Part 2: Prognosis in Penetrating Brain Injury

Transcription

1 Part 2: Prognosis in Penetrating Brain Injury J Trauma. 2001;51:S44 S86. INTRODUCTION AND METHODOLOGY Part 2 of this document presents early clinical indicators that may be prognostic of outcome among patients with penetrating brain injury (PBI). This portion of the document is derived from the same principles of clinical epidemiology described in Part 1 of this Guideline s methodology section, and evaluates the pertinent literature on prognosis in patients with PBI qualitatively. The clinical factors used to determine prognosis for this section of the Guideline were determined by clinical assessments derived from measures with proven reliability (e.g., pupillary reflex or Glasgow Coma Scale [GCS] score). In this context, reliability means that different people with different backgrounds will reach the same conclusion about what they observe most of the time. Fortunately, good studies testing the reliability of these measures have been carried out, and will be discussed in the section on assessment that follows. Only reliable measures were included in this document. These include the following: clinical assessments (GCS score, pupillary size and reactivity, intracranial pressure [ICP], and neuroimaging), factual data (demographics, e.g., age), epidemiology (cause and mode of injury and caliber of weapon), and systemic measures (hypotension, coagulopathy, respiratory depression). Demographic factors such as race and gender were excluded because the literature lacked data on these topics. If we liken clinical assessments to diagnostic tests, and particularly to their role as predictors of poor outcome, we must be able to determine whether the diagnostic test has sensitivity, specificity, and positive or negative predictive value. An explanation of these factors is described below in a bayesian contingency table (Fig. 1). In a diagnostic test, sensitivity and specificity measure the appropriateness of the test. However, in prognosis, the most important aspects of this model are positive and negative predictive value. Positive predictive value represents the number of patients who had the clinical sign or prognostic indicator and had a poor outcome. For this to be meaningful and useful, a minimum positive predictive value of 70% or greater is thought to be necessary. However, this is inappropriate in PBI, since the prevalence of death is so high. Therefore, a different and more precise measure of prediction is used, as discussed below. To analyze prognostic parameters in nonpenetrating traumatic brain injury (TBI), a probability of at least 70% of a given outcome has been used as a cut-off point in defining parameters of clinical importance. 1 According to the bayesian statistical approach, the probability of a defined outcome in analyzing prognostic features should be viewed with respect to the prior probability, i.e., the distribution of outcome in the population studied. Since the prevalence of death (the primary outcome measure) in PBI is already above 70% in most series, when assessing positive predictive values of various prognostic indicators, a cut-off point of 70% is not relevant to this population. Therefore, instead of using a high positive predictive value as the standard of clinical importance, we used the odds ratio. The odds ratio is an estimate of relative risk for the outcome of interest if the studied factor is present as opposed to the factor being absent. Data from each article were used to determine the odds ratio for each prognostic indicator. For example, an odds ratio of 1 indicates no increase in risk if the factor is present, and an odds ratio of 3 indicates a three times greater risk of the same outcome. A 95% confidence interval was also calculated for each odds ratio. Clinical significance was defined as an odds ratio whose 95% confidence interval did not include the number 1. Prognosis studies (including prognosis with treatment) can have strength or weakness just like studies of therapeutic effectiveness. In the strongest studies, the patients should: Be seen at a uniform time in their disease (e.g., at time of admission to treating facility). Be followed prospectively for a designated block of time (e.g., for the first 24 hours after injury). Have their outcomes measured definitively and reliably (e.g., death). Be part of a continuous or defined cohort of at least 25 patients (e.g., from an ongoing, prospectively collected database). Be examined for extraneous prognostic variables, such as underlying disease or age (e.g., use of appropriate statistics such as multivariate analysis). Published studies that report on prognosis cannot be assessed using the same paradigm as those for therapeutic effectiveness. Prognosis studies are observational in nature and can never be randomized controlled trials, which are meant to compare treatments. However, for consistency in assessing levels of evidence in reaching conclusions about prognosis, the same designations (Class I, II, and ) as those for therapeutic effectiveness were identified. The Guideline development group agreed on the following definitions: Class I articles with all of the above characteristics. Class II articles exhibiting four out of the five characteristics (including prospectively collected data). Class articles with three or fewer of any of the above characteristics. Using this classification scheme, all published articles were evaluated and are listed in the evidentiary tables within S44 August Supplement 2001

2 Prognosis in Penetrating Brain Injury that use prognosis studies, therefore, the authors reached conclusions rather than made recommendations. Fig. 1. Bayesian contingency table. each Guideline section. It should be noted that an observational study, such as a case series that might be Class I by the above criteria, would only be Class if it were being used as an article on therapeutic effectiveness. These designations are meant to indicate scientific rigor and, thus, strength of evidence. However, unlike therapeutic effectiveness, articles on prognosis cannot be transposed from classification to recommendation because recommendations concern treatment, not prognosis. In the Guideline sections on assessment Outcome Distribution and Mortality in Penetrating Brain Injury: Implications for Prognostic Analysis In prognostic analysis, features are analyzed with respect to defined outcome categories. Outcome in nonpenetrating and penetrating brain injury series is generally reported according to the Glasgow Outcome Scale (GOS). 2 To analyze prognostic features of non-pbi, the GOS is commonly dichotomized into unfavorable (dead/vegetative/severe disability) versus favorable (moderate disability/good recovery) outcome. The outcome distribution in series of nonpenetrating TBI is typically U-shaped, 3 implying an inherent dichotomy in outcome consistent with the differentiation into unfavorable versus favorable (Fig. 2A). However, in PBI, the outcome distribution differs significantly. As is shown below, an overwhelming majority of patients have an unfavorable outcome; mostly, death. Thus, the appropriate outcome measure for PBI is mortality, not GOS score. It should be noted that the majority of studies addressed here deal with patients admitted to the hospital and do not account for prehospital deaths. The outcome distribution for PBI, shown in Figure 2A, is derived from a meta-analysis of results reported from five unselected series of civilian gunshot wounds. 4 8 Full data are given in the evidentiary tables. A large number of deaths and low incidence of patients in the categories vegetative or severe disability characterize the figure. Consequently, the shape of the curve is similar to that of a hockey stick. In contrast, the outcome distribution in nonpenetrating brain injury is more Fig. 2. TBI versus PBI outcome distribution. Volume 51 Number 2 S45

3 Table 1 Percentage of Unfavorable Outcome in Penetrating Brain Injury (Civilian) Authors, Year No. of Patients Time of Outcome Percentage Unfavorable Outcome (%) Mortality (%) Unselected series Clark et al., Discharge Grahm et al., Discharge or later Kaufman et al., GOS score discharge Nagib et al., GOS score 1 y Hernesniemi, Variable, long term Jacobs et al., GOS score discharge Selection Selected series Aldrich et al., Discharge GCS score 8 Kaufman et al., 190 GOS score discharge GCS score (series A) Levy et al., GOS score 6 mo TSAH TSAH, traumatic subarachnoid hemorrhage. typically U-shaped (Fig. 2B). However, among survivors, the percentage of patients with favorable outcome ( moderate disability plus good recovery ) is equal: 74%. Consequently, the main difference in outcome between penetrating and nonpenetrating brain injury is the difference in mortality. Series of PBI that focus on patients with a low GCS score show much higher mortality rates (Table 1, selected series 4,6 12 ) than for nonpenetrating TBI. Series that include patients dead on arrival or dead at the scene of the accident show mortality up to 93% (Table 2). 4 7,9,13 23 Mortality rates in unselected series that include a wide range of GCS scores are shown in Table 2 with rates varying between 34% and 79%, explained by inclusion of patients without dural penetration, 11 high number of mild injuries (GCS score of 13 15), 15 or low percentage of suicide as a cause of injury. 18 The rate of mortality and unfavorable outcome is considerably lower in nonpenetrating TBI. The most appropriate comparison of outcome can be made between patients with PBI versus nonpenetrating TBI from the Traumatic Coma Data Bank. In PBI, 10 overall mortality is 88% and the percentage of unfavorable outcome is 97% at discharge as compared with 32.5% mortality and 74.5% favorable outcome in nonpenetrating TBI at discharge. 24 Because of the high death rate in PBI and relatively low number of patients in the other bad outcome groups, in contrast to nonpenetrating PBI, death is a more appropriate outcome measure for prognostic analysis than the dichotomized GOS. Table 2 Mortality Rate in Civilian Penetrating Brain Injury Authors, Year Year of Study No. of Patients Mortality Rate (%) Remarks Patients admitted to neurosurgical centers Byrnes et al., August 1969 to April Clark et al., January 1984 to August Grahm et al., mo period Series includes 7 patients without dural penetration. Mortality rate excluding these patients: 63%. Hernesniemi, Hubschmann et al., Jacobs et al., January 1990 December % of deaths within first 6 h. Kaufman et al., June 1980 June Kaufman et al., Including patients without dural penetration. Kennedy et al., January 1987 June Excluding patients dying before CT scan examination. Large number of mild injuries: 70 of 192 patients with GCS score of Kunz et al., Includes patients without dural penetration. Lillard, Exclusion of patients considered near death. Mancuso et al., Low number of suicides: 6 of 40 (15%). Shaffrey et al., Shoung et al., Suddaby et al., Series including patients dead at the scene of accident or dead on arrival Cavaliere et al., ? Siccardi et al., ? Hernesniemi, S46 August Supplement 2001

