Impact of traumatic subarachnoid hemorrhage on outcome in nonpenetrating head injury

Transcription

1 J Neurosurg 83: , 1995 Impact of traumatic subarachnoid hemorrhage on outcome in nonpenetrating head injury Part I: A proposed computerized tomography grading scale KARL A. GREENE, M.D., PH.D., FREDERICK F. MARCIANO, M.D., PH.D., BLAKE A. JOHNSON, M.D., RONALD JACOBOWITZ, PH.D., ROBERT F. SPETZLER, M.D., AND TIMOTHY R. HARRINGTON, M.D. Divisions of Neurological Surgery and Neuroradiology, Barrow Neurological Institute, St. Joseph s Hospital and Medical Center, Phoenix, Arizona; and Department of Mathematics, Arizona State University, Tempe, Arizona The presence of traumatic subarachnoid hemorrhage (tsah) on admission computerized tomography (CT) scans obtained from patients suffering from severe, nonpenetrating head injury has been shown to be associated with a worse outcome than the injury alone would warrant. However, no previous study has provided a simple means of relating the amount of tsah, its location, or other abnormal findings on initial head CT scans to outcome in patients with nonpenetrating head injury. In this study, admission head CT scans from 252 patients with tsah, treated at a single institution, were reviewed to ascertain thickness of the tsah; its location; evidence of mass lesion(s); shift of midline structures ( 5 mm vs. 5 mm); basal cistern effacement; and cortical sulcal effacement. The CT scans were then organized into Grades 1 to 4 with 1 indicating thin tsah ( 5 mm); 2, thick tsah ( 5 mm); 3, thin tsah with mass lesion(s); and 4, thick tsah with mass lesion(s). A stepwise regression analysis of CT features ranked them in descending order of contribution to Glasgow Outcome Scale (GOS) scores at the time of discharge from acute hospitalization as follows: basal cistern effacement, thickness of tsah, cortical sulcal effacement, presence of mass lesion(s), and location of tsah. A shift of midline structures was not found to be a significant variable. Further analysis comparing CT grades and admission postresuscitation Glasgow Coma Scale (GCS) scores was highly significant. Patients with lower CT grades had better admission GCS values and discharge GOS scores than those with higher CT grades. From their experience, the authors conclude that their CT grading scale is simple and reliable and relates significantly to outcome at the time of discharge from acute hospitalization. KEY WORDS traumatic subarachnoid hemorrhage nonpenetrating head injury outcome grading scale T HE presence of subarachnoid hemorrhage (SAH) on computerized tomography (CT) scans obtained at admission in patients with severe penetrating and nonpenetrating head injuries 14,26 has been shown to confer a worse prognosis than the injury alone. 14,26,31 The mechanism by which SAH results in worse outcomes is unclear. Although the factors that influence outcome in severe nonpenetrating head injury are multifactorial, 12,13,16,17,29,32,46 it has been postulated that vasospasm in the presence of SAH produces ischemia in this group of patients. 8,14,32 The use of grading systems based on easily obtainable radiographic findings is a helpful adjunct for predicting prognosis in a variety of neurological disorders 16,20,22,39 including nonpenetrating head injury. 17,30 The existing grading systems for head injury used to classify patients according to findings on admission CT and correlate these J. Neurosurg. / Volume 83 / September, 1995 findings with outcome do not take into account the presence, location, or amount of SAH. 30 As part of a clinical study of nonpenetrating head injury that compared the acute hospital courses of patients with traumatic SAH (tsah) with those of a matched control group, we developed a simple and reliable grading system for tsah that relates significantly to clinical outcome at the time of discharge from acute hospitalization. Clinical Material and Methods Patient Population Between January 1988 and December 1991, 3414 patients were admitted to the Barrow Neurological Institute of St. Joseph s Hospital and Medical Center with traumat- 445

2 K. A. Greene, et al. TABLE 1 Distribution by GCS ADMIT of 3157 patients with nonpenetrating TBI with or without associated tsah on admission CT* No. of Patients No. of Patients GCS ADMIT Without tsah With tsah (69%) 111 (31%) (78%) 76 (22%) (84%) 44 (16%) (94%) 124 (6%) total 2802 (89%) 355 (11%) * GCS ADMIT = admission postresuscitation Glasgow Coma Scale (GCS) 41 score; TBI = traumatic brain injury; tsah = traumatic subarachnoid hemorrhage; CT = computerized tomography. ic brain injuries. Retrospectively, clinical data regarding the mechanism of injury, admission characteristics and findings on head CT, clinical variables related to hospital course, and outcome at the time of discharge from acute hospitalization were recorded. These data were then entered into a computer database. Using this database, all patients with tsah on admission head CT scans (3157 patients) were identified. Individuals with penetrating head injuries and tsah were then removed from this group for a total of 355 (11%) (Table 1). FIG. 1. Photograph showing a data sheet used for acquiring information about traumatic subarachnoid hemorrhage and other features present on admission computerized tomography (CT) scans of the head. Admission head CT scans were available in 252 cases (71%) of patients with nonpenetrating head injury and tsah. A CT grading system was designed to classify tsah in a manner similar to that of Fisher, et al., 16 although it differed by the inclusion of mass lesions known to be present in severe nonpenetrating head injuries. 