Vous êtes sur la page 1sur 16

See the corresponding editorial in this issue, pp 1246–1247.

 
J Neurosurg 119:1248–1254, 2013 ©AANS, 2013 

Intracranial  pressure  monitoring  in  severe  head  injury:  compliance 


with  Brain  Trauma  Foundation  guidelines  and effect on outcomes: a 
prospective study 
Clinical article 
P D 
eeP iMiTra 


alving 


KiaDa 

, M.D., , M.D., P 

l .D., yDia e 
fsTaThios l 
aM 

, M.D., K 
araManos h 
owarD 


, elzBerg M.D., P 
, eDro 

M.D., g. K 

enji 
eixeira i 
naBa 

, M.D., 
, M.D., anD 

eMeTrios 


eMeTriaDes 

, M.D., P 

.D. 
Division of Acute Care Surgery (Trauma, Emergency Surgery and Surgical Critical Care), Department of Surgery, Keck School 
of Medicine, Los Angeles County + University of Southern California Medical Center, Los Angeles, California 
Object.  The  Brain  Trauma  Foundation  (BTF)  has  established  guidelines  for  intracranial  pressure  (ICP)  monitor-  ing  in 
severe traumatic brain injury (TBI). This study assessed compliance with these guidelines and the effect on outcomes. 
Methods.  This  is  a  prospective,  observational  study  including  patients  with severe blunt TBI (Glasgow Coma Scale score ≤ 
8,  head  Abbreviated  Injury  Scale  score  ≥  3)  between  January  2010  and  December  2011. Demographics, clinical characteristics, 
laboratory  profile,  head  CT  scans,  injury  severity  indices,  and  interventions  were  collected.  The  study  population was stratified 
into  2  study  groups:  ICP  monitoring  and  no  ICP  monitoring.  Primary  outcomes  included  compliance  with  BTF  guidelines, 
overall  in-hospital  mortality,  and  mortality  due  to  brain herniation. Sec- ondary outcomes were ICU and hospital lengths of stay. 
Multiple regression analyses were deployed to determine the effect of ICP monitoring on outcomes. 
Results.  A  total  of  216  patients  met  the  BTF  guideline  criteria  for  ICP  monitoring. Compliance with BTF guide- lines was 
46.8%  (101  patients).  Patients  with  subarachnoid  hemorrhage  and  those  who  underwent  craniectomy/  craniotomy  were 
significantly  more  likely  to  undergo  ICP  monitoring.  Hypotension,  coagulopathy, and increasing age were negatively associated 
with  the  placement  of  ICP  monitoring  devices.  The  overall  in-hospital mortality was significantly higher in patients who did not 
undergo  ICP  monitoring  (53.9%  vs  32.7%,  adjusted  p  =  0.019).  Similarly,  mortality  due  to  brain  herniation  was  significantly 
higher  for  the  group  not  undergoing  ICP monitoring (21.7% vs 12.9%, adjusted p = 0.046). The ICU and hospital lengths of stay 
were significantly longer in patients subjected to ICP monitoring. 
Conclusions. Compliance with BTF ICP monitoring guidelines in our study sample was 46.8%. Patients man- aged 
according to the BTF ICP guidelines experienced significantly improved survival. 
(http://thejns.org/doi/abs/10.3171/2013.7.JNS122255) 

ey 

orDs 
• intracranial pressure monitoring • mortality • brain herniation • Brain Trauma Foundation guidelines • 
traumatic brain injury T 
RAUMATIC tality settings.7,14,22 and brain morbidity The injury incidence both is the in and major 
civilian risk cause of and intracranial military of mor- 
Abbreviations  used  in  this  paper:  AIS  =  Abbreviated  Injury  Scale;  AOR  =  adjusted  odds  ratio;  AUC  =  area under the curve; 
BTF  =  Brain  Trauma  Foundation;  CPP  =  cerebral  perfusion  pres-  sure;  GCS  =  Glasgow  Coma  Scale;  ICH  =  intracranial 
hemorrhage; ICP = intracranial pressure; INR = international normalized ratio; IPH = intraparenchymal hemorrhage; ISS = Injury 
Severity  Score;  LOS  =  length  of  stay;  NTDB  =  National  Trauma  Data  Bank;  PT  =  prothrombin  time;  PTT  =  partial 
thromboplastin  time;  SAH  =  subarachnoid  hemorrhage;  SBP  =  systolic  blood  pressure;  SDH  =  subdural  hematoma;  TBI  = 
traumatic brain injury. 

1248  J  Neurosurg  /  Volume  119  /  November  2013  hypertension after severe traumatic brain injury (TBI) 


have  been  previously  documented.11–13,15,18,22  Thus,  ac-  cording  to  the  Brain  Trauma  Foundation 
(BTF)  guide-  lines,  intracranial  pressure  (ICP)  monitoring  is  consid-  ered  to  be  the  standard  of  care  in 
comatose  patients  who  have  sustained severe head injury.1 Previous studies, however, have demonstrated 
inconsistent  compliance  with  BTF  guidelines  for  ICP  monitoring.6,20  The  reason  for  the  marked 
inconsistency  with  BTF  guidelines  relates  to  conflicting  clinical  outcomes  and  a  lack  of  randomized 
controlled trials. 
We  set  out  to  prospectively  investigate  compliance  with  the  BTF  guidelines  for  ICP  monitoring  in a 
large urban trauma center. We hypothesized that compliance 
 
