Intracranial pressure monitoring in severe head injury: compliance with Brain Trauma Foundation guidelines and effect on outcomes: a prospective study

Transcription

1 See the corresponding editorial in this issue, pp J Neurosurg 119: , 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 Peep Talving, M.D., Ph.D., Efstathios Karamanos, M.D., Pedro G. Teixeira, M.D., Dimitra Skiada, M.D., Lydia Lam, M.D., Howard Belzberg, M.D., Kenji Inaba, M.D., and Demetrios Demetriades, M.D., Ph.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) monitoring 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 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. Secondary 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 guidelines 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 managed according to the BTF ICP guidelines experienced significantly improved survival. ( Key Words intracranial pressure monitoring mortality brain herniation Brain Trauma Foundation guidelines traumatic brain injury Traumatic brain injury is the major cause of mortality and morbidity both in civilian and military settings. 7,14,22 The incidence and risk of intracranial 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 pressure; 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. hypertension after severe traumatic brain injury (TBI) have been previously documented ,15,18,22 Thus, according to the Brain Trauma Foundation (BTF) guidelines, intracranial pressure (ICP) monitoring is considered 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 1248 J Neurosurg / Volume 119 / November 2013

2 Intracranial pressure monitoring in severe blunt TBI with BTF ICP monitoring guidelines is associated with improved outcomes. Methods After obtaining approval from the institutional review 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 criteria 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, The decision to place an ICP monitoring device was at the neurosurgeon s discretion. Exclusion 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 age, sex, blood pressure on admission, GCS score on admission, Injury Severity Score (ISS), AIS for each body region (head, chest, abdomen, and extremity), type of intracranial injury, ICP values in patients undergoing intervention, intracranial hemorrhage (ICH) treatment modalities, and neurosurgical documentation for omission of ICP monitoring. A standardized data collection sheet was used, allowing the treating physicians to choose from a variety of standardized and clinically pertinent choices as to why an ICP monitoring device was not placed. In case the predetermined choices were not sufficient, the 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 undergoing ICP monitoring. All subsequent analyses were performed 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). J Neurosurg / Volume 119 / November 2013 Statistical Analysis Continuous variables were dichotomized using clinically 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 categorical variables using the Fisher exact or Pearson chi-square tests as appropriate. The Shapiro-Wilk Test for normality was deployed for continuous variables; normally distributed variables were compared using the Student t-test while nonnormally distributed variables were compared using the Mann-Whitney U-test. To identify independent predictors of ICP monitoring, a forward stepwise logistic regression was deployed using variables at a p < 0.2 level after univariate analysis. Only the statistically significant variables are reported. Subsequent univariate analyses for in-hospital mortality and mortality due to brain herniation were performed. Independent predictors of in-hospital mortality and mortality 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 regression was performed using as independent variables the placement of the ICP monitoring device, adjusting for differences between the groups at p < The regression calculated the predicted probability (propensity score) of being subjected to ICP monitoring using variables that independently predicted the placement of an ICP monitoring device. Propensity score matching is a technique that tries to estimate the effect of a treatment by accounting for the covariates that predict receiving the treatment. It was performed in an effort to reduce the bias due to confounding factors that might be involved in comparing outcomes among patients who received the treatment with outcomes of those who did not. The overall in-hospital mortality and mortality due to brain herniation were assessed for each study group using logistic regression to adjust for factors that were significant at p < The propensity score was also inserted in the regression as a covariate. Adjusted odds ratios (AORs) with 95% CIs were derived from the logistic regression. An AOR < 1.00 (95% CI) implies that the factor decreases significantly the odds of developing the outcome, whereas an AOR > 1.00 (95% CI) increases significantly the odds of developing 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 compared using an independent t-test or Mann-Whitney U- test and subsequently linear regressions adjusting for differences that were significant at p < To correct for mortality bias, the same tests were deployed after exclusion of deaths. Values are reported as the mean ± SEM for continuous variables and as percentages for categorical variables. All analyses were performed using SPSS for Windows (version 12.0, SPSS, Inc.). Results Overall, 216 patients who sustained a severe TBI met the BTF guidelines for ICP monitoring. The epidemiological 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. Almost half of the patients (43.5%) had a GCS score of 3 on admission and 44.4% had a head AIS score of 5. A total of 46.8% of patients who met the BTF criteria underwent ICP monitoring (n = 101). A Becker ventriculostomy EMDS II (Medtronic Corp.) was placed in 60 patients (59%), and a transducer-tipped pressure/temperature 1249