17 In addition, several other characteristics on admission head CT scans known to influence outcome in severe nonpenetrating head injury were recorded and analyzed in this population. 13,23,41,44,46 A sample data sheet is shown in Fig. 1. Clinical Variables Several clinical variables were extracted from the database for correlational analysis with the population of 252 tsah patients. Variables included age, sex, injury severity score (ISS), 7 and the presence or absence of a variety of complicating factors occurring within the first 24 hours of admission, such as hypoxia (PO 2 60 mm Hg for 5 minutes and 30 minutes), hypotension (systolic blood pressure 80 mm Hg for 5 minutes and 30 minutes), seizures, cardiopulmonary arrest, and the presence or absence of alcohol or other drugs. Admission postresuscitation Glasgow Coma Scale (GCS) 41 values as well as elevated ( 20 mm Hg for 15 minutes) intracranial pressure (ICP), nonneurological injuries, neurosurgical procedures, and neurological status at the time of discharge from acute hospitalization based on Glasgow Outcome Scale (GOS) 21 scores were also recorded for all patients. Also noted was whether an intracranial mass lesion visualized on admission head CT scan was surgical or nonsurgical. Surgical mass lesions were defined as follows: 1) extraaxial blood with a thickness of 5 mm or greater and associated with cortical sulcal effacement and/or CT evidence of perimesencephalic cistern effacement suggestive of impending transtentorial uncal herniation; or 2) intraparenchymal hemorrhage or contusion associated with a midline shift greater than 5 mm and/or CT evidence of perimesencephalic cistern effacement suggestive of impending transtentorial uncal herniation. Surgical mass lesions were evacuated as expeditiously as possible and only when the patient was medically and hemodynamically stable. Nonsurgical mass lesions were simply observed over the course of a patient s hospitalization. Data Collection Using available admission head CT scans of all patients with nonpenetrating head injuries and tsah, the information needed for completion of the head CT forms was extracted and documented independently by a neuroradiologist and a neurosurgeon. Location of the tsah on admission head CT scans was determined by using the criteria established by Fisher, et al. 16 Thickness of the tsah was measured manually using an absolute-centimeter scale that was printed on each head CT radiograph. Midline shift was determined using the septum pellucidum as the midline structure, and the volume of each mass lesion was not quantified. From this prospective review of head CT features, a total CT grade was assigned to each CT scan, consisting of the following features: Grade 1 indicated thin tsah 446 J. Neurosurg. / Volume 83 / September, 1995

3 Computerized tomography grading of SAH in blunt head injury TABLE 2 Regression model evaluating admission CT features using discharge GOS score as the dependent variable* Model Features Regression Coefficient basal cistern effacement thickness of tsah cortical sulcal effacement mass lesion location of tsah *r 2 = 0.52, mean square error = CT = computerized tomography; GOS = Glasgow Outcome Scale; 21 tsah = traumatic subarachnoid hemorrhage. ( 5 mm); Grade 2, thick tsah ( 5 mm); Grade 3, thin tsah with mass lesion(s) and a shift less than or equal to 5 mm (3A) or greater than 5 mm (3B); Grade 4, thick tsah with mass lesions and shift less than or equal to 5 mm (4A) or greater than 5 mm (4B). The distinctions between CT Grades 3A and 3B, and between 4A and 4B were later modified when stepwise regression analysis revealed that shift of midline structures was not a significant variable in an equation relating admission CT features to GOS 21 at the time of hospital discharge (Table 2). The grading scale was changed and Grades 3A and 3B, and 4A and 4B were collapsed into Grades 3 and 4, respectively. Location of the tsah was modified so that a distinction was made between those regions involving the basal and vertical cisterns and sylvian and basofrontal interhemispheric fissures (basal region), and those involving the remaining portions of the interhemispheric fissure, areas directly above and below the tentorium, and cerebral convexities (hemispheric region). Based on this distinction, the location was categorized in a hierarchical manner and assigned values of 1 through 3 such that 1 referred to the presence of tsah in the hemispheric region alone, 2 referred to the basal region alone, and 3 referred