to emphasize that the diagnosis of brain

death and the subsequent withdrawal of ventilation is
part of good medical practice and is required regardless
of any need for organ donation. Indeed in many
instances the patient is medically unsuitable to become
an organ donor. Nevertheless, brain-dead patients do
present the best source of kidneys for donation
(although these can be removed after terminal cardiac
arrest), and are the only source of hearts, lungs and
livers. It is accepted that it is ethically acceptable to
maintain the blood pressure and fluid intake of braindead patients in order to optimize the condition of
organs to be transplanted. There is, however, some
controversy about the elective ventilation of patients in
order that they might become organ donors (i.e.
patients who would not otherwise be ventilated).
Investigations in England (Feest et al., 1990), in Wales
(Salih et al., 1991) and in New South Wales (Hibberd,
Pearson and McCasker, 1992) suggest that such a
practice could lead to considerably more organs being
available. Concerns about the lawfulness of interventions that are not in the best interests of the patient
who is a potential donor has led to this practice being
stopped in the UK, but there is debate as to how this
legal problem might be circumvented (Riad et al., 1995).
447
Meanwhile, surveys in intensive care units in England

and Wales (Gore, Hindes and Rutherford, 1989; Gore,
Taylor and Wallwork, 1991; Gore, Cable and Holland,
1992) and in one neurosurgical unit (Gentleman, Easton
and Jennett, 1990) showed that more than a quarter of
potentially brain-dead patients are never tested for
brain death while 30% of relatives refuse consent for
organ donation when testing has confirmed that their
relative is brain-dead.
The mechanisms to be followed in arranging organ
donation will vary from place to place, but clear
guidelines should be available, as in the Code of
Practice of the Health Departments of Great Britain and
Northern Ireland (1983). Most will require that the
management of the patient up to and including the
tests for brain death should be in the hands of the
clinicians who are not part of the transplantation
team. While some places require additional tests to
confirm brain death when organ donation is under
consideration, this seems illogical, in that it suggests
that less than reliable criteria are being used to reach a
decision to discontinue ventilation when organ donation is not an issue.
22.6
The vegetative state
The management of patients in this state, with no

evidence of a functioning cerebral cortex, has given rise
to considerable debate in medical, ethical and legal
circles in recent years. There is a growing consensus
that once there is no prospect of recovery survival in
this state is of no benefit to the patient, and that
withdrawal of life-sustaining treatment may therefore
be morally and legally justified. However, particularly
after trauma, the vegetative state may be temporary
and before such a decision is made it is essential that
both the diagnosis and prognosis be reliable.
The diagnosis remains a clinical one depending on
doctors and nurses being sure that there is no
sustained, reproducible purposeful activity or
response to external stimuli, and no words uttered or
commands obeyed. Reflex responses to light, sound
and pain need to be discounted, and spontaneous
emotional behavior patterns unrelated to relevant
events. No investigations can reliably confirm the
diagnosis. CT scanning will show progressive atrophy
but similar degrees have been found in demented
patients who retain some conscious behavior. However, the absence of atrophy would indicate the
possibility of recovery. EEG recordings vary from
isoelectric to near normal, but are usually unresponsive to external stimuli. Absence of a cortical
response to somatosensory stimuli is common but
neither its absence nor its presence confirms or
excludes this diagnosis. PET scanning has shown
CMRO2 for glucose in the cortex and basal ganglia to
448
be less than 50% of normal in vegetative patients, a

level associated with deep barbiturate coma.
The commonest pathological finding is severe diffuse axonal injury, but some patients have severe
ischemic brain damage in the cerebral cortex and basal
ganglia, while others have both types of lesion
(McLellan et al., 1986; McLellan, 1989). There is
considerable overlap between the findings in vegetative and severely disabled patients. Vegetative survival is the long-term outcome when the aggregate of
damage in the cortex and subcortical structures is such
that there is no longer the critical amount of surviving
or connected cortex needed for consciousness.
Estimates of the incidence and prevalence of vegetative survivors varies considerably according to how
soon after injury patients are considered to be in a
vegetative state, because many of those vegetative for
months after insult will die or recover in the course of
the next few months (Table 22.5).
Head injuries accounted for some 40% of patients
found to be vegetative 36 months after an acute insult
in surveys in Japan and the Netherlands. However, an
American review of over 700 published series of
patients vegetative 1 month after insult revealed 70%
due to head injuries in both adults and children (MultiSociety Task Force, 1994). In the US Trauma Coma
Data Bank, 14% were vegetative on discharge from the
neurosurgical unit but only 5% at last contact many
months later (Levin et al., 1991). In the three-countries
data bank 10% were vegetative at 1 month, 2% at 6
months and only 1% at 1 year (Braakman, Jennett and
Minderhoud, 1988). A more recent study in four UK
units showed only 6% of severe injuries were vegetative 1 month after injury and 2% at 6 months (Murray
et al., 1993). The potential for recovery from the.
vegetative state is greater after traumatic cases than
others. However, a distinction should be made between
recovery of a limited degree of consciousness and the
restoration of useful function. Most patients whose
recovery begins many months after being vegetative
either do not speak or are capable of only occasional

monosyllabic utterances or of obeying simple commands. Some, however, regain some degree of independence, and a few do become able to function on a
day-to-day basis without support the definition of
independence on the Glasgow Outcome Scale (=
moderate or good). By this standard 10% of the patients
vegetative 1 month after severe injury in the threecountries study were independent by 1 year, but such
recoveries were limited to patients under 40 years of
age and were twice as frequent in those under 20 years
(Braakman, Jennett and Minderhoud, 1988). After 3 and
6 months in a vegetative state 20% and 16% of patients
in this series became conscious but none became
independent. In the aggregate of 754 published cases of
patients vegetative 1 month after head injury a quarter
of both adults and children became independent at 1
year but for those vegetative at 3 and 6 months, only
children had a reasonable chance of becoming independent. It is now widely accepted that, although
occasional patients regain some limited consciousness
after a year, it is reasonable to consider the vegetative
state as permanent after 1 year following treatment.
Declarations to this effect have been made by, inter
alia, the American Medical Association (1990), the
American Neurological Association (1993), the British
Medical Association (1993), Medical Council of New
Zealand (1993) and the Multi-Society Task Force
(1994). Some consider 6 months long enough to wait in
adults, particularly in those over 50 years of age.
Whilst almost 50% of patients vegetative a month
after head injury are dead by the end of a year, those
still alive then may survive for long periods. Several
have been recorded for 1520 years and some even
longer. Death is eventually usually due to pulmonary
or urinary tract infection.
In view of the potential for recovery in the early
months it is important to maintain full supportive care
and active efforts at rehabilitation until the condition
is considered to be irreversible. It is then usual to
Table 22.5 One-year outcome of patients vegetative 1, 3 and 6 months after head injury (Source: derived from data in
Multi-Society Task Force, 1994)
Dead
(%)
Vegetative
(%)
Severely
disabled (%)
Independent
(%)
Conscious
(%)
Vegetative at 1 month
Adults (n = 434)
Children (n = 106)
33
9
15
29
28
35
24
27
52
62
Vegetative at 3 months
Adults (n = 218)
Children (n = 50)
35
14
30
30
19
24
16
32
35
56
Vegetative at 6 months
Adults (n = 123)
Children (n = 28)
32
14
52
54
12
21
4
11
16
32
CEREBRAL HEMISPHERES
decide not to treat infection with antibiotics and to

have a do not resuscitate order. In spite of this,
prolonged survival is not unusual and the question of
discontinuing tube feeding arises. A number of medical bodies and more widely based ethical committees
have declared that this is a morally acceptable course
of action, amounting to the withdrawal of treatment
that is futile in that it brings no benefit to the patient.
Many courts in the US, including the Supreme Court,
have sanctioned this, declaring that tube feeding is
medical treatment, as have the High Court, Appeal
Court and the House of Lords in England (Jennett,
1992b; Dyer, 1993). While the US bench recommends
that such decisions now be made by doctors in
consultation with families, a recent House of Lords
Committee (1994) considers that each such case
should come for judicial review and suggests a new
form of Court to deal with such cases. It is, however,
only a small proportion of families who request such
treatment withdrawal even though they accept that
there is no prospect of recovery (Tresch et al., 1991).
22.7 Neurophysical sequelae in conscious