Intracranial pressure monitoring in severe blunt TBI 
with BTF ICP monitoring guidelines is associated with 
ing, a forward stepwise logistic regression 
was deployed improved outcomes. 
using variables at a p < 0.2 level after univariate analysis. Only the statistically significant variables are 
reported. Methods 
Subsequent univariate analyses for in-hospital mortality and mortality due to brain herniation were 
performed. After obtaining approval from the institutional re- view board, we conducted a prospective 
observational study of trauma patients with severe blunt TBI (Glasgow Coma Scale [GCS] score ≤ 8 and 
head Abbreviated Injury Scale [AIS] score ≥ 3) who met the BTF inclusion crite- ria for ICP monitoring 
and were admitted to the surgical ICU at Los Angeles County and University of Southern California 
Medical Center between January 01, 2010, and December 30, 2011. The decision to place an ICP moni- 
toring device was at the neurosurgeon’s discretion. Exclu- sion criteria included the pediatric population 
(age < 18 years), patients who were moribund, and those who were not expected to improve prior to the 
decision of whether an ICP monitoring device would be placed. Placement of ICP monitoring devices 
occurred in all instances in the first 24 hours after admission. 
Demographic and clinical data collected included 
Independent  predictors  of  in-hospital  mortality  and  mor-  tality  due  to  brain herniation were derived from 
forward  stepwise  regression  models  using  variables  from  each univariate model that were different at p < 
0.2.  A  variance  inflation  factor  ≥  5  for  each variable entered in the models was considered as evidence of 
multicolinearity. 
To correct for the differences between the groups (ICP monitoring vs no ICP monitoring), logistic regres- 
sion was performed using as independent variables the placement of the ICP monitoring device, adjusting 
for dif- ferences between the groups at p < 0.05. The regression calculated the predicted probability 
(propensity score) of being subjected to ICP monitoring using variables that independently predicted the 
placement of an ICP moni- toring device. Propensity score matching is a technique that tries to estimate 
the effect of a treatment by account- ing for the covariates that predict receiving the treatment. age, sex, 
blood pressure on admission, GCS score on ad- 
It was performed in an effort to reduce the 
bias due to mission, Injury Severity Score (ISS), AIS for each body 
confounding factors that might be involved in 
compar- region (head, chest, abdomen, and extremity), type of in- 
ing outcomes among patients who received 
the treatment tracranial injury, ICP values in patients undergoing in- 
with outcomes of those who did not. The 
overall in-hos- tervention, intracranial hemorrhage (ICH) treatment mo- 
pital mortality and mortality due to brain 
herniation were dalities, and neurosurgical documentation for omission of 
assessed for each study group using logistic 
regression to ICP monitoring. A standardized data collection sheet was 
adjust for factors that were significant at p < 
0.05. The used, allowing the treating physicians to choose from a 
propensity score was also inserted in the 
regression as variety of standardized and clinically pertinent choices 
a covariate. Adjusted odds ratios (AORs) with 
95% CIs as to why an ICP monitoring device was not placed. In 
were derived from the logistic regression. An 
AOR < 1.00 case the predetermined choices were not sufficient, the 
(95% CI) implies that the factor decreases 
significantly physician could write down the reason for not placing an ICP monitoring device. 
The  study  population  was  stratified  into  2 study arms: patients subjected to ICP monitoring and those 
not  under-  going  ICP  monitoring.  All  subsequent  analyses  were  per-  formed  comparing  these  groups. 
Elevated  ICP  was  defined  as  higher  than  20  mm  Hg  for  more  than  15  minutes  based  on  the  BTF 
guidelines. 
Primary  outcomes  included  compliance  with  BTF  guidelines,  overall  in-hospital  mortality,  and 
mortality due to brain herniation. Secondary outcomes were ICU and hospital lengths of stay (LOSs). 
the  odds  of  developing the outcome, whereas an AOR > 1.00 (95% CI) increases significantly the odds of 
develop-  ing  the  outcome.  A  confidence  interval  that  crosses  1.00  implies that the factor does not predict 
the  outcome.  The hospital and ICU LOSs and ventilator days were com- pared using an independent t-test 
or  Mann-Whitney  U-  test  and  subsequently  linear  regressions  adjusting  for  dif-  ferences  that  were 
significant  at  p  <  0.05.  To  correct  for  mortality  bias,  the  same  tests  were  deployed  after  exclu-  sion  of 
deaths. 
Values are reported as the mean ± SEM for continu- ous variables and as percentages for categorical 
variables. 
Statistical Analysis 
All analyses were performed using SPSS for Windows (version 12.0, SPSS, Inc.). Continuous variables 
were dichotomized using clini- cally relevant cut points: age (≤ 55 years vs > 55 years), systolic blood 
pressure (SBP) on admission (< 90 mm Hg vs ≥ 90 mm Hg), international normalized ratio (INR, < 1.3 vs 
≥ 1.3), ISS (≤ 15, 16–24, ≥ 25), AIS score (≥ 3 vs < 3), and heart rate on admission (> 120 bpm vs ≤ 120 
bpm). The 2 groups were compared for differences in categori- cal variables using the Fisher exact or 
Pearson chi-square 
Results Overall, 216 patients who sustained a severe TBI met the BTF guidelines for ICP monitoring. The 
epidemio- logical and clinical characteristics of the study population are shown in Table 1. Hypotension 
was present in 5.6% of the patients while 29.2% were tachycardic on admission. tests as appropriate. The 
Shapiro-Wilk Test for normal- 
Almost half of the patients (43.5%) had a 
GCS score of 3 ity was deployed for continuous variables; normally dis- 
on admission and 44.4% had a head AIS score 
of 5. tributed variables were compared using the Student t-test 
A total of 46.8% of patients who met the 
BTF criteria while nonnormally distributed variables were compared 
underwent ICP monitoring (n = 101). A 
Becker ventricu- using the Mann-Whitney U-test. 
lostomy EMDS II (Medtronic Corp.) was placed 
in 60 pa- To identify independent predictors of ICP monitor- 
tients (59%), and a transducer-tipped pressure/temperature 
J Neurosurg / Volume 119 / November 2013 
1249 
 