4 Prognosis in Penetrating Brain Injury Table 3 Time to Death in Hospitalized Patients Authors, Year Dead within 12 h (%) Dead within 24 h (%) Dead within 48 h (%) Aldrich et al., Byrnes et al., Kaufman et al., Timing of Outcome Measure in PBI Implications for Classification of Data Death after PBI usually occurs soon after injury. Cavalieri et al. 22 describe 95% of deaths occurring within 3 hours and 97% within 24 hours of injury. Siccardi et al. report 92% of deaths occurring within 3 hours. In the study reported by Grahm et al., 6 77% of patients died before admission to the intensive care unit; in the study by Shaffrey et al., 19 53%. Hernesniemi 5 describes 76% of all deaths occurring in the first 24 hours, and Jacobs et al. 9 also found that 93% of deaths occurred in the first 24 hours. The percentages of death occurring within 12, 24, or 48 hours after injury, as reported in three studies, are shown in Table 3. The observation that, of the patients who die, the majority of deaths occur within the first 24 hours (70%) implies that the majority of outcomes in any given study occurs within the first 24 to 48 hours. For this reason, measurement of early mortality is relevant for prognostic purposes. Civilian versus Military Penetrating Injuries Implications for Prognostic Analysis PBI occurs in both military and civilian situations. However, major differences exist in mechanism of injury and outcome distribution in these two settings, requiring that they be analyzed separately. In the civilian setting, PBI is commonly caused by gunshot wounds; in the military setting, PBI is commonly caused by low-velocity shrapnel. Full outcome distribution according to the GOS has only been described in two series of penetrating brain injures sustained during military action. Aarabi 25 analyzed the outcome in 435 patients injured during the Iran-Iraq War and both Brandvold et al. 26 and Levi et al. 27 report separately on essentially the same patient population with injuries sustained in ongoing military hostilities in Lebanon. The outcome distribution of these two patient series is shown in Figure 3. Fig. 3. Outcome distribution in military PBI. Volume 51 Number 2 S47

5 When comparing this figure to the composite outcome distribution in penetrating brain injuries in the civilian situation (Fig. 2A), it is evident that the outcome in injuries sustained during military action is significantly different from civilian penetrating brain injuries. Military PBI series differ from civilian series in both mechanism of injury and severity. A number of issues explain the differences in reported outcomes between military and civilian PBI. In civilian injuries, gunshot wounds are the primary cause of PBI. In addition, among civilians, suicide is the most common cause of gunshot wounds to the brain, implying contact injuries, with high-energy transmission into the brain. A large proportion of deaths in civilian injuries occurs during the first few hours after injury. In the military setting, penetrating injuries that reach medical attention are predominantly caused by shell and shrapnel injuries. Because of the extremely damaging nature of the cerebral injuries inflicted by the high velocity of bullets commonly found in war injuries, the majority of those suffering battlefield gunshot wounds to the head presumably never reach medical care. This skews the surviving military PBI population toward lower velocity shrapnel wounds. The priorities of field triage under battle conditions means that there is no control for those patients who are found to have a low probability of survival and therefore are not prioritized for rapid transport. Although evacuation times from the front line to military hospitals has been considerably improved in the past two decades, the average time to arrival at the hospital is considerably longer than in civilian situations. Brandvold et al. 26 describe a mean evacuation time in the military setting of 2.3 hours. In civilian injures, the average time from injury to hospital is less than 30 to 45 minutes. Additionally, the exigencies of war mean that the determinants of the methods and applications of field resuscitation and triage are difficult to control and compare with the civilian situation. There is also a much higher likelihood of wound contamination in the battlefield. The conditions under which early surgical treatment is rendered are generally unique and more limited with respect to treatment facilities in the civilian setting. For instance, although computed tomographic (CT) scanning was generally available in civilian emergency centers in the 1970s, the first reports describing routine use of CT scanning in the military situation date from conflicts in the early 1980s. Also, injuries in the military setting occur primarily in young men in otherwise excellent physical condition; this contrasts with the civilian setting, where a broader age range is represented. Death rates are similarly much lower for patients who reach care for penetrating injuries occurring in the military setting than in civilian penetrating brain injuries. The reported case fatality rate from World War II is 10% to 13%. 28,29 Meirowski observed a 10% mortality reporting among 673 penetrating brain wounds from Korea. On composite analysis of 2,187 consecutive penetrating wounds of the brain from Vietnam, Hammon 30 describes a hospital mortality of 11%. This excludes, however, patients with injuries of extreme severity in whom the management policy was unoperated expectant. When these patients were included, an overall mortality of approximately 30% was demonstrated. This rate is very similar to the 26% mortality described in the series reporting on injuries sustained during the Lebanese conflict. The death rate reported by Aarabi for the Iran-Iraq War is considerably lower at 16.3%. However, this series concerns a selected patient population referred to a tertiary center for definitive treatment. The mean time from injury to arrival at this center was 49 hours (range, hours). Evidentiary Table: Outcome Distribution of Penetrating Brain Injury Military Injuries Authors, Year Data Class Description of Study Aarabi, Analysis of outcome in 435 patients in whom full data were available from a series of 690 patients with missile head wounds, sustained during the Iran-Iraq War, evacuated to Nemazee hospital in Shiraz. Inclusion Criteria 435 of 690 patients evacuated to Nemazee hospital with missile head wounds and dural penetration. 255 patients were excluded from the study (58 patients for inadequate information on admission GCS score, 64 patients for inadequate information about type of missile injury, and 132 patients for insufficient follow-up). Mean time from injury to arrival at Nemazee hospital was 49 hours (range, h). Time of Outcome and Outcome Distribution 364 of the 435 patients were followed for at least 6 mo. No. Dead No. Vegetative No. with Severe Disability No. with Moderate Disability No. with Good Recovery Total S48 August Supplement 2001

6 Prognosis in Penetrating Brain Injury Evidentiary Table. Continued. Authors, Year Data Class Description of Study Brandvold et al., Retrospective analysis of 113 patients evacuated to Rambam Maimonides Medical Center with penetrating craniocerebral injuries sustained in ongoing military hostilities in Lebanon. Essentially the same patient population as described by Levi et al. Levi et al., Retrospective study on 116 patients with penetrating injury, occurred during involvement of Israeli defense forces in Lebanon from June 1982 to June Essentially same patient population as described by Brandvold et al. Inclusion Criteria All patients with penetrating head injury evacuated to Rambam Maimonides Medical Center, the primary evacuation center for those wounded in Israeli-occupied territory. Median time from injury to arrival at Rambam Maimonides Medical Center emergency department was 2.0 h. Initially, the series contained 116 patients, but 3 were removed after review of charts, because there was no penetrating injury. All patients with penetrating head injury sustained during ongoing hostilities in Lebanon evacuated to Rambam Maimonides Medical Center. Series includes patients reaching the facility alive. Time of Outcome and Outcome Distribution Outcome at discharge GOS score discharge No. Dead No. Vegetative No. with Severe Disability No. with Moderate Disability No. with Good Recovery Total Volume 51 Number 2 S49

7 PROGNOSTIC INDICATORS The following variables will be discussed in the clinical prognostic indicator section. Unless otherwise stated, conclusions will discuss civilian prognostic indicators (Table 4). Process Of the original 65 articles, 24 were found to have adequate information for inclusion in this section on prognostic clinical features. Cross-referencing and expertise available among the authors added an additional 19 published studies. Among the 43 studies addressing clinical prognostic features, 5 report on the relation between age and outcome in civilian PBI. Nine studies address the relation between cause of injury and outcome in civilian gunshot wounds. Three studies report on the relation between mechanism of injury in the military setting and outcome. Three studies provide data on the prognostic value of the mode of injury. Eight studies investigate the relation between caliber and outcome. Table 4 Prognostic Indicators Variable Type Predictor Variables Demographics Age Epidemiology Cause of injury Mode of injury Caliber of weapon Systemic measures Hypotension Coagulation disturbances Respiratory distress Neurologic measures Level of consciousness and GCS score Pupillary examination Intracranial pressure Neuroimaging features Missile track Evidence of increased intracranial pressure Presence of hemorrhage or mass lesion S50 August Supplement 2001

8 Prognosis in Penetrating Brain Injury DEMOGRAPHIC FACTORS: AGE I. Conclusion Increasing age correlates with increased mortality after penetrating brain injury (Class ). II. Overview Age has been shown to be one of the most important prognostic factors in evaluating prognosis after nonpenetrating TBI. The mean age reported in 10 civilian series that describe age in PBI ranges from 25 to 35 years (Table 5). Four series, three of which exclude patients dead on arrival, report the age distribution in decades (Table 6). Reported series on PBI among patients admitted to the hospital, however, include relatively few patients in older age groups. In the series by Clark et al. 4 describing 76 patients with civilian gunshot wounds, only six patients above 65 years of age were included. Suddaby et al. 21 reported on 59 patients, and only three patients were over the age of 60. The majority of patients in those series that exclude patients dead on arrival are in the third decade (20 29 years), and only 15 patients from these two series totaling 236 patients are over the age of 50. The age distribution in those series that include patients dead at the scene of accident or dead on arrival have many more patients over the age of 50.. Process (Methodology) Of the 43 studies that address clinical prognostic features, 5 report on the relation between age and outcome in civilian PBI. In the military setting, PBI is largely restricted to young men; consequently, analysis of the prognostic value of age in military PBI is impractical. IV. Scientific Foundation Only two studies describe a significant relation between age and outcome. Both Kaufman et al. 7 and Siccardi et al. 23 show a significantly higher mortality among patients over 49 Table 5 Mean Age in Reported Series of Penetrating Brain Injury (Civilian) Authors, Year No. of Patients Mean Age (y) Range (y) Grahm et al., Hubschmann et al., Kaufman et al., Kaufman et al., Lillard, Nagib et al., Shaffrey et al., Shoung et al., mo 78 y Suddaby et al., Jacobs et al., mo 95 y Hernesniemi, Levy et al., mo 97 y years of age (odds ratio, ). Suddaby et al. 21 describe a slightly better outcome among patients below 40 years of age, but the finding was not statistically significant. Similarly, Jacobs et al. 9 report a slightly higher average age among patients dying (34.5 vs. 29), but the differences are not statistically significant. Hernesiemi 5 states that the influence of age is not statistically significant despite a trend towards unfavorable outcome and higher mortality among patients over 45 years of age. V. Summary Although age has been shown to be one of the most significant prognostic indicators in nonpenetrating TBI, its significance in PBI is not as clear. In civilian series, the mean age of patients ranges between the second and third decades. Most series have relatively few patients in the older age groups. Although patients older than 50 constitute a relatively low percentage of patients in most series, statistical analysis reveals that age greater than 50 is associated with an increased mortality. Table 6 Age Distribution in Penetrating Brain Injury Age (y) Series Including Patients Dead At the Scene of the Accident or Dead On Arrival Cavalieri et al., (n 179) (%) Siccardi et al., (n 314) (%) Series Excluding Patients Dead On Arrival Byrnes et al., (n 93) (%) Kaufman et al., (n 143) (%) Volume 51 Number 2 S51