to the presence of tsah in both the basal and hemispheric regions. Statistical Analysis The combination of CT features that included basal cistern effacement, thickness of tsah, presence of mass lesion(s), shift of midline structures, location of tsah, and cortical sulcal effacement were subjected to stepwise regression analysis with backward elimination and limit of inclusion of 0.05 using the discharge GOS score as the dependent variable. Identical methods were used to evaluate the contribution to outcome of admission postresuscitation GCS score, CT grade, and patient age. Interrater reliability was performed on all CT characteristics and compared statistically using standardized paired z-tests. Binomial tests of equal outcome proportions were used for evaluating statistical significance in those situations for which the null hypothesis assumed equal outcome proportions for two different interventions. 19 Finally, chisquare analysis and the Fisher s Exact test were performed when appropriate. Statistical significance was set at the 0.05 level. J. Neurosurg. / Volume 83 / September, 1995 Results The mean age of patients in this cohort was 34 years (standard deviation 21.4 years), with ages ranging from 1 month to 102 years. There were 193 males and 59 females. Interrater reliability between neuroradiologist and neurosurgeon was 94%. The stepwise regression analysis, which used the GOS score at the time of discharge from acute hospitalization as the dependent variable, resulted in exclusion of shift of midline structures as a significant variable relating to outcome. In terms of the contribution to GOS scores at the time of discharge from acute hospitalization, basal cistern effacement was the most significant variable in the regression model (Table 2), followed by thickness of tsah, cortical sulcal effacement, presence of mass lesion(s), and location of tsah. Because of the comparatively small contribution made by location of tsah to GOS scores, this variable was excluded from the final CT grading scale. The relationship between admission postresuscitation GCS score and CT grade is shown in Fig. 2 left. This relationship is significant ( 2 = 37.57, p ), suggesting that patients with lower CT grades had higher admission GCS scores and that patients with higher CT grades had lower admission GCS scores. In a similar manner, the relationship between CT grade and discharge GOS values is shown in Fig. 2 right. The significant relationship between these two variables ( 2 = , p 0.001) suggests that patients with a lower CT grade have a better outcome as reflected by the discharge GOS score and that patients with higher CT grades have a worse outcome. No significant relationship was found between CT grade and ISS on admission. An additional regression model evaluating the contributions of admission postresuscitation GCS score, CT grade, and patient age to GOS classification at the time of discharge from acute hospitalization revealed that the admission GCS score was the most significant variable, followed by CT grade (Table 3). Patient age contributed the least to the discharge GOS score in this patient population. Within the first 24 hours of admission, hypoxia and hypotension occurred as complicating factors in 15 and 24 patients with tsah, respectively. These factors were correlated with cortical sulcal effacement and basal cistern effacement (Table 4). Elevations in ICP at the time of ventriculostomy placement and 6 hours after admission also demonstrated a significant association with basal cistern effacement ( 2 = 5.66, p 0.025; and 2 = 7.40, p 0.01, respectively). Table 5 shows the stratification of outcomes based on admission postresuscitation GCS scores and CT grades. More than 25% of the 252 patients in this study with good outcomes based on their discharge GOS category had mild head injuries according to their admission postresuscitation GCS scores, which ranged from 13 to 15. Interestingly, of those patients with more moderate head injuries (GCS values of on admission), those with mass lesions had moderate-to-poor outcomes (72% of patients with CT Grade 3 and 75% of patients with CT Grade 4). Because of the variability in the relationships noted between CT grade and discharge GOS classification, further evaluation of groups at the extremes (CT Grade 1 and 447

