Classification and Prediction of Outcome in Traumatic

The Journal of International Medical Research 2009; 37: 983 – 995

Classification and Prediction of Outcome in Traumatic Brain Injury Based on Computed Tomographic Imaging GW ZHU1,2, F WANG2,3

AND

WG LIU1,2

1

Department of Neurosurgery, Second Affiliated Hospital, College of Medicine, 2Institute of Brain Medicine, and 3Department of Neurosurgery, First Affiliated Hospital, College of Medicine, Zhejiang University, Hangzhou City, Zhejiang Province, China

Traumatic brain injury (TBI) is a common and potentially devastating problem. The classification of TBI is necessary for accurate diagnosis and the prediction of outcomes. The increased use of early sedation, intubation and ventilation in more severely injured patients has decreased the value of the Glasgow Coma Scale for the purposes of classification. An alternative is the classification of TBI according to morphological criteria based on computed tomography (CT) investigations. This article reviews the KEY WORDS: TRAUMATIC

BRAIN INJURY

current classification and prediction of outcomes in TBI based on CT imaging. Classifications based on the presence or absence of intracranial local lesions, diffuse injury, signs of subarachnoid or intraventricular haemorrhage and fractures or foreign bodies are considered, and their predictive value is discussed. Future studies should address the complicated issue of how optimally to combine CT characteristics for prognostic purposes and how to improve on currently used CT classifications to predict outcomes more accurately.

(TBI); COMPUTED PREDICTION

Introduction Traumatic brain injury (TBI) is a common and potentially devastating problem. The classification of TBI is necessary to make accurate diagnoses and predict outcome, and requires patients to be grouped according to specific characteristics. In clinical practice, the severity of TBI is generally classified as severe, moderate or mild according to the level of consciousnesss as measured using the Glasgow Coma Scale (GCS).1 The increased use of early sedation, intubation and ventilation in more severely injured patients has decreased the value of

TOMOGRAPHY

(CT); CLASSIFICATION;

the full GCS for the purposes of TBI classification, because patients need to be conscious and able to respond verbally.2 – 5 An alternative in such patients is the classification of TBI according to morphological criteria based on computed tomography (CT) investigations. Although TBI can also be classified using magnetic resonance imaging characteristics, which may be more sensitive for detecting small white matter lesions in later phases,6,7 CT examination remains the investigation of choice to identify the presence and extent of structural damage in the acute phase after

983

Downloaded from imr.sagepub.com by guest on October 18, 2016

GW Zhu, F Wang, WG Liu Computed tomographic imaging in traumatic brain injury TBI. The CT scan provides essential diagnostic information with therapeutic implications for operative intervention or intracranial pressure (ICP) monitoring, and also provides objective prognostic information. This article reviews the current classification and prediction of outcome in TBI based on CT imaging.

Classifications of TBI The CT scan has proved to be of great value in the assessment of gross pathological findings at the time of injury. This has led to a better understanding of the mechanism of head injury and has significantly improved clinical care, reducing both morbidity and mortality.8,9 The relevance of CT scanning for the purposes of classification and prediction has increased with the growing difficulties in reliably assessing clinical severity, due to the use of early sedation, intubation and ventilation in more severely injured patients.2 – 5 Such classifications may be based on the presence or absence of intracranial local lesions, diffuse injury, signs of subarachnoid or intraventricular haemorrhage, or fractures or foreign bodies.

INTRACRANIAL LOCAL LESIONS Intracranial local lesions may be caused by haemorrhage or oedema and can result in ischaemia or infarct, direct impingement on other vital structures or herniation of different parts of the brain.10 Any lesions producing a mass effect should, therefore, always be urgently evacuated. Because haemorrhages frequently progress, repeat imaging is usually indicated, especially if changes in neurological status occur. The location of the lesions helps determine the risk and the appropriate management. Intracranial lesions can be classified in various ways; studies correlating outcome with various anatomical parameters of

intracranial lesions are shown in Table 1.11 – 18

DIFFUSE BRAIN INJURY Diffuse brain injury usually manifests on CT imaging with no visible focal lesions, but with compression of the basal cisterns and/or the third ventricle, together with midline shift.15 Patients with non-focal diffuse lesions have been shown to have a high risk of intracranial hypertension and an increased mortality rate,15 although mortality is lower than in patients with mass lesions.11,19 Studies correlating outcome with signs of diffuse brain injury are shown in Table 2.15, 20 – 22 Since its introduction in 1991, the Marshall CT classification15 of diffuse brain injury has become widely used for descriptive purposes in patients with diffuse injuries and as a major predictor of outcome in TBI.