9 VI. Evidentiary Table: Age Authors, Year Year of Study Description of Study Class Conclusions Siccardi et al., Retrospective study of 314 consecutive patients with civilian gunshot wounds, of whom 228 died at the scene of the accident. Mortality is significantly higher among patients over 49 y of age (p 0.01); OR: 3.45* Age Kaufman et January 1980 al., June 1982 Descriptive and prognostic (retrospective) study of 143 patients with civilian gunshot injuries. No probability analysis performed because series includes patients dead at the scene of the accident (n 314). Survival rates related to age, being good among patients under 10 y of age and poorer among those over 40. Age Suddaby et Retrospective analysis of 49 al., patients with low-velocity gunshot wounds. Pprior, 66%; PPV, 95%; OR, 11.0* (n 143). Slightly better outcome among patients below 40 y of age, but not statistically significant difference (p 0.88). Alive Age (y) GOS Discharge Unfavorable Favorable Mortality (Dead) Pprior, 61%; PPV, 67%; OR, 1.33 (n 49). Jacobs et al., January December 1992 Retrospective analysis of 57 patients in whom full data were available from 72 cases admitted from January 1990 to December 1992 with transcranial gunshot wounds. Focus of the study is on cost and consequences. Hernesniemi, Retrospective study of 90 patients admitted to neurosurgical department with penetrating gunshot wounds. Average age is slightly higher among patients dying (34.5 vs. 29). No statistically significant difference. Authors states that in this series the influence of age on outcome is not important. However, evidence presented suggests higher incidence of unfavorable outcome among patients over 45 y of age and mortality among patients over 60 (n 22) is 95%. GOS variable, long term: Age (y) Unfavorable Favorable Mortality: Age (y) Pprior, 79%; PPV, 88%; OR, 2.74 (n 90). OR, odds ratio; PPV, positive predictive value; Pprior, prior probability (the outcome distribution in the population studied). S52 August Supplement 2001

10 Prognosis in Penetrating Brain Injury EPIDEMIOLOGY: CAUSE OF INJURY I. Conclusion Suicide is correlated with a higher rate of mortality than other causes of PBI (Class II). II. Overview In civilians with PBI, gunshot wounds almost always are the cause of the injury. These gunshot wounds may be accidental, self-inflicted (suicide), or caused by assault/homicide. The incidence of suicide as a cause of civilian PBI varies considerably among the series; from 5% to 88%. Clark et al. 4 report suicide to be significantly more common in Caucasians. An overview of the incidence of suicide in civilian gunshot wounds is given in Table 7. In military injuries, shrapnel/shell fragments and debris predominate over missile wounds. In the series reported by Aarabi, 25 86% of injuries were caused by shell fragments. The series reported by Brandvold et al. 26 reported 82% higher mortality after gunshot wounds then in patients with shrapnel wounds.. Process (Methodology) Among the 43 studies that address clinical prognostic features, 9 report on the relation between cause of injury and outcome among civilians with gunshot wounds. Three studies report on the relation between mechanism of injury in the military setting and outcome. IV. Scientific Foundation Civilian Injuries Of the nine studies reporting on the results between suicide and outcome, seven describe a higher mortality and poorer outcome for patients who attempted suicide compared with patients injured accidentally or by assault. In the series reporting a relation between suicide and mortality, odds ratios varying from 1.63 to 5.83 were calculated. In a smaller retrospective series of 56 patients, no such relation could be found. 20 In the series reported by Cavalieri et al., including patients dead at the scene, no relation was found between suicide and mortality. 22 Although mortality is higher in suicide than in other causes, there remains the possibility that in some centers, suicide is treated less aggressively than other causes of PBI. In one survey of attitudes of neurosurgeons, over 60% of neurosurgeons responding stated that suicide patients were much less likely to receive aggressive resuscitation efforts Table 7 Reported Incidence of Suicide As Cause of Injury in Civilian Gunshot Wounds Authors, Year No. Suicide/Total No. Percentage Suicide Cavaliere et al., / Clark et al., /76 46 Grahm et al., / Hernesniemi, /90 87 Jacobs et al., /57 72 Kaufman et al., / Levy et al., /31 19 Lillard, /83 18 Mancuso et al., /40 13 Nagib et al., /55 38 Shaffrey et al., /62 60 Shoung et al., /56 63 Siccardi et al., / Suddaby et al., /49 88 than other types of patients. 31 Therefore, suicide may be a significant indicator of poor outcome not because there are some physiologic variables that relate to poor outcome, but because this method of injury changes the behavior or expectations of the clinicians attending the patient. It is clear from the clinical psychology literature that patients who recover from attempted suicide frequently report relief over survival and express gratitude to their caregivers. 32 Thus, the finding that suicide has been shown to be a predictor of poor outcome should not deter clinicians from intervening. Military Injuries Aarabi 25 reports no statistically significant difference in outcome between gunshot wounds and shell injuries. Brandvold et al., 26 however, report a statistically highly significant increased mortality among patients with gunshot wounds, with an odds ratio of towards mortality. These divergent results may be explained by patient selection. In the series reported by Aarabi, 25 the mean time to admission was 49 hours. Consequently, patients with more severe injuries may have been excluded. V. Summary Although the reported incidence varies among series, suicide is a major cause of PBI among civilians, predominantly because of gunshot wounds to the head. Suicide is associated with higher mortality than other causes of PBI. In the military setting, PBI is caused mainly by shell and shrapnel fragments, whereas military gunshot wounds are associated with poorer outcomes. Volume 51 Number 2 S53

11 VI. Evidentiary Table: Cause of Injury Authors, Year Years of Study Description of Study Class Conclusions Civilian Grahm et al., Prospective study over an 18-mo period of 100 patients with civilian gunshot wounds. II Prognostic analysis on caliber, GCS, pupils, OCR, bihemispheric versus unilateral injury, and cause of injury. Prognosis is poorer among patients who attempted suicide. GOS discharge or later a : Unfavorable Favorable Suicide Assault/unknown/other Pprior, 68%; PPV, 75%; OR, Lillard, Retrospective analysis in 5-y period, describing 83 patients with PBI. Nagib et al., Retrospective study of prognosis and management of civilian gunshot wounds during a 5-y period from Retrospective study of 314 Siccardi et al., consecutive patients with civilian gunshot wounds, of whom 228 died at the scene of the accident. Shoung et Retrospective study on 56 patients al., with craniocerebral gunshot wounds. Mortality: Suicide Assault/unknown/other Pprior, 60%; PPV, 70%; OR, b Predictors of outcome: level of consciousness, cause of injury, caliber, missile track, multilobar involvement. Poorer prognosis among patients who attempted suicide. Suicide 8 7 Assault/accident Pprior, 43%; PPV, 53%; OR, Related to poorer outcome: suicide level of consciousness, multilobar unilateral injury, bilateral injury. GOS score at 1 y: GOS score of 1/2 GOS score of 3/4/5 Suicide 16 5 Other Pprior, 53%; PPV, 76%; OR, b Mortality: Suicide 15 6 Other Pprior, 51%; PPV, 71%; OR, 4.03 b (n 55). Mortality is higher in suicide, age 49, bilateral lesions or multilobar unilateral lesions, intraventricular hemorrhage. Mortality in attempted suicide is 95%; in other causes, 88%. Overall mortality is 53.6%. In cases of attempted suicide, no significant difference in mortality in respect to overall population. Suicide Accident/assault 13 8 Prior, 54%; PPV, 49%; OR, 0.58 (n 56). S54 August Supplement 2001

12 Prognosis in Penetrating Brain Injury VI. Continued Authors, Year Years of Study Description of Study Class Conclusions Hernesniemi, Retrospective study on 90 patients admitted to neurosurgical department with penetrating gunshot wounds. Prognostic factors analyzed: age, level of consciousness, caliber, missile track, cause of injury. High percentage of suicide in this series. Outcome in suicide is unfavorable. GOS score at 1 mo or later: Unfavorable Favorable Suicide 69 9 Other cause 9 3 Pprior, 87%; PPV, 88%; OR, Jacobs et al., January December 1992 Cavaliere et al., Retrospective analysis of 57 patients in whom full data were available from 72 cases admitted from January 1990 to December 1992 with transcranial gunshot wounds. The study focuses on cost and consequences. Retrospective study of 178 patients with civilian gunshot wounds. Mortality: Suicide Other cause 7 5 Pprior, 79%; PPV, 82%; OR, 3.27 (n 90). Relation to mortality of Caucasian race, GCS score 4, lateral trajectory. High percentage of suicide as cause of injury (72%). Suicide 35 6 Other cause 8 8 Pprior, 75%; PPV, 85%; OR, b High overall mortality (93%). Mortality even higher in lesions crossing midline and massive intraventricular hemorrhage. No relation between suicide and mortality. Kaufman et January 1980 al., June 1982 Descriptive and prognostic (retrospective) study on 143 patients with civilian gunshot injuries. Suicide Homicide 41 2 No probability analysis performed because series included patients dead on arrival. OR, 0.83 (n 169). Cause of injury: assault 57%, suicide 43%. Mortality higher in attempted suicide Suicide Assault Pprior, 66%; PPV, 78%; OR, 2.75 b (n 143). Military Arabi, February 1981 August 1988 Analysis of outcome in 435 patients in whom full data were available from a series of 690 patients with missile head wounds, sustained during the Iran-Iraq War, evacuated to Nemazee hospital in Shiraz. Outcome: GWS Shell Bad Pprior, 18%; PPV, 22%; OR, 1.3. Good Volume 51 Number 2 S55

