Department of Neurosurgery, The MD Anderson Cancer Center, Houston, Texas

Transcription

1 J Neurosurg 121: , 2014 AANS, 2014 Predictors of outcome in civilians with gunshot wounds to the head upon presentation Clinical article Loyola V. Gressot, M.D., 1 Roukoz B. Chamoun, M.D., 2 Akash J. Patel, M.D., 1 Alex B. Valadka, M.D., 3 Dima Suki, Ph.D., 4 Claudia S. Robertson, M.D., 1 and Shankar P. Gopinath, M.D. 1 1 Department of Neurosurgery, Baylor College of Medicine, Houston, Texas; 2 Department of Neurosurgery, The University of Kansas, Kansas City, Kansas; 3 Seton Brain and Spine Institute Neurosurgery, Austin; and 4 Department of Neurosurgery, The MD Anderson Cancer Center, Houston, Texas Object. Prediction of outcome from initial presentation after a gunshot wound to the head (GSWH) is essential to further clinical decision making. The authors goals are to report the survival and functional outcomes of these patients, to identify prognostic factors, and to propose a scoring system that can predict their outcome. Methods. The records of 199 patients admitted with a GSWH with dural penetration between 1990 and 2008 were retrospectively reviewed. The inclusion criterion was a CT scan available for review. Patients declared brain dead on presentation were excluded, which yielded a series of 119 patients. Statistical analysis was performed using a logistic regression model. Results. Fifty-eight (49%) of the 119 patients died. Twenty-three patients (19%) had a favorable outcome defined as a 6-month Glasgow Outcome Scale (GOS) score of moderate disability or good recovery, 35 (29%) had a poor outcome (GOS of persistent vegetative state or severe disability), and 3 (3%) were lost to follow-up. Significant prognostic factors for mortality were age older than 35 years, nonreactive pupils, bullet trajectory of bihemispheric (excluding bifrontal), and posterior fossa involvement compared with unihemispheric and bifrontal. Factors that were moderately associated with higher mortality included intracranial pressure (ICP) above 20 mm Hg and Glasgow Coma Scale (GCS) score at presentation of 3 or 4. Upon multivariate analysis, the significant factors for mortality were bullet trajectory and pupillary response. s found to be significant for good functional outcome were admission GCS score greater than or equal to 5, pupillary reactivity, and bullet trajectory of unihemispheric or bifrontal. Factors moderately associated with good outcome included age of 35 years or younger, initial ICP 20 mm Hg or lower, and lack of transventricular trajectory. In the multivariate analysis, significant factors for good functional outcome were bullet trajectory and pupillary response, with age moderately associated with improved functional outcomes. The authors also propose a scoring system to estimate survival and functional outcome. Conclusions. Age, pupils, GCS score, and bullet trajectory on CT scan can be used to determine likelihood of survival and good functional outcome. The authors advocate assessing patients based on these parameters rather than pronouncing a poor prognosis and withholding aggressive resuscitation based upon low GCS score alone. ( Key Words gunshot wound to the head Glasgow Coma Scale outcome penetrating head injury traumatic brain injury Gunshot wounds to the head (GSWHs) involving the brain are often devastating and can be a challenge to treat. Recently, gun violence has received increased scrutiny by the media. In one series, GSWHs accounted for 35% of all deaths from head injury. 8 With guns representing the leading weapon used in homicides in the US, caring for patients with GSWHs continues to plague many hospitals ( cius2008/). Because of historically poor neurological outcomes, treatment paradigms remain controversial. Up to 66% of Abbreviations used in this paper: AIS = Abbreviated Injury Scale; GCS = Glasgow Coma Scale; GOS = Glasgow Outcome Scale; GSWH = gunshot wound to the head; ICP = intracranial pressure. J Neurosurg / Volume 121 / September 2014 patients sustaining a GSWH die at the scene. 8 Mortality rates as high as 93% have been reported, with survivors commonly exhibiting devastating neurological outcomes. 19 Despite aggressive management, outcome depends on both primary and secondary injury. Previous studies performed at our institution have demonstrated that secondary cerebral ischemia is associated with worse neurological outcomes in patients with GSWHs as well as closed head injury. 5,20 The clinical course of many patients is marked by both early and late complications. These grim statistics raise the question of which patients benefit from aggressive management. Several predictors of survival have been identified, with widely accepted negative prognostic indicators including low Glasgow Coma Scale (GCS) score upon arrival, bihemi- 645

