RESEARCH—HUMAN—CLINICAL STUDIES

TOPIC RESEARCH—HUMAN—CLINICAL STUDIES

Are Initial Radiographic and Clinical Scales

Associated With Subsequent Intracranial Pressure
and Brain Oxygen Levels After Severe Traumatic
Brain Injury?
Michael Katsnelson, MD*
Larami Mackenzie, MD*‡
Suzanne Frangos‡
Mauro Oddo, MD‡
Joshua M. Levine, MD*‡§
Bryan Pukenas, MDjj
Jennifer Faerber, PhD¶
Chuanhui Dong, PhD#
W. Andrew Kofke, MD‡§
Peter D. le Roux, MD‡
Departments of *Neurology, ‡Neuro-
surgery, §Anesthesiology and Critical Care,
jjRadiology, and ¶Biostatistics, Center for
Clinical Epidemiology and Biostatistics,
University of Pennsylvania Medical Center,
Philadelphia, Pennsylvania; #Department
of Neurology, Miller School of Medicine,
University of Miami, Miami, Florida
Correspondence:
Peter D. le Roux, MD,
Department of Neurosurgery,
University of Pennsylvania,
235 S 8th Street, 4th Floor,
Philadelphia, PA 19106.
E-mail: lerouxp@uphs.upenn.edu
Received, March 2, 2011.
Accepted, October 14, 2011.
Published Online, November 9, 2011.
BACKGROUND: Prediction of clinical course and outcome after severe traumatic brain
injury (TBI) is important.
OBJECTIVE: To examine whether clinical scales (Glasgow Coma Scale [GCS], Injury
Severity Score [ISS], and Acute Physiology and Chronic Health Evaluation II [APACHE II])
or radiographic scales based on admission computed tomography (Marshall and Rot-
terdam) were associated with intensive care unit (ICU) physiology (intracranial pressure
[ICP], brain tissue oxygen tension [PbtO2]), and clinical outcome after severe TBI.
METHODS: One hundred one patients (median age, 41.0 years; interquartile range
[26-55]) with severe TBI who had ICP and PbtO2 monitoring were identified. The
relationship between admission GCS, ISS, APACHE II, Marshall and Rotterdam scores and ICP,
PbtO2, and outcome was examined by using mixed-effects models and logistic regression.
RESULTS: Median (25%–75% interquartile range) admission GCS and APACHE II
without GCS scores were 3.0 (3-7) and 11.0 (8-13), respectively. Marshall and Rotterdam
scores were 3.0 (3-5) and 4.0 (4-5). Mean ICP and PbtO2 during the patients’ ICU course
were 15.5 6 10.7 mm Hg and 29.9 6 10.8 mm Hg, respectively. Three-month mortality
was 37.6%. Admission GCS was not associated with mortality. APACHE II (P = .003),
APACHE-non-GCS (P = .004), Marshall (P , .001), and Rotterdam scores (P , .001) were
associated with mortality. No relationship between GCS, ISS, Marshall, or Rotterdam
scores and subsequent ICP or PbtO2 was observed. The APACHE II score was inversely
associated with median PbtO2 (P = .03) and minimum PbtO2 (P = .008) and had
a stronger correlation with amount of time of reduced PbtO2.
CONCLUSION: Following severe TBI, factors associated with outcome may not always
predict a patient’s ICU course and, in particular, intracranial physiology.
KEY WORDS: Acute Physiology and Chronic Health Evaluation, Brain tissue oxygen tension, Computed
tomography, Glasgow Coma Scale, Injury Severity Score, Intracranial pressure, Marshall CT classification,
Rotterdam CT Score, Traumatic brain injury
Neurosurgery 70:1095–1105, 2012 DOI: 10.1227/NEU.0b013e318240c1ed www.neurosurgery-online.com

T
Copyright ª 2011 by the
Congress of Neurological Surgeons raumatic brain injury (TBI) remains a
significant cause of morbidity and mor-
tality worldwide. Each year in the United
States, about 2 million people sustain TBI,
ABBREVIATIONS: APACHE II, Acute Physiology
and Chronic Health Evaluation II; CPP, Cerebral
SANS LifeLong Learning and
NEUROSURGERY offer CME for subscribers perfusion pressure; GCS, Glasgow Coma Scale;
that complete questions about featured GCSa, Glasgow Coma Scale on admission; ICP,
articles. Questions are located on the SANS intracranial pressure; IQR, interquartile range; ISS,
website (http://sans.cns.org/). Please read Injury Severity Score; PbtO2, brain tissue oxygen
the featured article and then log into SANS tension; TBI, traumatic brain injury
for this educational offering.
and of these, 500 000 require hospital care.1
In particular, severe TBI (Glasgow Coma
Scale [GCS] # 8) is associated with 30%
mortality and significant disability among
most survivors. However, there are few effec-
tive treatments for severe TBI,2-11 and today
management is centered on the early evacua-
tion of mass lesions and identification and
treatment of secondary cerebral insults that
evolve in the hours and days after TBI.12,13
This includes treatment of raised intracranial
pressure (ICP) and, more recently, at some

NEUROSURGERY VOLUME 70 | NUMBER 5 | MAY 2012 | 1095

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

KATSNELSON ET AL
centers, efforts to optimize brain tissue oxygen tension
(PbtO2).14-17
To best manage severe TBI and to treat ICP and PbtO2
requires admission to a level I trauma center with specialized
intensive care and neurosurgical facilities and an invasive
intracranial monitor. In many countries, capacity limits triage
of all severe TBI patients to level I centers, and decisions about
care, including placement of monitors, are based on expected
prognosis. Several studies have described TBI prognostic models
based on admission characteristics,18-24 but there is less study on
how to predict which patients may benefit from specialized
intensive care or from information provided by an intracranial
monitor.25 In particular, there is little study on how to select
patients whose management may be informed by a PbtO2 monitor.
The decision to place an intracranial monitor is based largely
on the admission GCS: in patients with a GCS # 8, an ICP
monitor is recommended.15 This same recommendation is used
to select patients for a PbtO2 monitor.26,27 However, not all
patients with a GCS # 8 are the same, because there is significant
pathophysiological heterogeneity among these patients.28 In
addition, the evolution of prehospital care and the use of early
sedation, intubation, and ventilation in TBI patients have
reduced the ability of the GCS to accurately classify all patients
and so decide who requires an intracranial monitor.29-31
Alternatively, TBI patients may be classified according to
morphological criteria based on admission computed tomo-
graphic (CT) head scan.32 These CT classification systems are
accepted for descriptive purposes, and various studies have
confirmed their association with TBI outcome.33-36 There also
are many clinical classification systems that attempt to measure
disease severity for adult patients admitted to intensive care
units (eg, Acute Physiology and Chronic Health Evaluation II
[APACHE II])37 or anatomic scoring systems that provide an
overall score for patients with multiple injuries (eg, Injury
Severity Score [ISS]38). Whether these various clinical or
radiographic scales can be used to select patients for intracranial
monitors after TBI is not well defined.
In this study, we examined the relationship between admission
clinical (APACHE, ISS, and GCS) and radiographic scales (Marshall
and Rotterdam) and subsequent physiological data obtained from
intracranial monitors during a patient’s intensive care unit (ICU)
course after severe TBI. The primary objective of our study was
to examine whether initial clinical or radiographic classification of
TBI was associated with the subsequent development of increased
ICP or reduced PbtO2; ie, could we “predict” which patients were
likely to develop abnormal intracranial physiology.
METHODS
Study Population
Patients who were admitted to the Hospital of the University of
Pennsylvania, a level I trauma center, for TBI were retrospectively
identified from a prospective observational database (Brain Oxygen
Monitoring Outcome study) with institutional review board approval.
1096 | VOLUME 70 | NUMBER 5 | MAY 2012
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
The patients were selected by using the following inclusion criteria: (1)
admission within 8 hours from their initial injury and between October
2001 and March 2008; (2) age $ 18 years; (3) an admission GCS # 8 or
who deteriorated to a GCS # 8 shortly after admission; (4) PbtO2
monitoring . 24 hours; and (5) an admission head CT. Exclusion
criteria included the following: (1) fixed and dilated pupils upon
admission; (2) penetrating brain injury; (3) previous TBI or central
nervous system disease; and (4) cardiopulmonary instability (systolic
blood pressure ,90 mm Hg and SaO2 ,90) after initial resuscitation.
Clinical Grading of TBI Severity on Admission
The severity of each patient’s injury was graded according to 3
scales.38-40 (1) The postresuscitation GCS was used.39 (2) The ISS was
recorded for each patient at admission, and each injury was assigned an
injury scale in each of the 6 body regions (head and neck, face, chest,
abdomen, extremities and pelvis, external soft tissue). The highest scores
from 3 of the most affected areas were selected, squared, and summed for
the final ISS.38 (3) The APACHE II score was calculated retrospectively
for all patients from admission and initial resuscitation data.40 An
Internet-based APACHE II calculator (iicumedicus.com) was used to
derive the score for each patient. The APACHE II score also was
calculated without its GCS component.
Head CT Classification
The radiographic severity of TBI for each patient was classified
according to both the Marshall32 and Rotterdam41 systems based on
the admission head CT scan. Two neurologists and a radiologist
independently graded the studies in a blinded fashion. The 3 scores were
compared, and a consensus score was reached in the following manner.
The score with at least 2 raters in agreement or a regrading of the scan if
all 3 raters were in disagreement until at least 2 of the 3 raters agreed was
determined to be the consensus score.
Physiological Monitors
Intracranial pressure, brain temperature, and PbtO2 were monitored
continuously by use of a Camino and Licox monitors (Integra
Neuroscience, Plainsboro, New Jersey). Intracranial monitors were
inserted at the bedside through a burr hole into the frontal lobe and
secured with a triple-lumen bolt. Intracranial monitors were placed into
white matter that appeared normal on admission head CT scan and
ipsilateral to the worst pathology. A noncontrast head CT scan obtained
within 24 hours of insertion confirmed probe placement, and PbtO2
monitor function was confirmed by an increase in PbtO2 following an
oxygen challenge (FiO2 of 1.0 for 5 minutes). Cerebral perfusion pressure
(CPP) was calculated as the difference between mean arterial pressure
(MAP) (transuded at the level of the heart) and ICP (CPP = MAP 2
ICP). Intracranial monitors were removed once ICP was normal (,20
mm Hg) for .24 hours without specific treatment. The first 2 hours of
PbtO2 data were not used in analysis to allow for probe stabilization.
Subsequently, physiological parameters were collected from the written
ICU flow sheets, where values were recorded every 15 minutes. Heart
rate, arterial line blood pressure, central venous pressure, and arterial
oxygen saturation (SaO2) also were recorded continuously in each patient
using a bedside monitor (Component Monitoring System M1046-
9090C: Hewlett Packard, Andover, Massachusetts).
www.neurosurgery-online.com