3 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 ± ± age >55 yrs 27 (26.7%) 46 (40.0%) male 80 (79.2%) 81 (70.4%) admission physiology mean SBP in mm Hg 142 ± ± hypotension (SBP <90 mm Hg) 2 (2.0%) 10 (8.7%) mean heart rate 104 ± ± tachycardia (heart rate >120 bpm) 26 (25.7%) 37 (32.2%) mean respiratory rate 18 ± 1 18 ± injury severity indices mean ISS 25 ± 1 25 ± ISS (12.9%) 19 (16.5%) ISS (31.7%) 27 (23.5%) ISS (55.4%) 69 (60.0%) head AIS Score 3 42 (41.6%) 35 (30.4%) head AIS Score 4 22 (21.8%) 21 (18.3%) head AIS Score 5 37 (36.6%) 59 (51.3%) median GCS score 4 (3 8) 4 (3 8) GCS Score 3 39 (38.6%) 55 (47.8%) chest AIS 3 38 (37.6%) 36 (31.3%) abdomen AIS 3 11 (10.9%) 8 (7.0%) extremity AIS 3 18 (17.8%) 9 (7.8%) specific head injuries brain contusion 75 (74.3%) 82 (71.3%) SDH 55 (54.5%) 59 (51.3%) SAH 59 (58.4%) 40 (34.8%) IPH 37 (36.6%) 22 (19.1%) epidural hematoma 18 (17.8%) 17 (14.8%) midline shift 6 (5.9%) 8 (7.0%) loss of basal cisterns 31 (30.7%) 39 (33.9%) cerebral edema 74 (73.3%) 91 (79.1%) loss of gray/white differential 35 (34.7%) 28 (24.3%) fixed, dilated pupils on admission 23 (22.8%) 35 (30.4%) admission laboratory values mean PTT 30.4 ± ± mean PT 15.7 ± ± mean INR 1.21 ± ± INR (27.7%) 41 (35.7%) 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%) mean probability of receiving ICP monitoring 0.64 ± ± 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) (Integra LifeSciences Corp.) was placed in 41 patients (41%). All patients subjected to fiberoptic monitor placement remained with the fiberoptic monitor the entire critical phase of care. However, 58% of patients subjected to ventriculostomy 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 1250 J Neurosurg / Volume 119 / November 2013

4 Intracranial pressure monitoring in severe blunt TBI TABLE 2: Independent predictors of ICP monitoring* Step of Forward Logistic Regression Analysis Variable Cumulative R 2 AOR (95% CI) Adjusted p Value 1 decompressive craniectomy 4 hrs ( ) < extremity AIS score ( ) increasing age ( ) increasing PTT on admission ( ) best GCS score w/in 24 hrs of admission ( ) hypotension on admission ( ) SAH ( ) * 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 (95% CI ), p < (89.6%), followed by decompressive surgery (13.9%) and expectation of a rapid improvement (4.3%). Patients who sustained a subarachnoid hemorrhage (SAH) or an intraparenchymal hemorrhage (IPH) were more likely to receive an ICP monitoring device. Patients with SAH who did not receive ICP monitoring accounted for 34.8% of the study population, whereas patients with SAH who received ICP monitoring accounted for 58.4% (p = 0.001). Patients with an IPH who did not receive ICP monitoring accounted for 19.1% of the population, whereas patients with an IPH who received ICP monitoring accounted for 36.6% of the entire study sample (p = 0.006). No other differences in the incidence of type of intracranial lesions were noted between the groups (Table 1). The independent predictors for ICP monitoring device placement are summarized in Table 2. Patients undergoing decompressive craniectomy or patients with an extremity AIS score 3 were more likely to be subjected to ICP monitoring (AOR 3.85 [95% CI ] adjusted p < and AOR 3.01 [95% CI ] adjusted p = 0.033, respectively). Patients who were older (AOR 0.97 [95% CI ], adjusted p = 0.001), those with an increased partial thromboplastin time (PTT) on admission (AOR 0.96 [95% CI ], adjusted p = 0.021), those with a higher GCS score within first 24 hours of admission (AOR 0.81 [95% CI ], adjusted p = 0.002), or those who were hypotensive on admission (AOR 0.13 [95% CI ], adjusted p = 0.034) were significantly less likely to undergo ICP monitoring. We noted no difference in mortality for the ICP monitoring group when the 2 devices were compared (Table 3). For the patients subjected to ICP monitoring, the highest mean ICP noted in the study sample was 33.4 ± 2.3 mm Hg (± SEM, range mm Hg) and the lowest ICP was 5.9 ± 1.1 mm Hg (range 0 99 mm Hg). Patients undergoing ICP monitoring who experienced episodes of sustained elevated ICP were treated in all instances (n = 64). Elevation of the head of the bed was used in 90.6% of the cases, sedation in 89.1%, furosemide injection in 73.4%, pentobarbital coma in 23.4%, hypertonic saline infusion in 62.5%, mannitol bolus in 54.7%, hyperventilation in 51.6%, decompressive craniectomy in 15.6%, hypothermia in 14.1%, and paralysis in 3.1% of cases. Intracranial pressure monitoring was noted to be an independent predictor of overall in-hospital mortality (AOR 0.13 [95% CI ], adjusted p = 0.029), along with a head AIS score of 5 (AOR [95% CI ]), early initiation of nutrition (AOR 0.27 [95% CI ]), older age, best GCS score within the first 24 hours, and presence of SDH on CT. Intracranial pressure monitoring was also found to be an independent predictor of mortality due to brain herniation with an AOR of 0.31 (95% CI , adjusted p = 0.037), along with a head AIS score of 5 (AOR [95% CI ], adjusted p < 0.001), loss of basal cisterns on CT (AOR 5.01 [95% CI ]), best GCS score within the first 24 hours, and presence of IPH. The R 2 and the area under the curve (AUC) for the model were and for overall mortality and and for mortality due to brain herniation, respectively (Table 4). The overall complication rate throughout hospitalization was 5%. The incidence of pneumonia, acute kidney injury, and deep venous thrombosis/pulmonary embolism was 3.7% (8 of 216), 0.9% (2 of 216), and 0.9% (2 of 216), respectively; all of these complications occurred in the ICP study group. No incidents of acute respiratory distress syndrome and/or septic shock were noted. 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) Parameter Ventriculostomy (n = 60) Fiberoptic Monitor (n = 41) OR (95% CI) p Value AOR (95% CI)* Adjusted p Value* overall in-hospital mortality 22 (36.7%) 11 (26.8%) 1.58 ( ) ( ) mortality due to brain herniation 8 (13.3%) 5 (12.2%) 1.11 ( ) ( ) * Controlled for age, sex, IPH, SAH, PTT on admission, and presence of reactive pupils on admission. J Neurosurg / Volume 119 / November