2 L. V. Gressot et al. spheric injuries, ventricular involvement, and nonreactive pupils. 14 Although there are many reports in the neurosurgical literature detailing various prognostic factors for mortality and outcome, there have been no attempts to create a useful system to synthesize this information. Our primary goal was to report the 6-month rates of survival and functional outcome of patients presenting with GSWHs at our facility. Secondary goals were to identify prognostic factors for favorable outcome and to propose a scoring system for the rapid calculation of likely patient outcomes based on initial CT scan. Methods Patient Population Between 1990 and 2008, 199 patients were admitted to Ben Taub General Hospital, a Level 1 trauma center, with GSWHs after surviving initial resuscitation. This yielded a series of 119 consecutive patients whose prognosis was in question at presentation and whose records were obtained and retrospectively reviewed. Early in our experience, patients in extremis with profound hypotension or moribund appearance did not undergo CT scanning. Those who appeared brain dead upon admission were admitted as potential organ donors. Inclusion criteria included GSWH with dural penetration and availability of an admission head CT scan. The goal of this study was to identify factors that affect prognosis. Therefore, patients who did not survive initial resuscitation in the emergency department and patients who were known to be deceased at presentation from either brain death or cardiovascular death and were admitted as prospective organ donors were excluded. The data for this series were obtained from a prospectively maintained patient database, IRB H Demographic information with regard to age, sex, and race was collected. Information collected from initial emergency department evaluation included GCS score; presence or absence of respiratory distress as defined by prehospital intubation, agonal respirations, respiratory rate greater than 30 or less than 8 breaths/minute, or oxygen saturation (SpO 2 ) less than 90%; systolic blood pressure; pupillary equality and response; and Abbreviated Injury Scale (AIS) score. Initial noncontrast head CT scans were obtained and independently reviewed for the purpose of this study by 3 separate evaluators, including the senior author (S.P.G.), and were assessed for bullet trajectory and classified as unihemispheric, bifrontal, bihemispheric not purely bifrontal, or posterior fossa. There were no discrepancies among the reviewers regarding type of bullet injury sustained. Presence of intraventricular hemorrhage, degree of midline shift, ventricular size, and appearance of cisterns as well as presence and type of discrete hematoma were also noted. Patients treatment, surgical procedures, discharge condition, and functional outcome by Glasgow Outcome Scale (GOS) classification at 6-month follow-up were also recorded. Outcome classification was determined by direct interviews at clinic follow-up when possible and by phone if not. Statistical Analysis Frequencies and descriptive analyses were performed on the variables under study. The logistic regression model was used to identify factors associated with mortality and functional status 6 months following the GSWH. Univariate and multivariate odds ratios and their 95% confidence intervals were obtained. A p value 0.05 was considered significant. The IBM PASW version 17.0 was used for statistical analyses. Acute Care All patients included in this series received initial stabilization according to a standardized protocol that included support of ventilation by intubation if indicated and hemodynamic support as needed to maintain cerebral perfusion pressure higher than 60 mm Hg with fluids and vasopressors. The head of the bed was elevated at 30. Intracranial pressure (ICP) was monitored with a targeted ICP of less than 20 mm Hg and was aggressively treated with CSF drainage, sedation, paralysis, mannitol, and moderate hyperventilation. If these measures were not sufficient, barbiturate coma was implemented. Normothermia was maintained. Electrolyte disturbances and coagulopathy were corrected. All patients received broadspectrum antibiotic therapy prophylactically, with prolonged therapy for documented infections. Indications for emergency surgery included evacuation of a large intracranial hematoma with a greater than 5-mm midline shift, refractory intracranial hypertension, and wound debridement for infection control when visible brain herniation from the bullet wound was present. It is our policy to continue aggressive treatment of patients