General Management
Each patient was fully resuscitated according to Advanced Trauma Life
Support guidelines from the American College of Surgeons42 and all were
managed in the Neurosurgical and Trauma Intensive Care Unit
according to a local policy adapted from published recommendations
for severe TBI and ICU care.15,43,44 This included early evacuation of
mass lesions in the operating room. Each patient was endotracheally
intubated and mechanically ventilated with the head of the bed initially
elevated to 30. FiO2 and minute ventilation were adjusted to maintain
SaO2 . 93%, and to avoid PaO2 , 60 mm Hg and PbtO2 , 20 mm Hg.
PaCO2 was targeted between 35 and 45 mm Hg except when ICP was
elevated, in which case PaCO2 was maintained at between 30 and 40 mm
Hg. Patients were sedated with propofol for the first 24 hours, and then
analgesia and sedation were maintained with morphine or fentanyl, and
lorazepam. Euvolemia was achieved with 0.9% normal saline. Thera-
peutic targets were adjusted to avoid ICP . 20 mm Hg and CPP , 60
mm Hg. Packed red blood cells were transfused if the hemoglobin was
less than 8 to 10 g/dL. All patients received anticonvulsants (phenytoin)
for 1 week. Anticonvulsants then were stopped unless there were seizures
when patients received long-term Keppra. Tylenol and external cooling
were used to maintain normothermia (35-37C) when possible.
ICP Management
Therapy aimed to keep ICP , 20 mm Hg and CPP . 60 mm Hg.
Whenever ICP increased (.20 mm Hg, .2 minutes), initial
management consisted of head of bed elevation, sedation (lorazepam),
analgesia (fentanyl), and intermittent cerebrospinal fluid drainage if
a ventricular drain was already inserted. If ICP remained . 20 mm Hg
for .10 minutes, despite initial therapies, osmotherapy was adminis-
tered (mannitol bolus 0.5-1 g/kg), provided serum osmolality was ,320
and serum sodium was , 155 mmol/L. Hypertonic saline (7.5%) was
introduced in May 2006 for use. Phenylephrine (10-100 mg/min) was
administered when CPP was # 60 mm Hg for .15 minutes.
Subsequent-tier therapies for ICP that remained elevated included
optimized hyperventilation (PaCO2  30 mm Hg), propofol, pentobar-
bital, or decompressive craniectomy. Induced hypothermia was incor-
porated into the ICP management protocol in October 2006.
PbtO2 Management
Compromised brain oxygen (PbtO2 , 20 mm Hg) was managed
according to its cause.45,46 Initial interventions included transient
increased FiO2 (60-100%) or increased CPP. Pulmonary-associated
PbtO2 reductions were corrected through ventilator settings. Metabolic
delivery (mean arterial blood pressure and volume status), ICP, and
oxygen demand (pain, fever, seizures, shivering) were optimized next.
A blood transfusion was administered to achieve a hemoglobin
concentration $ 10 mg/dL if these measures failed.47 A decompressive
craniectomy was performed when, despite maximal medical measures,
PbtO2 was , 20mm Hg for . 15 minutes or progressively declined.
Data Analysis
The following intracranial physiological parameters were obtained from
ICU flow sheets and calculated for each patient: (1) elevated ICP
including maximum and median ICP; (2) minimum and median PbtO2;
and (3) incidence of compromised PbtO2 (,20 mm Hg) and the total
amount of time in minutes PbtO2 was below threshold (PbtO2 , 20,
PbtO2 , 15, PbtO2 , 10).14,46,48-53 The mean, maximum and median
NEUROSURGERY
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
INITIAL CLINICAL SCALES AFTER SEVERE TBI
ICP values were calculated from the duration of a patient’s admission.
The duration of compromised PbtO2 was calculated by use of the simple
averaging method. PbtO2 values were plotted and the points connected
by straight lines. The time coordinates where these lines crossed our
predetermined PbtO2 levels were determined, and the total times spent
below each threshold were calculated to the nearest minute. Both
cumulative (total time PbtO2 , 20, PbtO2 , 15, PbtO2 , 10) and
nonoverlapping (PbtO2 , 10, 10 # PbtO2 , 15, 15 # PbtO2 , 20)
intervals were recorded. The duration of elevated ICP .20 mm Hg
or .25 mHg was calculated by use of a similar method. A P value of
#.05 was considered statistically significant.
Outcome
Patient outcome was determined as favorable (patient had no, little, or
moderate disability) or unfavorable (patient died, or was in a vegetative
state or had severe disability) using the Glasgow Outcome Scale by
telephone follow-up or medical record review at 3 months after TBI
(6 2 weeks). In addition, 30-day and 3-month mortality was recorded.
Statistical Analysis
All analyses were performed using SAS version 9.2 (SAS Institute
Inc., Cary, North Carolina). Patient clinical characteristics and
physiological variables were compared between patient groups with
favorable vs unfavorable outcome and survivors vs nonsurvivors to
determine whether there were any significant differences. The various
physiological variables then were compared with admission clinical
grading scales and head CT classification. Means and standard
deviation (SD) or median and interquartile range (IQR) were used to
report the admission clinical scores and physiological measures
according to distribution of the data, whereas percentage was used to
describe the categorical variables. Spearman partial correlation analysis
was conducted to examine the relationships between admission clinical
scores and physiological measures, after controlling for age and sex.
Mann-Whitney U test was used to compare the differences in
admission clinical scores or physiological assessments between
mortality and favorable outcome groups. Stepwise logistic regression
analysis was conducted to identify the significant predictors of
mortality and favorable outcome, whereas stepwise linear regression
was used to search for the significant predictors at admission (clinical
or radiographic scales) of physiological measures. In the linear
regression analysis, physiological measures were log-transformed to
reduce the skewness and kurtosis, except median PtbO2.
RESULTS
Patient Characteristics
One hundred one patients, including 76 males and 25
females (median age, 41; IQR 26-55 years), who underwent
ICP and PbtO2 monitoring were included in our study. Their
demographic, clinical, and radiographic data are summarized
in Table 1. The median (IQR) GCS on admission (GCSa) was
3 (3-7); 54.5% of the patients had a GCS of 3 and 94.1%had
a GCSa # 8. The remaining 6 (5.9%) patients’ GCS scores
deteriorated to # 8 shortly after admission. The median
(IQR) ISS, APACHE II, and APACHE noGCS scores were 30
(26-38), 21 (18-25), and 11 (8-13), respectively. Median
Rotterdam and Marshall scores were 4 and 3 (Table 1).
VOLUME 70 | NUMBER 5 | MAY 2012 | 1097
KATSNELSON ET AL
TABLE 1. Admission Demographic and Clinical Characteristics in
101 Patientsa
Characteristics n or (Median)
Sex
Male 76
Female 25
Age, y (41)
Glasgow Coma Scale (admission) (3)
3 55
4 2
5 3
6 14
7 14
8 7
10 4 4.0
11 1 1.0
15 1 1.0
Injury Severity Score (30)
APACHE II (21)
APACHE noGCS (11)
Rotterdam consensus (4)
2 7
3 17
4 36
5 26
6 15
Marshall consensus (3)
1 2
2 12
3 37
4 7
5 42
6 1
a
IQR, interquartile range; APACHE II, Acute Physiology and Chronic Health
Evaluation II; noGCS, without Glasgow Coma Scale.
Overall, 99 (98.0%) patients had an intracranial abnormality
on their admission head CT scan. This included 43 (42.6%)
who had a parenchymal lesion, 44 (43.6%) who had abnormal
perimesencephalic cisterns, and 7 (6.9%) who had midline
shift .5 mm.
Intracranial Physiology
Mean ICP for all patients was 15.5 6 10.7 mm Hg and the mean
maximum ICP was 34.0 6 19.2 mm Hg. At least 1 episode of
elevated ICP (.20 mm Hg) was observed in 81 (80.2%) patients.
On average, 1367.3 6 2192.4 cumulative minutes of elevated ICP
was observed in each patient. Average median and minimum PbtO2
were 29.9 6 10.8 mm Hg and 12.3 6 10.0 mm Hg, respectively.
Overall, a total of 88372 minutes of compromised PbtO2 (, 20
mm Hg) was recorded during monitoring in all patients (this
includes brain hypoxia [,10 mm Hg]). The mean and median
duration of compromised PbtO2 was 875 6 1689 minutes and
334.1 (IQR 43.1-991.8) per patient. At least 1 episode of
compromised PbtO2 was observed in 81 (80.2%) patients, and
1098 | VOLUME 70 | NUMBER 5 | MAY 2012
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
an episode of brain hypoxia was identified in 43 (42.6%) patients.
On average, 207 6 761.4 cumulative minutes of brain hypoxia
were observed in each patient.
% or (IQR)
Relationship of Admission Clinical and CT Scales to
75.3
Subsequent Intracranial Monitoring
24.8
(26-55) Table 2 summarizes the relationship, after adjustment for age
(3-7) and sex, between the initial assessment scales and intracranial
54.5 physiological measures (ICP, elevated ICP, and duration of
2.0 compromised PbtO2). The admission GCS was inversely
3.0
associated with the Rotterdam CT score (P = .03), but not with
13.9
13.9
the Marshall CT Score, ICP measurements, or amount of time of
6.9 compromised PbtO2. No association between ISS and CT score,
ICP, or time of PbtO2 compromise was observed. We did not
observe a significant association between duration of elevated ICP
(.20 or .25 mm Hg) and the various clinical and radiographic
(26-38) grading scales, although there was a tendency for ICP .25 mm
(18-25)
Hg to be associated with APACHE (P = .08) and Rotterdam
(8-13)
(4-5) scores (P= .07). The APACHE II score was inversely associated
6.9 with median PbtO2 (P = .03) and minimum PbtO2 (P = .008)
16.8 and had a stronger correlation with amount of time of reduced
35.6 PbtO2: PbtO2,20 (P = .001), PbtO2,15 (P , .001), PbtO2,10
25.7 (P = .003), 10#PbtO2,15 (P = .001), and 15#PbtO2,20
14.9 (P = .009). The Apache II noGCS was inversely associated with
(3-5)
2.0
median PbtO2 (P = .03) and minimum PbtO2 (P = .01) and had
11.9 a positive correlation with amount of time of decreased PbtO2:
36.6 PbtO2,20 (P , .001), PbtO2,15 (P = .001), PbtO2,10
6.9 (P = .003), 10#PbtO2,15 (P = .002) and 15#PbtO2,20
41.6 (P = .003), ie, the longer the duration of abnormal PbtO2, the
1.0 worse the APACHE score. A relationship between the APACHE
II score or the APACHE II noGCS and ICP or CT classification
was not found. Similarly, no relationship was observed between
the Rotterdam or Marshall scores and ICP or duration of
compromised PbtO2 recorded during a patient’s ICU time.
Among the 21 patients who never developed compromised PbtO2
(,20 mm Hg), the maximum APACHE score was 31, whereas it
was 35 in patients who had an episode of compromised PbtO2.
Multiple linear regression analysis then was conducted to
identify the admission predictors associated with intracranial
physiological measures. The results are shown in Table 3. Age
was inversely associated with both median ICP (P = .02) and
maximum ICP (P , .001). Female sex was associated with
higher minimum PbtO2 (P = .02). APACHE II was inversely
associated with minimum PbtO2 (P = .002) and directly
associated with the amount of time that PbtO2 was ,20 mm
Hg (P = .04), ,15 mm Hg (P , .001), between 10 and 15 mm
Hg (P = .001), and , 10 mm Hg (P , .001). No significant
associations were found between GCS, ISS, Rotterdam, and
Marshall scores, and subsequent ICP or PbtO2 measurements in
multivariate analysis. Patients who underwent surgery (for a mass
lesion, ie, Marshall 5) were separately examined. Marshall 5 was
associated with decreased median and maximal ICP but not
PbtO2 (Table 3).
www.neurosurgery-online.com
 INITIAL CLINICAL SCALES AFTER SEVERE TBI