survivors
The true frequency of various kinds of deficit at
various intervals after injuries of differing severity is
difficult to determine, because most reports are based
on series of patients who have been referred to
neurologists, otologists or ophthalmologists because
of persisting complaints. Moreover many deficits are
temporary and resolve during the first few months
after injury.
Persisting neurological deficits in the limbs are
common after severe injuries. Three main patterns of
neurological disability were found in a series of over
300 patients followed for 20 years after injuries severe
enough to have caused unconsciousness or posttraumatic amnesia of a week or more (Roberts, 1979).
Hemiparesis was the main disability in 40%, although
it was slight in the majority of these. Some 20% had a
brain-stem syndrome with asymmetrical cerebellar
and pyramidal signs. In 5% the state was described as
Table 22.6
449
athetoid pseudobulbar, a combination of bilateral

pyramidal and extrapyramidal signs. About 25% of
this series had no neurological deficit; the others had
abnormalities outside this classification.
Neurophysical disabilities 6 months after severe
head injury in 150 Glasgow survivors who regained
consciousness (Jennett et al., 1981) were most often
related to dysfunction in the cerebral hemispheres and
in the cranial nerves (Table 22.6). The brain-stem and
pseudobulbar syndromes described by Roberts (1979)
were seldom encountered, but his finding that a
quarter of the patients had no neurological abnormalities was confirmed.
22.8
Cerebral hemispheres
Pathological evidence indicates that the brunt of the

impact damage from blunt injury falls on the cerebral
cortex (contusions) and on the subcortical white
matter (shearing lesions). Secondary ischemic damage
is common in fatal cases, most often affecting the
cortex and the basal ganglia, and this probably also
affects some survivors. Many patients who remain
disabled have had an intradural hematoma, which
caused secondary focal brain damage in one cerebral
hemisphere. Of 372 consecutive head injuries of all
severities discharged from the Newcastle neurosurgical unit 15% had hemiparesis when they left hospital,
which was severe in 4% (Cartlidge and Shaw, 1981).
More than half these patients had no deficit 2 years
later, and only two of the originally severe deficits
were still severe. In a large series of patients with
compound depressed fracture in Glasgow focal signs
of damage to the cerebral hemisphere were found
soon after injury in 20%, but only half of these had
residual deficits 6 months after injury. Almost twothirds of patients recovering after evacuation of an
acute intracranial hematoma in Glasgow had hemiparesis soon after operation, but 6 months later only
half of them still had a deficit.
In 935 cases of severe head injury in the International Data Bank study cerebral hemisphere damage was clinically evident during the acute stage in
Neurophysical sequelae (%) at 6 months after severe injury (coma 6 hours) (Source: from Jennett et al., 1981)
All cases
(n = 150)
After intracranial
hematoma (n = 77)
No intracranial
hematoma (n = 73)
Any cerebral hemisphere dysfunction
65
62
67
Cranial nerve palsy

All cases
As only sign
37
13
38
10
36
15
14
Ataxia
450
89%; a third of these (28% of all cases) had evidence of

bilateral hemisphere damage. In the 150 Glasgow
survivors 6 months after severe injury, 49% had
hemiparesis and 29% dysphasia, while 21% had both
(Jennett et al., 1981). Hemianopia occurred in 5%,
usually in association with other signs of hemisphere
damage. About half the 150 patients had had an
intracranial hematoma removed.
22.9
Cranial nerve deficits
In the 150 severe patients assessed 6 months after

severe injury, cranial nerve palsies were found in 32%
and in 14% of the series these were the only persisting
signs (Jennett et al., 1981).
22.9.1
ANOSMIA
Loss of sense of smell occurs in about 5% of all

patients admitted to hospital with a head injury. The
incidence is higher (20%) in patients who have been
unconscious, but a fifth of one large series of patients
with anosmia had never been unconscious (Zusho,
1982). The site of injury was occipital in 30%, facial in
20%, frontal in 19% and temporoparietal in 15%. Less
than half the cases have a fracture, but anosmia
occurred in about 50% of patients with CSF rhinorrhea
from an anterior fossa fracture; after surgical repair
80% were anosmic. Recovery rates of 1550% are
quoted, a variation that may reflect how carefully the
deficit was sought soon after injury. Recovery after 3
months is rare, and long-delayed recovery suggests a
central lesion.
The significance of anosmia for the patient can
easily be underestimated. Anosmia can rightly form a
basis for compensation, not only for the loss of many
of the pleasures of life but for interference with
occupation (e.g. in cooks, food handlers and tasters of
wine and tea); also for loss of the ability to detect
dangerous smells (e.g. of escaping gases or of burning). The most reliably recognized test odors are
coffee, tar, oil of lemon, and almond (benzaldehyde).
22.9.2
VISUAL PATHWAYS
These may be affected anywhere from the retina to the

calcarine cortex. Penetrating injuries may affect any
part of the system and are the only frequent cause of
lesions to the optic radiation. In a prospective survey
of 363 patients with midfacial fractures 56 (15%) had
temporary or permanent visual loss (Al-Qurainy et al.,
1991). In a review of 1800 cases of head injury in
hospital optic nerve lesions were detected in 3.6%
(Rowbotham, 1964). Blunt injuries affect the intracanalicular part of the optic nerve five to 25 times
more often than the chiasm. With optic nerve lesions
in the canal it is often difficult to show a fracture, but

there is frequently a nearby orbital or anterior fossa
fracture. Autopsy has shown hemorrhagic, ischemic
and shearing lesions in the optic nerve (Heinze, 1969).
Usually there is complete monocular blindness of
immediate onset (with an unreacting pupil). Recovery
is rare and the disc usually becomes pale within 34
weeks. The mechanism of damage is probably ischemic and no more than 2040% show any recovery.
Chiasmal lesions cause bitemporal hemianopia,
probably owing to ischemia of the vulnerable central
part of the chiasm; like nerve lesions these are usually
present immediately after injury and neither progress
nor improve. Lesions of the calcarine cortex are not
uncommon, but temporary hemianopia or cortical
blindness may be overlooked. Cortical blindness is
often delayed for hours or days after injury and may
last for only hours or days. The delay in onset and the
temporary nature of the disorder suggests hypoxia or
ischemia. Cortical blindness that is marked and
persistent is often associated with aphasia and
agnosia.
22.9.3 DISORDERS OF THE OCULOMOTOR NERVES
AND CONNECTIONS
In the acute stage after injury, temporary abnormalities of eye movements are common. Patients who
are in coma for hours or days may have dysconjugate
roving or reflex (vestibulo-ocular) eye movements,
which return to normal as consciousness is regained,
probably reflecting transient dysfunction in the brain
stem rather than structural lesions that will lead to
sequelae.
Diplopia is common after recovery from the acute
stage of head injury. Often, the problem lies in the
orbit and need not indicate intracranial damage, nor
even involvement of the cranial nerves. Even minimal
dislocation of the globe or mechanical restriction of
movement can produce ocular imbalance, as a result
of orbital fractures, blood, edema or the escape of air
or CSF into the orbit, and the ocular muscles or their
nerve supply may also be involved. When no definite
mechanical or neurological lesion can be found this
symptom frequently responds to orthoptic treatment,
as it is probably due to breakdown of an existing latent
squint.
In a review of 170 ocular nerve palsies due to
trauma the sixth nerve was affected in 34%, the third
in 30%, the fourth in 15% and more than one in 22%
(Rucker, 1966). Third-nerve palsy may be the result of
impact injury or (more often) of tentorial herniation.
Impact lesions are most often in the superior orbital
fissure. Recovery is the rule, but upward movement
may remain restricted; aberrant regeneration may
result in lid elevation when eye movement is initiated.
DELAYED COMPLICATIONS
Sixth-nerve palsy is usually associated with fracture of

the petrous temporal or sphenoid bones, but can occur
because of phenytoin intoxication. Recovery is usual.
When squint persists from any nerve palsy after 6
months, muscle shortening surgery may improve
appearance and it may also restore binocular vision.
22.9.4
SEVENTH AND EIGHT CRANIAL NERVES
These nerves or their end-organs are frequently

damaged by petrous fractures (Healy, 1982; Toglia and
Katinsky, 1976). Most are longitudinal or horizontal,
the fracture running parallel to the long axis of the
petrous bone in front of the nerves, but often damaging the middle ear. Transverse fractures occur with
more severe injuries, and run at right angles to the
petrous axis, often disrupting the bony and membranous labyrinth, the inner ear and the facial nerve.
(a)
Facial palsy
With transverse fracture the nerve is disrupted and the

paralysis is usually immediate and complete and is
often permanent (Potter and Braakman, 1976). With
longitudinal fractures paralysis is often delayed by
23 days, and is incomplete and temporary, with
recovery over a period of 68 weeks. Surgical decompression is frequently recommended but most clinicians consider that intervention is seldom justified,
since most delayed palsies recover (at least partially)
while most immediate lesions remain permanent
whatever is done.
(b)
of Corti. It resembles damage due to high-intensity