sustaining a GSWH and wait at least 7 10 days or longer prior to discussing code status with the patient s family as most patients are very young. In cases in which the family wishes to withdraw care earlier and is adamant about its decision, we respect the family s wishes; this was the case in only 5 patients. Results Of the 119 patients included in this study, the median age was 26 years (range years). There were 104 males (87%) and 15 (13%) females (Table 1). Sixty-one injuries (51%) occurred during the decade between 1990 and 1999 and 58 injuries (49%) occurred between 2000 and No significant association was found between sex and mortality (Table 2). The 6-month mortality rate for all comers was 49%, or 58 of the 119 patients included in this series. Three patients (3%) were lost to follow-up but survived to hospital discharge. Factors significantly associated with mortality are summarized in Table 2. Twenty-three patients (19%) had a favorable outcome defined as a GOS score at 6-month follow-up of moderate disability or good recovery. Thirty-five patients (29%) survived but had a poor outcome, defined as GOS score of persistent vegetative state or severe disability. Factors significantly associated with improved functional outcome at 6 months are summarized in Table 3. The GCS score upon presentation to the emergency department was classified as either 3 4 or The breakpoint was determined by recursive partitioning analysis. In the univariate analysis, a GCS score of 3 or 4 tended 646 J Neurosurg / Volume 121 / September 2014

3 Predictors of outcome in GSWHs upon presentation TABLE 1: Distribution of various demographic measurements Parameter to be associated with higher mortality (OR 2.01 [95% CI ]; p = 0.064) (Table 2). Patients with a GCS score of 3 or 4 at presentation had a 60% mortality rate, and patients with a GCS score of 5 15 had a 39% mortality rate. A higher initial GCS score of 5 15 was also associated with good functional outcome at 6-month follow-up (OR 2.60 [CI ]; p = 0.049). Patients with an initial GCS score of 3 or 4 had an 11% rate of good outcomes and an 89% rate of poor outcomes or death, whereas patients with an initial GCS score of 5 15 had a 27% rate of good outcomes and a 73% rate of poor outcomes or death. The age distribution of patients was positively skewed, with most injuries occurring in the 2nd and 3rd decades. Age did not have a continuous effect upon statistical analysis and thus was grouped for analysis based upon the upper end of mean ages of GSWH series reported in the existing literature. 6,16,19 Mortality was increased in patients older than age 35 years and was 66.7% in these patients. Patients 35 years and younger had a 42.7% mortality. An age of 36 years and older was associated with higher mortality in the univariate analysis (OR 2.53 [95% CI ]; p = 0.037). Patients with an age of 35 years or younger had a significantly increased likelihood of having a good outcome (OR 4.52 [95% CI ]; p = 0.051). Patients 35 years or younger had 23.6% rate of good outcome and a 76.4% rate of poor outcomes or death, and patients older than 35 years had a 6.7% rate of good outcome and a 93.3% rate of poor outcomes or death. The status of nonreactive pupils upon arrival to the emergency department was significantly associated with increased mortality (OR 3.08 [95% CI ]; p = 0.004). Patients with at least 1 reactive pupil had a 35% mortality rate and patients with bilateral nonreactive pupils had a 63.6% mortality rate. Reactive pupils were also associated with good functional outcomes at 6-month follow-up (OR 4.50 [95% CI ]; p = 0.006). Patients with reactive pupils had a 30% rate of good outcome, and J Neurosurg / Volume 121 / September 2014 Value* age (yrs) median 26 range sex male 104 (87) female 15 (13) race Asian 4 (3) black 42 (35) Hispanic 46 (39) white 26 (22) unknown 1 (0.8) time period (51) (49) * Values are number of patients (%) unless stated otherwise. TABLE 2: Factors associated with increased mortality: univariate analysis* No. Deceased (%) Univariate bullet track unihemispheric/ 24/70 (34.3) bifrontal bihemispheric/ 34/49 (69.4) < posterior fossa age (yrs) 35 38/89 (42.7) 36 20/30 (66.7) pupils reactive 21/60 (35.0) nonreactive 35/55 (63.6) initial ICP (mm Hg) 20 10/41 (24.4) 21 11/22 (50) GCS score in ED /55 (60) /64 (39) sex male 51/104 (49) female 7/15 (46.7) respiratory distress yes 30/54 (55.6) no 19/41 (46.3) systolic BP normal 31/59 (52.5) hypotensive 4/8 (50) hypertensive 20/42 (47.6) AIS score 25 48/96 (50) /23 (43.5) pupil equality equal 43/81 (53.1) unequal 10/27 (37.0) transventricular trajectory yes 23/39 (59) no 35/80 (43.8) ventricle size normal 17/29 (58.6) dilated 4/8 (50) compressed 37/82 (45.1) midline shift (mm) 0 25/46 (54.3) /33 (48.5) /27 (37) >15 7/13 (53.8) (continued) 647