P. Talving et al. 
TABLE 1: Univariate analysis of the clinical characteristics of patients meeting the BTF guidelines* 
Characteristic ICP Monitoring (n = 101) No ICP Monitoring (n = 115) p Value 
demographics 
mean age in yrs 40.1 ± 1.9 48.0 ± 2.4 0.011 age >55 yrs 27 (26.7%) 46 (40.0%) 0.044 male 80 (79.2%) 81 (70.4%) 0.160 
admission physiology 
mean SBP in mm Hg 142 ± 3 137 ± 3 0.267 hypotension (SBP <90 mm Hg) 2 (2.0%) 10 (8.7%) 0.040 mean heart rate 104 ± 3 
105 ± 3 0.811 tachycardia (heart rate >120 bpm) 26 (25.7%) 37 (32.2%) 0.370 mean respiratory rate 18 ± 1 18 ± 1 1.000 injury 
severity indices 
mean ISS 25 ± 1 25 ± 1 1.000 ISS ≤15 13 (12.9%) 19 (16.5%) 0.565 ISS 16–24 32 (31.7%) 27 (23.5%) 0.221 ISS ≥25 56 
(55.4%) 69 (60.0%) 0.581 head AIS Score 3 42 (41.6%) 35 (30.4%) 0.153 head AIS Score 4 22 (21.8%) 21 (18.3%) 0.609 head 
AIS Score 5 37 (36.6%) 59 (51.3%) 0.039 median GCS score 4 (3–8) 4 (3–8) 0.762 GCS Score 3 39 (38.6%) 55 (47.8%) 0.216 
chest AIS ≥3 38 (37.6%) 36 (31.3%) 0.389 abdomen AIS ≥3 11 (10.9%) 8 (7.0%) 0.343 extremity AIS ≥3 18 (17.8%) 9 (7.8%) 
0.038 specific head injuries 
brain contusion 75 (74.3%) 82 (71.3%) 0.649 SDH 55 (54.5%) 59 (51.3%) 0.683 SAH 59 (58.4%) 40 (34.8%) 0.001 IPH 37 
(36.6%) 22 (19.1%) 0.006 epidural hematoma 18 (17.8%) 17 (14.8%) 0.582 midline shift 6 (5.9%) 8 (7.0%) 0.790 loss of basal 
cisterns 31 (30.7%) 39 (33.9%) 0.663 cerebral edema 74 (73.3%) 91 (79.1%) 0.338 loss of gray/white differential 35 (34.7%) 28 
(24.3%) 0.102 fixed, dilated pupils on admission 23 (22.8%) 35 (30.4%) 0.221 admission laboratory values 
mean PTT 30.4 ± 0.7 40.3 ± 3.7 0.010 mean PT 15.7 ± 0.3 16.9 ± 0.8 0.082 mean INR 1.21 ± 0.03 1.40 ± 0.09 0.058 INR ≥1.3 
28 (27.7%) 41 (35.7%) 0.243 early nutrition 85 (84.2%) 66 (57.4%) <0.001 decompressive craniectomy/craniotomy in 1st 24 hrs 
42 (41.6%) 18 (15.7%) <0.001 decompressive craniectomy/craniotomy in 1st 4 hrs 32 (31.7%) 17 (14.8%) 0.003 mean 
probability of receiving ICP monitoring 0.64 ± 0.02 0.34 ± 0.02 <0.001 
* Mean values are presented as the mean ± SEM. Median values are presented as the median (range). All other non–p values are 
the number of patients (%). PT = prothrombin time. 

fiberoptic  catheter  Camino  Advanced  Monitor  (bolt)  (In-  tegra  LifeSciences  Corp.)  was  placed  in  41 
patients  (41%).  All  patients  subjected  to  fiberoptic  monitor  placement  re-  mained  with  the  fiberoptic 
monitor the entire critical phase 
1250  J  Neurosurg  /  Volume  119  /  November  2013  of  care.  However,  58%  of  patients  subjected  to 
ventriculos-  tomy  were  changed  subsequently  to  a  fiberoptic  monitoring  device.  The  most  common 
reason for not placing an ICP monitoring device was the treating physician’s decision 
 