noise, affecting the high-frequency range. It may be
temporary and missed without early testing. In
severely injured patients who cannot cooperate, brainstem evoked response audiometry can be helpful;
even if sedation or anesthesia is required to perform
this test the response is not affected (Hall, Huang-fu
and Gennarelli, 1982).
Conductive hearing loss is much more common
because it can occur after less severe injuries, is often
temporary and is related to hemotympanum or a
lacerated drum. However, the recognition of ossicular
chain damage is important because this can often be
corrected surgically, if deafness persists.
22.10
Delayed complications
Head injury is so common that it is inevitable that some

patients who develop one or other of a wide variety of
neurological conditions will be found to have previously sustained an injury. Sometimes, when there is a
valid causal relationship (e.g. epilepsy or meningitis),
the association with injury may not be recognized by
the clinician. The patient may not mention the head
injury because the patient sees no connection with his
present complaint, or the clinician may dismiss a
known head injury as unrelated to the present
condition either because it had been relatively mild or
was so long ago. However, epilepsy declares itself more
than 4 years after injury in 25% of cases of traumatic
epilepsy, while meningitis related to a basal skull
fracture can occur 10 years or more after injury.
Vestibular dysfunction
Transverse fractures that have caused disruption of

the labyrinth and utricle usually result in severe
vertigo and spontaneous nystagmus for 612 weeks,
until there is compensation. Testing series of patients
months after head injury reveals many with eighthnerve dysfunction. Some have clinically evident nystagmus, but many more patients have abnormalities
when electronystagmography is employed. These
data come from series of patients sent to otologists
with persisting symptoms long after injury. The
severity of injury is seldom stated and it is difficult to
assess the frequency of these abnormalities in unselected head injuries of different severities.
(c)
451
Hearing loss
Many patients with severe head injury have some

hearing loss, usually sensorineural, and often associated with a transverse fracture; it may be bilateral.
Sensorineural impairment also occurs without a fracture, probably due to concussive damage to the organ
22.10.1
POST-TRAUMATIC HYDROCEPHALUS
The availability of CT scanning is making known the

frequency and extent of ventricular dilatation after
various kinds of head injury and the natural history of
this condition (Cardoso and Galbraith, 1985). Three
kinds occur; that caused by wasting of the white
matter after severe injury (ex vacuo); that caused by
acute impairment of the circulation of the CSF,
probably secondary to bleeding (obstructive); and
chronic communicating hydrocephalus (normal pressure). A firm diagnosis of normal pressure hydrocephalus is made only when a patient who has already
recovered to a considerable degree from injury develops new symptoms (mental impairment and disorder
of gait); diagnosis depends on these characteristic
clinical features and on evidence of retarded CSF
circulation. In a large study of hydrocephalus of all
kinds after head injury, only a quarter of those
diagnosed as having normal-pressure hydrocephalus
responded well to shunting (Zandler and Foroglou,
1976).
452
22.10.2
LATE TRAUMATIC EPILEPSY
This is by far the most frequent of delayed complications, although it occurred in only about 5% of all
patients admitted to hospital in the UK after nonmissile head injury (Jennett, 1975). An epilepsy rate of
2.5% was recorded for a large series of patients in
Olmstead County, but this included some who did not
attend hospital and many who were not admitted
(Annegers et al., 1980). After some types of injury the
risk is much higher (see below). Of 150 severe injuries
followed for more than a year after injury, 17% had
epilepsy (Jennett et al., 1981); this would certainly
have been higher had the follow-up been longer. The
incidence was twice as great in those who had had a
hematoma and in those with severe disability; 20 of 22
severely disabled patients with epilepsy had either an
intracranial hematoma or a depressed fracture.
The significance of epilepsy for the patient depends
on whether he/she has other disabling sequelae and
on how it impinges on his/her particular life-style.
Epilepsy was the only physical disability in almost
half the patients in whom it occurred after severe
injury. Many are young men on the threshold of their
careers whose future options can be appreciably
limited by the occurrence of epilepsy, even by the
threat that it may develop. Many patients regard the
restriction on car driving that epilepsy entails as one of
the most disabling aspects of this complication, even
for those who are not vocational drivers.
(a)
Time of onset
Seizures in the first week are recognized as a distinct

category (early epilepsy). The proportion of patients
regarded as having begun to suffer from traumatic
epilepsy within a year of injury depends on whether
early fits are counted and on how long the patients are
followed. In a study of 481 patients with late epilepsy
(Jennett, 1975), 56% had their first late fit in the first
year (27% within 3 months of injury). When there had
been early epilepsy the late fits more often began in
the first year (74%). About a quarter of cases in this
series had their first late fit more than 4 years after
injury.
(b)
Type of fit
About 40% of patients with late epilepsy have at least

some fits with focal features. Over 70% of patients
have attacks in which they become unconscious. A
fifth of patients have seizures with temporal lobe
features and when these attacks begin they may not be
recognized for some time as being epileptic in origin.
Petit mal has not been encountered after injury.
(c)
Persistence of fits
It is crucial for the patients future to know whether or

not fits are likely to persist once the first late seizure
has occurred. There are several references to posttraumatic epilepsy having died out, but most were of
wartime missile injuries; this led to the concept that
seizures were a manifestation of a certain stage in the
healing process, and therefore temporary. Remission
of epilepsy is a safer term than cessation; 2 years
without fits is a reasonable (if arbitrary) definition of
remission. Clinicians frequently recommend discontinuation of anticonvulsant drugs in adults after 2
years without fits. However, even a remission of 2 or
more years is frequently followed by reappearance of
post-traumatic fits. It has to be accepted that once a
patient suffers even one late fit there is a high
probability that he/she will continue to have epilepsy,
although this may be relatively well controlled by
anticonvulsants and there may be remissions.
(d)
Prediction of fits
Because this complication occurs relatively seldom

and may not develop until years after injury, there is a
premium on the ability to predict the likelihood of its
occurrence, in order to advise the patient about his
future, and to guide lawyers concerned with claiming
compensation for injury.
The findings of a study of over 800 patients with
traumatic epilepsy following non-missile head injury
have now been confirmed by several observers in
other countries (Jennett, 1975). Three factors increase
the risk of late epilepsy significantly: an acute
intracranial hematoma evacuated within 2 weeks of
injury, an early fit (within the first week) and a
compound depressed fracture of the vault (Table 22.7).
The risk is greater after surgery for an intradural (45%)
than after an extradural hematoma (22%). When a
significant intracerebral hematoma is detected by CT
scanning but surgical evacuation has not been necessary, the epilepsy rate is only 23%.
After a compound depressed fracture, the risk of late
epilepsy varies according to four risk factors: PTA
Table 22.7 Factors increasing the incidence of late

epilepsy (Source: from Jennett, 1975)
No hematoma
Hematoma
No early epilepsy
Early epilepsy
No depressed fracture
Depressed fracture
27/854
45/128
29/868
59/238
27/832
76/447
3
35
3
25
3
17
MENTAL SEQUELAE
453
Figure 22.1 Risks of late epilepsy after compound depressed fracture with varying combinations of factors where three are
known. (Reproduced with permission from Jennett, Epilepsy after Non-Missile Head Injury. Published by Heinemann,
1975.)
exceeding 24 hours, early epilepsy, dural tearing and

focal signs. Various combinations of the risk factors
enable the clinician to identify patients whose risks of
late epilepsy vary between 3% and more than 60% on
discharge from hospital after injury (Figure 22.1). Highrisk combinations occur relatively seldom and some
40% of patients with depressed fracture can now be
reassured that their risk of epilepsy is less than 5%.
In patients with neither a depressed fracture nor an
acute intracranial hematoma the risk of epilepsy is low,
unless there has been an early fit. This applies whether
or not there has been prolonged unconsciousness
(PTA > 24 h). The risk of late epilepsy is increased even
when early epilepsy was confined to a single fit, and
whether the first (or only) early fit was in the first hour
after injury or later during the first week. Although
Table 22.8
children are somewhat less liable to develop late

epilepsy after an early fit, the risk is still significant. The
risks of late epilepsy are summarized in Table 22.8,
which can be regarded as an algorithm if no
hematoma, was there a depressed fracture; if neither,
then was there early epilepsy? The EEG is not helpful in
predicting late epilepsy (Terespolsky, 1972).
(e)
Prophylactic anticonvulsant therapy
There is increasing doubt about the effectiveness of

anticonvulsants in preventing, as distinct from temporarily suppressing, traumatic epilepsy. An attempt to
establish therapeutic levels of phenytoin within 24
hours of injury by intravenous and intramuscular
injections, followed by oral maintenance checked by
Summary of risks of late epilepsy
Operated 45%
Not operated 23%
Intradural
Compound depressed fracture
Early epilepsy + PTA > 24 hours