5 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 R 2 AOR (95% CI) Adjusted p Value overall in-hospital mortality 1 head AIS Score ( ) < early nutrition ( ) increasing age ( ) best GCS score w/in 24 hrs of admission ( ) SDH ( ) ICP monitoring ( ) mortality due to brain herniation 1 head AIS Score ( ) < loss of basal cisterns ( ) best GCS score w/in 24 hrs of admission ( ) presence of IPH ( ) ICP monitoring ( ) * 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 (95% CI ), p < Mortality due to brain herniation: AUC (95% CI ), p < higher in patients not subjected to ICP monitoring after adjusting 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 craniectomy/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). Similarly, mortality due to brain herniation was significantly higher in patients not subjected to ICP monitoring (21.7% vs 12.9% (adjusted p = [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 after 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 cisterns (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 evidence 1 ). 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 device, allowing continuous ICP monitoring. The BTF guidelines are based on retrospective and limited prospective observational data; thus, the ICP compliance varies widely in neurocritical care, and the outcomes 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 monitoring. It is worth noting, however, that these guidelines do not constitute a universal protocol and thus treating physicians 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 providing 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 center relying on ICP/CPP-targeted therapy, the prevalence of ICP monitoring was 67%. 4 Shafi et al. performed a National 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 criteria 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 independent 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 coagulopathy, those with an elevated GCS score, or those pre J Neurosurg / Volume 119 / November 2013

6 Intracranial pressure monitoring in severe blunt TBI TABLE 5: Outcome measures Variable ICP Monitoring (n = 101) No ICP Monitoring (n = 115) OR (95% CI) p Value AOR (95% CI)* Adjusted p Value* overall in-hospital mortality 33 (32.7%) 62 (53.9%) 0.42 (0.24 to 0.72) (0.03 to 0.74) mortality due to brain herniation 13 (12.9%) 25 (21.7%) 0.53 (0.26 to 1.11) (0.10 to 0.87) 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 ± ± ( to 5.41) < ( 9.46 to 1.69) <0.001 mean hospital LOS in days 19.4 ± ± ( to 5.42) < ( to 2.08) after exclusion of deaths mean ICU LOS in days 21.7 ± ± ( to 4.13) < ( to 2.78) mean hospital LOS in days 25.8 ± ± ( to 2.09) < ( 9.56 to 2.34) * 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. J Neurosurg / Volume 119 / November 2013 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 comparing 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 incidence of SAH and IPH. Finally, patients with admission coagulopathy were less likely to undergo ICP monitoring 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. Patients 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 universally applied and each patient is treated based on physician 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 mortality due to brain herniation when compared with ventriculostomy (Table 3). We performed a regression analysis to elucidate independent 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 outcome of 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 including head injury severity (Table 5). Similar to our findings, previous examinations have observed survival advantage in ICP-targeted therapy, but with less marked survival effect. 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 interpreted the BTF guidelines to be inadequate to identify patients who benefit from ICP monitoring. Furthermore, these investigators suggested that the interventions in patients subjected to ICP monitoring may be associated with worsened outcomes. Such interventions include ICP monitoring device insertion in a coagulopathic state, utilization of vasoactive medication, osmotic diuresis with mannitol, furosemide utilization, hyperventilation, and use of paralytics. We documented all of the above interventions in our study group subjected to ICP monitoring; nevertheless, despite all these potentially harmful interventions, these patients 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 neurosurgeon 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 1253