4 L. V. Gressot et al. TABLE 2: Factors associated with increased mortality: univariate analysis* (continued) No. Deceased (%) Univariate cisterns present 8/24 (33) compressed 40/76 (52.6) absent 9/18 (50) hematoma yes 51/87 (58.6) no 19/47 (40.4) craniotomy yes 30/80 (37.5) no 28/39 (71.8) * BP = blood pressure; ED = emergency department. Strongly or moderately associated with increased mortality. patients with nonreactive pupils had a 9.1% rate of good outcome. Initial ICP was reported for 63 patients. Patients with an initial ICP of 21 mm Hg or higher had a mortality rate of 50% compared with 24.4% in patients with initial ICP 20 mm Hg or lower. High ICP was moderately associated with mortality in the univariate analysis (OR 2.56 [95% CI ]; p = 0.09). Lower initial ICP was also moderately associated with improved functional outcome (OR 3.69 [95% CI ]; p = 0.062). Patients without elevated ICP had a 39% rate of good outcomes compared with a rate of 14% in patients with high initial ICP. Bullet track on initial CT scan was classified as involving a single hemisphere, bilateral frontal lobes only, bilateral hemispheres not purely involving the frontal lobes, or posterior fossa (Fig. 1). No significant difference regarding outcome was found between patients sustaining unihemispheric injuries and those with purely bifrontal injuries; thus, these groups were combined. Similarly no significant difference was found between the patients sustaining posterior fossa injuries and those with bihemispheric (excluding bifrontal) injuries, and thus these were combined. Bullet track involving a bihemispheric (not purely bifrontal) trajectory or involving the posterior fossa was found to be significantly associated with mortality (OR 4.06 [95% CI ]; p 0.001). Patients with unihemispheric involvement had the lowest mortality (32.7%), followed by patients with bifrontal involvement (38.9%), patients with bihemispheric involvement (67.6%), and finally patients with posterior fossa involvement (75%) (Table 4). Bullet track with bifrontal or unihemispheric trajectory was associated with favorable outcomes (OR [95% CI ]; p = 0.002). Patients with unihemispheric or bifrontal involvement had higher rates of good outcomes at 30.8% and 33.3%, respectively. Patients with bihemispheric or posterior fossa involvement had lower rates of good outcomes, at 2.7% and 0%, respectively (Table 4). Although patients with transventricular bullet trajectory had a 59% mortality rate compared with 43.8% in TABLE 3: Factors associated with increased 6-month good functional outcome: univariate analysis No. of Good Outcomes (%) Univariate bullet track* unihemispheric/ 22/70 (31.4) bifrontal bihemispheric/ 1/49 (2) posterior fossa age (yrs)* 35 21/89 (23.6) /30 (6.7) pupils* reactive 18/60 (30) nonreactive 5/55 (9.1) initial ICP* 20 16/41 (39) /22 (14) GCS score in ED* 3 4 6/55 (11) /64 (27) sex male 20/104 (19.2) female 3/15 (20) respiratory distress yes 5/54 (9.3) no 6/41 (14.6) systolic BP normal 12/59 (20.3) hypotensive 1/8 (12.5) hypertensive 4/42 (9.5) AIS score 25 17/96 (17.7) 26 6/23 (26.1) pupil equality equal 17/81 (21) unequal 5/27 (18.5) transventricular trajectory* yes 4/39 (10.3) no 19/80 (23.8) ventricle size normal 4/29 (13.8) dilated 2/8 (25) compressed 17/82 (20.7) midline shift (mm) 0 7/46 (15.2) 1 4 8/33 (24.2) /27 (22.2) >15 2/13 (15.4) (continued) 648 J Neurosurg / Volume 121 / September 2014

5 Predictors of outcome in GSWHs upon presentation TABLE 3: Factors associated with increased 6-month good functional outcome: univariate analysis (continued) No. of Good Outcomes (%) Univariate cisterns present 7/24 (29.2) compressed 11/76 (24.5) absent 5/18 (27.8) hematoma yes 14/87 (16.1) no 10/47 (21.3) craniotomy* yes 20/80 (25) no 3/39 (7.7) * Strongly or moderately associated with functional outcome at 6 months. J Neurosurg / Volume 121 / September 2014 Fig. 1. Bullet trajectories as seen on noncontrast axial CT scans. A: Left unihemispheric involvement. B: Bifrontal involvement. C: Bihemispheric involvement. D: Posterior fossa involvement. patients without a transventricular trajectory, this was not statistically significant (p = 0.17). Lack of transventricular trajectory was moderately associated with good functional outcome (OR 2.87 [95% CI ]; p = 0.074). Surgical intervention was performed in 80 patients and was associated with lower mortality (OR 4.48 [95% CI ]; p = 0.001). Patients undergoing a craniotomy had a mortality rate of 37.5%, whereas patients who did not receive surgical intervention had a mortality rate of 71.8%. Surgical intervention was also associated with better functional