Other pairs of risk factors*
Only one risk factor
No risk factors
Neither hematoma nor depressed fracture
Early epilepsy 26%

No early epilepsy < 2%
Acute intracranial hematoma
Extradural
*Risk factors: PTA > 24 h, early epilepsy, dural tearing, focal signs
Operated 22%
> 50%
2040%
520%
< 3%
454
blood levels for 1 year, proved difficult and no

convincing benefit could be shown (Young, Rapp and
Perrier, 1975). Another trial of phenytoin for postcraniotomy cases, including many with head injury,
showed a modest reduction in epilepsy at 1 year.
Therapy was then discontinued and by 2 years there
was no significant difference between those who had
been treated during the first year and those who had
not (North et al., 1983). More recent studies on
prophylaxis after intracranial surgery for non-traumatic conditions raise further doubts about the efficacy of prophylaxis with phenytoin or carbamazepine,
and they reveal a considerable incidence of side effects
(Foy, Chadwick and Rajgopalan, 1992). Another controlled trial of phenytoin showed no reduction in posttraumatic seizures (Temkin et al., 1990).
22.11
Mental sequelae
Coma represents the initial disorder of mental functioning after severe head injury, and its depth and
duration indicate the severity of diffuse brain damage.
Once the patient comes out of coma (opens eyes,
speaks or obeys) he remains in a state of disordered
consciousness for a much longer period than he was in
coma, and is always amnesic for this period (posttraumatic amnesia, PTA). Most patients in coma for 6
hours or more have a PTA of a week or more, half of
them of a month or more. Duration of PTA correlates
with ultimate outcome only patients with a months
PTA remain severely disabled, but a quarter of these
patients with a long PTA make a good recovery.
However, in patients whose PTA exceeds 3 weeks it is
almost always possible to detect impairment of
performance on some tests of cognitive function 6
months after injury, and some measurable deficit is
often permanent. Changes in personality are more
frequent than altered intellectual function and
although they can be equally disabling they are less
readily measured. The most frequently encountered
mental sequelae are probably related to widespread
rather than to focal brain damage. This is consistent
with the wide distribution of initial axonal lesions and
of secondary hypoxic damage in the brain. If disorders
of language and of visuospatial perception are regarded as neurophysical, mental sequelae related to focal
brain damage are not common. Some patients, however, do develop features characteristic of frontal lobe
damage, while the frequency of memory disorder may
be related to the predominance of damage to the
temporal lobes. After blunt head injury, however,
damage is seldom confined to one lobe, or even to one
side of the brain; it is therefore unwise to overemphasize the localized lesions. What matters is brain
damage that persists. Neuropsychological deficits
1518 months after injury correlate more closely with
late MRI abnormalities than with those seen on MRI or

CT soon after injury (Wilson et al., 1988).
Few head-injured patients develop major psychoses
but many have reactions such as depression or anxiety
as well as behavioral disorders. It seems preferable to
use the descriptive term mental to cover these
various features and to avoid the terms psychological, functional or psychiatric, each of which
implies a specific (and unverifiable) explanatory
interpretation.
22.12 Deficits of intellectual (cognitive)

function
The availability of large numbers of tests of IQ in
general, and of subtests for different psychological
functions, has led to many reports about the range of
abnormalities that can be found after head injury.
Many of these tests depend heavily on verbal ability
and these put at a disadvantage both patients with
lesions in the left hemisphere and those with educational limitations affecting their pretraumatic vocabulary and capacity to manipulate language. Performance on IQ tests tends to reflect what has been
accumulated over a lifetime by way of intellectual
habits, motivation and cultural expectations. This is
the so-called crystallized intellectual ability, as contrasted with the psychophysiological adequacy of the
brain to solve new problems at present (fluid
intelligence). It can, however, be difficult to judge
what the functional capacity of an individuals brain
was before injury. School performance gives some
guide while occupational status provides a crude
measure for adults. Vocabulary allows an approximate
retrospective assessment because simple verbal tests
tend to show little impairment after injury, probably
because they test overlearned skills. Non-verbal tests
(e.g. part of the Wechsler Adult Intelligence Scale
WAIS and the Ravens Progressive Matrices) depend
on visuospatial ability and on motor performance. But
they also test the ability to reason at the time of testing;
their value lies in their independence from educational and cultural influences prior to injury. On the
other hand, they may be affected by focal brain
damage to the non-dominant cerebral hemisphere that
has caused perceptual and psychomotor deficits.
The balance between focal and general brain damage differs according to the kind of injury. We owe
much of our knowledge about the effects of strictly
localized brain damage to the detailed psychological
follow-up and testing of military head injuries caused
by missile fragments. By contrast, all the evidence
from pathological examination of brains that have
sustained blunt injury indicates that the damage is
usually widespread, although there may be accentuation in certain areas. A study that compared the
DEFICITS OF INTELLECTUAL (COGNITIVE) FUNCTION
psychological test deficits in patients with intracranial

tumor, with cerebrovascular accidents and with craniocerebral trauma showed more similarities than
differences between these different groups of patients
(Reitan, 1973). However, those with head injuries
whose brain damage was judged by neurological
examination to have been unilateral showed cognitive
deficits that were indicative of bilateral pathology
more often than did patients in the two other
diagnostic groups. Psychological deficits indicative of
lesions in the opposite side of the brain from the
primary impact were often found in another series of
brain-damaged patients who were tested many years
after injury (Smith, 1974). Studies of Vietnam veterans
have shown that the more global the cognitive deficit
the greater the importance of the volume of tissue
loss, whereas lesion location is significant for focal
deficits (Grafman et al., 1986).
Verbal abilities are not only less severely affected in
the early stages than are performance IQ tests but they
also recover more rapidly; verbal scores have usually
largely recovered (it they are going to) within 36
months, whereas performance IQ may go on improving for a year or more. Performance tests are more
severely impaired, probably because they depend on a
wider range of cerebral activities and on the integration of these. They also reflect other aspects of higher
mental function, such as motivation and attention,
speed of performance and perseverance as well as the
ability to organize complex tasks over a period of
time. Most complaints by patients and their relatives
are in the areas of fluid intellectual function and in
memory. Routine IQ testing often fails to demonstrate
abnormalities in patients who are clearly not performing normally at home, because the tests commonly
used were not designed to discover these kinds of
alteration in mental activity. Patients may also perform better during the brief period and relatively
structured situation of psychological testing than they
do in real life; this applies particularly to patients with
frontal damage.
Recently psychologists have begun to focus on
various general aspects of mental activity and to devise
appropriate means of testing these (Brooks, 1984;
Levin, Benton and Grossman, 1982). Tests of attention
and vigilance include reaction time to visual and
auditory stimuli, and recognizing and checking off
repeated letters or words in lists. These may be applied
for varying periods (to show fatigue effects) and with
the addition of various degrees of distraction. There is
evidence that the ability to screen out irrelevant
information in order to focus on the task at hand may be
one of the mental skills that patients with diffuse brain
damage lose. Other tests that require the integration of
many different aspects of brain function and may
therefore be sensitive indicators of widespread brain
455
damage are the recognition of faces, the completion of

half-finished pictures and the recognition of anomalies
in sketches of various life situations. These tests all
depend heavily on the integrity of the non-dominant
parietal lobe and may be impaired by local damage in
that location.
The ability to learn new tasks is another aspect of
brain activity that it is appropriate to measure. This
may provide a better indication of the state of the
brain than the capacity to reproduce previously
overlearned material, or to carry out simple tasks, or
to solve problems one at a time in a test situation.
Indeed, many patients can continue to undertake
activities that were previously routine for them,
including their work, but are unable to tackle new
tasks or to learn new skills. There is a similarity
between the effects of head injury and of the normal
aging process; it is characteristic of the elderly that
they can perform well in a routine and familiar
environment, but react badly to new situations. If
recovery in the brain is partly a learning process it
might be expected that the ability to learn after a head
injury would correlate with the capacity to recover
function as a whole. Learning depends to some extent
on memory but this is a function so specifically
affected after head injury, sometimes out of proportion
to other cognitive defects, that it is considered separately later.
This raises the difficult question of the interdependence of different mental functions and the extent to
which one may be affected independently of others.
Because blunt head injury produces widespread damage, dysfunction in the brain as a whole is important,
but cognitive tests are mostly concerned with certain
focal deficits. Although focal dysfunction rarely occurs
without some deterioration of mental function as a
whole, one or other of these focal deficits (such as
memory) may be predominant. It is, however, important not to conclude mistakenly that a patient has
general intellectual deterioration when his difficulties
stem largely from a specific detect. Research on focal
deficits has largely depended on studies of missile
injuries and of patients recovering from ischemic
strokes, both of which can produce focal lesions
without involvement of the brain as a whole. But leftsided lesions can also lead to perceptual deficits
(difficulty in figure ground discrimination) and to
deficits in certain motor functions, such as copying
complex gestures.
22.12.1
LEFT HEMISPHERE LESIONS
Even when patients with clinically detectable dysphasia are excluded, patients with lesions in the dominant
hemisphere tend to have particular difficulty with
various cognitive tests (Grafman et al., 1986). In some
456
patients deficits in verbal skills that were not obvious