Intracranial pressure monitoring in severe blunt TBI 
TABLE 2: Independent predictors of ICP monitoring* 
Step of Forward Logistic 
Regression Analysis Variable Cumulative R2 AOR (95% CI) Adjusted p Value 
1  decompressive  craniectomy  ≤4  hrs  0.113  3.85  (1.82–8.14)  <0.001  2  extremity  AIS  score  ≥3  0.171  3.01  (1.09–8.32)  0.033  3 
increasing  age  0.213  0.97  (0.96–0.99)  0.001  4 increasing PTT on admission 0.256 0.96 (0.92–0.99) 0.021 5 best GCS score w/in 
24  hrs  of  admission  0.298  0.81  (0.71–0.93)  0.002  6  hypotension  on  admission  0.329  0.13  (0.02–0.86)  0.034  7 SAH 0.353 2.07 
(1.09–8.32) 0.033 
*  Other  variables  entered  in  the  model  were  sex,  presence  of  IPH  on  CT,  loss  of gray and white differential on CT, presence of 
reactive  pupils  during  the  first  physical  examination,  PT  and  INR  values  on  admission,  initiation of nutrition in the first 7 days, 
head  AIS  score  of  3,  head  AIS  score  of  5,  and  evacuation  of  a  mass  lesion  within  the  first  24  hours.  AUC  0.807  (95%  CI 
0.746–0.867), p < 0.001. 

(89.6%), followed by decompressive surgery (13.9%) and 


ICP was 5.9 ± 1.1 mm Hg (range 0–99 mm 
Hg). Patients expectation of a rapid improvement (4.3%). Patients who 
undergoing ICP monitoring who experienced 
episodes of sustained a subarachnoid hemorrhage (SAH) or an intrapa- 
sustained elevated ICP were treated in all 
instances (n = renchymal hemorrhage (IPH) were more likely to receive 
64). Elevation of the head of the bed was used 
in 90.6% an ICP monitoring device. Patients with SAH who did not 
of the cases, sedation in 89.1%, furosemide 
injection in receive ICP monitoring accounted for 34.8% of the study 
73.4%, pentobarbital coma in 23.4%, 
hypertonic saline population, whereas patients with SAH who received ICP 
infusion in 62.5%, mannitol bolus in 54.7%, 
hyperven- monitoring accounted for 58.4% (p = 0.001). Patients with 
tilation in 51.6%, decompressive craniectomy 
in 15.6%, an IPH who did not receive ICP monitoring accounted for 
hypothermia in 14.1%, and paralysis in 3.1% 
of cases. 19.1% of the population, whereas patients with an IPH who 
Intracranial pressure monitoring was 
noted to be an received ICP monitoring accounted for 36.6% of the entire 
independent predictor of overall in-hospital 
mortality study sample (p = 0.006). No other differences in the inci- 
(AOR 0.13 [95% CI 0.02–0.81], adjusted p = 
0.029), along dence of type of intracranial lesions were noted between 
with a head AIS score of 5 (AOR 79.81 [95% 
CI 12.95– the groups (Table 1). 
91.75]), early initiation of nutrition (AOR 0.27 
[95% CI The independent predictors for ICP monitoring de- 
0.01–0.43]), older age, best GCS score within 
the first 24 vice placement are summarized in Table 2. Patients un- 
hours, and presence of SDH on CT. 
Intracranial pressure dergoing decompressive craniectomy or patients with an 
monitoring was also found to be an 
independent predic- extremity AIS score ≥ 3 were more likely to be subjected 
tor of mortality due to brain herniation with 
an AOR of to ICP monitoring (AOR 3.85 [95% CI 1.82–8.14] adjust- 
0.31 (95% CI 0.10–0.93, adjusted p = 0.037), 
along with ed p < 0.001 and AOR 3.01 [95% CI 1.09–8.32] adjusted p 
a head AIS score of 5 (AOR 35.54 [95% CI 
8.22–53.75], = 0.033, respectively). Patients who were older (AOR 0.97 
adjusted p < 0.001), loss of basal cisterns on 
CT (AOR [95% CI 0.96–0.99], adjusted p = 0.001), those with an in- 
5.01 [95% CI 1.82–13.82]), best GCS score 
within the first creased partial thromboplastin time (PTT) on admission 
24 hours, and presence of IPH. The R2 and 
the area under (AOR 0.96 [95% CI 0.92–0.99], adjusted p = 0.021), those 
the curve (AUC) for the model were 0.881 
and 0.989 for with a higher GCS score within first 24 hours of admis- 
overall mortality and 0.474 and 0.887 for 
mortality due to sion (AOR 0.81 [95% CI 0.71–0.93], adjusted p = 0.002), 
brain herniation, respectively (Table 4). or 
those who were hypotensive on admission (AOR 0.13 
The overall complication rate throughout 
hospitaliza- [95% CI 0.02–0.86], adjusted p = 0.034) were significant- 
tion was 5%. The incidence of pneumonia, 
acute kidney ly less likely to undergo ICP monitoring. We noted no dif- 
injury, and deep venous 
thrombosis/pulmonary embolism ference in mortality for the ICP monitoring group when 
was 3.7% (8 of 216), 0.9% (2 of 216), and 
0.9% (2 of 216), the 2 devices were compared (Table 3). 
respectively; all of these complications occurred 
in the For the patients subjected to ICP monitoring, the 
ICP study group. No incidents of acute 
respiratory dis- highest mean ICP noted in the study sample was 33.4 ± 
tress syndrome and/or septic shock were 
noted. 2.3 mm Hg (± SEM, range 9–118 mm Hg) and the lowest 
The overall in-hospital mortality was significantly 
TABLE 3: Impact of type of ICP monitoring on outcomes for patients receiving an ICP monitoring device (n = 101) 
Ventriculostomy 
Fiberoptic Monitor Parameter 
(n = 60) 
(n = 41) OR (95% CI) p Value AOR (95% CI)* 
1251 Adjusted p Value* 
overall in-hospital mortality 22 (36.7%) 11 (26.8%) 1.58 (0.66–3.76) 0.388 0.65 (0.10–4.32) 0.651 mortality due to brain 
herniation 8 (13.3%) 5 (12.2%) 1.11 (0.34–3.66) 1.000 0.74 (0.17–3.24) 0.693 
* Controlled for age, sex, IPH, SAH, PTT on admission, and presence of reactive pupils on admission. 
J Neurosurg / Volume 119 / November 2013 
 