13 VI. Continued Authors, Year Years of Study Description of Study Class Conclusions Brandvold et al., June 1985 Retrospective analysis of 113 patients evacuated to Rambam Maimonides Medical Center with penetrating craniocerebral injuries sustained in ongoing military hostilities in Lebanon. Levi et al., Retrospective study with penetrating injury, occurred during involvement of Israeli defense forces in Lebanon from June 1982 to June Essentially same patient population as described under Brandvolt et al. GOS score at discharge: Unfavorable Favorable GSW 14 4 Shrapnel injury Pprior, 38%; PPV, 78%; OR, 8.26 b Mortality: GSW 13 5 Shrapnel injury Pprior, 28%; PPV, 72%; OR, * GOS score at discharge: Unfavorable Favorable GSW 13 2 Shrapnel injury Pprior, 37%; PPV, 87%; OR, 16.* Mortality: GSW 12 3 Shrapnel injury Pprior, 28%; PPV, 80%; OR, * OCR, oculocardiac reflex; Pprior, prior probability (the outcome distribution in the population studied); PPV, positive predictive value; OR, odds ratio; GSW, gunshot wound. a Outcome was determined at discharge; among patients who were vegetative or severely disabled at discharge, a mean follow-up of 9 mo was performed. b Indicates that the confidence interval for the odds ratio does not include 1. S56 August Supplement 2001

14 Prognosis in Penetrating Brain Injury EPIDEMIOLOGY: MODE OF INJURY I. Conclusion Perforating injuries correlate with a poorer outcome when compared with either penetrating or tangential brain injuries (Class ). II. Overview PBIs are differentiated into the following categories: penetrating (a foreign object penetrates skull and dura and remains lodged within the intracranial cavity); tangential (a foreign object glances off the skull, often driving bone fragments into the brain); and perforating (a through-andthrough injury, characterized by entry and exit wounds). Debris and bone fragments commonly are present within the brain in this latter type of injury. Differentiation as to the mode of injury (penetrating/ tangential/perforating) is found mainly in military series. The absence of reports of tangential injuries in the civilian population is remarkable. A possible explanation could be that such injuries are not recognized as a penetrating type of brain injury, but are classified as compound depressed skull fracture and included in series of patients with depressed skull fractures. The differentiation between a compound depressed skull fracture caused by a falling object and a tangential type of PBI caused by a projectile appears to be arbitrary. Major differences exist between military series concerning the incidence of penetrating, tangential, and perforating injuries. Aarabi 25 describes penetrating injuries in 266 of 435 (61%), tangential injuries in 128 of 435 (29%), and perforating injuries in 41 of 435 (9%). In the series described by Brandvold et al., 26 the majority of injuries are classified as penetrating; only one patient having a perforating injury, and 10 of 113 (9%) tangential. No studies report on the relation between mode of injury and other features. However, it seems that a relation between a perforating injury and a missile wound of relatively high velocity can be inferred from the ballistic data and the geometry of the calvarium.. Process (Methodology) Among the 43 studies addressing clinical prognostic features, 3 provided data on the prognostic value of the mode of injury. IV. Scientific Foundation Aarabi 25 reports lower mortality and more favorable outcome in patients with a tangential injury, though this was not statistically significant. Brandvold et al. 26 report no significant differences between outcome in penetrating and tangential types of injury. On analysis of the results reported, however, a significantly lower mortality can be shown in tangential types of injuries, with an odds ratio of 3.58 when compared with penetrating injuries. Aarabi 25 reports a significantly poorer outcome in perforating types of injuries. Odds ratios of 4.18 (unfavorable outcome) and 3.06 (mortality) can be calculated in the presence of perforating injuries when compared with the other modes of injury. The very low incidence (only one case) of penetrating injuries in the series reported by Brandvold et al. 26 and Levi et al. 27 limits their relevance. V. Summary Mode of injury can be divided into perforating, penetrating, and tangential. The predictive value of the mode of injury has been studied mainly in the military setting. Although the incidence of these modes of injury varies among studies, perforating injuries are associated with the highest mortality. There is a trend toward higher mortality in penetrating injuries compared with tangential injuries, although no study has shown a statistically significant increase in mortality. VI. Evidentiary Table: Mode of Injury Authors, Year Years of Study Description of Study Class Conclusions Arabi, February 1981 August 1988 Analysis of outcome in 435 patients in whom full data were available from a series of 690 patients with missile head wounds, sustained during the Iran-Iraq War, evacuated to Nemazee hospital in Shiraz. Brandvold et al., June 1985 Retrospective analysis of 113 patients evacuated to Rambam Maimonides Medical Center with penetrating craniocerebral injuries sustained in ongoing military hostilities in Lebanon. Levi et al., Retrospective study on 116 patients with penetrating injury, occurred during involvement of Israeli defense forces in Lebanon from June 1982 to June Essentially same patient population as described under Brandvolt et al. a Indicates that the confidence interval for the odds ratio does not include 1. Not related to outcome: side of injury, retention of foreign material, type of projectile. Related to outcome: mode of injury and GCS admission. Poor outcome in 19.9% of patients with penetrating injury; 15.6% in tangential injury and 48.8% in perforating injury. The mode of injury was an important factor in outcome, with shrapnel and tangential wounds faring better than gunshot wounds. Parameters related to outcome: GCS score, mechanism of injury, intracerebral track, multilobar involvement, transventricular injury, hypotension. GOS score at discharge. Volume 51 Number 2 S57

15 EPIDEMIOLOGY: CALIBER OF WEAPON I. Conclusion The effect of weapon caliber on outcome, independent of total kinetic energy, was not demonstrated in the published data. II. Overview The caliber of a missile is defined as the inner diameter of the barrel of the weapon in inches or millimeters. The amount of energy transmitted from a missile striking the head is dependent on its mass and velocity according to the formula: E 1/2 MV 2. Therefore, the velocity of a bullet will have a greater effect on the transmission of kinetic energy to the brain than its mass or caliber.. Process (Methodology) Among the 43 studies addressing clinical prognostic features, 8 investigated the relation between caliber and outcome. IV. Scientific Foundation The relation between missile caliber and patient outcome is not clear. Some studies report an increased mortality with increasing caliber of the weapon. 4,6,17 The reported differences, however, are not statistically significant. The low incidence of larger caliber injuries often precludes definite conclusions. Grahm et al. 6 report a 75% mortality in rifle injuries compared with 67% in low-caliber pistol injuries. The number of patients with rifle injuries, however, is low (n 4). V. Summary Several studies have related outcome to the weapon caliber but none showed statistical significance. Various factors complicate the analysis of the effect of caliber on the extent of injury and outcome. There is a large preponderance of low-caliber injuries in civilian gunshot wounds of the brain. In civilian gunshot wounds, suicide is a common cause of injury. This cause usually represents a close contact type of injury with high initial velocity of the missile and variable blast effect. Velocity varies significantly with a given caliber, depending on the circumstances surrounding the injury. The extent of damage is related not only to caliber and velocity but to increased tissue damage that may result from missile fragmentation or a tumbling effect (yaw) resulting from the initial impact. VI. Evidentiary Table: Caliber of Weapon Authors, Year Year of Study Description of Study Class Conclusions Lillard, Retrospective analysis in 5-y period, describing 83 patients with PBI. Shoung et al., Retrospective study on 56 patients with craniocerebral gunshot wounds. Clark et al., January 1984 August 1985 Retrospective analysis of 76 patients with civilian craniocerebral gunshot wounds. Hernesniemi, Retrospective study on 90 patients admitted to neurosurgical department with penetrating gunshot wounds. Mortality increases with caliber of weapon; shotgun wounds are particularly devastating because of blast effect and widespread tissue disruption. Caliber Shotgun 5 3 Highest mortality (83%) in hunting rifle injuries. Otherwise, no relation to caliber or type of weapon. Caliber 6.35-mm mm Hunting rifle 5 1 In this series, preponderance of small-caliber injuries (70.9%). Tendency for patients wounded with larger caliber weapons to worse outcome: 78% mortality vs. 54.5% in smaller caliber. No significant correlation between caliber of weapon and outcome. Caliber Unfavorable Favorable Shotgun Airgun S58 August Supplement 2001

16 Prognosis in Penetrating Brain Injury VI. Continued Authors, Year Year of Study Description of Study Class Conclusions Aldrich et al., January September 1987 Prospective study on prognosis in 151 patients with civilian gunshot injuries from the Traumatic Coma Data Bank. Nagib et al., Retrospective study on prognosis and management of civilian gunshot wounds during a 5-y period from 1978 to Grahm et al., Prospective study over an 18-mo period of 100 patients with civilian gunshot wounds. Siccardi et al., Retrospective study of 314 consecutive patients with civilian gunshot wounds, of whom 228 died at the scene of the accident. I All patients with large-caliber injuries died (n 4). However, there was no significant relation between caliber of weapon and mortality. II Caliber Caliber of bullet does not have prognostic value. Caliber Unfavorable Favorable Tendency to higher percentage of unfavorable outcome in large caliber injuries. GOS score at discharge or later b : Caliber Unfavorable Favorable , 9-mm magnum, , rifle The authors state that no statistically significant difference in the mortality rate exists when comparing bullets of a caliber of 7.65 mm or less with heavier missiles. Extracting data presented in the patients wounded with larger caliber bullets. Caliber 7.65-mm or less mm 22 1 a Indicates that the confidence interval for the odds ratio does not include 1. b Outcome was determined at discharge; among patients who were vegetative or severely disabled at discharge, a mean follow-up of 9 mo was performed. Volume 51 Number 2 S59