in ordinary speech become obvious when learning
and retention of verbal material are tested.
22.12.2
RIGHT HEMISPHERE LESIONS
Some patients with these lesions have clear topographical disorientation or have difficulty in recognizing faces. These deficits may in turn affect memory, in
that initial registration is impaired because of the
perceptual difficulty. In others the visuospatial difficulties may be so subtle as to require highly specialized tests to uncover them.
22.12.3
FRONTAL LOBE SYNDROMES
The importance of frontal lobe damage is being

increasingly recognized. Although its effect is primarily on behavior or personality there is a secondary
influence on cognitive performance in many types of
test (Stuss and Benson, 1984). Three main types of
behavioral change are seen with some correlation with
the location of the damage in the frontal lobe. Lesions
in the dorsolateral region affect the ability to plan and
to correct errors when undertaking complex tasks
(such as are tested by mazes). There is a tendency to
tackle problems with a fixed strategy with an inability
to innovate or change direction in response to the
demands of different tasks or failure to succeed with
one. The disinhibition or defect of social restraint that
is often considered a classical sign of frontal lobe
damage is associated with lesions in the basal or
basomedial region. Medial lesions cause lack of drive
and motivation, features that clearly impinge on
performance tests and on social life. Bilateral lobe
damage produces more marked abnormalities.
22.12.4
MEMORY DEFICITS
Considering the universality of PTA it is not surprising that some disorder of memorizing is a persistent
complaint of many patients and is often of concern
also to their families. Such reports should not always
be taken at their face value because what is loosely
described as a bad memory may prove to refer to more
generalized cognitive deficit or even to dysphasia
(forgetting names). Nonetheless, two-thirds of
patients 5 years after severe injury complained of their
memory in one study. The phenomenon of PTA
indicates that in recovery from unconsciousness the
capacity to lay down on-going memory is usually the
last function to return. This may be because it is one of
those processes that depends on the integration of
several aspects of brain function; it requires that the
mechanics of perception be intact and that attention be
adequate, so that images are clearly received. Little is
known about what is needed to ensure encoding of the
memory, its persistence, and its availability for

retrieval when required; but all are aspects of information-processing, a function that is consistently slow
after head injury (Lezak, 1979).
A distinction should be made between recent and
remote memory. It is a familiar feature of the elderly
demented patient, who cannot remember from day to
day or even hour to hour, that he/she can often
vividly recall his/her childhood; the same occurs after
head injury. Short- and long-term recent memory
should also be distinguished; after head injury it is
recall over the long term (over half an hour or more)
that is impaired. Even patients with devastating
deterioration on this time-scale may retain short-term
memory (e.g. repeating digits correctly). But even this
short-term memory may break down if too much
information is presented and the system is overloaded. These patients are also slow at learning
because this partly depends on memory, although
after much effort and extra time they may eventually
achieve a near-normal proficiency. Learning also
depends on motivation, attention, information processing and planning each of which is often affected
after head injury.
Failure of recall may be because the memory was
never imprinted, or has decayed, or cannot be
retrieved. In patients with prolonged retrograde
amnesia, there is usually recovery of much of the
memory of events that happened prior to the injury,
and this indicates that the problem was a defect of
retrieval. By contrast there is usually permanent loss
of memory for seconds or a minute or so immediately
prior to impact, the trace of those happenings presumably never having been imprinted. This is certainly the case with post-traumatic amnesia, which
remains stable and which does not yield to attempts to
uncover it by abreaction or drugs; such techniques can
sometimes accelerate the return of the more distant
events that are part of retrograde amnesia.
The question of how discrete memory loss can be
without there being parallel deficits in cognitive
function is a matter of dispute among psychologists. Of
87 moderate or severe injuries 25% had defective
auditory and pictorial memory despite normal Wechsler verbal and performance scores; the deficit was
reminiscent of alcoholic Korsokoff amnesia (Levin et al.,
1988). In the context of blunt head injury that causes
widespread brain damage, marked memory disorder is
usually associated with some impairment on standard
IQ tests. Some of these tests are themselves directly
affected by memory dysfunction and some include
specific memory subtests. Even when these are allowed
for, however, there are patients with severe memory
loss whose routine psychometric test results are
otherwise normal; they usually have discrete lesions
that affect the temporal lobes bilaterally. Non-traumatic
PERSONALITY CHANGE
examples are encephalitis and temporal lobectomy,

while bilateral temporal lobe damage is common in
head injury.
Several reports (Lezak, 1979; Brooks, 1989) from
series of patients with severe non-missile head injuries
reveal a good correlation between the duration of PTA
and the degree of persisting memory defect in patients
with more than a weeks PTA. But neither skull
fracture nor focal neurological signs (including dysphasia) are related to memory impairment, and
neither is the score on standard WAIS IQ tests.
Recovery of memory function (to the level of a stable
deficit) occurs relatively rapidly once the patient is out
of PTA but there is seldom any significant improvement after 6 months.
22.12.5
CONCLUSIONS ABOUT COGNITIVE TESTS
There is considerable overlap in the functional effects

of separate deficits in cognition, while behavioral
changes can also affect performance on tests. Brooks
and McKinlay (1983) have suggested a simplification
of Prigatonos classification of deficits, as follows:

learning and memory;

complex information processing;
perception and communication.
Although broad correlations can be found between

measures of severity of injury and the location of the
lesion with the cognitive deficits found, as well as
with their rate and degree of recovery, there are wide
variations between patients.
In seeking correlations between the site of damage
and cognitive deficits new methods of imaging of the
brain have been used. The degree of atrophy on CT scan
was found to relate well to performance IQ (especially
with left frontal atrophy), and to memory quotient
(with atrophy in either hemisphere); but atrophy was
not related to verbal IQ, representing crystallized
intelligence (Cullum and Bigler, 1986). Imaging has its
limitations, however, especially if it is assumed that
absence of abnormality indicates lack of damage.
Studies by MRI in Glasgow showed deep white-matter
lesions soon after injury in 30% of a series of patients in
whom CT showed such lesions in only 2% (Jenkins et
al., 1986). Neuropsychological tests were frequently
abnormal in patients whose lesions had shown only on
MRI. But lesions shown 1518 months after injury
correlate more closely with cognitive deficits than do
those evident soon after injury (Wilson et al., 1988). A
further caution about assuming focality comes from a
study that correlated the frequency of abnormalities on
language tests with the apparent laterality of the brain
damage (Levin, Grossman and Kelly, 1976). Language
deficits were found in only seven of 15 cases judged to
have predominantly left hemisphere damage, but in six
457
of ten whose lesions were considered to be mainly in

the non-dominant hemisphere. The greater importance
of the volume of brain tissue lost than the location of the
lesion after missile injuries has already been discussed.
That the recovery of head-injured patients depends
in part on their pretraumatic psychosocial status has
been recognized for some time. As a group these
patients include a disproportionate number who were
already risk-takers, heavy drinkers and unemployed.
These factors predict poor social recovery but may
also affect performance on cognitive tests. In the
Vietnam follow-up study the best predictor of recovery was the preinjury intellectual and educational
performance, as indicated by tests carried out at the
time of enlistment (Grafman et al., 1986).
The objective of psychometric testing may be
summarized as follows. It should allow an accurate
assessment of the patients cognitive behavioral and
affective strengths and weaknesses, and the implications for rehabilitation. Moreover, testing regimes
should not require prolonged sessions because few
severely head-injured patients can complete such tests.
Good correlation between cognitive tests and outcome
on the Glasgow scale have been shown in two studies
(Brooks, Hosie and Bond, 1986; Clifton et al., 1993). The
latter study identified four tests out of 19 that
correlated most closely with the Glasgow scale at 3
and 6 months after injury. These were Controlled Oral
Word Association, Grooved Pegboard, Trail Making
Part B, and ReyOsterrieth Complex Figure Delayed
Recall. Of these Grooved Pegboard accounted for 80%
of the variation in the outcome scale.
22.13
Personality change
This is the most consistent feature of mental change