P. Talving et al. 
TABLE 4: Independent predictors of overall in-hospital mortality and mortality due to brain herniation* 
Step of Forward Logistic 
Regression Analysis Variable Cumulative R2 AOR (95% CI) Adjusted p Value 
overall in-hospital mortality† 
1 head AIS Score 5 0.788 79.81 (12.95–91.75) <0.001 2 early nutrition 0.819 0.27 (0.01–0.43) 0.005 3 increasing age 0.842 1.07 
(1.02–1.12) 0.005 4 best GCS score w/in 24 hrs of admission 0.86 0.51 (0.31–0.85) 0.009 5 SDH 0.871 10.70 (1.69–67.81) 0.012 
6 ICP monitoring 0.881 0.13 (0.02–0.81) 0.029 mortality due to brain herniation‡ 
1  head  AIS  Score  5  0.284  35.54  (8.22–53.75)  <0.001  2  loss  of  basal  cisterns  0.376  5.01  (1.82–13.82)  0.002  3  best  GCS  score 
w/in  24  hrs  of  admission  0.423  0.73  (0.55–0.96) 0.023 4 presence of IPH 0.441 5.16 (1.70–15.65) 0.004 5 ICP monitoring 0.474 
0.31 (0.10–0.93) 0.037 
* Other variables entered in the models were age; sex; hypotension on admission; tachycardia on admission; extremity AIS score 
≥ 3; chest AIS score ≥ 3; ISS; other intracranial lesions on CT; admission PT, PTT and INR values; alcohol intoxication; head 
AIS score of 3; head AIS score of 4; GCS on admission of 3; and fixed dilated pupils on admission. † Overall mortality: AUC 
0.989 (95% CI 0.978–1.000), p < 0.001. ‡ Mortality due to brain herniation: AUC 0.887 (95% CI 0.841–0.933), p < 0.001. 

higher  in  patients  not  subjected  to  ICP  monitoring  after  ad-  justing  for  relevant confounders between the 
groups  (age,  presence  of  hypotension  on  admission,  head  AIS  score  of  5,  extremity  AIS  score  ≥  3, 
presence  of  IPH  or  SAH  on  CT,  PTT  on  admission,  early  nutrition,  decompressive  craniec- 
tomy/craniotomy  in  ≤  4  hours,  decompressive  craniectomy/  craniotomy  in  ≤  24  hours,  and probability of 
receiving  ICP  monitoring;  53.9%  vs  32.7%,  adjusted  p  =  0.019).  Simi-  larly,  mortality  due  to  brain 
herniation  was significantly higher in patients not subjected to ICP monitoring (21.7% vs 12.9% (adjusted 
p  =  0.046  [adjusting  for  the  cofounders  mentioned  above]).  Early  deaths  (<  48  hours)  were  equally 
distributed among the study groups. 
The  hospital  and  ICU  LOSs  were  significantly  longer  for  the  group  subjected  to  ICP  monitoring 
before and af- ter exclusion of deaths (Table 5). 
Discussion The BTF guidelines support ICP monitoring in all salvageable patients with 
severe TBI (GCS score of 3–8 after resuscitation) with an abnormal CT scan depicting ICH, brain edema, 
herniation, or compressed basal cis- terns (Level II evidence). In addition, patients with a GCS score of 
3–8 with no CT-identified lesion featuring 2 of 3 variables (unilateral/bilateral posturing, patient age ≥ 40 
years, and presence of hypotension) are also candidates for ICP monitoring (Level III evidence1). In brain 
injuries that are likely to require aggressive ICP management, a ventriculostomy is placed, which can also 
provide CSF drainage. Patients not likely to require CSF drainage are subjected to placement of a 
fiberoptic ICP monitoring de- vice, allowing continuous ICP monitoring. 
The BTF guidelines are based on retrospective and limited prospective observational data; thus, the 
ICP com- 
1252  J  Neurosurg / Volume 119 / November 2013 pliance varies widely in neurocritical care, and the out- 
comes  related  to  the  intervention  are  conflicting.2–4,8,10,20,21  The  BTF  provides  guidelines  for 
physicians  with  regard  to  the  type  of  patients  who  should  receive  ICP  monitor-  ing.  It  is  worth  noting, 
however,  that  these  guidelines  do  not  constitute  a  universal  protocol  and  thus  treating  physi-  cians 
commonly  use  their  own  experience  and  judgment  to  decide  which  patient  will  be  subjected  to  ICP 
monitoring.  Previous  surveys  of  ICP  monitoring  in  Europe  and  North  America  have  noted  utilization  of 
ICP  monitoring  in  50%–  75%  of  patients  with  severe  head  injury  in  institutions pro- viding neurocritical 
care.6,9,17,19,20,23 
Cremer  et  al.4  performed  a  retrospective  cohort  study  with  prospective  follow-up  (≥ 12 months) in 2 
trauma  centers  providing  supportive  therapy  only  compared  with  ICP/CPP-targeted  management  in  the 
severe  head  injury  population.  The  ICP/CPP-guided  therapy did not improve in-hospital survival, and the 
follow-up  Glasgow  Outcome  Scale  score  was  similar  in  both  study  groups.  At  the  cen-  ter  relying  on 
ICP/CPP-targeted  therapy,  the  prevalence  of  ICP  monitoring  was  67%.4  Shafi  et  al.  performed  a  Na- 
tional  Trauma  Data  Bank  (NTDB)  analysis and noted that ICP monitoring was applied in 43% of patients 
who  met  BTF  criteria.20  Likewise,  we  observed  the  prevalence  of  ICP  monitoring  in  patients  meeting 
BTF  guideline  cri-  teria  at  46.8%.  In  our  prospective  study  we  documented  reasons  for  omission  of  ICP 
monitoring,  which  included  neurosurgeon’s  discretion,  decompressive  craniotomy/  craniectomy 
precluding  the  need  for  an  ICP  device, and expectation of rapid neurological recovery. The indepen- dent 
predictors  of  ICP  monitoring  included  SAH,  early  decompressive  craniectomy,  and  severe  extremity 
injury  with  AORs  of  2.07,  3.85,  and  3.01,  respectively  (Table  2).  Patients  with  increasing  age,  those 
experiencing coagu- lopathy, those with an elevated GCS score, or those pre- 
 