17 SYSTEMIC MEASURES: HYPOTENSION I. Conclusion Hypotension is associated with increased mortality (Class ). II. Overview Hypotension is an important secondary insult after traumatic brain injury. Hypotension is defined as a systolic blood pressure of 90 mm Hg. In nonpenetrating TBI, hypotension is frequently observed, both at the scene of the accident 33 and at hospital admission. 34 In these studies, the occurrence of hypotension is significantly related to poorer outcome. Four studies report on the incidence of hypotension in civilian gunshot wounds to the brain. 7,10,11,21 Incidences are summarized in Table 8. The percentage of patients with hypotension at admission varies considerably from 10% to 50%. The studies are unclear about how many patients with hypotension at admission were moribund. In the military series of Levi et al., 27 hypotension occurred in 116 patients (19%).. Process (Methodology) Among the 43 studies that addressed clinical prognostic features, 6 (5 civilian, 1 military) described the relation between hypotension in PBI and outcome. IV. Scientific Foundation In 480 PBI patients, Kaufman et al. 11 report a statistically significant correlation between occurrence of hypotension Table 8 Incidence of Hypotension in Penetrating Brain Injury Authors, Year No. Hypotension/ Total No. Percentage With Hypotension Aldrich et al., / Kaufman et al., / Kaufman et al., / Suddaby et al., /49 10 and mortality. In the overall population, mortality was 34%, whereas in the presence of hypotension it was 62%. Two studies 7,13 report higher mortality among patients with hypotension, but do not report results of statistical significance. Aldrich et al., 10 analyzing civilian gunshot injuries in 151 patients, report a higher mortality among patients with hypotension, although statistical significance could not be demonstrated (p 0.6). Although this paper is a Class I study, it showed no difference in outcome when comparing those who experienced hypotension with those who were normotensive because one third of the study s patients were excluded. Conversely, in the study by Byrnes et al., 13 higher mortality was noted among patients with an increased blood pressure at admission ( 150 mm Hg). V. Summary Hypotension correlates with increased mortality (Class ). VI. Evidentiary Table: Hypotension Authors, Year Years of Study Description of Study Class Conclusions Byrnes et al., August April 1973 Review article of 93 civilian penetrating craniocerebral missile injuries in civil disturbances in Northern Ireland. Patients with PBI often had a low SBP ( 90 mm Hg), because of bleeding from scalp, cerebral arteries, or dural sinuses. Hypotension and hypertension ( 150 mm Hg) are related to poorer outcome; no exact numbers reported. Kaufman et al., January June 1982 Descriptive and prognostic (retrospective) study of 143 patients with civilian gunshot injuries. Dead (%) Alive (%) SBP 90 mm Hg SBP mm Hg SBP 150 mm Hg 93 7 Hypotension (SBP 90 mm Hg) was noted in 71 of 143 patients (50%) and related to poorer outcome. SBP 90 mm Hg SBP 90 mm Hg Pprior, 66%; PPV, 77%; OR, 2.75 a (n 143). S60 August Supplement 2001

18 Prognosis in Penetrating Brain Injury VI. Continued Authors, Year Years of Study Description of Study Class Conclusions Kaufman et al., Two series of patients with civilian gunshot wounds to the brain are described. Emphasis of the study is on patients with admission GCS scores of 3 5. The first series (series A) evaluates treatment results in 190 patients presenting over a 6-y period to the University of Southern California. These patients were subjected to a conservative approach. The second series (series B) describes 480 patients admitted to Cook County Hospital in Chicago from 1983 through Full data are reported only on patients with GCS scores of 3 5 (n 146). Levi et al., Retrospective study of 116 patients with penetrating injury that occurred during involvement of Israeli defense forces in Lebanon from June 1982 to June Essentially same patient population as described under Brandvolt et al. Aldrich et al., January September 1987 Prospective study on prognosis in 151 patients with civilian gunshot injuries from the Traumatic Coma Data Bank. Data on hypotension are given for the second series reported in this manuscript. Overall mortality in this series was 34%. Hypotension occurred in 10% of patients and was related to poorer outcome (mortality 62%), statistically significant (p 0.005). SBP 90 mm Hg SBP 90 mm Hg Pprior, 34%; PPV, 62%; OR, 3.68* (n 480 b ). Overall mortality in this series was 26%. Hypotension (SBP 90 mm Hg) was noted in 19% of patients from whom data on blood pressure were available. Mortality among patients with hypotension was 67% (12 of 18). I Hypotension was noted in 43% of patients; mortality was higher among patients with hypotension, but differences were not statistically significant. Suddaby et al., Retrospective analysis of 49 patients with lowvelocity gunshot wounds. SBP 90 mm Hg 37 7 SBP 90 mm Hg Pprior, 80%; PPV, 84%; OR, 1.64 (n 103). Hypotension (SBP 90 mm Hg) was seen in 5 of 48 patients (10%). Mortality was slightly higher in the presence of hypotension, but differences were not statistically significant. c SBP 90 mm Hg 4 1 SBP 90 mm Hg Pprior, 60%; PPV, 80%; OR, 2.88 (n 48). SBP, systolic blood pressure; Pprior, prior probability (the outcome distribution in the population studied); PPV, positive predictive value; OR, odds ratio. a Indicates that the confidence interval for the odds ratio does not include 1. b Data derived from article. c Mistake in original table of the article; date derived by default. Volume 51 Number 2 S61

19 SYSTEMIC MEASURES: COAGULOPATHY I. Conclusion Coagulopathy is associated with increased mortality, particularly at lower levels of the GCS (Class ). II. Overview Coagulopathy may complicate PBI. Coagulation studies available include prothrombin time, thrombin time, partial thromboplastin time, fibrinogen level, fibrin degradation products, and platelet count. Subsequent complications, such as intracerebral hematoma, may increase mortality.. Process Among the 43 studies that addressed clinical prognostic features, 2 investigated the occurrence of coagulation disturbances in PBI and its relation to outcome. IV. Scientific Foundation Two studies investigated the occurrence of coagulation disturbances in PBI and investigated their relationship to outcome. Kaufman et al. 7 investigated screening tests for disseminated intravascular coagulation with six tests: prothrombin time, partial thromboplastin time, thrombin time, fibrinogen level, fibrin split products, and platelet count. The study, however, has various flaws, precluding definite conclusions. In 47 of 143 patients, no coagulation tests were performed. In the remaining patients, the number of tests performed per patient varied considerably. Nevertheless, among nonsurvivors, 94% had at least one abnormal test versus 71% among survivors. The average hospital stay for those who lived was much longer if at least one coagulation test was abnormal. In the presence of at least one abnormal test, mortality was 62% versus 20% when coagulation status was normal. Abnormalities most frequently noted on coagulation testing were long thrombin time (72%), high fibrinogen split products (68%), and low fibrinogen (53%). Shaffrey et al. 19 reported on the coagulation profiles in 43 of 62 patients with civilian gunshot wounds. Overall mortality was 55 and mortality among patients with coagulopathy was 80%. Coagulopathy was significantly associated with mortality (odds ratio, 50). On multivariate logistic regression analysis, coagulopathy status was retained together with GCS score at admission in the final model, predicting mortality at 7 days (p 0.001). Coagulopathy status is particularly significant in predicting mortality at lower levels of GCS score; the effect diminishes as GCS score increases. V. Summary Only two studies reported on the predictive value of coagulopathy. Although both show an increased mortality in the presence of coagulopathy, in neither was the entire population studied. Nevertheless, it may be concluded that the presence of coagulopathy increases the mortality risk, especially in association with a low GCS score. VI. Evidentiary Table: Coagulation Disturbances Authors, Year Years of Study Description of Study Class Conclusions Kaufman et January 1980 al., June 1982 Descriptive and prognostic (retrospective) study on 143 patients with civilian gunshot injuries. Patients were screened for DIC with six tests: PT, PTT, TT, fibrinogen level, fibrin split products, and platelet count. In 47 of 143 patients, no coagulation tests were performed. In the others, the number of tests performed per patient varied considerably. Nevertheless, among nonsurvivors, 94% had at least one abnormal test vs. 71% among survivors. Average hospital stay for those who lived was much longer if at least one test was abnormal. Abnormalities most frequently noted on coagulation tests were long TT (72%), high fibrin split products (68%), and low fibrinogin (53%). Shaffrey et al., Retrospective study of 62 patients with gunshot wounds; multivariate analysis of predictors of mortality performed. Coagulation profiles were obtained on 43 of 62 patients (18 died before coagulation studies were obtained). At least one test abnormal Normal coagulation status 3 12 Pprior, 55%; PPV, 62%; OR, 6.5.* Overall mortality is 55%. 80% of patients with coagulopathy died. Coagulopathy was significantly associated with mortality, with an odds ratio of 50. On multivariate logistic regression analysis, coagulopathy status was retained, together with GCS score, on admission in the final model, predicting mortality at 7 days (p 0.001). Coagulopathy status particularly differentiates in predicting mortality at lower levels of GCS score; the effect diminishes as GCS score increases. DIC, disseminated intravascular coagulation; PT, prothrombin time; PTT, partial thromboplastin time; TT, thrombin time; Pprior, prior probability (the outcome distribution in the population studied); PPV, positive predictive value; OR, odds ratio. S62 August Supplement 2001