after blunt head injury (Brooks, 1988). In some
instances the patients behavior is clearly abnormal,
but in others the change is noticeable only to relatives
or close associates; unless they are questioned systematically the doctor may mistakenly believe that the
patient has made a complete recovery.
Categorizing a concept such as personality is
difficult but it is helpful to consider three aspects of
behavior.
Drive is usually reduced and the apathy that results
may be described as laziness, or simply slowness.
Circumstances may, however, enable a person to carry
out his/her work satisfactorily, particularly it this is
done in a structured environment. Yet when he/she
comes home at night he/she may fail to follow
previous leisure pursuits, preferring to dream the
evening away in an armchair. In the early stages this
lack of drive may be an obstacle to successful
rehabilitation, but later it may be dealt with by a near
relative acting as a daily goad.
458
Affect most often changes in the direction of

depression, which affects half the patients 27 years
after injury. Lack of drive and lack of insight may,
however, lead patients to passive acceptance of their
condition. This may lead them to underestimate their
disabilities and to claim that they are better than they
really are. More florid aspects of disturbed affect are
seen in patients who experience emotional lability.
Inexplicable bouts of crying, or less often of laughter,
may occur; patients with insight can explain that these
represent the outward signs of an emotion that is not
mirrored by a corresponding inner feeling. In that
event they are more distressing to the onlooker than to
the patient. Occasionally a relative will say that a
patient is better behaved or easier to live with since
suffering a head injury. This will usually be when a
previously aggressive individual is now quieter than
before.
Social restraint and judgment are qualities that
individuals exercise in varying degrees, according to
their personality traits and their cultural background.
But when a person who is normally well behaved
socially, who is sensitive to the needs of others and in
control of those inner feelings of dislike and frustration that everyone experiences from time to time,
becomes tactless, talkative and hurtful, there is no
doubt about the change. Such patients may be no
more than a harmless nuisance to those around them,
but they may be subject to outbursts of rage that are
not only out of character but frightening to the
onlooker. It is sometimes questioned whether these
might represent episodes of temporal lobe epilepsy,
but they can seldom be so explained. More often they
result from some trivial frustration that would previously not have led that particular patient to respond
in this fashion. The whole picture of lack of social
restraint is often referred to as childish behavior,
reminiscent of a child not yet trained by years of social
schooling by parents, relatives, and teachers.
A simple means of scaling personality change used
by Brooks (1988) was a five-point scale between two
opposing adjectives (Table 22.9). The amount of these
changes were related to the subjective burden recorded by the relatives. More personality change was
recorded at 6 and 12 months than at 3 months, perhaps
because relatives no longer denied the changes.
Preservation of insight into personality and cognitive
changes is associated with better prospects for slow
improvement over 23 years, with possible benefit
from behavioral modification techniques.
22.13.1
RELATIONSHIP TO PREVIOUS PERSONALITY
It is useful to obtain an account from relatives about

the patients pretraumatic characteristics soon after
injury, when a more unbiased version may be given
than when there has been time for reflection about the
consequences of injury. Even without the prospect of
possible compensation, relatives are apt later to
idealize the patients previous psychosocial status,
and this can make it difficult to assess the degree of
change. Prior personality can best be assessed using a
formal questionnaire; one of the inventories used for
self-report may be modified for use by a relative,
whose opinion on how the patient would have
answered the various introspective questions can then
be recorded.
Sometimes the personality change after injury takes
the form of exaggeration of that patients pretraumatic
personality traits; or it may be a reversal of them for
example, a person previously quiet, cautious, and
kind may become the opposite. It has been suggested
that such patients may have been unduly dependent
for their previous model behavior on the exercise of
marked degree of restraint, probably dependent on
the frontal lobes. However, there is no consistent
relationship between premorbid personality and the
kind of change that follows trauma. Nor is there often
a clear relationship between the type of change and
the site of brain damage, although patients with
frontal damage without prolonged coma sometimes
show a degree of change that is more marked than
would be expected from the severity of the diffuse
damage.
Table 22.9 Assessing personality change by five-point scale between

pairs of adjectives (Source: derived from Brooks, 1988)
Talkative Quiet
Even-temper quick-temper
Relies on others self reliant
Affectionate cold
Likes company dislikes company
Irritable easy going
Unhappy happy
Excitable calm
Energetic lifeless
Down to earth out of touch

Rash cautious
Listless enthusiastic
Mature childish
Sensitive insensitive
Cruel kind
Generous mean
Unreasonable reasonable
Stable changeable
REFERENCES
22.13.2 REACTIVE AFFECTIVE (PSYCHIATRIC)

SYMPTOMS
To suffer a head injury, even a brief concussion, is a

significant experience for anyone. When the incident
is mild the patient recovers sufficiently rapidly to
remember the scene of the accident, the crowd around
him/her, the ambulance, the accident department, and
admission to hospital. By contrast, the more severely
injured patient wakes up in hospital, often after
several days or sometimes weeks about which his/her
mind remains forever blank. The patient finds around
him/her relatives who, unbeknown to the patient,
have been fearing for his/her life but who are now
concerned for his/her sanity. There may be major
physical problems, either related to the brain damage
or to associated injuries. But insight into the situation
as a whole and its implications for the future seldom
develops for some weeks or months. For the moment
living day to day is enough. Only when the patient
goes home do he/she and his/her family realize the
magnitude of the effects that a severe brain injury has
on life as a whole. At this stage improvement can
usually still be recognized on a week-to-week scale
but as this process slows down the probability of
permanent disability becomes increasingly apparent.
Both the patient and his/her family then enter a
new phase of reaction to the situation. Some of the
more severely affected patients are so blunted or
euphoric that they do not appreciate their plight. But
others at this stage become not only aware but also
distressed by their condition. They may react to this by
frustration and anger, placing blame for their shortcomings either on their relatives or on the doctors and
therapists who are trying to help them. Others again
become depressed but many deal with the situation by
denial of disability, particularly of cognitive and
memory deficits that are all too obvious to others. The
relatives may likewise react with frustration, depression or denial, and the psychodynamics of the family
can become crucial once the patient returns home
(Livingstone, Brooks and Bond, 1985a, b; Brooks,
Campsie and Symington, 1986; Brooks et al., 1987).
Prior counseling of families can prepare them for the
nature and the time scale of the problems that they
may have to face. It has long been realized that it is
the mental disability (cognitive and memory deficits
and personality change) that has the most serious
consequences for social reintegration. This is because
mental handicaps tend to evoke secondary or reactive
psychiatric symptoms both in the patient and in his/
her family, and these can aggravate the situation.
The ability to adapt and to cope with new environmental stresses is one of the mental capacities that
head injury most consistently impairs. It is never an
easy matter to adjust a whole lifestyle to a sudden and
459
catastrophic change such as that which commonly

results from severe head injury. But after head injury
this difficulty is compounded by the mental component of this change. It is this that makes parallels with
severe physical disability (such as paraplegia) so
inappropriate. The youth of the patient is another
factor that influences reaction to injury; the average
age of survivors after severe head injury is under 30,
and many are in their late teens or early 20s. The
problems of adolescence or of early married life are
then compounded with those of brain damage, which
may threaten the completion of education and preparation for a career.
22.14
References
Ad Hoc Committee of the Harvard Medical School. (1968) Report on