TABLE 2. Physiological Measures and Their Correlation With Admission Clinical Scoresa
Physiological Measures Correlation Between Admission Clinical Scores and Physiological Measures
Mean 6 SD Spearman APACHE Rotterdam Marshall
Measures Median (IQR) Correlationb GCSa ISS APACHE II noGCS Consensus Consensus
Median ICP mm Hg 15.5 6 10.7 Score 2.041 .110 .073 .052 .150 2.001
14 (11-17) P value .69 .28 .47 .61 .14 1.00
Max ICP mm Hg 34.0 6 19.2 Score 2.010 2.105 .093 .094 .088 2.097
30 (21-44) P value .92 .30 .36 .35 .39 .34
ICP .20 1367.3 6 2192.4 Score 2.015 .036 .151 .169 .167 2.064
Min 521 (67-1991) P value .89 .72 .14 .10 .10 .53
ICP .25 581.6 6 1200.1 Score 2.064 2.006 .179 .179 .185 2.047
Min 92 (0-696) P value .53 .95 .08 .08 .07 .65
Median PbtO2 mm Hg 29.9 6 10.8 Score .013 2.037 2.224 2.218 .049 .008
30.7 (24.5-35.6) P value .90 .72 .03 .03 .63 .94
Min PbtO2 mm Hg 12.3 6 10.0 Score .120 2.002 2.266 2.251 2.138 2.065
10.8 (5.2-17.0) P value .24 .98 .008 .01 .17 .52
PbtO2,20 min 875.0 6 1689.0 Score 2.034 .087 .343 .349 .104 .121
334.1 (43.1-991.8) P value .74 .39 .001 ,.004 .31 .23
PbtO2 ,15 min 407.2 6 1161.9 Score 2.136 .103 .369 .342 .087 .129
66.5 (0-381.2) P value .18 .31 ,.001 .001 .39 .20
PbtO2,10 min 207 6 761.4 Score 2.168 .055 .300 .292 .039 .098
0 (0-105.2) P value .10 .59 .003 .003 .70 .34
PbtO2 10-15 min 200.1 6 521.7 Score 2.105 .135 .323 .304 .079 .098
34.9 (0-166.9) P value .30 .18 .001 .002 .44 .33
PbtO2 15-20 min 467.8 6 790.8 Score .030 .117 .261 .298 .105 .043
141.0 (8.9-516.5) P value .77 .25 .009 .003 .30 .67
Rotterdam consensus 4.3 61.1 Score 2.219 2.078 .100 .043 1.00 .591
4 (4-5) P value .03 .45 .32 .67 ,.001
Marshall consensus 3.8 61.2 Score 2.157 2.110 .192 .174 .591 1.00
3 (3-5) P value .12 .28 .06 .08 ,.001
a
ICP, intracranial pressure; PbtO2, brain tissue oxygen tension; GCSa, Glasgow Coma Score on admission; ISS, Injury Severity Score; APACHE II, Acute Physiology and Chronic
Health Evaluation II; APACHE no GCS, Acute Physiology and Chronic Health Evaluation II without the GCS component; Min, minimum; Max, maximum.
b
Adjusted for age and sex.