4 K. A. Greene, et al. FIG. 2. Bar graphs depicting the relationships between admission postresuscitation Glasgow Coma Scale 41 (left) and discharge Glasgow Outcome Scale 20 (right) scores and percentages of patients within each computerized tomography (CT) grade of traumatic subarachnoid hemorrhage. discharge GOS 5; CT Grade 2 and discharge GOS 5) was performed. Patients having CT grades 1 and 2 and GOS scores of 5 had mean ISSs of and , respectively (LD 50 = 40 for ages years). When compared with the other outcome scores within these CT grades, the ISS was significantly greater for GOS 5 patients. Thus patients with CT Grades 1 and 2 and GOS 5 most likely died from nonneurological causes. A total of 182 (72%) of the 252 patients with tsah had surgical or nonsurgical intracranial mass lesions (CT Grades 3 and 4) (Table 6). Eighty-one (45%) of the 182 mass lesions were surgical; 60 (74%) of these were operatively evacuated within the first 3 hours of admission and all surgical mass lesions were evacuated within 24 hours of admission. Of the 252 patients evaluated in this study, 31 (12%) had a midline shift greater than 5 mm on admission CT, and 30 of these patients (97%) had intracranial mass lesions (Table 7). Nineteen (63%) of the 30 patients with intracranial mass lesions and a shift greater than 5 mm had surgical mass lesions; 18 of these surgical mass lesions (95%) were evacuated within 3 hours of admission, and all 19 patients had their mass lesions evacuated within 24 hours of admission. In terms of outcome at the time of discharge from acute hospitalization, only those groups of patients with good or vegetative outcomes showed significant differences in the proportion of patients with surgical versus nonsurgical mass lesions (Table 8). The majority of patients (89%) with good outcomes (Table 8) had nonsurgical mass lesions, whereas all seven patients with vegetative outcomes had surgical mass lesions that were evacuated within 3 hours of admission. There were no differences in the proportions of patients with surgical versus nonsurgical mass TABLE 3 Regression model evaluating GCS ADMIT, CT grade, and patient s age using discharge GOS score as dependent variable* Model Feature Regression Coefficient GCS ADMIT CT grade patient s age *r 2 = 0.59, mean square error = GCS ADMIT = admission postresuscitation Glasgow Coma Scale 41 score; CT = computerized tomography; GOS = Glasgow Outcome Scale. 21 lesions who died during the course of their acute care hospitalization (Table 8). Discussion In this retrospective study of 252 patients with nonpenetrating head injury and the presence of tsah on admission head CT, we have demonstrated a significant association between a simple CT grading system for tsah and outcome based on GOS scores at the time of discharge from acute hospitalization. Although no long-term followup data were available for the majority of patients, results from the Traumatic Coma Data Bank for severe head injury revealed a difference of less than 4% between the rates of mortality at time of discharge and at a later followup date. 14 This suggests that minimal variability occurs over time in the outcome extremes (good recovery vs. death) for follow-up GOS scores in a cohort of patients with mild, moderate, and severe head injuries such as those included in our study. The scoring system on which a given CT grade is based has a high interrater reliability factor (94%). Furthermore, with the exception of the shift of midline structures variable, all of the admission head CT characteristics used to arrive at the final CT grade significantly relate to outcome on regression analysis (Table 2). Admission CT Features Our CT grading system was designed to compare those TABLE 4 Relationships between hypoxia and hypotension occurring within 24 hours of admission and cortical sulcal and basal cistern effacement demonstrated on admission CT* Complicating Factor/CT Feature 2 p Value hypoxia/cortical sulcal effacement NS hypoxia/basal cistern effacement NS hypotension/cortical sulcal effacement 8.77 < hypotension/basal cistern effacement < hypoxia or hypotension/cortical sulcal effacement 8.20 < hypoxia or hypotension/basal cistern effacement < * CT = computerized tomography; NS = not significant. 448 J. Neurosurg. / Volume 83 / September, 1995

5 Computerized tomography grading of SAH in blunt head injury TABLE 5 Relationship between GCS ADMIT and GOS DISCH for each CT grade* GCS ADMIT CT Grade & No. of No. of No. of No. of GOS DISCH Patients Patients Patients Patients CT Grade 1 good 2 (20%) 3 (38%) 4 (100%) 28 (90%) moderate 1 (10%) 1 (12%) 0 (0%) 2 (6%) poor 7 (70%) 4 (50%) 0 (0%) 1 (3%) CT Grade 2 good 1 (11%) 0 (0%) 0 (0%) 6 (100%) moderate 1 (11%) 1 (50%) 0 (0%) 0 (0%) poor 7 (78%) 1 (50%) 0 (0%) 0 (0%) CT Grade 3 good 1 (3%) 4 (14%) 5 (28%) 21 (88%) moderate 1 (3%) 10 (34%) 6 (33%) 3 (12%) poor 30 (94%) 15 (52%) 7 (39%) 0 (0%) CT Grade 4 good 0 (0%) 2 (13%) 2 (25%) 9 (53%) moderate 3 (8%) 1 (7%) 4 (50%) 6 (35%) poor 36 (92%) 12 (80%) 2 (25%) 2 (12%) * GCS ADMIT = admission postresuscitation Glasgow Coma Scale 41 score; GOS DISCH = Glasgow Outcome Scale 21 category at time of discharge from acute hospitalization; CT = computerized tomography; good = good outcome; moderate = moderate disability outcome; poor = poor outcome (severe disability, vegetative, or dead). Outcome designations are based on criteria of Jennett and Bond 21 for summation of GOS discharge categories. Percentages are calculated from total number of patients within a given GCS ADMIT /CT grade. variables known to influence clinical outcome after a severe head injury with those associated with tsah. As expected (Table 2), basal cistern effacement was the most powerful contributor to outcome in this study. 