20 Prognosis in Penetrating Brain Injury SYSTEMIC MEASURES: RESPIRATORY DISTRESS I. Conclusion Respiratory distress is associated with increased mortality in PBI (Class ). II. Overview Respiratory distress is variously defined in the PBI literature as a respiratory rate of less than 10, apnea, or respiratory depression; however, studies do not contain a clear and consistent definition for respiratory distress. A significant number of patients arrive at hospital already intubated, and the reasons for intubation in the field are often difficult to ascertain. For the purpose of this section, we have chosen a respiratory rate of less than 10 breaths/min or impairment of the patient s airway related to a level of coma that required intubation as the criteria for our definition of respiratory distress.. Process (Methodology) Among the 43 studies that address clinical prognostic features, 2 were found to contain significant data regarding the relation between respiratory distress and outcome. IV. Scientific Foundation Kaufman et al., 11 reporting on a series of 480 patients admitted to Cook County Hospital in Chicago from 1983 to 1992, describe apnea or respiratory depression in 89 patients (21%) on presentation at hospital. These patients had a 75% mortality and significantly poorer outcome than patients without respiratory distress (p 0.005). Jacobs et al. 9 reported that 57 of 72 patients with transcranial gunshot wounds had an odds ratio towards mortality of 7.3 when their respiratory rate was less than 10 breaths/min (p 0.004). V. Summary Although a standardized definition of respiratory distress is lacking, the presence of respiratory distress on arrival at hospital (respiratory rate 10 breaths/min or airway impairment requiring intubation) is associated with an increased mortality in PBI. VI. Evidentiary Table: Respiratory Distress Authors, Year Years of Study Description of Study Class Conclusions Kaufman et al., Two series of patients with civilian gunshot wounds to the brain are described. Emphasis of the study is on patients with admission GCS score of 3 5. The first series (series A) evaluates treatment results in 190 patients presenting over a 6-y period to the University of Southern California. These patients were subjected to a conservative approach. The second series (series B) describes 480 patients admitted to Cook County Hospital in Chicago from 1983 through Full data are only reported on patients with GCS score of 3 5 (n 146). Jacobs et al., January 1990 December 1992 Retrospective analysis of 57 patients in whom full data were available from among 72 cases admitted from January 1990 to December 1992 with transcranial gunshot wounds. Focus of the article is on cost and consequences. Some data on prognostic predictive value of respiratory depression are reported in series B. Apnea or respiratory depression was documented in 89 patients (21%). These patients had a 75% mortality and significantly poorer outcome than patients without respiratory distress ( 2 : p 0.05). Respiratory depression plus No respiratory depression Dead Alive Pprior, 35%; PPV, 75%; OR, 8.98* (n 480). Univariate statistical analysis showed that race, mechanism, trajectory, and respiratory status were each associated with an approximate sevenfold increase in mortality. Odds ratio towards mortality for a respiratory rate 10 breaths/min was 7.3 (p 0.004). Pprior, prior probability (the outcome distribution in the population studied); PPV, positive predictive value; OR, odds ratio. * Indicates that the confidence interval for the odds ratio does not include 1. Volume 51 Number 2 S63

21 NEUROLOGIC MEASURES: LEVEL OF CONSCIOUSNESS AND GLASGOW COMA SCALE I. Conclusions In civilian patients, low GCS correlates with higher mortality and unfavorable outcome (Class I). In military injuries, fewer patients have a low GCS score. A strong correlation also exists between low GCS score and unfavorable outcome in military series (Class ). II. Overview Since the introduction of the GCS in 1974 by Teasdale and Jennett, 35 this score has become widely accepted as an objective measure of level of consciousness. Prior to this time, various grading systems, often ill defined, were used across the world. In published reports on PBI before 1975, most authors distinguish four grades of consciousness: alert, drowsy, purposeful reaction to pain, and deep coma. In nonpenetrating TBI, patients with a GCS score 8 are classified as having severe injuries, and patients with GCS scores of 9 of 12 are classified as moderate. Several studies focused specifically on those with severe injuries. 10,11 The distributions of the level of consciousness observed in the series reported are shown in Tables 9 and 10. In civilian series, some variation in the severity of injury exists, probably reflecting different admission referral policies. All series are characterized by the relatively high number of patients in grade 4 (approximately 35%) or with a GCS score of 3 to 5 (38 61%) in unselected series. Few patients are included with a GCS score of 6 to 8; the percentage of patients who may be considered in coma is 53% to 57%, and the percentage of patients with a GCS score 8 is 37% to 70%. The percentage of patients with milder injuries (GCS score of 13 15) in various series ranges from 15% to 51%. In the series reported by Kennedy et al., 15 51% of patients had a GCS score of 14 to 15. The incidence of various categories of GCS score reported in military series is different from civilian series. The percentage of patients with GCS score of 3 to 5 is lower (11% and 35%) and the percentage of patients with GCS score of 13 to 15 is higher (63% and 45%). No results of interobserver reliability in the use of GCS in PBI have been reported, but there is no reason to suppose that results might be different from those observed in nonpenetrating TBI. Some studies in nonpenetrating TBI demonstrated a high intra- and interobserver reliability in the use Table 9 Level of Consciousness in Penetrating Brain Injury Byrnes, 1974 (n 89) Byrnes et al., (n 89) Hubschmann et al., (n 82) Shoung et al., (n 56) Alert (%) Drowsy (%) Reacts to pain (%) Deep coma (%) of GCS, and also when applied by medical personnel with limited experience However, a study by Marion et al. 40 suggested considerable variability in how the GCS score was assessed. Furthermore, the use of paralytic agents and drugs or alcohol can significantly compromise the reliability of the GCS.. Process (Methodology) Among the 43 studies that addressed clinical prognostic features, 8 report on the relation between level of consciousness and outcome in the civilian setting, 11 on relation between GCS score at admission and outcome, and 4 on the relation between level of consciousness or GCS score and outcome in military injuries. IV. Scientific Foundation Civilian Series All civilian studies reporting on level of consciousness or GCS score in PBI show a strong correlation between GCS score and outcome. Mortality among patients in deep coma is uniformly high, and a mortality of 100% is reported in three series. In the civilian series that report both GCS and GOS, only 5 of 490 patients with GCS score of 3 to 5 had a favorable outcome. Among patients alert at hospital admission, mortality is low (0 10%), and among patients with GCS score of 13 to 15, mortality and unfavorable outcome was 0%. Prognosis for patients in a coma (GCS score 8) are summarized in Tables 11 13,14,17,20,41 and ,25,18,26 Military Series As in civilian injuries, the three military studies reporting on GCS score and outcome in PBI show a strong correlation between GCS score and outcome. Both in the Iranian series 25 and in the Israeli series 26, a positive predictive value of 61% was calculated among patients with a GCS score 8. Among Table 10 GCS Score in Penetrating Brain Injury GCS Score Aldrich et al., (n 151) Grahm et al., (n 100) Kaufman et al., (n 141) Mancuso et al., (n 40) Nagib et al., (n 55) 3 5 (%) (%) (%) (%) S64 August Supplement 2001

22 Prognosis in Penetrating Brain Injury Table 11 Level of Consciousness: Predictive Value of Coma to Mortality Authors, Year Years of Study Prediction of Mortality of Series Mortality (Deep Coma) Odds Ratio Byrnes et al., August 1969 April 1973 Mortality Carey et al., September 1968 September 1969 Mortality Hubschmann et al., Mortality Lillard, Mortality Shoung et al., Mortality Table 12 GCS Score versus Outcome: Analysis of GCS Score < 8 versus GCS Score > 9 Authors, Year Years of Study Prediction of Mortality of Series Mortality (GCS Score 8) Odds Ratio Arabi, Unfavorable outcome Mortality Brandvold et al., Unfavorable outcome Mortality Grahm et al., ? Unfavorable outcome Mortality Kaufman et al., Mortality Mancuso et al., Mortality a Nagib et al., Unfavorable outcome Mortality a a 0 set to 1 calculation. patients with a GCS score of 13 to 15, a positive predictive value to survival was calculated as 97% and 94%. The results of the study reported by Aarabi 25 should, however, be interpreted with care because it concerns a selected series of patients with a mean admission time of 49 hours after injury (range, hours), which is different from most civilian series. V. Summary In civilian series, large proportions of patients are admitted with a GCS score of 3 to 5. However, in military series, fewer patients have a low admission GCS score, and more have a GCS score of 13 to 15. Nevertheless, in both civilian and military PBI, the level of consciousness as assessed by the GCS is the strongest indicator of outcome and mortality. VI. Evidentiary Table: Level of Consciousness Authors, Year Year of Study Description of Study Class Conclusions Byrnes et al., August April 1973 Review study of 93 civilian penetrating craniocerebral missile injuries in civil disturbances in Northern Ireland. Features related to outcome (dead/alive): hypotension, level of consciousness, pupil reactivity. Alert 3 23 Drowsy/irritable 6 12 Only reacts to pain 19 5 Deep Coma 25 0 Coma 44 5 Coma 9 35 Pprior, 57%; PPV, 88%; OR, 34.2 a (n 93). Volume 51 Number 2 S65

23 VI. Continued Authors, Year Year of Study Description of Study Class Conclusions Carey et al., September 1968 September 1969 Description on neurosurgical treatment and treatment results of 131 U.S. soldiers sustaining craniocerebral missile wounds in Vietnam. 25 patients appeared to have fatal injuries and received no further care; 19 had only extracranial injuries. Results are reported on the operative treatment of 89 patients with brain penetration. Hubschmann et Retrospective study of 82 civilian al., patients with craniocerebral gunshot injuries. Lillard, Retrospective analysis in 5-y period, describing 83 patients with penetrating brain injury. Overall mortality is 12%. A progressive raise in the mortality rate was seen, with deterioration of overall neurologic function. Grade I, alert 3 33 Grade II, drowsy 6 12 Grade, purposeful response to 4 14 pain Grade IV, decerebrate, or 3 4 unresponsive to pain Coma 7 18 Coma 9 35 Pprior, 12%; PPV, 28%; OR, 1.51 (n 89). Related outcome: level of consciousness presence/ absence release phenomena. However, no data on this parameter described. Level of consciousness related to outcome. Grade I, alert 0 14 Grade II, obtundation 3 18 Grade, unresponsive to all but 4 4 noxious stimuli Grade IV, comatose 39 0 Coma 43 4 Coma 3 32 Pprior, 57%; PPV, 91%; OR, a (n 82). Predictors of outcome: level of consciousness, cause of injury, caliber, missile track, multilobar involvement. Poorer prognosis among patients with attempted suicide. Mortality is related to the level of consciousness. Class I, alert without neurologic 0 13 deficit Class II, alert with neurologic deficit 1 9 Class, somnolent or confused 7 18 Class IV, decreased level of 13 5 consciousness, responding with purpose to painful stimuli Class V, comatose with no or only 3 4 decerebrate or decorticate response (GCS score of 3 5) or unresponsive to pain Coma 28 7 Coma 8 40 Pprior, 43%; PPV, 80%; OR, 20 a (n 83). S66 August Supplement 2001