Intracranial pressure monitoring in severe blunt TBI 
TABLE 5: Outcome measures 
ICP Monitoring 
No ICP Monitoring Variable 
(n = 101) 
(n = 115) OR (95% CI) p Value AOR (95% CI)* 
1253 Adjusted p Value* 
overall in-hospital mortality 33 (32.7%) 62 (53.9%) 0.42 (0.24 to 0.72) 0.002 0.15 (0.03 to 0.74) 0.019 mortality due to brain 
herniation 13 (12.9%) 25 (21.7%) 0.53 (0.26 to 1.11) 0.107 0.34 (0.10 to 0.87) 0.046 
Variable 
ICP Monitoring (n = 101) 
No ICP Monitoring (n = 115) 
Mean Difference 
(95% CI) p Value 
Adjusted Mean Difference (95% CI)* 
Adjusted p Value* 
mean  ICU LOS in days 16.8 ± 1.3 8.2 ± 1.0 −8.62 (−11.83 to −5.41) <0.001 −6.04 (−9.46 to −1.69) <0.001 mean hospital LOS in 
days 19.4 ± 1.6 10.1 ± 1.2 −9.26 (−13.10 to −5.42) <0.001 −7.14 (−11.14 to −2.08) 0.001 after exclusion of deaths 
mean ICU LOS in days 21.7 ± 1.5 13.0 ± 1.8 −8.74 (−13.36 to −4.13) <0.001 −2.78 (−11.58 to −2.78) 0.007 mean hospital LOS 
in days 25.8 ± 1.8 18.2 ± 2.1 −7.61 (−13.13 to −2.09) <0.001 −2.34 (−9.56 to −2.34) 0.022 
*  Controlled  for  age,  presence  of  hypotension  on  admission,  head  AIS  of  5, extremity AIS ≥ 3, presence of IPH or SAH on CT, 
PTT  on  admission,  early  nutrition,  decompressive  craniectomy/craniotomy  in  ≤  4  hours  and  decompressive 
craniectomy/craniotomy in ≤ 24 hours, and probability of receiving ICP monitoring. 