27,41,44,46 This feature, more than any other, is correlated with elevations in ICP. 30,41 Basal cistern effacement in severely head injured patients with tsah in the present study was significantly associated with ICP elevations at the time of ventriculostomy placement and 6 hours after admission. Sustained elevations in ICP have been shown by the Traumatic Coma Data Bank to be associated with worse outcomes in severe nonpenetrating head injury. 29 The presence of cortical sulcal effacement following severe nonpenetrating head injury has been shown to be associated with basal cistern effacement and the occurrence of hypoxia and/or hypotension by the Traumatic Coma Data Bank; 14 these factors may also influence outcome. In addition, we found an association between basal cistern and cortical sulcal effacement and the occurrence of hypoxia and/or hypotension (Table 4). The presence of thick subarachnoid blood also plays a significant role in the outcomes reviewed by this study. Although hydrocephalus can develop in association with SAH, fewer than 3% of the patients in this study actually developed posttraumatic hydrocephalus. Some investigators have speculated that tsah may act in ways similar to aneurysmal SAH with regard to the risk of vasospasm. 14,26,32 We were unable to verify this hypothesis in our study because of an inadequate number of angiographic studies (only 16) performed in the tsah patient population. J. Neurosurg. / Volume 83 / September, 1995 TABLE 6 Surgical and nonsurgical intracranial mass lesions in 182 patients with tsah* No. of No. of Patients With Patients With CT Surgical Nonsurgical Grade Mass Mass Total p Value 3 43 (42%) 60 (58%) 103 NS 4 38 (48%) 41 (52%) 79 NS * Intracranial mass lesion = epidural hematoma, subdural hematoma, or hemorrhagic contusion; tsah = traumatic subarachnoid hemorrhage. Two-tailed binomial test of equal outcome proportions. NS = not significant. The presence of an intracranial mass lesion is known to be an important factor in determining outcome in patients with severe head injury. 17 This was true in the present study as well, although other admission CT features appeared more influential, perhaps because of the policy of prompt evacuation of surgical mass lesions at our institution. Of interest in this regard is the absence of a significant contribution on the part of shift of midline structures on admission head CT to outcome in this study. Previous reports of admission head CT findings associated with outcome in nonpenetrating head injury have demonstrated that midline shift plays a significant role. 30,46 However, the vast majority of intracranial mass lesions present in patients with nonpenetrating head injury at our institution are evacuated within the first 3 hours of admission. This is true for the population of patients with tsah and mass lesions with midline shift greater than 5 mm in the present study (Table 7): 18 (95%) of 19 patients in this category had their mass lesions removed within 3 hours of admission. Prompt evacuation of mass lesions in this cohort of patients may have diminished significantly the contribution of midline shift on admission head CT to outcome in this study. Although we recorded the presence of intraventricular hemorrhage (IVH), this feature reflects primarily the severity of head injury, 13,23 but is unassociated with elevations in ICP. 23 Diffuse axonal injury 3 is also a significant factor in relationship to outcome in nonpenetrating head injury. 2 Such lesions are more readily visualized on magnetic resonance (MR) images 47,48 and their imaging characteristics on admission CT, as well as both early and late TABLE 7 Surgical and nonsurgical intracranial mass lesions in 30 patients with tsah associated with midline shift greater than 5 mm* No. of No. of Patients With Patients With CT Surgical Nonsurgical Grade Mass Mass Total p Value 3 6 (75%) 2 (25%) 8 NS 4 13 (59%) 9 (41%) 22 NS * Intracranial mass lesion = epidural hematoma, subdural hematoma, or hemorrhagic contusion; tsah = traumatic subarachnoid hemorrhage. Fisher s exact test (two-tailed, p = 0.424). NS = not significant. 449

6 K. A. Greene, et al. TABLE 8 Outcomes at time of discharge for 182 patients with tsah and associated surgical or nonsurgical intracranial mass lesion (CT Grades 3 and 4)* No. of No. of Patients With Patients With Surgical Nonsurgical GOS DISCH Mass Mass Total p Value 1 5 (11%) 39 (89%) 44 < (56%) 15 (44%) 34 NS 3 20 (63%) 12 (37%) 32 NS 4 7 (100%) 0 (0%) 7 < (46%) 35 (54%) 65 NS total 81 (45%) 101 (55%) 182 NS * tsah = traumatic subarachnoid hemorrhage; intracranial mass lesion = epidural hematoma, subdural hematoma, or hemorrhagic contusion; CT = computerized tomography. Glasgow Outcome Scale 21 category at time of discharge from acute hospitalization: 1 = good recovery; 2 = moderate disability; 3 = severe disability; 4 = vegetative; 5 = dead. Two-tailed binomial test of equal outcome proportions. NS = not significant. MR images, are significantly associated with the degree of neurological disability. 24,25,45 Although diffuse axonal injury may be present on admission in patients with focal or diffuse brain injury, 36,37 there is evidence that indicates that focal lesions may be more severe in the absence of diffuse axonal injury. 