24 Prognosis in Penetrating Brain Injury VI. Continued Authors, Year Year of Study Description of Study Class Conclusions Shoung et al., Retrospective study on patients with craniocerebral gunshot wounds. Hernesniemi, Retrospective study on 90 patients admitted to neurosurgical department with penetrating gunshot wounds. Level of consciousness related to mortality. Grade I, alert, no neurologic deficit 2 13 Grade II, drowsy 3 8 Grade, coma 9 5 Grade IV, deep coma with brain stem 16 0 clinical impairment Coma 25 5 Coma 5 21 Pprior, 45%; PPV, 83%; OR, 21 a (n 56). Related to outcome are level of consciousness and missile tract. GOS variable, long term: Unfavorable (Dead) Favorable 5 Alive Alert 10 3 Pprior, prior probability (the outcome distribution in the population studied); PPV, positive predictive value; OR, odds ratio. a Indicates that the confidence interval for the odds ratio does not include 1. VII. Evidentiary Table: GCS Score At Admission Authors, Year Years of Study Description of Study Class Conclusions Arabi, February 1981 August 1988 Analysis of outcome in 435 patients in whom full data were available from a series of 690 patients with missile head wounds, sustained during the Iran-Iraq War, evacuated to Nemazee hospital in Shiraz. Aldrich et al., January 1984 September 1987 Prospective study on prognosis among 151 patients with civilian gunshot injuries from the Traumatic Coma Data Bank. I GCS score at admission is related to outcome (GOS) and focal neurologic deficit. Selected series of patients with a mean admission time of 49 h after injury (range, h). Mortality: GOS score at 6 mo or more: Admission GCS score Unfavorable Favorable Probability analysis GCS score 8 vs. GCS score 8. Unfavorable outcome: Pprior, 21%; PPV, 62%; OR, a Mortality: Pprior, 16%; PPV, 51%; OR, a Overall mortality 88%. Related to outcome: GCS score at admission. GCS score after resuscitation and intensive care. Admissions GCS score GOS score at discharge: Unfavorable Favorable No probability analysis performed because patient population studied included only patients with GCS score 8or deterioration to that level within 48 h of injury. Volume 51 Number 2 S67

25 VII. Continued Authors, Year Years of Study Description of Study Class Conclusions Brandvold et al., June 1985 Retrospective analysis of 113 patients evacuated to Rambam Maimonides Medical Center with penetrating craniocerebral injuries sustained in ongoing military hostilities in Lebanon. Prognostic features related to outcome: mechanism of injury, GCS score, hypotension, CT scan characteristics, and infection. Mortality in GCS score of 2 4 is 80%, GCS score of 5 12 is 12%, GCS score of is 6%. GCS score: GOS score at discharge: Unfavorable Favorable Probability analysis GCS score 8 vs. GCS score 8. Pprior, 37%; PPV, 77%; OR, a Mortality: Pprior, 28%; PPV, 62%; OR, a Clark et al., January 1984 August 1985 Retrospective analysis of 76 patients with civilian craniocerebral gunshot wounds. Grahm et al., Not specified Prospective study over an 18-mo period of 100 patients with civilian gunshot wounds. GCS score: Probability analysis GCS score 8 vs. GCS score 8. Overall mortality 62%. Suicide more common in Caucasian males; homicide in blacks. No patients with an admission GCS score of 3 survived. Prognostic variables are GCS score, ventricular injury, bihemispheric wounds. Admission GCS score is lower among patients with poor outcome ( ) in comparison with patients with good outcome ( ). The difference is statistically significant (p 0.001). No exact data reported. II 58% of patients had an admission GCS score of 3 5; 8%, 6 8; 12%, 9 12; and 22%, Lower GCS score is associated with poorer prognosis. A relation was described between the admission GCS score and abnormalities in pupillary reactivity and/ or oculocephalic eye movements. GCS score GOS discharge or later b Unfavorable Favorable Probability analysis GCS score 8 vs. GCS score 8. Pprior, 68%; PPV, 97%; OR, 240. a Mortality: Pprior, 59%; PPV, 85%; OR, a GCS score S68 August Supplement 2001

26 Prognosis in Penetrating Brain Injury VII. Continued Authors, Year Years of Study Description of Study Class Conclusions Kaufman et January 1980 al., June 1982 Kaufman et al., Descriptive and prognostic (retrospective) study of 143 patients with civilian gunshot injuries. Two series of patients with civilian gunshot wounds to the brain are described. Emphasis of the study is on patients with admission GCS score of 3 5. The first series (series A) evaluates treatment results in 190 patients presenting over a 6-y period to the University of Southern California. These patients were subjected to a conservative approach. The second series (series B) describes 480 patients admitted to Cook County Hospital in Chicago from 1983 through Full data are reported only on patients with GCS score of 3 5 (n 146). January 1987 June 1982 Kennedy et Retrospective analysis of al., June 1990 prognostic features among 192 patients with gunshot and shotgun injuries. Levi et al., Retrospective study of June 1985 patients with penetrating injury; occurred during involvement of Israeli defense forces in Lebanon. Mancuso et Retrospective study of 40 al., patients with craniocerebral gunshot wounds. Mainly descriptive study. Overall mortality is 66%. Low GCS score is related to poor outcome, but a few unexpected recoveries are described among patients in whom a large debridement was performed. GCS score Probability analysis GCS score 8 vs. GCS score 8. Pprior, 66%; PPV, 87%; OR, 219. a Logistic regression analysis performed on series A shows admission GCS score, TSAH, pupils to be predictive of mortality. Predictive of morbidity are GCS score, bihemispheric injury, diffuse fragmentation. Mortality: GOS score at discharge: GCS score Unfavorable Favorable Logistic regression analysis on series B showed relation to outcome of GCS score, bihemispheric injury, pupil reactivity, and TSAH. Mortality: GOS score at discharge: GOS score at discharge Unfavorable Favorable No probability analysis performed because series includes only patients with GCS score of 3 5. Related to outcome: GCS score, brain extrusion, but interaction with GCS score, CT scan: single-lobe injury better than multiple-lobe injuries. Related to outcome: GCS score, mechanism of injury, intracerebral track, multilobar injury, transventricular injury, and hypotension. Poor outcome among patients with GCS score 8. GCS score Pprior, 40%; PPV, 84%; OR, 112 a,c Volume 51 Number 2 S69

27 VII. Continued Authors, Year Years of Study Description of Study Class Conclusions Retrospective study of prognosis Nagib et al., and management of civilian gunshot wounds during a 5-y period from 1978 to Related to poorer outcome: suicide, level of consciousness, multilobar unilateral injury, bilateral injury. GCS score GOS score at 1 y: Unfavorable Favorable Probability analysis GCS score 8 vs. GCS score 8. Pprior, 58%; PPV, 79%; OR, a Retrospective study of 62 Pprior, 51%; PPV, 74%; OR, a,c Shaffrey et al., patients with gunshot wounds; multivariate analysis of predictors of mortality is performed. Suddaby et Retrospective analysis of 49 al., patients with low-velocity gunshot wounds. January 1990 Jacobs et al., Retrospective analysis of December 1992 patients in whom fill data were available from 72 cases admitted from January 1990 to December 1992 with transcranial gunshot wounds. Focus of the study is on cost and consequences. Mortality: GCS score Overall mortality is 55%. In GCS score of 3, 87.5%; GCS score of 4 7, 56.3%; GCS score 7, 0%. Multivariate logistic regression analysis of clinical variables shows GCS score at admission plus coagulopathy status to be the best predictor of mortality at 7 days. Coagulopathy status makes a difference in predicted mortality at lower levels of GCS score, but its value diminishes as GCS score increases. CT scan parameters retained in the step-up logistic regression procedure are volume of contused brain and crossing of horizontal or sagittal plane. The odds ratio for admission GCS score of 3 to mortality is 266; for admission GCS score of 4 7, % were suicide cases. Overall mortality 61%. Related to poorer prognosis: GCS score 7 and nonreactive pupils. Mortality: GOS score at discharge: GCS scale Unfavorable Favorable Relation to mortality of Caucasian race, GCS score 4, lateral trajectory. Odds ratio to unfavorable outcome in GCS score 4is 33.7 (p 0.003). Pprior, prior probability (the outcome distribution in the population studied); PPV, positive predictive value; OR, odds ratio; TSAH, traumatic subarachnoid hemorrhage. a Indicates that the confidence interval for the odds ratio does not include 1. b Outcome was determined at discharge; among patients vegetative or severely disabled at discharge a mean follow-up of 9 mo was performed. c 0 set to 1 for calculation. S70 August Supplement 2001