outcomes (OR 4.02 [95% CI ]; p = 0.034). Patients in the operative group had a 25% rate of favorable outcomes, while patients in the nonoperative group only had a 7.7% rate. Patients in the operative group had high mean GCS score upon presentation: 6.5 ± 3.3 versus 4.9 ± 2.5 (p = 0.012). Respiratory distress on presentation was associated with 55.6% mortality. Patients without respiratory distress upon presentation had a 46.3% mortality rate. This was not statistically significant (p = 0.45). Systolic blood pressure at presentation, associated injury score, pupillary equality, ventricular size, extent of midline shift, cistern effacement, and presence of a hematoma were not significantly different between groups. Multivariate analysis was performed using variables that were found to be significant in the univariate analysis. Because the decision to proceed with surgery was a treatment decision as opposed to a variable determined by patient presentation, whether the patient underwent a craniotomy was not included in the multivariate analysis. In the multivariate analysis for variables associated with mortality, bullet track involving bihemispheric (not bifrontal) trajectory or the posterior fossa was significantly associated with increased mortality (OR 4.57 [95% CI ]; p = 0.001) (Table 5). Reactive pupils were significantly associated with decreased mortality (OR 0.3 [95% CI ]; p = 0.012). Age and GCS score were not significantly associated with increased mortality in the multivariate analysis. In the multivariate analysis for functional outcomes, bullet track involving unihemispheric or bifrontal trajectory was associated with improved functional outcomes (OR 26.9 [95% CI ]; p = 0.002) (Table 6). Nonreactive pupils were associated with decreased rate of good functional outcomes (OR 0.15 [95% CI ]; p = 0.006). Lower age was moderately associated with improved functional outcomes in the multivariate analysis (OR 1.05 [95% CI ]; p = 0.109). Glasgow Coma Scale score did not reach statistical significance for 6-month functional outcome in the multivariate analysis. Proposed Scoring System to Predict Outcome We propose a scoring system for the rapid assessment of patients with a GSWH at the time of the first CT scan. The goal is to predict likely outcome at presentation. This system, presented in Table 7, is based on variables found to be statistically significant and include age, GCS score upon arrival, pupillary reactivity, and bullet trajectory. TABLE 4: Outcome of patients grouped by missile trajectory Trajectory Dead No. of Patients (%) Poor Outcome Favorable Outcome Lost to Follow-Up Total unihemispheric 17 (32.7) 17 (32.7) 16 (30.8) 2 (3.8) 52 bifrontal 7 (38.9) 4 (22.2) 6 (33.3) 1 (5.6) 18 bihemispheric 25 (67.6) 11 (29.7) 1 (2.7) 0 37 posterior fossa 9 (75) 3 (25)

6 L. V. Gressot et al. TABLE 5: Multivariate analysis of factors associated with mortality Multivariate bullet track involving posterior fossa/ bihemispheric* age GCS score in ED reactive pupils* * Associated with mortality. Possible total scores range from 0 to 5. Scores of 0 1 have approximately 25% mortality, 2 has approximately 50% mortality, and 3 5 imply approximately 75% mortality. In terms of functional outcome, a score of 0 is associated with approximately 50% chance of good outcome, 1 with approximately 30% good outcome, 2 with approximately 10% good outcome, and 3+ with essentially no chance of a good outcome (Table 8). While initial ICP and operative intervention were either significant variables or trended to significance for patient outcome, these were not included in the scoring system because this is not information readily available at initial presentation. Discussion Overall Mortality and Functional Outcome Gunshot wounds to the head continue to be devastating injuries. Our institutional policy is to aggressively treat all patients who have sustained a GSWH, and we undertook this study to validate whether our outcomes supported this. Mortality rates reported in the literature range from 7.7% to 93%. 1,2,6,7,9,11,14,16 18,21 23 In our series, we found an overall mortality rate of 48.7%. Around 19% of all comers and 40% of those who survived ultimately had a favorable functional outcome at 6-month follow-up. The significant prognostic factors we identified were age, GCS score on presentation, pupillary reactivity, and bullet trajectory. We integrated these factors into a simple scoring system that would be relatively easy to remember and would allow prediction of outcome upon presentation. GCS Score It is well documented in the literature that a lower TABLE 6: Multivariate analysis of factors associated with improved functional outcomes at 6 months Multivariate bullet track unihemispheric/bi frontal* younger age* GCS in ED nonreactive pupils* * Strongly or moderately associated with functional outcomes. TABLE 7: Scoring system* Score age >35 yrs +1 GCS Score 3 or 4 +1 nonreactive pupils +1 trajectory posterior fossa or bihemispheric +2 * Points are summated for each of these factors. If none of these factors are present, a score of 0 is assigned. The total score has prognostic value for mortality and functional outcome. GCS score upon presentation is associated with higher mortality and worse functional outcomes. 