1 In addition, the focal lesions that coexist with diffuse axonal injury on admission imaging studies appear to have a greater impact on outcome when compared to diffuse axonal injury alone, particularly when tsah is present. 38 Furthermore, diffuse axonal injury is often unassociated with elevations in ICP. 18 Because too few MR imaging and autopsy studies were performed on patients in this study, we are unable to comment on the contribution of diffuse axonal injury to outcome in this cohort of 252 patients with tsah on admission head CT. Computerized Tomography Grade, Neurological Injury, and Admission Injury Severity The association between admission postresuscitation GCS score and CT grade (Fig. 2) suggests that patients with higher CT grades tend to have more severe neurological injuries. Other factors may significantly influence the severity of neurological injury assessed on admission, and the motor score obtained on admission may be more reflective of the severity of the injury. Whether this is true awaits further study because the design of our database would not permit the extraction of such data for analysis from the admission GCS values. Computerized Tomography Grade and Outcome The relationship between admission CT grade and outcome at the time of discharge from acute hospitalization is strikingly clear (p ) (Fig. 2 right). There is clearly a trend of worsening outcomes coinciding with higher CT grades. Although the relationship is clearly significant, the variance not explained by the correlation coefficient between the two variables probably reflects the complexity of factors that contribute to outcome in nonpenetrating head injury. 30 An analysis of ISS with regard to admission CT grade reveals no relationship between these two. However, in those CT grades in which little linearity between CT grade and GOS score exist, for example GOS Category 5 (Fig. 2 right), patients with high admission ISSs comprise a substantial proportion of patients with CT Grades 1 and 2. Other factors such as age clearly contribute to outcome differences. However, CT grade contributed substantially more to outcome than age when regression analysis was performed in conjunction with admission GCS and age using the GOS score at the time of discharge from acute hospitalization as the dependent variable (Table 3). Although other factors may also play a significant role in determining overall outcome of patients with tsah in nonpenetrating head injury, our grading system remains a simple, reliable, and significant predictor of outcome at the time of discharge from acute hospitalization. Implications of the Study As noted previously, some authors have postulated that tsah may exert its negative effect in a manner similar to that of aneurysmal SAH with regard to the risk of developing cerebral ischemia from vasospasm. Unlike nontraumatic, nonaneurysmal SAH, vasospasm following severe nonpenetrating head injury is not uncommon 4,32,40 and may result in symptomatic ischemia and stroke. 32 In aneurysmal SAH, the amount of subarachnoid blood visualized on admission CT relates significantly to the risk of cerebral vasospasm. 16,22 In nonpenetrating head injury, vasospasm may be related to more than simply the presence and quantity of subarachnoid blood. 11,31,32,40 Studies using transcranial Doppler (TCD) ultrasonography for the detection of vasospasm in patients with nonpenetrating severe head injury do not reliably correlate the incidence of vasospasm with the amount of tsah on admission CT. 11,32 Perhaps this is because high velocities measured by TCD studies in severe head injury reflect hyperemia rather than vasospasm. Studies that incorporate the use of cerebral angiography, cerebral blood flow, and TCD in severe head injury might be useful in addressing this issue. In a recent study, Martin and colleagues 31 showed that the risk of development of posttraumatic vasospasm in a series of 101 patients with nonpenetrating brain injury was associated with the presence of tsah, subdural hematoma, IVH, and contusions and/or intraparenchymal hemorrhage on admission head CT scan. Vasospasm did not develop in patients with normal scans or scans that revealed either cerebral edema alone or epidural hemorrhage. Patients who had all four features associated with an increased risk of vasospasm always developed vasospasm. Patients who had tsah in addition to one of the other remaining three CT features had a greater risk of developing vasospasm than patients with tsah alone. This finding indicates that one must take into account the addition of other CT features in nonpenetrating head trauma when attempting to develop an adequate grading system for tsah rather than creating a simple extrapolation of grading systems used in aneurysmal SAH. 16 Furthermore, complicated CT classification schemes in head injury, such as the currently available system outlined by 450 J. Neurosurg. / Volume 83 / September, 1995