Clinical Outcome elevated ICP (.20 mm Hg or 25 mm Hg) was not associated

At 3 months, 38 (37.6%) patients had died, and 52 (51.5%)
had a favorable outcome. Table 3 shows the associations between
admission clinical and radiographic scores and intracranial
monitoring values with outcome. An association between GCS
and ISS with mortality or favorable outcome was not observed.
Those who died (median, 24; IQR 20-26) or those who had an
unfavorable outcome (median, 24; IQR 20-26) had higher
APACHE II scores than those who survived (median, 21; IQR
17-24; P =.003) or had a favorable outcome (median, 20: IQR
16.5-22.5; P = .002). A similar relationship was observed when
APACHE noGCS was examined. Greater Rotterdam and
Marshall scores were both significantly associated with mortality
and unfavorable outcome (Table 4).
Relationship Between ICP, PbtO2, and Outcome
Median and maximum ICP were not associated with mortality
or favorable outcome (Table 4). In addition, the duration of
with outcome. By contrast, several PbtO2 parameters demon-
strated a significant relationship with outcome: lower PbtO2
was associated with both mortality and unfavorable outcome
(Table 4). Median PbtO2 (27.6; IQR 22.2-32.3) was less in
patients who died than survivors (median, 32.4; IQR 24.9-36.7;
P = .01), and in those with an unfavorable outcome (median
32.5; IQR 25.5-37.8 vs 27.7; IQR 24.1-33; P = .03) than in
those with a favorable outcome. Lower minimum PbtO2 (median
6.5; IQR 2.2-12.1) was observed in those who died than
survivors (median 12; IQR 6.8-19.4; P , .001). There was
a tendency for minimum PbtO2 to be less in patients with an
unfavorable outcome than those with a favorable outcome
(median 9.7; IQR 5-13.7 vs 11.8; IQR 5.6-18.8; P = .07).
The duration of compromised PbtO2 (,20 mm Hg) was
significantly longer in patients who died (median 606.8; IQR
194.9-1358.3 minutes) than survivors (191.3; IQR 0-961.4;
P = .01). This was true also for unfavorable outcome (median
587.9; IQR 171.4-1335 vs favorable outcomes median 180.7;

NEUROSURGERY VOLUME 70 | NUMBER 5 | MAY 2012 | 1099

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

KATSNELSON ET AL

TABLE 3. Multiple Linear Regression Analysis Showing the Admission Predictors Associated With Physiological Measuresa
Physiological Measuresb Associated Admission Predictorc Beta (95% CI) P R2
Median ICP Age 2.0042 (2.0089, 2.0004) .073 .11
Rotterdam consensus .1047 (.0205, .0890) .015
Marshall score = 5 2.1985 (3945, 2.0035) .046
Max ICP Age 2.0085 (2.0136, 2.0033) ,.001 .24
Female 2.1924 (2.4114, .0266) .084
Marshall score = 5 2.2347 (2.4305, 2.0390) .019
Median PbtO2 APACHE noGCS 2.5829 (21.0756, 2.0902) .021 .05
Min PbtO2 Female .4466 (.0837, .8094) .016 .14
APACHE II 2.0512 (2.0828, 2.0197) .002
PbtO2 ,20 APACHE II .1165 (.0079, .2251) .036 .04
PbtO2 ,15 APACHE II .1959 (.0914, .3032) ,.001 .12
PbtO2 ,10 Female 2.9906 (22.1503, .1692) .093 .13
APACHE II .1743 (.0736, .2750) ,.001
PbtO210-15 Age 2.0246 (2.1882, .0022) .072 .10
APACHE II .1725 (.0689, .2761) .001
PbtO215-20 None
a
ICP, intracranial pressure; PbtO2, brain tissue oxygen tension; GCSa, Glasgow Coma Score on admission; ISS, Injury Severity Score; APACHE II, Acute Physiology and Chronic
Health Evaluation II; APACHE noGCS, Acute Physiology and Chronic Health Evaluation II without the GCS component; Min, minimum; Max, maximum; Beta, regression
coefficient; CI, confidence interval.
b
log transformed except median PbtO2.
c
With a P value ,.1 in the stepwise regression analysis.

TABLE 4. Univariate Analysis to Test the Associations of Admission Clinical Scores and Physiological Measures With Mortality and Favorable
Outcomea
Mortality Favorable Outcome
No (n = 63) Yes (n = 38) No (n = 49) Yes (n = 52)
b
Variable Median (IQR) Median (IQR) P Median (IQR) Median (IQR) Pb
GCSa 5 (3-7) 3 (3-6) .17 3 (3-6) 5.5 (3-7) .17
ISS 30 (26-38) 29 (26-38) .78 30 (26-38) 30 (26-38) .95
APACHE II 21 (17-24) 24 (20-26) .003 24 (20-26) 20 (16.5-22.5) .002
APACHE noGCS 9 (7-13) 12 (9-15) .004 12 (9-14) 9.5 (7-12) .003
Rotterdam consensus 4 (3-4) 5 (4-6) ,.001 5 (4-5) 4 (3-4) .001
Marshall consensus 3 (3-5) 5 (4-5) ,.001 5 (3-5) 3 (3-5) ,.001
Median ICP, mm Hg 14 (11-16) 15 (10-18) .13 14 (10-18) 14.5 (12-17) .66
Max ICP, mm Hg 32 (23-43) 24.5 (19-46) .42 28 (19-48) 31 (23-39.5) .58
ICP .20 min 479 (135-1230) 582 (22-2556) .86 512 (22-2133) 546 (153-1305) .68
ICP .25 min 74 (0-502) 241 (0-1132) .42 133 (0-793) 70 (0-572) .81
Median PbtO2, mm Hg 32.4 (25.9-36.7) 27.6 (22.2-32.3) .01 27.7 (24.1-33) 32.5 (26.5-37.8) .03
Min PbtO2, mm Hg 12 (6.8-19.4) 6.5 (2.2-12.1) ,.001 9.7 (5-13.7) 11.8 (5.6-18.8) .07
PbtO2 , 20 min 191.3 (0-961.4) 606.8 (194.9-1358.3) .01 587.9 (171.4-1335) 180.7 (6.8-533.7) .007
PbtO2 , 15 min 20.6 (0-237.8) 148.6 (20-810.7) .001 109.4 (10.7-500.0) 19 (0-261.1) .03
PbtO2 , 10 min 0 (0-25.8) 23.2 (0-366.9) .004 1.5 (0-207.6) 0 (0-82.5) .13
PbtO210-15 min 13.8 (0-117.4) 92.6 (10.7-213.4) .008 67.9 (6.7-209.0) 11.2 (0-113.2) .02
PbtO215-20 min 105.8 (0-555.8) 235.4 (60-508.1) .27 299.2 (60-604.7) 89.7 (0-360.2) .02
a
IQR, interquartile range; ICP, intracranial pressure; PbtO2, brain tissue oxygen tension; GCSa, Glasgow Coma Score on admission; ISS, Injury Severity Score; APACHE II, Acute
Physiology and Chronic Health Evaluation II; APACHE no GCS, Acute Physiology and Chronic Health Evaluation II without the GCS component; Min, minimum; Max, maximum.
b
Based on Mann-Whitney test.

1100 | VOLUME 70 | NUMBER 5 | MAY 2012 www.neurosurgery-online.com

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.