6,7,9 11,14,16,18,22 In our series the chi-square test for trend was statistically significant for mortality. Grahm et al. and Martins et al. argued that aggressive surgical intervention should only be undertaken in a patient with GCS score of 3 5 if there is a significant hematoma present, provided the patient has been adequately resuscitated. 6,14 Clark et al. reported a 100% mortality rate for patients presenting with a GCS score of 3. 4 Though our findings agree that lower GCS score upon presentation is a negative prognostic indicator, we did have 5 patients who presented with a GCS score of 3 who had a favorable outcome. All 5 of these patients had unihemispheric injuries, but only 3 had reactive pupils. Four of them underwent craniotomy: 3 for evacuation of a subdural hematoma and 1 for intraparenchymal hematoma. One patient with a GCS score of 4 had a good outcome. He had a bifrontal injury and nonreactive pupils. Two patients with a GCS score of 5 had a good outcome, and both had reactive pupils. One had a bifrontal injury and the other had a unihemispheric injury. Data regarding alcohol level and presence of sedative/hypnotic drugs in these patients were not available for review. While alcohol and drug intoxication can lower GCS score, these data are commonly unavailable to the treating physician immediately at initial patient presentation; there is also usually some uncertainty about drugs administered to these patients during their transfer, especially when they are intubated in the field. 3 At our institution, it has been our philosophy to treat all patients aggressively to give them a chance for survival. We believe that although low GCS score upon presentation is an important factor to consider in the initial evaluation of a GSWH patient, it is not a sufficient measure of a patient s prognosis and should not be used as the sole criterion to deny aggressive treatment. Thus, our data reflect this treatment philosophy, and our outcomes reflect aggressive surgical management in all patients, except those without evidence of brain function at the time of presentation. Age and Sex Similar to other series, we found that most GSWH patients are in their 2nd and 3rd decades. 14 In accordance with other studies, we found that older patients had worse outcomes. 9,22 Kaufman et al. found that women had better survival than men and postulated that this may be due to 650 J Neurosurg / Volume 121 / September 2014

7 Predictors of outcome in GSWHs upon presentation TABLE 8: Score versus mortality and functional outcome Score a lower likelihood of women completing suicide. 9 We did not find a relationship between sex and outcome in our series. Pupils In accordance with other reports, bilaterally nonreactive pupils were a negative prognostic indicator in our series for both mortality and functional outcome. 9,12,23 Some authors argue that pupillary response is an equivocal finding because most patients with a low GCS score will have nonreactive pupils. 6 Pupillary response appears to be most enlightening in patients with GCS scores of 3 5 at presentation as an indicator of possible survivability, and in these cases it is usually a critical point in clinical decision making. 13 Based on these data, many advocate nonaggressive treatment in patients with low GCS score and nonreactive pupils. Although our data show that these patients have worse outcomes, with a 64% mortality, 9% had a good outcome at 6 months. In our mind, this is enough to justify a more optimistic surgical philosophy. J Neurosurg / Volume 121 / September 2014 Rate mortality % 2 50% % good outcome 0 50% 1 30% 2 10% 3 5 0% Bullet Trajectory Bullet trajectory has been shown to impact both mortality and morbidity from penetrating brain injuries, with bihemispheric and posterior fossa injuries demonstrating worse outcomes. 6,7,9 12,14,16,18,21,22 Martins et al. reported a 96.2% mortality rate for bihemispheric injuries and a 100% mortality rate for posterior fossa injuries. 14 Nathoo et al. report a much lower mortality rate of 19.2% for their series of patients with GSWHs involving the posterior fossa. 