7 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. Conclusions Compliance with the BTF ICP monitoring guidelines in our study sample was 46.8%. Compliance with these guidelines was associated with improved survival. Disclosure The authors report no conflict of interest concerning the materials 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. 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 (Erratum in J Neurotrauma 25: , 2008) 2. Bulger EM, Nathens AB, Rivara FP, Moore M, MacKenzie EJ, Jurkovich GJ: Management of severe head injury: institutional variations in care and effect on outcome. Crit Care Med 30: , 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: , Cremer OL, van Dijk GW, van Wensen E, Brekelmans GJ, Moons KG, Leenen LP, et al: Effect of intracranial pressure monitoring and targeted intensive care on functional outcome after severe head injury. Crit Care Med 33: , Fakhry SM, Trask AL, Waller MA, Watts DD: Management of brain-injured patients by an evidence-based medicine protocol improves outcomes and decreases hospital charges. J Trauma 56: , Hesdorffer DC, Ghajar J, Iacono L: Predictors of compliance with the evidence-based guidelines for traumatic brain injury care: a survey of United States trauma centers. J Trauma 52: , 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 patients. J Neurosurg 99: , Idris Z, Ghani RI, Musa KI, Ibrahim MI, Abdullah M, Nyi NN, et al: Prognostic study of using different monitoring modalities in treating severe traumatic brain injury. Asian J Surg 30: , Jeevaratnam DR, Menon DK: Survey of intensive care of severely head injured patients in the United Kingdom. BMJ 312: , Lane PL, Skoretz TG, Doig G, Girotti MJ: Intracranial pressure monitoring and outcomes after traumatic brain injury. Can J Surg 43: , Lee TT, Galarza M, Villanueva PA: Diffuse axonal injury (DAI) is not associated with elevated intracranial pressure (ICP). Acta Neurochir (Wien) 140:41 46, 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 implications. J Neurosurg 65: , Miller MT, Pasquale M, Kurek S, White J, Martin P, Bannon K, et al: Initial head computed tomographic scan characteristics have a linear relationship with initial intracranial pressure after trauma. J Trauma 56: , 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 Trauma 64: , 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: , 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: , Patel HC, Menon DK, Tebbs S, Hawker R, Hutchinson PJ, Kirkpatrick PJ: Specialist neurocritical care and outcome from head injury. Intensive Care Med 28: , 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, Sahjpaul R, Girotti M: Intracranial pressure monitoring in severe traumatic brain injury results of a Canadian survey. Can J Neurol Sci 27: , Shafi S, Diaz-Arrastia R, Madden C, Gentilello L: Intracranial pressure monitoring in brain-injured patients is associated with worsening of survival. J Trauma 64: , Stocchetti N, Penny KI, Dearden M, Braakman R, Cohadon F, Iannotti F, et al: Intensive care management of head-injured patients in Europe: a survey from the European Brain Injury Consortium. Intensive Care Med 27: , Thurman DJ, Alverson C, Dunn KA, Guerrero J, Sniezek JE: Traumatic brain injury in the United States: a public health perspective. J Head Trauma Rehabil 14: , Wilkins IA, Menon DK, Matta BF: Management of comatose head-injured patients: are we getting any better? Anaesthesia 56: , 2001 Manuscript submitted December 20, Accepted July 22, Please include this information when citing this paper: published online August 23, 2013; DOI: / JNS Address correspondence to: Peep Talving, M.D., Ph.D., Department of Surgery, Division of Acute Care Surgery, (Trauma, Emergency Surgery and Surgical Critical Care), Keck School of Medicine, LAC + USC Medical Center, 2051 Marengo St., Rm. IPT-C5L100, Los Angeles, CA peep.talving@surgery.usc.edu J Neurosurg / Volume 119 / November 2013