IQR 6.8-533.7 minutes; P = .007). Similar results, ie, longer
duration of time PbtO2 was below threshold, was observed
for PbtO2 ,10 mm Hg and between 10 and 20 mm Hg
(Table 3).
DISCUSSION
We examined 101 severe TBI patients and compared clinical
(GCS, ISS, APACHE II, APACHE II noGCS) and radiographic
scales (Rotterdam and Marshall) obtained at admission with the
patient’s intracranial physiology during ICU care (ICP and
PbtO2). The APACHE score was associated with PbtO2 but not
ICP. The other clinical and radiographic scales did not have
a relationship with ICP or PbtO2. Worse APACHE, Rotterdam,
and Marshall scales, but not the GCS and ISS, were associated
with outcome. These data suggest that all patients with severe
TBI should be considered for intracranial monitoring and that
patient selection for aggressive ICU care may be best made by
understanding the morphological consequences of the injury
(Marshall or Rotterdam score) and the overall physiological
impact (eg, APACHE score). These findings will require
validation in a prospective study and with other clinical scales.
A second important finding from this study is the relationship
between the initial Rotterdam score and mortality, because it
represents validation of this scale in an independent series.
Methodological Limitations
Our study has several potential limitations. First, we examined
the data retrospectively, and patients were selected because their
GCS was # 8; this may bias our findings. The results also may
differ for patients with moderate or mild TBI and cannot be
extrapolated toward them. Second, the study included patients
from a single center and so the findings may not apply at other
institutions. Third, the sample size, 101 patients, is relatively small
and so may result in type 2 errors. Therefore, the findings will
need confirmation in a larger cohort. Furthermore, most of our
patients were ventilated and sedated before arrival in hospital and
this may alter the association between the GCS, clinical findings,
and outcome. Fourth, CPP was not used as an outcome variable in
our analysis, although it was calculated for all our patients and
included in management algorithms. Fifth, several treatment
modalities (eg, hypertonic saline and induced hypothermia) were
introduced during the course of our study and this may affect
our results. Sixth, in outcome analysis we examined short-term
outcome and mortality. Although this is useful and accepted in
ICU studies, it may not adequately describe TBI outcome where
more long-term functional and cognitive measures are important.
Seventh, our analysis cannot account for withdrawal of care in the
patient who was monitored and does not include patients who
were not selected for aggressive care or monitoring. This may bias
both the outcome and intracranial physiology findings. In
addition, the study is not a pure observational study in that
abnormal ICP, CPP, and PbtO2 were treated. We do not have data
to examine the effect of therapy (eg, the therapeutic intensity level)
NEUROSURGERY
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
INITIAL CLINICAL SCALES AFTER SEVERE TBI
on the relationship between the admission findings and sub-
sequent ICU physiology. It is conceivable that this could influence
our findings. Finally, we calculated the APACHE score from
admission findings rather than the first 24 hours of care. We also
used data from the admission CT scan. Other studies34,35 have
shown that the worst CT scan obtained during a patient’s course
has greater predictive value. Although this may influence our
findings, we believe it is more practical because decisions about
intracranial monitoring need to be made early in a patients’ course
and it was our aim to examine whether admission data can be used
to “predict” who needs an intracranial monitor. We also limited
our analysis to select scales. Whether the results will be similar
with other scales, eg, subsequent APACHE score iterations or the
Simplified Acute Physiology Score will require further study.
Despite these limitations, our findings suggest a combination of
anatomic and physiological information about a patient may
provide the best insight into their expected course and outcome
after severe TBI. In addition, the results emphasize the well-
recognized need to better define and classify TBI.
Prediction of TBI Outcome and Clinical Course
Traumatic brain injury’s heterogeneity is a potential barrier to
find effective therapies. Today the GCS is the primary means by
which TBI is classified and patients selected for management or
inclusion in clinical trials. The GCS, however, was designed in
the 1970s, in part, to help triage patients for a CT scan, and,
although useful in prognostic models, it does not provide
complete information about the pathoanatomic subtypes of TBI
or the pathophysiological mechanisms for a specific interven-
tion. Furthermore, reliable acute evaluation of the GCS may be
limited by medical interventions or intoxication.29,54 The
overall clinical severity of TBI also is related to extracranial
injury that often is assessed with the abbreviated injury score or
the ISS.38 However, there is no consensus on the prognostic
value of major extracranial injury in TBI patients18,55,56 Ideally,
a TBI classification system should help select patients with the
potential to benefit from a specific therapy from pathoanatomic,
pathophysiological, and prognostic standpoints.28 This may be
difficult because factors that predict outcome may not
necessarily be associated with the potential for a patient to
benefit from a specific intervention, ie, we cannot assume that
factors that differentiate death from survival also differentiate
good from poor outcome or response to a specific therapy. This
ability to predict clinical course and outcome, however, is
a cornerstone of clinical medicine and particularly important in
such a varied condition as TBI.
Factors such as age and level of consciousness are primarily
prognostic factors. Data about the morphological consequences of
the injury (head CT scan) also are prognostic but allow selection for
surgery, ie, mass lesion evacuation. In our opinion, measures that
describe pathophysiology (eg, ICP, edema volume, microdialysis
analysis of metabolism, blood flow, tissue oxygenation both systemic
and brain, or coagulation among others) may have greater potential
VOLUME 70 | NUMBER 5 | MAY 2012 | 1101
KATSNELSON ET AL
to select patients who may benefit from a specific intervention. Our
results, in part, support this contention in that the APACHE,
Rotterdam, and Marshall scores all were associated with outcome,
whereas only the APACHE score showed an association with PbtO2.
None of the various scales we used were associated with ICP. This
does not mean that ICP is not important, but rather may reflect
treatment effect and our use of permissive intracranial hypertension,
ie, we “tolerated” small increases in ICP if PbtO2 was normal.
Although GCS was not associated with outcome in our study, this
may represent the cohort we studied or the influence of prehospital
care. In particular, we can only conclude that for patients with
a GCS # 8, the GCS cannot further discriminate outcome. Age
and GCS, both well-described factors associated with outcome, are
included in the APACHE model. Our sample size is small and so
we were not able to identify any subgroups in our analysis that did
not warrant monitoring. Our findings imply it may be difficult to
predict a patient’s intracranial pathophysiology during their ICU
course after TBI based on “prognostic” factors. Therefore, it may be
reasonable to place an intracranial monitor in all patients after severe
TBI and then use that pathophysiological information gathered
over time to select and target specific therapies. Our study is not
designed to determine how best to predict outcome after TBI; this
has been well studied in the past.19,21,23,57,58 However, because the
true GCS often may be made less reliable by prehospital care, our
data suggest that a combination of head CT morphological data
(eg, Rotterdam score) and a full understanding of the patient’s
entire pathophysiology (eg, APACHE score) after severe TBI may
provide the best insight into expected outcome at admission.
CT Classification
A head CT scan is the investigation of choice in acute
evaluation after severe TBI and provides an objective appraisal
of structural injury. In particular, the head CT provides
diagnostic information to decide on operative intervention and
also provides objective information for prognosis. The 2 most
common radiographic TBI classifications used are the Marshall
and Rotterdam classifications32,41; both are strongly related to
outcome.20,23,57 Combination of individual CT characteristics
in a model, such as in the Rotterdam CT score, may provide
better discrimination between patients with good vs poor
outcomes.23,36 The presence of traumatic subarachnoid hem-
orrhage is a strong predictor for outcome and mortality in
TBI59-62 but is not included in the Marshall classification. The
Marshall classification also differentiates between patients with
evacuated vs nonevacuated mass lesions. This reflects a clinical
decision and is not a CT parameter. The importance of CT
scanning to classify patients and help predict outcome has
increased, because early sedation, intubation, and ventilation in
patients with severe and even moderate TBI may influence
reliable clinical assessment.29-31 In this study, both CT scales
were associated with outcome. This is an important finding of
this study and represents an independent validation of the
predictive power of admission CT scales (Rotterdam and
1102 | VOLUME 70 | NUMBER 5 | MAY 2012
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
Marshall) for morbidity and mortality. However, neither was
associated with ICP or PbtO2.
Acute Physiology and Chronic Health Evaluation
We found that the APACHE II score, among the other clinical
scales we examined, was the only one associated with subsequent
abnormalities in intracranial physiology. APACHE II was developed
as a predictive tool and incorporates GCS with other physiological
variables for patients that require ICU admission.40 In ICU
medicine, standardized mortality rates are generally adjusted for
age and sex and for baseline characteristics, based on scales such as
Apache II, Trauma Injury Severity Score, or Simplified Acute
Physiology Score II/III, among others. The applicability of these
scores for TBI outcome assessment is unclear.63 In our study, the
APACHE score was not associated with ICP measurements but was
associated with both descriptive (median and minimum PbtO2) and
quantitative measure of PbtO2 compromise (PbtO2,20 mm Hg,
,15 mm Hg, or ,10 mm Hg). When other variables were
controlled for, APACHE II still was associated with minimum
PbtO2 and most time measures of brain hypoxia. Because the GCS
may be a confounding variable in the APACHE II score, we also
used APACHE II without the GCS. APACHE II noGCS had
similar associations in correlation and univariate analysis and was
associated with median PbtO2 in multivariate analysis. Both
measures were associated with outcome. The APACHE II
incorporates many physiological nonneurological measures
(eg, A-a gradient, pH, PaO2, respiratory and heart rate) and
also reflects the patient’s premorbid condition. Therefore,
a higher score could imply impaired global oxygen delivery,
increased systemic aerobic demand, or poor oxygen extraction
associated with altered VO2max, lung disorders, or other
pathologies27,64 that all may subsequently reduce PbtO2. Our
results suggest that this, ie, an understanding of the full
physiological impact of the injury (APACHE), rather than
how it affects brain function alone (eg, GCS or CT score), may
better predict brain oxygenation over time.
Why Predict Who Needs a Monitor?
Severe TBI outcome remains poor and often is associated with
secondary brain injury.20,65 Early identification of secondary
brain injury therefore can lead to its earlier treatment and the
potential to prevent it; this requires neurological monitoring. The
analysis of how information derived from neuromonitoring alters
management or changes outcome remains relatively crude,
although many studies show an association of high ICP, low
CPP, and decreased PbtO2 with poor outcome.66-68 An ICP
monitor is regarded as the “gold standard” monitor, and although
there is no level I evidence that demonstrates that its use improves
outcome after severe TBI, a recent meta-analysis suggests that
adherence to Severe TBI Guidelines and an ICP monitor is
associated with favorable outcome after severe TBI.69,70 However,
some observational studies imply that ICP monitor use may
aggravate outcome.71,72 Several observational clinical series
suggest that a PbtO2 monitor can complement an ICP
www.neurosurgery-online.com