28 Prognosis in Penetrating Brain Injury NEUROLOGIC MEASURES: PUPILLARY SIZE AND LIGHT REFLEX I. Conclusion The presence of bilateral fixed and dilated pupils is highly predictive of mortality (Class ). II. Overview An abnormal pupillary light reflex may be an indirect indicator of cerebral herniation or possible brain stem injury. Other causes of an abnormal pupillary response include direct orbital trauma. Measurement of pupillary reactivity includes measurement of the light reflex and size of the pupils. An overview of the incidence of pupillary abnormalities in series of patients with penetrating brain injury is given in Table 13. The series are characterized by a high percentage of patients with fixed and dilated pupils. In the PBI literature, no studies have been conducted that examine the interobserver or intraobserver reliability on scoring of pupillary light reflex and size equality of pupils. A high interobserver reliability has been reported in the assessment of ocular signs in coma. 42 In nonpenetrating TBI, a relation between absence of pupillary reactivity and obliteration of the basal cisterns, as documented by CT scan investigation, has been demonstrated. 43,44 Other authors reporting on nonpenetrating TBI show an association of pupillary reactivity with the GCS score, 45,46 focal lesions, 46 and a history of early hypoxic or hypotensive insults. 47 Among patients with PBI, Byrnes et al. 13 describe a relation between pupillary reactivity/size and level of consciousness. Similarly, Table 13 Distribution Pupils in Penetrating Brain Injury Byrnes et al., (%) Series B, Kaufman et al., (%) Suddaby et al., (%) Normal Unequal 23 9 Fixed dilated Table 14 Pupillary Reactivity/Size and GCS Score Alert Drowsy Reacts to pain Deep coma Normal Unequal Fixed, dilated Grahm et al. 6 report a relation between ocular signs and GCS score (Table 14).. Process (Methodology) Among the 43 studies that address clinical prognostic features, 5 report on the prognostic value of pupillary reactivity/size in PBI, all in civilian injuries. IV. Scientific Foundation Clinical studies investigating the prognostic value of pupillary light reactions have examined this parameter using various methodologies. No study of PBI has rigorously measured the pupil size and reactivity; none give exact definitions of pupillary inequality. Five studies report on the prognostic value of pupillary reactivity/size in PBI. All authors describe a relation between absence of pupillary reactivity and mortality. Kaufman et al., 7 studying 143 patients, and Shaffrey et al., 19 in a retrospective study of 62 patients, show a strong relation between unreactive pupils and outcome (odds ratio, for mortality in the presence of nonreactive pupils). Neither Suddaby et al. 21 nor Byrnes et al. 13 could demonstrate a statistically significant relation between inequality of pupils and outcome. V. Summary Although standardized definitions are lacking, the incidence of pupillary abnormalities ranges between one third and one half of patients admitted with PBI. Pupillary abnormalities appear to correlate with a patient s level of consciousness. Unequal or fixed and dilated pupils correlate with increased mortality. VI. Evidentiary Table: Pupillary Examination Authors, Year Years of Study Description of Study Class Conclusions Byrnes et al., August April 1973 Review of study of 93 civilians with penetrating craniocerebral missile injuries in civil disturbances in Northern Ireland. Pupillary reactivity is related to the level of consciousness and mortality. No multivariate analysis was performed to determine the relative importance of pupillary reactivity vs. level of consciousness. Pupillary reactivity/size: Fixed, dilated 24 1 Unequal 8 13 Normal Probability analysis: nonreacting vs. one or both reacting. Pprior, 52%; PPV, 96%; OR, 44. a Volume 51 Number 2 S71

29 VI. Continued Authors, Year Years of Study Description of Study Class Conclusions Kaufman et January 1980 al., June 1982 Kaufman et al., Descriptive and prognostic (retrospective) study of 143 patients with civilian gunshot injuries. Two series of patients with civilian gunshot wounds to the brain are described. Emphasis of the study is on patients with admission GCS score of 3 5. The first series (series A) evaluates treatment results in 190 patients presenting over a 6-y period to the University of Southern California. These patients were subjected to a conservative approach. The second series (series B) describes 480 patients admitted to Cook County Hospital in Chicago from 1983 through Full data are reported only on patients with GCS score of 3 5 (n 146). Shaffrey et al., Retrospective study of 62 patients with gunshot wounds; multivariate analysis of predictors of mortality is performed. Suddaby et Retrospective analysis of 49 patients with al., low-velocity gunshot wounds. Bilateral unreactive pupils strongly correlated to mortality. Both nonreacting nonreacting 4 4 Reactive pupils 8 30 Probability analysis: nonreacting vs. one or both reacting. Pprior, 70%; PPV, 99%; OR, a (n 116). Series A: logistic regression analysis showed two models to be the best predictors of mortality: GCS score plus SAH GCS score plus pupil (p ) No exact numbers reported. Series B: dilated pupils, nonreactive pupils, unequal pupils are all related to poor outcome. Only four patients with fixed bilaterally dilated pupils survived. In a linear logistic regression model, pupillary reactivity is retained at a significance of p Bilateral nonreactive pupils are strongly correlated to mortality. Mortality (%) Odds ratio Both nonreactive nonreactive Normal pupils 4.6 Absence of pupillary reactivity is correlated to unfavorable outcome. Unequal, reacting pupils are not related to outcome. Pupillary reactivity: Unfavorable Favorable Non-reactive 22 0 Pupils both reacting 7 18 Pprior, 62%; PPV, 100%; OR, 56.6 a,b Equality of pupils: Unfavorable Favorable Equal, reacting 5 11 Unequal, reacting 2 7 Pprior, prior probability (the outcome distribution in the population studied); PPV, positive predictive value; OR, odds ratio; SAH, subarachnoid hemorrhage. a Indicates that the confidence interval for the odds ratio does not exceed 1. b 0 set to 1 for calculation. S72 August Supplement 2001

30 Prognosis in Penetrating Brain Injury NEUROLOGIC MEASURES: INTRACRANIAL PRESSURE I. Conclusion High ICP is predictive of higher mortality (Class II). II. Overview Brain injury, whether from a closed or penetrating brain injury, is associated with increased ICP. Monitoring ICP can be performed using intracranial pressure monitors (intraparenchymal or subdural) or ventriculostomy. ICP values are used not only to treat elevated ICP primarily but also in conjunction with the mean arterial pressure, to maximize cerebral perfusion pressure. Most reports give incidental information regarding ICP and outcome 7,13 and the frequency of ICP monitoring is low in PBI. In the studies by Kaufman et al. 7 and Aldrich et al., 10 only 14% and 23% of patients, respectively, received ICP monitoring.. Process (Methodology) Among the 43 articles addressing clinical prognostic features, 1 study reported on the prognostic value of ICP monitoring and 1 other mentioned the use of ICP monitoring but did not report its prognostic value. IV. Scientific Foundation Only a few studies report on experience with intracranial pressure monitoring among patients with penetrating brain injury. Aldrich et al. 10 found the average ICP during the first 72 hours among patients who died was 62 mm Hg versus 31 mm Hg in those who survived. The percentage of time that the ICP was 20 mm Hg in the first 72 hours was likewise substantially higher in the group that died (75% vs. 26%). Linear regression analysis showed that high ICP was the second most important predictor of mortality. V. Summary Few reports provide data relating ICP and outcome. Although there are a number of methods for reporting ICP, available data suggest that elevated ICP within the first 72 hours predicts higher mortality. VI. Evidentiary Table: Intracranial Pressure Authors, Year Years of Study Description of Study Class Conclusions Aldrich et al., January 1984 September 1987 Kaufman et al., January June 1982 Prospective study on prognosis in 151 patients with civilian gunshot injuries from the Traumatic Coma Data Bank. Descriptive and prognostic (retrospective study on 143 patients with civilian gunshot injuries). II ICP was monitored in 35 patients; high ICP was associated with high mortality. Average ICP among patients dying during the first 72 h was 62 mm Hg vs. 31 mm Hg in survivors. Percentage of time that ICP was greater than 20 mm Hg in the first 72 h was likewise substantially higher in the group that died (75% vs. 26%). Linear regression analysis showed that high ICP was the second most important predictor of mortality. ICP monitors were inserted in 20 patients (14%). Among patients operated for debridement, ICP monitoring often revealed elevated pressures. No data on ICP or predictive value of ICP reported. Volume 51 Number 2 S73

31 FEATURES ON CT SCANNING Categories of Pathologic CT Scan Features The prognostic significance of various pathologic features found on CT scan examination will be discussed. These features have been divided into the following categories: Missile Track Bihemispheric lesions Multilobar injuries Ventricular involvement Evidence of increased intracranial pressure Cisternal effacement Midline shift Presence of hemorrhage or mass lesions Intracranial hematoma Intraventricular hemorrhage Mass lesion/contusions Subarachnoid hemorrhage FEATURES ON CT SCANNING Missile Track Bihemispheric lesions Multilobar injuries Ventricular involvement I. Conclusions Bihemispheric lesions relate to increased mortality (Class II). Compared with unilobar injuries, multilobar injuries are strongly associated with mortality (Class ). Injuries with ventricular involvement have an increased mortality rate (Class ). II. Overview CT scan examination permits an accurate detection of the path a missile has followed through the brain. The extent and location of resulting brain damage can be estimated, and the findings related to outcome. Aspects of the missile track analyzed for prognostic value include the presence of bihemispheric lesions, the number of planes crossed, multilobar involvement, and ventricular involvement. Bihemispheric injuries are defined as injuries in which the missile track crosses the midsagittal plane, causing injury to both cerebral hemispheres. The path of the missile can be described in reference to three planes: the midsagittal plane, the horizontal plane at the level of the third ventricle, and the midcoronal plane (Fig. 4). Multilobular involvement is defined as CT scan evidence of damage to more than one lobe of the cerebrum. Ventricular involvement refers to CT scan evidence of the missile path having penetrated the ventricular system. The reported incidence of bihemispheric lesions in the studies varies from 33% to 63%. The highest incidence is seen in those studies including patients dead at the scene of accident or Fig. 4. CT wound profile of a penetrating brain trauma in a young man. The entrance wound is in the right temporal region. Multiple retained bone and metal fragments are visible in the right hemisphere, and a small foreign body is seen in the trigone of the right lateral ventricle. There is extensive intracerebral and intraventricular hematoma in association with pneumocephalus. The projectile has crossed the sagittal, axial, and coronal planes of the cerebral hemispheres and after penetrating the parietal bone has been stopped by the scalp. This victim was comatose at admission and died within 2 hours. on arrival at the hospital. 22,23 It may, therefore, be concluded that bihemispheric lesions are more common in the most severely injured patients. Multilobar involvement in PBI is frequent, with a reported incidence varying from 59% to 89%. 8,21 Ventricular involvement is described less frequently in 37% of cases in the military series reported by Brandvold et al. 26 and in 57% of patients with civilian brain injuries described by Clark. 4 No studies report a direct relation between the missile track and other prognostic features. The study by Nagib et al., 8 however, states that all patients with bilateral lesions had a GCS score 6. An overlap exists between the number of planes crossed and the presence of bihemispheric lesions. A trajectory through the midsagittal plane implies bihemispheric injury.. Process (Methodology) Of the 65 original studies, 16 were found to have adequate information on the prognostic value of CT scan examination. The relation between presence or absence of bihemispheric lesions and outcomes is described in 12 of the 65 studies analyzed. Two studies additionally describe injuries crossing the horizontal or coronal plane (i.e., lateral trajecto- S74 August Supplement 2001