senting  with  hypotension  on  admission  were  less  likely  to  be  subjected  to  ICP  monitoring  (AOR  0.97, 
0.96,  0.81,  and  0.13,  respectively).  The  univariate  analysis  compar-  ing  the  patients  who  received  ICP 
monitoring  with  those  who  did  not  revealed  some  differences  in  their  basic  characteristics,  which, 
however,  did  not  impact  outcomes.  Even though patients with hypotension were significantly more likely 
not  to  receive  an  ICP  monitoring  device,  the  mean  SBP  did  not  differ  between  the  groups  (142  ±  3  mm 
Hg  vs  137  ±  3  mm  Hg,  p  =  0.267).  In  addition,  injury  severity  indices  were  not  statistically  different 
between  the  groups.  Likewise,  the  severity  of  head  injury based on head AIS and GCS score was equally 
distributed  between  the  groups.  When  specific  injury  patterns  of  TBI  were  analyzed, again, there was no 
difference  except  in  the  in-  cidence  of  SAH and IPH. Finally, patients with admission coagulopathy were 
less  likely  to  undergo  ICP  monitor-  ing  placement,  not  because  of  the  severity  of  injury  but  to  avoid 
potential  bleeding  due  to  the  intervention.  The  main  reason  for  not  receiving  an  ICP  monitoring  device 
was  at  the  discretion  of  the  neurosurgical  attending.  Pa-  tients  who  were  moribund  or  were  assessed  to 
have  poor  outcome  on  admission  were  excluded  from  the  study  per  study  design.  In  addition,  a  small 
percentage  of patients who did not receive ICP monitoring were expected by the neurosurgeon to improve 
rapidly.  That  contrasts  with  the  idea  that  ICP  monitoring  was  not  used  in  sicker patients and emphasizes 
the  fact  that  BTF  guidelines  are  not  uni-  versally  applied  and  each  patient is treated based on phy- sician 
preference.  In  41%  of  all  ICP  devices  a  fiberoptic  ICP  monitoring  device,  that  is,  a  fiberoptic  monitor, 
was  placed  with  no  difference  in  overall mortality and mortal- ity due to brain herniation when compared 
with ventricu- lostomy (Table 3). 
We  performed  a  regression  analysis  to  elucidate  inde-  pendent  predictors  for  both  overall  and  head 
injury–related  outcomes.  The  deployed  model  noted  multiple  sig  nificant  independent  predictors  for 
overall  in-hospital  mor  tality  including  devastating  head  injury  (head  AIS  score  of  5)  and  subdural 
hematoma (SDH) with AORs for poor out- 
come  of  79.81  and  10.70,  respectively  (Table  4).  We  noted  that  patients  subjected  to  ICP  monitoring 
experienced  a  significant  overall  mortality  reduction  effect  of  69%  after  adjustment  for  most  clinically 
relevant  confounders  includ-  ing  head  injury  severity  (Table  5).  Similar  to  our  findings,  previous 
examinations  have  observed survival advantage in ICP-targeted therapy, but with less marked survival ef- 
fect.3,5,15,16  However,  a  recent  NTDB-based  investigation  by  Shafi  et  al.  observed  a  significantly 
worsened  survival  in  patients  subjected  to  ICP  monitoring.20  These  authors  in-  terpreted  the  BTF 
guidelines  to  be  inadequate  to  identify  patients  who  benefit  from  ICP  monitoring.  Furthermore,  these 
investigators  suggested  that  the  interventions  in  pa-  tients subjected to ICP monitoring may be associated 
with  worsened  outcomes. Such interventions include ICP moni- toring device insertion in a coagulopathic 
state,  utilization  of  vasoactive  medication,  osmotic  diuresis  with  mannitol,  furosemide  utilization, 
hyperventilation,  and  use of para- lytics. We documented all of the above interventions in our study group 
subjected  to  ICP  monitoring;  nevertheless,  de-  spite  all  these  potentially  harmful  interventions,  these pa- 
tients experienced significantly improved outcomes. 
We  noted  extended  ICU  and  hospital  LOSs  when  ICP  monitoring  was  instituted  before  and  after 
exclusion of deaths. This finding may relate to improved survival in patients subjected to the intervention. 
Our  study  has  multiple  limitations  including  the  lack  of  randomization,  long-term  follow-up  of 
outcome,  and  functional  impairment  measures.  First,  we  attempted  to  compensate  for  the  lack  of 
randomization  with  multiple  regression  models  adjusting  for  an  extensive  number  of  clinically  relevant 
confounders  introducing  propensity  score  into  our  analysis.  Second,  the  patients  were  selected  for  all 
interventions  at  the  discretion  of  the  attending  neu-  rosurgeon  without  any  strict  protocol.  Finally,  it  is 
likely  that  some  of  those  patients  in  the  group  not  receiving  ICP  monitoring  were  being  treated 
differently—and  perhaps less intensively—than those in the ICP monitoring group. Nevertheless, our data 
were collected prospectively and our study is, to the best of our knowledge, the first to 
J Neurosurg / Volume 119 / November 2013 
 