monitor,14,26,49,50,73 identify secondary brain injury when ICP is
normal,49,50,73-76 and that PbtO2-based care combined with ICP/
CPP-based care is associated with a greater likelihood of favorable
outcome after severe TBI than ICP/CPP-based care alone.16,17,77,78
In the present study, lower PbtO2, and a longer duration of
compromised PbtO2 was associated with worse outcome.
There are several unanswered questions about the use of PbtO2
monitors, among them, who should receive a PbtO2 monitor?
Selecting a patient for a monitor implies that information provided
by the monitor translates into better care. However, the value of
any monitor depends on the accuracy of interpretation of the data
provided, as has been demonstrated with pulmonary artery
catheters.79-81 It would be ideal to predict a patient’s response
to treatment similar to how the culture and sensitivity is used in
infectious diseases. Response to treatment, eg, reduction of elevated
ICP or correction of compromised PbtO2 after acute brain injury
appears to be associated with better outcome.16,27,82 However,
while neuromonitoring has become more sophisticated over the
past decade, no rules have been developed, as yet, to predict
treatment response in the ICU.19 The ability to predict outcome,
who to monitor, or what therapies may work after severe TBI is
important because it enables appropriate use of resources and
allows physicians to guide family expectations. Prediction research
in TBI has largely focused on a diagnostic approach to assess the
probability of structural brain damage or an intracranial hema-
toma, or when to recommend CT scanning in patients with mild
TBI. For patients with moderate and severe TBI, research has
centered on clinical outcome prediction where prognostic models
using baseline admission characteristics to predict unfavorable
outcome at 6 months after injury are well described.21,23,57,58
Although there are gaps in our knowledge, eg, genomics and
biomarkers, these models have been validated.83 There has been far
less study on “prediction” of which TBI patients need specialized
ICU care, should receive multimodality monitoring, or how they
will respond to treatment; that which has been done suggests it is
difficult to do so using baseline characteristics.25 Given secondary
brain injury’s temporal (and to an extent causal) relationship to
primary brain injury, use of the one to reliably predict the other is
an attractive option that may alter the disease’s management. Our
results show that, other than the APACHE score that provides
insight into the entire physiological effect of the injury, admission
clinical or radiographic scales do not reliably predict ICU course
(ICP and PbtO2). This finding emphasizes that it is important to
differentiate between the concepts of prognostic factors and factors
that relate to a patient’s ICU course and potential to benefit from
an intervention.
CONCLUSION
Predictive models provide individual risk estimates and thus can
set a baseline for clinical audits and benchmarking. Our results
suggest that, apart from the APACHE score, the initial clinical and
radiographic scales, while associated with mortality and outcome,
do not “predict” ICP or PbtO2 in patients with severe TBI.
NEUROSURGERY
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
INITIAL CLINICAL SCALES AFTER SEVERE TBI
Instead, we suggest that when an ICP monitor is to be placed,
a PbtO2 monitor also should be placed regardless of the initial
clinical and radiological admission scores. A high APACHE score
may be used to guide this selection. We believe that, as the
monitoring and treatment of secondary brain injury evolves, the
information about pathophysiology provided by the monitors,
and not the initial clinical scores, will guide and target treatment
and ultimately may be more predictive of outcome.
Disclosures
Dr le Roux was supported in part by research grants from the Integra
Foundation, Integra Neurosciences, and the Mary Elisabeth Groff Surgical and
Medical Research Trust. Dr le Roux is a member of the Integra Speaker’s Bureau.
All other authors have no financial disclosures to report.
REFERENCES
1. Faul MX, Wald MM, Coronado VG. Traumatic Brain Injury in the United States:
Emergency Department Visits, Hospitalizations, and Deaths. Atlanta, GA: Centers
for Disease Control and Prevention; 2010.
2. Bullock MR, Lyeth BG, Muizelaar JP. Current status of neuroprotection trials for
traumatic brain injury: lessons from animal models and clinical studies.
Neurosurgery. 1999;45(2):207-217; discussion 217-220.
3. Birmingham K. Future of neuroprotective drugs in doubt. Nat Med. 2002;8(1):5.
4. Clifton GL, Miller ER, Choi SC, et al. Lack of effect of induction of hypothermia
after acute brain injury. N Engl J Med. 2001;344(8):556-563.
5. Narayan RK, Michel ME, Ansell B, et al. Clinical trials in head injury.
J Neurotrauma. 2002;19(5):503-557.
6. Edwards P, Arango M, Balica L, et al. Final results of MRC CRASH, a randomised
placebo-controlled trial of intravenous corticosteroid in adults with head injury-
outcomes at 6 months. Lancet. 2005;365(9475):1957-1959.
7. Maas AI, Murray G, Henney H III, et al. Efficacy and safety of dexanabinol in
severe traumatic brain injury: results of a phase III randomised, placebo-controlled,
clinical trial. Lancet Neurol. 2006;5(1):38-45.
8. Roberts I, Schierhout G, Alderson P. Absence of evidence for the effectiveness of
five interventions routinely used in the intensive care management of severe head
injury: a systematic review. J Neurol Neurosurg Psychiatry. 1998;65(5):729-733.
9. Muir KW. Glutamate-based therapeutic approaches: clinical trials with NMDA
antagonists. Curr Opin Pharmacol. 2006;6(1):53-60.
10. Temkin NR, Anderson GD, Winn HR, et al. Magnesium sulfate for neuro-
protection after traumatic brain injury: a randomised controlled trial. Lancet
Neurol. 2007;6(1):29-38.
11. Robertson CS, Valadka AB, Hannay HJ, et al. Prevention of secondary ischemic
insults after severe head injury. Crit Care Med. 1999;27(10):2086-2095.
12. Enriquez P, Bullock R. Molecular and cellular mechanisms in the pathophysiology
of severe head injury. Curr Pharm Design. 2004;10(18):2131-2143.
13. Marshall LF. Head injury: recent past, present, and future. Neurosurgery. 2000;47
(3):546-561.
14. Maloney-Wilensky E, Gracias V, Itkin A, et al. Brain tissue oxygen and outcome after
severe traumatic brain injury: a systematic review. Crit Care Med. 2009;37(6):2057-2063.
15. Brain Trauma Foundation. American association of neurological Surgeons,
Congress of neurological Surgeons, Joint Section on Neurotrauma and Critical
Care, AANS/CNS, et al: guidelines for the management of severe traumatic brain
injury. J Neurotrauma. 2007;24(suppl 1):S1-S106.
16. Spiotta AM, Gracias VH, Garuffe AM, Kofke WA, Maloney-Wilensky E, Troxel AB,
Levine JM, Le Roux PD. Brain tissue oxygen-directed management and outcome in
patients with severe traumatic brain injury. J Neurosurg. 2010;113(3):571-580.
17. Narotam PK, Morrison JF, Nathoo N. Brain tissue oxygen monitoring in
traumatic brain injury and major trauma: outcome analysis of a brain tissue
oxygen-directed therapy. J Neurosurg. 2009;111(4):672-682.
18. Perel P, Edwards P, Wentz R, Roberts I. Systematic review of prognostic models in
traumatic brain injury. BMC Med Inform Decis Mak. 2006;6:38.
19. Lingsma HF, Roozenbeek B, Steyerberg EW, Murray GD, Maas AI. Early
prognosis in traumatic brain injury: from prophecies to predictions. Lancet Neurol.
2010;9(5):543-554.
VOLUME 70 | NUMBER 5 | MAY 2012 | 1103
KATSNELSON ET AL
20. McHugh GS, Engel DC, Butcher I, et al. Prognostic value of secondary insults in
traumatic brain injury: results from the IMPACT study. J Neurotrauma. 2007;24
(2):287-293.
21. Steyerberg EW, Mushkudiani N, Perel P, et al. Predicting outcome after traumatic
brain injury: development and international validation of prognostic scores based
on admission characteristics. PLoS Med. 2008;5(8):e165; discussion e165.
22. Perel P, Arango M, Clayton T, et al. Predicting outcome after traumatic brain
injury: practical prognostic models based on large cohort of international patients.
BMJ. 2008;336(7641):425-429.
23. Maas AI, Steyerberg EW, Marmarou A, et al. IMPACT recommendations for
improving the design and analysis of clinical trials in moderate to severe traumatic