examining the definition of brain death. Journal of the American Medical
Al-Qurainy, I. A., Titterington, D. M., Dutton, G. N. et al. (1991) Mid-facial
fractures and the eye: the development of a system for detecting patients at
risk of eye injury. British Journal of Oral and Plastic Maxillo-facial Surgery, 29,
363367.
American Medical Association Council on Ethical and Judicial Affairs (1990)
Persistent vegetative state and the decision to withdraw or withhold life
support. Journal of the American Medical Association, 263, 426430.
American Neurological Association Committee on Ethical Affairs (1993)
Persistent vegetative state. Annals of Neurology, 33, 386390.
Annegers, J. F., Grabow, J. D., Groover, R. V. et al. (1980). Seizures after head
trauma: a population study. Neurology, 30, 683689.
Barlow, P. and Jennett, B. (1991) Decisions to limit treatment in a neurosurgical
unit: an aspect of audit mortality. Scottish Medical Journal, 36, 109111.
Barlow, P. and Teasdale, G. (1986) Prediction of outcome and the management
of head injuries: the attitudes of neurosurgeons. Neurosurgery, 19, 989991.
Beecher, H. K. (1969) After the definition of irreversible coma. New England
Benzel, E. C., Gross, C. D., Haddon, T. A. et al. (1989) The apnoea test for the
determination of brain death. Journal of Neurosurgery, 71, 191194.
Bergner, M., Bobbit, R. A., Pollard, W. E. et al. (1976) The sickness impact
profile: validation of a health status measure. Medical Care, 14, 5767.
Braakman, R., Jennett, B. and Minderhoud, J. M. (1988) Prognosis of the posttraumatic vegetative state. Acta Neurochirurgica (Vienna), 95, 4952.
Bond, M. R. (1976) Assessment of the psychosocial outcome of severe head
injury. Acta Neurochirurgica, 34, 5770.
British Medical Association (1993) Guidelines on treatment decisions for
patients in the persistent vegetative state. Annual Report 1993, Appendix 7,
British Medical Association, London.
Brooks, N. (ed.) (1984) Closed Head Injury: Psychological, Social and Family
Consequences, Oxford University Press, Oxford.
Brooks, N. (1988) Personality changes after severe head injury. Acta
Brooks, N. (1989) Cognitive deficits, in Rehabilitation of the Head Injured Adult,
2nd edn, (ed. E. Griffiths et al.), F. A. Davis, Philadelphia, PA.
Brooks, N., Campsie, L. and Symington, C. (1986) The five year outcome of
severe blunt head injury. Journal of Neurology, Neurosurgery and Psychiatry,
49, 764770.
Brooks, D. N., Hosie, J., Bond, M. R. et al. (1986) Cognitive sequelae of severe
head injury in relation to the Glasgow Outcome Scale. Journal of Neurology,
Brooks, N., Campsie, L., Symington, C. et al. (1987) The effects of severe head
injury upon patient and relatives within seven years of injury. Journal of
Head Trauma Rehabilitation, 2, 113.
Brooks, N. and McKinlay, W. W. (1983) Personality and behaviour changes
after severe blunt injuries a relatives view. Journal of Neurology,
Cardoso, E. and Galbraith, S. (1985) Post-traumatic hydrocephalus: a
retrospective review. Surgical Neurology, 23, 261264.
Cartlidge, N. E. F. and Shaw, D. A. (1981) Head Injury, W. B. Saunders, London.
Clifton, G. L., Hayes, R. L., Levin, H. S. et al. (1992) Outcome measures for
clinical trials involving traumatically brain-injured patients: report of a
conference. Neurosurgery, 31, 975978.
Clifton, G. L., Kreutzer, J. S., Choi, S. C. et al. (1993) Relationship between
Glasgow Outcome Scale and neuropsychological measures after brain
injury. Neurosurgery, 33, 3439.
460
Conference of Medical Royal Colleges and their Faculties in the UK (1976)

Diagnosis of brain death. British Medical Journal, ii, 11871188.
Conference of Medical Royal Colleges and their Faculties of the UK (1979)
Diagnosis of death. British Medical Journal, i, 322.
Conference of Medical Royal Colleges and their Faculties in the UK (1988)
Report of the Working Party on Organ Transplantation in Neonates, Department
of Health and Social Security, London.
Cullum, C. M. and Bigler, E. D. (1986) Ventricle size, cortical atrophy and the
relationship with neuropsychological status in closed head injury: a
quantitative analysis. Journal of Clinical Experimental Neuropsychology, 8,
437452.
Dyer, C. (1993) Law Lords rule that Tony Bland does not create a precedent.
Emanuel, L. L., Barry, M. I., Stoeckle, J. D. et al. (1991) Advance directives for
medical care: a case for greater use. New England Journal of Medicine, 324,
889895.
Feest, T. G., Riad, H. N., Collins, C. H. et al. (1990) Protocol for increasing
organ donation after cerebrovascular deaths in a district general hospital.
Lancet, 335, 11331135.
Foy, P. M., Chadwick, D. W. and Rajgopalan, N. (1992) Do prophylactic
anticonvulsant drugs alter the pattern of seizures after craniotomy? Journal
of Neurology, Neurosurgery and Psychiatry, 55, 753757.
Gentleman, D., Easton, J. and Jennett, B. (1990) Brain death and organ
donation in a neurosurgical unit: an audit of recent practice. British Medical
Journal, 301, 12031206.
Gore, S. M., Cable, D. J. and Holland, A. J. (1992) Organ donation from
intensive care units in England and Wales: two year confidential audit of
deaths in intensive care. British Medical Journal, 204, 349355.
Gore, S. M., Hindes, C. J. and Rutherford, A. J. (1989) Organ donation from
intensive care units in England. British Medical Journal, 299, 11931197.
Gore, S. M., Taylor, R. M. R. and Wallwork, J. (1991) Availability of
transplantable organs from brain stem dead donors in intensive care units.
Grafman, J., Salazar, A., Weingartner, H. et al. (1986) The relationship of brain
tissue volume and lesion location to cognitive deficit. Journal of Neuroscience,
6, 301307.
Hall, J. W., Huang-Fu, M. and Gennarelli, T. A. (1982) Auditory function in
acute severe head injury. Laryngoscope, 92, 883890.
Hall, K., Cope, D. N. and Rappaport, M. (1985) Glasgow Outcome Scale and
Disability Rating Scale: comparative usefulness in following recovery in
traumatic head injury. Archives of Physical Medicine and Rehabilitation, 66,
3537.
Health Departments of Great Britain and Northern Ireland. (1983) Cadaveric
Organs for Transplantation: A Code of Practice Including the Diagnosis of Brain
Death, HMSO, London.
Healy, G. B. (1982) Hearing loss and vertigo secondary to head injury. New
Heinze, J. (1969) Cranial nerve avulsion and other neural injuries in road
accidents. Medical Journal of Australia, 2, 12461269.
Hibberd, A. C., Pearson, I. Y. and McCasker, C. J. (1992) Potential for cadaveric
organ retrieval in New South Wales. British Medical Journal, 304, 13391343.
House of Lords (1994) Report of the Select Committee on Medical Ethics (HL
Paper 211), HMSO, London.
Hunt, S. M., McKenna, S. P. and McEwan, J. (1981) Nottingham Health Profile:
subjective health status and medical consultations. Social Science and
Medicine, 15a, 221229.
Jenkins, A., Teasdale, G., Hadley, M. D. M. et al. (1986) Brain lesions detected
by Magnetic Resonance Imaging in mild and severe head injuries. Lancet, ii,
445446.
Jenkinson, C., Coulter, A. and Wright, L. (1993) Short form 36 (SF36) health
survey questionnaire: normative data for adults of working age. British
Jennett, B. (1975) Epilepsy After Non-Missile Head Injuries, 2nd edn, Heinemann, London.
Jennett, B. (1992a) Severe head injuries: ethical aspects of management. British
Journal of Hospital Medicine, 47, 354357.
Jennett, B. (1992b) Letting vegetative patients die. Ethical and lawful and
brings Britain into line. British Medical Journal, 305, 13051306.
Jennett, B. and Bond, M. (1975) Assessment of outcome after severe brain
damage. Lancet, i, 480484.
Jennett, B. and Plum, P. (1972) Persistent vegetative state after brain damage.
Lancet, i, 734737.
Jennett, B. and Teasdale, G. (1981) Management of Head Injuries, F. A. Davis,
Philadelphia, PA.
Jennett, B., Snoek, J., Bond, M. R. et al. (1981) Disability after severe head
injury: observations on the use of the Glasgow outcome Scale. Journal of
Levin, H. S., Benton, L. and Grossman, R. G. (1982) Neurobehavioral
Consequences of Closed Head Injury, Oxford University Press. New York.
Levin, H. S., Grossman, R. G. and Kelly, P. J. (1976) Aphasic disorder in
patients with closed head injury. Journal of Neurology, Neurosurgery and
Levin, H. S., Goldstein, F. C., High, W. M. et al. (1986) Residual memory deficit
in relation to intellectual functioning after closed head injury. Journal of
Levin, H. S., Saydjari, C., Eisenberg, H. et al. (1991) Vegetative state after
closed head injury: in the traumatic coma data bank. Archives of Neurology,
48, 580585.
Lezak, M. D. (1979) Recovery of memory and learning function following
traumatic brain injury. Cortex, 15, 6372.
Livingston, M. G., Brooks, D. N. and Bond, M. R. (1985a) Three months after
severe head injury: psychiatric and social impact on relatives. Journal of
Livingston, M. G., Brooks, D. N. and Bond, M. R. (1985b) Patient outcome in
the year following severe head injury: psychiatric and social functioning of
relatives. Journal of Neurology, Neurosurgery and Psychiatry, 48, 876881.
McEwan, J. (1983) The Nottingham Health Profile: a measure of perceived
health, in Measuring the Social Benefit of Medicine, (ed. G. Teeling-Smith),
Office of Health Economics, London.
McLellan, D. R. (1989) Neuropathological basis of a poor outcome after nonmissile head injury. MD thesis, University of Glasgow.
Le Coma Traumatique, (eds I. Papo, F. Cohadon and M. Massarotti), Livinia
Editrice, Padua.
Marshall, L. F, Gautille, T., Klauber, M. R. et al. (1991) The outcome of severe
closed head injury. Journal of Neurosurgery, 75, S28S36.
Medical Consultants to the Presidents Commission (1981) Guidelines for the
determination of death. Journal of the American Medical Association, 246,
21842186.
Medical Council of New Zealand (1993) Persistent Vegetative State and the
Withdrawal of Food and Fluids, Report from Bioethics Research Centre,
University of Otago, Medical Council of New Zealand, Wellington.
Multi-Society Task Force on the Persistent Vegetative State (1994) Medical
aspects of the persistent vegetative state. New England Journal of Medicine,
330, 14991508; 15721579.
Murray, L. S., Teasdale, G. M., Murray, G. D. et al. (1993) Does prediction of
outcome alter patient management? Lancet, 34, 14871491.
North, J. B., Penhall, R. K., Hanieh, A. et al. (1983) Phenytoin and postoperative epilepsy. Journal of Neurosurgery, 58, 672677.
OBrien, M. D. (1990) Criteria for diagnosing brain stem death. British Medical
Journal, 301, 108109.
Oddy, M., Coughlan, T., Tyerman, A. et al. (1985) Social adjustment after
closed head injury. A further follow-up seven years after injury. Journal of
Pallis, C. (1990) Brain stem death, in Handbook of Clinical Neurology, vol. 57
Head Injury, (ed. R. Braakman), Elsevier, Amsterdam.
Potter, J. M. and Braakman, R. (1976) Injury to the facial nerve, in Handbook of
Clinical Neurology, vol. 24, (eds P. J. Vinken and G. W. Bruyn), North
Holland, Amsterdam, pp. 105117.
Rappaport, M., Hall, K. M., Hopkins, K. et al. (1982) Disability Rating Scale for
severe head trauma: coma to community. Archives of Physical Medicine and
Rehabilitation, 63, 118123.
Reitan, R. M. (1973) Psychological testing after craniocerebral injury, in
Neurological Surgery, (ed. J. R. Youmans), W. B. Saunders, Philadelphia, PA,
vol. 2, pp. 10401048.
Riad, H., Nicholls, A., Neuberger, J. et al. (1995) Elective ventilation of
potential organ donors an ethical debate. British Medical Journal, 310,
714718.
Roberts, A. H. (1979) Severe Accidental Head Injury: An Assessment of Longterm
Prognosis, Macmillan, London.
Robson, J. G. (1981) Brain death. British Medical Journal, 283, 505.
Rowbotham, G. F. (1964) Acute Injuries of the Head: Their Diagnosis, Treatment,
Complications and Sequels, E. & S. Livingstone, Edinburgh.
Royal College of Physicians (1995) Criteria for the diagnosis of brainstem
death. Journal of the Royal College of Physicians, 29, 381382.
Rucker, C. W. (1966) The causes of paralysis of the third, fourth and sixth
cranial nerves. American Journal of Ophthalmology, 61, 12931298.
Salih, M. A. M., Harvey, I., Frankel, S. et al. (1991) Potential availability of
cadaver organs for transplantation. British Medical Journal, 302, 10531055.
Smith, E. (1974) Influence of side of impact on cognitive impairment
persisting long after severe closed head injury. Journal of Neurology,
Stuss, D. T. and Benson, D. F. (1984) Neuropsychological studies of the frontal
lobes. Psychiatry Bulletin, 95, 328.
Task Force for the Determination of Brain Death in Children (1987) Guidelines
for the determination of brain death in children. Archives of Neurology, 44,
587588.
Tate, R. L., Lulham, J. M., Broe, G. A. et al. (1989) Psychosocial outcome for the
survivors of severe blunt head injury: the results from a consecutive series
of l00 patients. Journal of Neurology, Neurosurgery and Psychiatry, 52,
11281134.
Temkin, N. R., Dikman, S. S., Wilensky, A. J. et al. (1990) A randomised double
blind study of phenytoin for the prevention of post-traumatic seizures. New
Terespolsky, P. S. (1982) Post-traumatic epilepsy. Forensic Science, 1, 147165.
REFERENCES
Thomsen, I. V. (1984) Late outcome of very severe blunt head trauma: a 1015
year second follow-up. Journal of Neurology, Neurosurgery and Psychiatry, 47,
260268.
Toglia, J. U. and Katinsky, S. (1976) Neuro-otological aspects of closed head
injury, in Handbook of Clinical Neurology, vol. 24, (eds P. J. Vinken and G. W.
Bruyn), North Holland, Amsterdam, pp. 119140.
Tresch, D. D., Sims, F. H., Duthie, E. H. et al. (1991) Patients in a persistent
vegetative state: attitudes and reactions of family members. Journal of the
American Geriatrics Society, 39, 1721.
Ware, J. E. (1993) Measuring patients views: the optimum outcome measure.
SF 36: a valid, reliable assessment of health from the patients point of view.
461
Wilson, J. T. L., Wiedmann, K. D., Hadley, D. M. et al. (1988) Early and