17 This series had a relatively high mean GCS score upon presentation of 11.8 and did not include any patient with a GCS score lower than 6. Transventricular injuries have also been reported to have high rates of mortality, though this did not reach statistical significance for mortality in our study and only trended toward significance for good functional outcome. 9,13,16,18 Grahm et al. advocated nonoperative management of patients with bihemispheric or multilobar involvement of the dominant hemisphere if they have a GCS score of 6 8 upon presentation unless a significant hematoma is also present. 6 Our series corroborated that patients with bihemispheric and posterior fossa injuries have worse outcomes, but identified a subset of patients with purely bifrontal bihemispheric involvement that achieved satisfactory recoveries (Table 3). Improved outcomes in patients with a pure bifrontal injury were also observed by Stone et al. in a relatively small cohort of 8 patients with bifrontal injuries. 23 While it is widely accepted that bihemispheric injuries are ominous, the outcomes of patients with purely bifrontal injuries in our series was equal to those of patients with unihemispheric injuries. We propose that bifrontal injuries be seen as a separate entity from other bihemispheric injuries when evaluating a patient s likely prognosis. None of our patients with posterior fossa involvement had a good outcome. Only 1 patient with bihemispheric (not pure bifrontal) involvement had a good outcome, and this patient presented with a GCS score of 7 and reactive pupils. Other Factors Though associated injury, respiratory arrest upon arrival, and hypotension upon arrival have been associated with mortality in other series, they were not significant in our series. 7,9,15 Surgical intervention was associated with improved mortality and morbidity in our series. Whether this result is significantly related to patient selection or not is difficult to establish in a retrospective series. Levy et al. showed that craniotomy for patients with a GCS score of 3 5 improved with regard to mortality but not morbidity. 13 In a later paper selecting for patients presenting with a GCS score of 6 15, Levy showed that operative intervention was only significantly associated with morbidity in patients with an initial GCS score of 12 15, though operative intervention was associated with decreased mortality for all patients. 12 All of these studies, including ours, have a selection bias in that patients who underwent surgery were those who were thought to be able to survive the surgery. In many of the previously published series, patients with poorer clinical status did not undergo surgery. Scoring System Multiple authors have developed treatment algorithms based upon their findings. 6,14,23 Turina et al. proposed the War Head Injury Score for predicting mortality in military penetrating craniocerebral trauma based upon GCS score and Injury Severity Score. 24 They also applied their scoring system to patients who suffered an acute penetrating head injury out of the war zone and found a weaker correlation when compared with military injuries. To our knowledge, we propose the first scoring system for civilian GSWH patients to predict mortality and likely functional outcome upon presentation (Tables 7 and 8). Limitations Our series consisted of a retrospective review of data collected from a prospectively maintained database. This series has the obvious inherent disadvantages of any retrospective review. The scoring system we propose was based on our results. If similar results are obtained from other centers, this may become part of the treatment guidelines. To confirm its reliability, it needs to be validated through other prospective studies, which we plan to perform at our center. 651

8 L. V. Gressot et al. Conclusions Patients with a gunshot wound to the brain continue to present a challenge to the treating physician. Patients suffering this type of injury continue to be plagued by high mortality rates and poor neurological outcomes. In this paper, we report a mortality rate of 49% and a 19% rate of good functional outcome at 6 months. We have found that age, pupils, GCS score, bullet trajectory, and initial ICP are factors prognostic of mortality. We propose a scoring system that can be used to determine the likelihood of survival and good functional outcome at presentation. In our view, although many of these patients have poor neurological outcomes and commonly have long, complex clinical courses, our data do not support a nihilistic approach to treating these patients. About 19% of these patients have good outcomes, and we propose a system to identify patients who are likely to do well, allowing for candid, informed discussion with family members. Disclosure This study is supported by NIH Grant P01-NS 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: Gopinath, Gressot, Chamoun. Acquisition of data: Gressot. Analysis and interpretation of data: Gressot, Chamoun, Patel. Drafting the article: Gressot, Chamoun, Patel. 