brain injury. Neurotherapeutics. 2010;7(1):127-134.
24. Murray GD, Butcher I, McHugh GS, et al. Multivariable prognostic analysis in
traumatic brain injury: results from the IMPACT study. J Neurotrauma. 2007;24
(2):329-337.
25. Hukkelhoven CW, Steyerberg EW, Habbema JD, Maas AI. Admission of patients
with severe and moderate traumatic brain injury to specialized ICU facilities:
a search for triage criteria. Intensive Care Med. 2005;31(6):799-806.
26. Bratton SL, Chestnut RM, Ghajar J, et al. Guidelines for the management of
severe traumatic brain injury. X. Brain oxygen monitoring and thresholds.
J Neurotrauma. 2007;24(suppl 1):S65-S70.
27. Maloney-Wilensky E, Le Roux P. The physiology behind direct brain
oxygen monitors and practical aspects of their use. Childs Nerv Syst. 2010;26
(4):419-430.
28. Saatman KE, Duhaime AC, Bullock R, Maas AI, Valadka A, Manley GT.
Classification of traumatic brain injury for targeted therapies. J Neurotrauma.
2008;25(7):719-738.
29. Balestreri M, Czosnyka M, Chatfield DA, et al. Predictive value of Glasgow Coma
Scale after brain trauma: change in trend over the past ten years. J Neurol Neurosurg
Psychiatry. 2004;75(1):161-162.
30. Buechler CM, Blostein PA, Koestner A, Hurt K, Schaars M, McKernan J.
Variation among trauma centers’ calculation of Glasgow Coma Scale score: results
of a national survey. J Trauma. 1998;45(3):429-432.
31. Moskopp D, Stahle C, Wassmann H. Problems of the glasgow coma scale with
early intubated patients. Neurosurg Rev. 1995;18(4):253-257.
32. Marshall LF, Klauber MR, Van Berkum Clark M, et al. A new classification of
head injury based on computerized tomography. J Neurosurg. 1991;75(suppl):
S14-S20.
33. Hukkelhoven CW, Steyerberg EW, Habbema JD, et al. Predicting outcome after
traumatic brain injury: development and validation of a prognostic score based on
admission characteristics. J Neurotrauma. 2005;22(10):1025-1039.
34. Lobato RD, Cordobes F, Rivas JJ, et al. Outcome from severe head injury related
to the type of intracranial lesion. A computerized tomography study. J Neurosurg.
1983;59(5):762-774.
35. Servadei F, Murray GD, Penny K, et al. The value of the "worst" computed
tomographic scan in clinical studies of moderate and severe head injury. European
Brain Injury Consortium. Neurosurgery. 2000;46(1):70-75; discussion 75-77.
36. Chesnut RM, Maas AR. Guidelines for the management and prognosis of severe
traumatic brain injury part II: early indicators of prognosis in severe traumatic
brain injury. Neurotrauma. 2000;17:556-627.
37. Knaus WA, Zimmerman JE, Wagner DP, Draper EA, Lawrence DE. APACHE-
acute physiology and chronic health evaluation: a physiologically based classifi-
cation system. Crit Care Med. 1981;9(8):591-597.
38. Baker SP, O’Neill B, Haddon W Jr, Long WB. The injury severity score: a method
for describing patients with multiple injuries and evaluating emergency care.
J Trauma. 1974;14(3):187-196.
39. Teasdale G, Jennett B. Assessment of coma and impaired consciousness.
A practical scale. Lancet. 1974;2(7872):81-84.
40. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of
disease classification system. Crit Care Med. 1985;13(10):818-829.
41. Maas AI, Hukkelhoven CW, Marshall LF, Steyerberg EW. Prediction of outcome
in traumatic brain injury with computed tomographic characteristics: a comparison
between the computed tomographic classification and combinations of computed
tomographic predictors. Neurosurgery. 2005;57(6):1173-1182; discussion 1173-1182.
42. American College of Surgeons Committee on Trauma. Advanced Trauma Life
Support Course for Doctors. Chicago, IL: American College of Surgeons; 1997.
43. Brain Trauma Foundation. Guidelines for the Management and Prognosis of Severe
Traumatic Brain Injury. New York, NY: Brain Trauma Found; 2000.
1104 | VOLUME 70 | NUMBER 5 | MAY 2012
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
44. Richard C. Critical care management of traumatic brain injury. In: Win HR, ed.
Youmans Neurological Surgery. 5ed. Philadelphia, PA: Saunders; 2004:5103-5144.
45. Wilensky EM, Bloom S, Leichter D, et al. Brain tissue oxygen practice guidelines
using the LICOX CMP monitoring system. J Neurosci Nurs. 2005;37(5):278-288.
46. Nortje J, Gupta AK. The role of tissue oxygen monitoring in patients with acute
brain injury. Br J Anaesth. 2006;97(1):95-106.
47. Smith MJ, Stiefel MF, Magge S, et al. Packed red blood cell transfusion increases
local cerebral oxygenation. Crit Care Med. 2005;33(5):1104-1108.
48. Valadka AB, Goodman JC, Gopinath SP, Uzura M, Robertson CS. Comparison of
brain tissue oxygen tension to microdialysis-based measures of cerebral ischemia in
fatally head-injured humans. J Neurotrauma. 1998;15(7):509-519.
49. van den Brink WA, van Santbrink H, Steyerberg EW, et al. Brain oxygen tension
in severe head injury. Neurosurgery. 2000;46(4):868-876; discussion 876-868.
50. van Santbrink H, Maas AI, Avezaat CJ. Continuous monitoring of partial pressure
of brain tissue oxygen in patients with severe head injury. Neurosurgery. 1996;38
(1):21-31.
51. Meixensberger J, Kunze E, Barcsay E, Vaeth A, Roosen K. Clinical cerebral
microdialysis: brain metabolism and brain tissue oxygenation after acute brain
injury. Neurol Res. 2001;23(8):801-806.
52. Valadka AB, Hlatky R, Furuya Y, Robertson CS. Brain tissue PO2: correlation
with cerebral blood flow. Acta Neurochir Suppl. 2002;81:299-301.
53. Maas AI, Fleckenstein W, de Jong DA, van Santbrink H. Monitoring cerebral
oxygenation: experimental studies and preliminary clinical results of continuous
monitoring of cerebrospinal fluid and brain tissue oxygen tension. Acta Neurochir
Suppl (Wien). 1993;59:50-57.
54. Stocchetti N, Pagan F, Calappi E, et al. Inaccurate early assessment of neurological
severity in head injury. J Neurotrauma. 2004;21(9):1131-1140.
55. Sarrafzadeh AS, Peltonen EE, Kaisers U, Kuchler I, Lanksch WR, Unterberg AW.
Secondary insults in severe head injury–do multiply injured patients do worse? Crit
Care Med. 2001;29(6):1116-1123.
56. Lefering R, Paffrath T, Linker R, Bouillon B, Neugebauer EA. Head injury and
outcome–what influence do concomitant injuries have? J Trauma. 2008;65(5):
1036-1043; discussion 1043-1044.
57. Maas AI, Steyerberg EW, Butcher I, et al. Prognostic value of computerized
tomography scan characteristics in traumatic brain injury: results from the
IMPACT study. J Neurotrauma. 2007;24(2):303-314.
58. Tbi Impact. International Mission for Prognosis and Analysis of Clinical Trials in
TBI. Available at: www.tbi-impact.org.
59. Eisenberg HM, Gary HE Jr, Aldrich EF, et al. Initial CT findings in 753 patients
with severe head injury. A report from the NIH Traumatic Coma Data Bank. J
Neurosurg. 1990;73(5):688-698.
60. Greene KA, Jacobowitz R, Marciano FF, Johnson BA, Spetzler RF,
Harrington TR. Impact of traumatic subarachnoid hemorrhage on outcome in
nonpenetrating head injury. Part II: relationship to clinical course and outcome
variables during acute hospitalization. J Trauma. 1996;41(6):964-971.
61. Kakarieka A, Braakman R, Schakel EH. Clinical significance of the finding of
subarachnoid blood on CT scan after head injury. Acta Neurochir (Wien). 1994;
129(1-2):1-5.
62. Servadei F, Murray GD, Teasdale GM, et al. Traumatic subarachnoid
hemorrhage: demographic and clinical study of 750 patients from the European
brain injury consortium survey of head injuries. Neurosurgery. 2002;50(2):
261-267; discussion 267-269.
63. Grmec S, Gasparovic V. Comparison of APACHE II, MEES and Glasgow coma
scale in patients with nontraumatic coma for prediction of mortality. Acute
physiology and chronic health evaluation. Mainz emergency evaluation system.
Crit Care. 2001;5(1):19-23.
64. Oddo M, Nduom E, Frangos S, et al. Acute lung injury is an independent risk
factor for brain hypoxia after severe traumatic brain injury. Neurosurgery. 2010;67
(2):338-344.
65. Chesnut RM, Marshall LF, Klauber MR, et al. The role of secondary brain
injury in determining outcome from severe head injury. J Trauma. 1993;34(2):
216-222.
66. Lannoo E, Van Rietvelde F, Colardyn F, et al. Early predictors of mortality
and morbidity after severe closed head injury. J Neurotrauma. 2000;17(5):
403-414.
67. Vik A, Nag T, Fredriksli OA, et al. Relationship of "dose" of intracranial
hypertension to outcome in severe traumatic brain injury. J Neurosurg. 2008;109
(4):678-684.
www.neurosurgery-online.com