P. Talving et al. 
document  reasons  for  omission  of  ICP  monitoring  with  respective  outcomes.  Our  findings  support  the 
need for prospective randomized trials to settle the debate. 
10. Lane PL, Skoretz TG, Doig G, Girotti MJ: Intracranial pressure monitoring and outcomes after traumatic brain injury. Can J 
Surg 43:442–448, 2000 11. Lee TT, Galarza M, Villanueva PA: Diffuse axonal injury (DAI) is not associated with elevated 
intracranial pressure (ICP). Acta Conclusions Compliance with the BTF ICP monitoring guidelines in our study sample was 
46.8%. Compliance with these guidelines was associated with improved survival. 
Neurochir (Wien) 140:41–46, 1998 12. Lobato RD, Sarabia R, Rivas JJ, Cordobes F, Castro S, Muñoz MJ, et al: Normal 
computerized tomography scans in severe head injury. Prognostic and clinical management implica- tions. J Neurosurg 
65:784–789, 1986 13. Miller MT, Pasquale M, Kurek S, White J, Martin P, Bannon K, et al: Initial head computed tomographic 
scan characteris- Disclosure 
tics have a linear relationship with initial intracranial pressure 
The  authors  report no conflict of interest concerning the mate- rials or methods used in this study or the findings specified in 
this paper. 
Author  contributions  to  the  study  and  manuscript  preparation  include  the  following.  Conception  and  design:  Talving, 
Demetriades.  Acquisition  of  data:  Karamanos,  Skiada.  Analysis  and  interpretation  of  data:  Karamanos,  Inaba.  Drafting  the 
article:  Talving,  Karamanos,  Lam.  Critically  revising  the  article:  all  authors.  Reviewed  submitted  version  of  manuscript:  all 
authors.  Approved  the  final version of the manuscript on behalf of all authors: Talving. Statistical analysis: Karamanos, Teixeira. 
Study supervision: Talving, Demetriades. 
after trauma. J Trauma 56:967–973, 2004 14. Myburgh JA, Cooper DJ, Finfer SR, Venkatesh B, Jones D, Higgins A, et al: 
Epidemiology and 12-month outcomes from traumatic brain injury in Australia and New Zealand. J Trau- ma 64:854–862, 2008 
15. Narayan RK, Kishore PR, Becker DP, Ward JD, Enas GG, Greenberg RP, et al: Intracranial pressure: to monitor or not to 
monitor? A review of our experience with severe head injury. J Neurosurg 56:650–659, 1982 16. Palmer S, Bader MK, Qureshi 
A, Palmer J, Shaver T, Borzatta M, et al: The impact on outcomes in a community hospital setting of using the AANS traumatic 
brain injury guidelines. J Trauma 50:657–664, 2001 References 
1. Bratton SL, Chestnut RM, Ghajar J, McConnell Hammond FF, Harris OA, Hartl R, et al: Guidelines for the management of 
severe traumatic brain injury. VIII. Intracranial pressure thresholds. J Neurotrauma 24 (Suppl 1):S55–S58, 2007 (Er- ratum in J 
Neurotrauma 25:276–278, 2008) 2. Bulger EM, Nathens AB, Rivara FP, Moore M, MacKenzie EJ, Jurkovich GJ: Management 
of severe head injury: insti- tutional variations in care and effect on outcome. Crit Care Med 30:1870–1876, 2002 3. Chesnut RM, 
Temkin N, Carney N, Dikmen S, Rondina C, Videtta W, et al: A trial of intracranial-pressure monitoring in traumatic brain 
injury. N Engl J Med 367:2471–2481, 2012 4. Cremer OL, van Dijk GW, van Wensen E, Brekelmans GJ, 
17. Patel HC, Menon DK, Tebbs S, Hawker R, Hutchinson PJ, Kirkpatrick PJ: Specialist neurocritical care and outcome from 
head injury. Intensive Care Med 28:547–553, 2002 18. Poca MA, Sahuquillo J, Báguena M, Pedraza S, Gracia RM, Rubio E: 
Incidence of intracranial hypertension after severe head injury: a prospective study using the Traumatic Coma Data Bank 
classification. Acta Neurochir Suppl 71:27–30, 1998 19. Sahjpaul R, Girotti M: Intracranial pressure monitoring in se- vere 
traumatic brain injury—results of a Canadian survey. Can J Neurol Sci 27:143–147, 2000 20. Shafi S, Diaz-Arrastia R, Madden 
C, Gentilello L: Intracra- nial pressure monitoring in brain-injured patients is associ- ated with worsening of survival. J Trauma 
64:335–340, 2008 21. Stocchetti N, Penny KI, Dearden M, Braakman R, Cohadon F, Moons KG, Leenen LP, et al: Effect of 
intracranial pressure monitoring and targeted intensive care on functional outcome 
Iannotti F, et al: Intensive care management of head-injured patients in Europe: a survey from the European Brain Injury after 
severe head injury. Crit Care Med 33:2207–2213, 2005 
Consortium. Intensive Care Med 27:400–406, 2001 
5. Fakhry SM, Trask AL, Waller MA, Watts DD: Management of brain-injured patients by an evidence-based medicine pro- 
22. Thurman DJ, Alverson C, Dunn KA, Guerrero J, Sniezek JE: Traumatic brain injury in the United States: a public health 
tocol improves outcomes and decreases hospital charges. J 
perspective. J Head Trauma Rehabil 14:602–615, 1999 
Trauma 56:492–500, 2004 
23. Wilkins IA, Menon DK, Matta BF: Management of 
comatose 6. Hesdorffer DC, Ghajar J, Iacono L: Predictors of compliance 
head-injured patients: are we getting any better? 
Anaesthesia with the evidence-based guidelines for traumatic brain injury 
56:350–352, 2001 care: a survey of United States 
trauma centers. J Trauma 52: 1202–1209, 2002 7. Hukkelhoven CW, Steyerberg EW, Rampen AJ, Farace E, Habbema JD, 
Marshall LF, et al: Patient age and outcome following severe traumatic brain injury: an analysis of 5600 
Manuscript submitted December 20, 2012. patients. J 
Neurosurg 99:666–673, 2003 
Accepted July 22, 2013. 8. Idris Z, Ghani RI, Musa KI, 
Ibrahim MI, Abdullah M, Nyi 
Please include this information when citing this paper: 
pub- NN, et al: Prognostic study of using different monitoring 
lished online August 23, 2013; DOI: 
10.3171/2013.7.JNS122255. modalities in treating severe traumatic brain injury. Asian J 
Address correspondence to: Peep Talving, M.D., Ph.D., 
Depart- Surg 30:200–208, 2007 
ment of Surgery, Division of Acute Care Surgery, 
(Trauma, Emer- 9. Jeevaratnam DR, Menon DK: Survey of intensive care of se- 
gency Surgery and Surgical Critical Care), Keck School of 
Medicine, verely head injured patients in the United Kingdom. BMJ 
LAC + USC Medical Center, 2051 Marengo St., Rm. 
IPT-C5L100, 312:944–947, 1996 
Los Angeles, CA 90033. email: peep.talving@surgery.usc.edu. 
1254 J Neurosurg / Volume 119 / November 2013 

Vous aimerez peut-être aussi