late magnetic resonance imaging and neuropsychological outcome after
head injury. Journal of Neurology, Neurosurgery and Psychiatry, 51,
391396.
Young, B., Rapp, R. P., Perrier, D. et al. (1975) Early post-traumatic epilepsy
prophylaxis. Surgical Neurology, 4, 339342.
Zandler, E. and Foroglou, G. (1976) Post-traumatic hydrocephalus, in
Handbook of Clinical Neurology, vol. 24, (eds P. J. Vinken and G. W. Bruyn),
North Holland, Amsterdam, pp. 231253.
Zusho, H. (1982) Post-traumatic anosmia. Acta Otolaryngologica, 108,
9092.

Rilley&#39;s Head Injury Complete

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Rilley&#39;s Head Injury Complete

Transféré par

Droits d'auteur :

Formats disponibles

HEAD INJURY

The illustrations on this page are black and white

MD, BMedSc, FRACS

Department of Neurosurgery, Royal Adelaide Hospital, Adelaide, Austalia

Division of Neurological Surgery, Medical College of Virginia, Richmond,

Published by Chapman & Hall, 26 Boundary Row, London SE1 8HN,

PART ONE: THE INJURY

Biomechanics of closed head injury

Impact to the head

Primary and secondary brain injury

Brain metabolism and cerebral blood flow

The basic principles of brain metabolism

Intracranial pressure and elastance

The problem: raised intracranial

Clinical examination and grading

Measuring cerebral blood flow and

Intracranial pressure monitoring

PART TWO: MEASURING AND

The definitive examination

Imaging the injury

Injury and cell function

Overview of CBF measurements

Electrical function monitoring

Magnetic resonance spectroscopy

From accident site to trauma center

Fluid, electrolyte and metabolic

Sedation and anesthesia

Management of intracranial pressure and

Fluid therapy in uncomplicated postoperative and post-traumatic states

Respiratory and cardiovascular support

PART THREE: TREATMENT

General aspects of trauma

The role of surgery for intracranial mass

Indications for evacuation of

Neuroprotection in head injury

Outcome after severe head injury

Outcome after severe head injury

The vegetative state

A Antibiotics recommended for infections of

Michael R. Fearnside MS, FRACS

Paul E. Bannan FRACS

J. E. Gilligan FANZCA, FFICANZCA

Peter C. Blumbergs FRACP, FRCPA

Donald M. Hadley PhD, FRCR

Ross Bullock MD, PhD

Bryan Jennett CBE, MD, FRCS

Kwan-Hon Chan MS, FRCS

Nigel Jones DPhil, FRACS

Peter J. Crawford BSc, FRCS

Peter J. Kirkpatrick BSc, MSc, FRCS(SN)

Guy L. Ludbrook FANZCA

Peter L. Reilly MD, BMedSc, FRACS

Brain injury remains one of the most difficult and

all neurosurgeons. We also wish to acknowledge the

Numbers in italics refer to tables; numbers in bold refer to illustrations

Agonist-gated ion channels 124, 125, 133,

Anal tone, lax 279

petrous 84, 156, 237, 451

primary 39, 40, 713, 385

Cerebral metabolism 914

testing cognition 157

Computed tomography (CT) contd

Core body temperature 376, 377

Dementia pugilistica 12, 65

Rilley's Head Injury Complete

Rilley's Head Injury Complete