68. Treggiari MM, Schutz N, Yanez ND, Romand JA. Role of intracranial pressure
values and patterns in predicting outcome in traumatic brain injury: a systematic
review. Neurocrit Care. 2007;6(2):104-112.
69. Stein SC, Georgoff P, Meghan S, Mirza KL, El Falaky OM. Relationship of
aggressive monitoring and treatment to improved outcomes in severe traumatic
brain injury. J Neurosurg. 2010;112(5):1105-1112.
70. Stein SC, Georgoff P, Meghan S, Mizra K, Sonnad SS. 150 years of treating severe
traumatic brain injury: a systematic review of progress in mortality. J Neurotrauma.
2010;27(7):1343-1353.
71. Cremer OL, van Dijk GW, van Wensen E, et al. Effect of intracranial pressure
monitoring and targeted intensive care on functional outcome after severe head
injury. Crit Care Med. 2005;33(10):2207-2213.
72. Shafi S, Diaz-Arrastia R, Madden C, Gentilello L. Intracranial pressure monitoring
in brain-injured patients is associated with worsening of survival. J Trauma. 2008;
64(2):335-340.
73. Chang JJ, Youn TS, Benson D, et al. Physiologic and functional outcome
correlates of brain tissue hypoxia in traumatic brain injury. Crit Care Med. 2009;37
(1):283-290.
74. Valadka AB, Gopinath SP, Contant CF, Uzura M, Robertson CS. Relationship of
brain tissue PO2 to outcome after severe head injury. Crit Care Med. 1998;26(9):
1576-1581.
75. Le Roux PD, Newell DW, Lam AM, Grady MS, Winn HR. Cerebral
arteriovenous oxygen difference: a predictor of cerebral infarction and outcome
in patients with severe head injury. J Neurosurgery. 1997;87(1):1-8.
76. Hillered L, Vespa PM, Hovda DA. Translational neurochemical research in acute
human brain injury: the current status and potential future for cerebral
microdialysis. J Neurotrauma. 2005;22(1):3-41.
77. McCarthy MC, Moncrief H, Sands JM, et al. Neurologic outcomes with cerebral
oxygen monitoring in traumatic brain injury. Surgery. 2009;146(4):585-590;
discussion 590-591.
78. Nangunoori RM, Stiefel MD, Park S, et al. Brain tissue oxygen based therapy and
outcome after severe traumatic brain injury: a systematic literature review
[published online ahead of print]. Neurosurgery. 2010.
79. Harvey S, Harrison DA, Singer M, et al. Assessment of the clinical effectiveness of
pulmonary artery catheters in management of patients in intensive care (PAC-
Man): a randomised controlled trial. Lancet. 2005;366(9484):472-477.
80. Richard C, Warszawski J, Anguel N, et al. Early use of the pulmonary
artery catheter and outcomes in patients with shock and acute respiratory
distress syndrome: a randomized controlled trial. JAMA. 2003;290(20):
2713-2720.
81. Sandham JD, Hull RD, Brant RF, et al. A randomized, controlled trial of the use of
pulmonary-artery catheters in high-risk surgical patients. N Engl J Med. 2003;348
(1):5-14.
82. Heuer GG, Smith MJ, Elliott JP, Winn HR, LeRoux PD. Relationship between
intracranial pressure and other clinical variables in patients with aneurysmal
subarachnoid hemorrhage. J Neurosurg. 2004;101(3):408-416.
NEUROSURGERY
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
INITIAL CLINICAL SCALES AFTER SEVERE TBI
83. Clinical Trials Unit. Clinical Randomisation of an Antifibrinolytic in Significant
Haemorrhage. Available at: http://crash2.lshtm.ac.uk/.
COMMENT
T his retrospective analysis of prospectively collected data in 101 patients
with severe traumatic brain injury (TBI) attempts to establish whether
initial radiographic and clinical rating scales are associated with intracranial
pressure (ICP) and brain oxygen levels during the course of hospitalization
in the intensive care unit. Not surprisingly, Acute Physiology and Chronic
Health Evaluation II (APACHE II) score, Rotterdam score, and Marshall
computed tomographic (CT) score were associated with mortality, although
admission Glasgow Coma Score (GCS) was not. However, the authors did
not find a significant correlation between admission GCS, Injury Severity
Score (ISS), Marshall CT score and subsequent intracranial pressure (ICP)
and brain tissue oxygen tension (PbtO2) values. The APACHE II score was
inversely correlated with PbtO2.
This study confirms the difficulty in trying to predict the course of
physiological parameters monitored in the intensive care unit based upon
the clinical and radiological information available on admission. None of
the clinical and radiological scales used was able to predict elevated ICP or
time spent above threshold ICP. As clinicians, we need to recognize that
our ability to foresee which patients will have elevated ICP or poor brain
oxygenation is limited. Because treatment of elevated ICP is a cornerstone
of the management of patients with severe TBI, we should adhere to the
published Guidelines that recommend monitoring all patients with TBI
that present with a GCS of 8 or less.1 Interestingly, the APACHE II score,
which assesses various parameters related to systemic physiology and
premorbid condition, was related to the observed episodes and duration
of brain tissue hypoxia. This suggests that patients with multisystem
injuries that may influence systemic oxygenation should receive special
consideration for brain tissue oxygen monitoring.
Guy Rosenthal
San Francisco, CA
1. Brain Trauma Foundation. American Association of Neurological Surgeons;
Congress of Neurological Surgeons; et al. Guidelines for the management of
severe traumatic brain injury. VI. Indications for intracranial pressure monitoring.
J Neurotrauma. 2007;24(suppl 1):S37-S44.
VOLUME 70 | NUMBER 5 | MAY 2012 | 1105