7 Computerized tomography grading of SAH in blunt head injury Marshall and coworkers, 30 have recently been shown to be inconsistently applied to individual patients in one multicenter study. 15 This study stressed the importance of establishing a reliable and simple method of reporting and grading tsah for use in future head-injury protocols. We believe that the simple CT grading scale for tsah outlined in this paper may eventually prove useful in this regard. Given the association between worse outcomes for patients with nonpenetrating head injury who have tsah on admission head CT, therapeutic modalities designed to limit the effects of secondary brain injury and neurotoxicity known to occur in nonpenetrating head injury 6,10,12 should be considered. The results noted in Table 5 reinforce this issue: a significant proportion of patients with moderate head injuries and mass lesions on admission CT scans had moderate-to-poor outcomes at the time of discharge from acute hospitalization. Hypervolemic hemidilutional therapy 5 and moderate hypothermia 28 could have been useful interventions in this population of patients. Although somewhat controversial at present, 9 the use of the calcium channel blocker nimodipine may be beneficial in patients with tsah. 15 Such intervention in a population of head-injured patients at risk for worse outcomes awaits further study. Conclusions In a cohort of 252 patients with nonpenetrating brain injuries and tsah on admission CT scans of the head, a CT grading scale was developed that incorporates thickness of subarachnoid blood, as well as the presence or absence of mass lesions. Using this simple grading system, we have shown a significant relationship between CT grade and outcome at the time of discharge from acute hospitalization. We propose that this simple and reliable grading system be used in prospective studies aimed at elucidating the pathophysiological process(es) responsible for worsening outcomes in patients with tsah, as well as in protocols designed to improve outcome in this patient population. Acknowledgments The authors wish to thank Shelley A. Kick, Ph.D., editor; Pamela A. Smith, medical photographer; and the editorial staff of the Neuroscience Publications Office at the Barrow Neurological Institute for their assistance in the preparation of this manuscript. References J. Neurosurg. / Volume 83 / September, Adams JH, Doyle D, Graham DI, et al: The contusion index: a reappraisal in human and experimental non-missile head injury. Neuropathol Appl Neurobiol 11: , Adams JH, Graham DI, Gennarelli TA, et al: Diffuse axonal injury in non-missile head injury. J Neurol Neurosurg Psychiatry 54: , Adams JH, Graham DI, Murray LS, et al: Diffuse axonal injury due to nonmissile head injury in humans: an analysis of 45 cases. Ann Neurol 12: , Arutiunov AI, Baron MA, Majorova NA: The role of mechanical factors in the pathogenesis of short-term and prolonged spasm of the cerebral arteries. J Neurosurg 40: , Awad IA, Carter LP, Spetzler RF, et al: Clinical vasospasm after subarachnoid hemorrhage: response to hypervolemic hemodilutional and arterial hypertension. Stroke 18: , Baker AJ, Moulton RJ, MacMillan VH, et al: Excitatory amino acids in cerebrospinal fluid following traumatic brain injury in humans. J Neurosurg 79: , Baker SP, O Neill B, Haddon W Jr, et al: The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 14: , Bouma GJ, Muizelaar JP, Choi SC, et al: Cerebral circulation and metabolism after severe traumatic brain injury: the elusive role of ischemia. J Neurosurg 75: , Braakman R: Multicentre trials on the efficacy of nimodipine on outcome after severe head injury. J Neurotrauma 10 (Suppl):S85, 1993 (Abstract) 10. Bullock R, Fujisawa H: The role of glutamate antagonists for the treatment of CNS injury. J Neurotrauma 9 (Suppl 2): S443 S462, Chan KH, Dearden NM, Miller JD: The significance of posttraumatic increase in cerebral blood flow velocity: a transcranial Doppler ultrasound study. Neurosurgery 30: , Chesnut RM, Marshall LF, Klauber MR, et al: The role of secondary brain injury in determining outcome from severe head injury. J Trauma 34: , Cordobés F, de la Fuente M, Lobato RD, et al: Intraventricular hemorrhage in severe head injury. J Neurosurg 58: , Eisenberg HM, Gary HE Jr, Aldrich EF, et al: Initial CT findings in 753 patients with severe head injury. A report from the NIH Traumatic Coma Data Bank. J Neurosurg 73: , The European Study Group on Nimodipine in Severe Head Injury: a multicenter trial of the efficacy of nimodipine on outcome after severe head injury. J Neurosurg 80: , Fisher CM, Kistler JP, Davis JM: Relation of cerebral vasospasm to subarachnoid hemorrhage visualized by computerized tomographic scanning. Neurosurgery 6:1 9, Gennarelli TA, Spielman GM, Langfitt TW, et al: Influence of the type of intracranial lesion on outcome from severe head injury. A multicenter study using a new classification system. J Neurosurg 56:26 32, Graham DI, Lawrence AE, Adams JH, et al: Brain damage in fatal non-missile head injury without high intracranial pressure. J Clin Pathol 41:34 