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: Gopinath. Statistical analysis: Suki. Administrative/technical/material support: Chamoun, Robertson. Study supervision: Gopinath, Robertson. References 1. Benzel EC, Day WT, Kesterson L, Willis BK, Kessler CW, Modling D, et al: Civilian craniocerebral gunshot wounds. Neu rosurgery 29:67 72, Cavaliere R, Cavenago L, Siccardi D, Viale GL: Gunshot wounds of the brain in civilians. Acta Neurochir (Wien) 94: , Chamoun RB, Robertson CS, Gopinath SP: Outcome in patients with blunt head trauma and a Glasgow Coma Scale score of 3 at presentation. Clinical article. J Neurosurg 111: , Clark WC, Muhlbauer MS, Watridge CB, Ray MW: Analysis of 76 civilian craniocerebral gunshot wounds. J Neurosurg 65:9 14, Gopinath SP, Robertson CS, Contant CF, Hayes C, Feldman Z, Narayan RK, et al: Jugular venous desaturation and outcome after head injury. J Neurol Neurosurg Psychiatry 57: , Grahm TW, Williams FC Jr, Harrington T, Spetzler RF: Civilian gunshot wounds to the head: a prospective study. Neurosurgery 27: , Jacobs DG, Brandt CP, Piotrowski JJ, McHenry CR: Transcranial gunshot wounds: cost and consequences. Am Surg 61: , Kaufman HH: Civilian gunshot wounds to the head. Neurosurgery 32: , Kaufman HH, Makela ME, Lee KF, Haid RW Jr, Gildenberg PL: Gunshot wounds to the head: a perspective. Neurosurgery 18: , Khan MB, Kumar R, Irfan FB, Irfan AB, Bari ME: Civilian craniocerebral gunshot injuries in a developing country: presentation, injury characteristics, prognostic indicators, and complications. World Neurosurg [in press], Kim TW, Lee JK, Moon KS, Kwak HJ, Joo SP, Kim JH, et al: Penetrating gunshot injuries to the brain. J Trauma 62: , Levy ML: Outcome prediction following penetrating craniocerebral injury in a civilian population: aggressive surgical management in patients with admission Glasgow Coma Scale scores of 6 to 15. Neurosurg Focus 8(1):E2, Levy ML, Masri LS, Lavine S, Apuzzo ML: Outcome prediction after penetrating craniocerebral injury in a civilian population: aggressive surgical management in patients with admission Glasgow Coma Scale scores of 3, 4, or 5. Neurosurgery 35:77 85, Martins RS, Siqueira MG, Santos MT, Zanon-Collange N, Moraes OJ: Prognostic factors and treatment of penetrating gunshot wounds to the head. Surg Neurol 60:98 104, Murano T, Mohr AM, Lavery RF, Lynch C, Homnick AT, Livingston DH: Civilian craniocerebral gunshot wounds: an update in predicting outcomes. Am Surg 71: , Nagib MG, Rockswold GL, Sherman RS, Lagaard MW: Civilian gunshot wounds to the brain: prognosis and management. Neurosurgery 18: , Nathoo N, Chite SH, Edwards PJ, van Dellen JR: Civilian infratentorial gunshot injuries: outcome analysis of 26 patients. Surg Neurol 58: , Ozkan U, Kemaloğlu S, Ozateş M, Aydin MD: Analysis of 107 civilian craniocerebral gunshot wounds. Neurosurg Rev 25: , Part 2: Prognosis in penetrating brain injury. J Trauma 51 (2 Suppl):S44 S86, Robertson CS, Gopinath SP, Goodman JC, Contant CF, Valadka AB, Narayan RK: SjvO2 monitoring in head-injured patients. J Neurotrauma 12: , Shoung HM, Sichez JP, Pertuiset B: The early prognosis of craniocerebral gunshot wounds in civilian practice as an aid to the choice of treatment. A series of 56 cases studied by the computerized tomography. Acta Neurochir (Wien) 74:27 30, Siccardi D, Cavaliere R, Pau A, Lubinu F, Turtas S, Viale GL: Penetrating craniocerebral missile injuries in civilians: a retrospective analysis of 314 cases. Surg Neurol 35: , Stone JL, Lichtor T, Fitzgerald LF: Gunshot wounds to the head in civilian practice. Neurosurgery 37: , Turina D, Sustić A, Tićac Z, Dirlić A, Krstulović B, Glavas A, et al: War head injury score: an outcome prediction model in War casualties with acute penetrating head injury. Mil Med 166: , 2001 Manuscript submitted August 28, Accepted May 29, Portions of this work were presented as an electronic poster at the American Association of Neurological Surgeons 78th Annual Meeting, Philadelphia, Pennsylvania, May 1 5, 2010, and the South Texas Chapter of the American College of Surgeons annual meeting, Houston, Texas, February 21 23, Please include this information when citing this paper: published online July 4, 2014; DOI: / JNS Address correspondence to: Shankar P. Gopinath, M.D., Department of Neurosurgery, Baylor College of Medicine, 6501 Fannin St. NC-100, Houston, TX shankarg@bcm.edu. 652 J Neurosurg / Volume 121 / September 2014