37, Hill AB: A Short Textbook of Medical Statistics. London: Hodder and Stoughton, 1977, pp Hunt WE, Kassell N, Pertuiset B, et al: Report of World Federation of Neurological Surgeons Committee on a Universal Subarachnoid Hemorrhage Grading Scale. J Neurosurg 68: , 1988 (Letter) 21. Jennett B, Bond M: Assessment of outcome after severe brain damage. A practical scale. Lancet 1: , Kistler JP, Crowell RM, Davis KR, et al: The relation of cerebral vasospasm to the extent and location of subarachnoid blood visualized by CT scan: a prospective study. Neurology 33: , LeRoux PD, Haglund MM, Newell DW, et al: Intraventricular hemorrhage in blunt head trauma: an analysis of 43 cases. Neurosurgery 31: , Levin HS, Amparo EG, Eisenberg HM, et al: Magnetic resonance imaging after closed head injury in children. Neurosurgery 24: , Levin HS, Williams DH, Valastro M, et al: Corpus callosal atrophy following closed head injury: detection with magnetic resonance imaging. J Neurosurg 73:77 81, Levy ML, Rezai A, Masri LS, et al: The significance of subarachnoid hemorrhage after penetrating craniocerebral injury: correlations with angiography and outcome in a civilian population. Neurosurgery 32: , Luerssen TG, Hults K, Klauber M, et al: Improved outcome as 451

8 K. A. Greene, et al. a result of recognition of absent or compressed cisterns on initial CT scans, in Hoff JT, Betz AL (eds): Intracranial Pressure VII. Berlin: Springer-Verlag, 1989, pp Marion DW, Obrist WD, Carlier PM, et al: The use of moderate therapeutic hypothermia for patients with severe head injuries: a preliminary report. J Neurosurg 79: , Marmarou A, Ward JD, Young HF, et al: Impact of ICP instability and hypotension on outcome in patients with severe head trauma. J Neurosurg 75 (Suppl):S59 S66, Marshall LF, Marshall SB, Klauber MR, et al: A new classification of head injury based on computerized tomography. J Neurosurg 75 (Suppl):S14 S20, Martin NA, Doberstein C, Khanna R, et al: Post traumatic cerebral arterial spasm. J Neurotrauma 10 (Suppl):S84, 1993 (Abstract) 32. Martin NA, Doberstein C, Zane C, et al: Posttraumatic cerebral arterial spasm: transcranial Doppler ultrasound, cerebral blood flow, and angiographic findings. J Neurosurg 77: , Rinkel GJ, Wijdicks EF, Vermeulen M, et al: Nonaneurysmal perimescencephalic subarachnoid hemorrhage: CT and MR patterns that differ from aneurysmal rupture. AJNR 12: , Rinkel GJE, Wijdicks EFM, Vermeulen M, et al: The clinical course of perimesencephalic nonaneurysmal subarachnoid hemorrhage. Ann Neurol 29: , Rinkel GJE, Wijdicks EFM, Vermeulen M, et al: Outcome in perimesencephalic (nonaneurysmal) subarachnoid hemorrhage: a follow-up study in 37 patients. Neurology 40: , Sahuquillo J, Vilalta J, Lamarca J, et al: Diffuse axonal injury after severe head trauma. A clinico-pathological study. Acta Neurochir 101: , Sahuquillo-Barris J, Lamarca-Ciuro J, Vilalta-Castan J, et al: Acute subdural hematoma and diffuse axonal injury after severe head trauma. J Neurosurg 68: , Shigemori M, Tokutomi T, Kuramoto S, et al: Diffuse axonal injury and early intracranial sequelae in severe head injury. Neurol Med Chir 31: , Spetzler RF, Martin NA: A proposed grading system for arteriovenous malformations. J Neurosurg 65: , Suwanwela C, Suwanwela N: Intracranial arterial narrowing and spasm in acute head injury. J Neurosurg 36: , Teasdale E, Cardoso E, Galbraith S, et al: CT scan in severe diffuse head injury: physiological and clinical correlations. J Neurol Neurosurg Psychiatry 47: , Teasdale G, Jennett B: Assessment of coma and impaired consciousness. A practical scale. Lancet 2:81 84, Toutant SM, Klauber MR, Marshall LF, et al: Absent or compressed basal cisterns on first CT scan: ominous predictors of outcome in severe head injury. J Neurosurg 61: , Uzzell BP, Dolinskas CA, Wiser RF, et al: Influence of lesions detected by computed tomography on outcome and neuropsychological recovery after severe head injury. Neurosurgery 20: , van Dongen KJ, Braakman R, Gelpke GJ: The prognostic value of computerized tomography in comatose head-injured patients. J Neurosurg 59: , Wilberger JE Jr, Rothfus WE, Tabas J, et al: Acute tissue tear hemorrhages of the brain: computed tomography and clinicopathological correlations. Neurosurgery 27: , Yokota H, Kurokawa A, Otsuka T, et al: Significance of magnetic resonance imaging in acute head injury. J Trauma 31: , 1991 Manuscript received December 20, Accepted in final form January 10, The findings of this study were presented, in part, at the Second International Neurotrauma Symposium of the International Neurotrauma Society, Glasgow, Scotland, July, 1993; the 11th Annual Neurotrauma Symposium of the Neurotrauma Society, November, 1993; and the 24th Annual Meeting of the Western Trauma Association, Crested Butte, Colorado, March, Address for Dr. Johnson: David Grant Medical Center, Travis Air Force Base, California. Address reprint requests to: Timothy R. Harrington, M.D., c/o Editorial Office, Barrow Neurological Institute, 350 West Thomas Road, Phoenix, Arizona J. Neurosurg. / Volume 83 / September, 1995