SUBARACHNOID AND INTRAVENTRICULAR HAEMORRHAGE The incidence of CT-documented traumatic subarachnoid haemorrhage (SAH) in patients with severe TBI is 23 – 63%.23 It is most common in patients with subdural haematoma or haemorrhage contusion and is associated with a poor prognosis.23 – 25 Traumatic intraventricular haemorrhage (IVH) is rare and is associated with poor outcomes that seem to be the consequence of associated injuries.26 The importance of SAH in TBI was shown by Kakarieka et al.,27 who demonstrated that the outcome of patients with traumatic SAH was significantly worse than that of patients whose first CT scan did not show subarachnoid bleeding. Traumatic SAH frequently occurs in patients with TBI, but is difficult to detect and grade.28

FRACTURES AND FOREIGN BODIES Fractures can be classified, according to their location, into fractures of the skull vault or

984


GW Zhu, F Wang, WG Liu Computed tomographic imaging in traumatic brain injury

TABLE 1: Studies of the classification and prediction of outcome in traumatic brain injury based on the presence and characteristics of intracranial local lesions on computed tomography (CT) Reference Gennarelli et al., 1982

Lobato et al., 198312

Lipper et al., 198513

11

CT classification criteria

Study conclusions

Categorized into epidural, acute subdural and other lesions, each of which was further subdivided into two GCS score ranges (3 – 5 and 6 – 8)

The type of lesion is, therefore, as important a factor in determining outcome, as is the GCS score, and both must be considered when describing severely headinjured patients The anatomical patterns had clinical and physiopathological significance and provided useful prognostic information. The patterns facilitated improved therapeutic decision-making in severely head-injured patients CT findings alone had a correct prediction rate of outcome of 69.7%, compared with 75.8% for CT findings and GCS score Patients with basal ganglia haematoma shared many features with patients with diffuse white matter injury, and had a worse prognosis than other traumatic intracranial haematomas When used in conjunction with traditional division of intracranial haemorrhage (extradural, subdural or intracerebral), this categorization allowed a much better assessment of the risk of intracranial hypertension and fatal versus non-fatal outcome Subdural haematomas were more prone to reabsorption, and intracerebral and extradural haematomas were more likely to increase in size or to appear as new lesions In patients with a mass lesion, the GCS score was the only significant prognostic factor in the epidural haematoma group. GCS score and the presence of subarachnoid haemorrhage were predictive factors in the acute subdural haematoma group. Outcomes were unfavourable in the majority of patients with intracerebral haematoma

Eight anatomical patterns: normal, pure extracerebral haematoma, single brain contusion, general brain swelling, large extracerebral haematoma, multiple brain contusion – unilateral or bilateral, and diffuse axonal injury Scale of severity of abnormalities, including size of traumatic lesions

Macpherson et al., 198614 Presence of haematoma in basal ganglia region

Marshall et al., 199115

Mass lesions categorized into evacuated and non-evacuated

Servadei et al., 199516

Evolution of lesions, defined as an increase or decrease in the size of an already present haematoma or the appearance of a totally new lesion Mass lesions categorized into three types of haematoma: epidural, acute subdural, and intracerebral

Ono et al., 200117

985



TABLE 1 (continued): Studies of the classification and prediction of outcome in traumatic brain injury based on the presence and characteristics of intracranial local lesions on computed tomography (CT) Reference Maas et al., 2005

CT classification criteria 18

Study conclusions

Prognostic score chart for the This simple CT score permitted a probability of mortality in patients; clearer differentiation of epidural mass lesion scores: 0, prognostic risk, particularly in when absent; 1 when present patients with mass lesions

GCS, Glasgow Coma Scale.

TABLE 2: Studies of the classification and prediction of outcome in traumatic brain injury based on the presence and characteristics of diffuse brain injury on computed tomography (CT) Reference

CT classification criteria

Gennarelli et al., 198220

Diffuse axonal injury was The amount of diffuse axonal categorized into three subgroups injury was directly proportional to based on duration of coma: the severity of injury (duration of < 15 min, 16 min – 6 h, and > 6 h coma and quality of outcome)a Classification as diffuse axonal The prognosis of diffuse axonal injury included small injury was worst within this intraparenchymal haemorrhages, severe head injury series apart intraventricular haemorrhage, from cases with subdural both intraparenchymal and haematoma intraventricular haemorrhages, brainstem haemorrhagic contusion, generalized brain swelling and ventriculocisternal collapse Midline shift and compression or The most important absence of the mesencephalic characteristics for the prediction cisterns were used to indicate of abnormal intracranial pressure diffuse injury and death were midline shift, compression or obliteration of the mesencephalic cisterns, and the presence of subarachnoid blood. Diffuse hemispheric swelling was also found to be associated with an early episode of either hypoxia or hypotension Diffuse head injury was divided This more accurate categorization into four subgroups using the of diffuse head injury permits status of the mesencephalic specific subsets of patients to be cisterns, the degree of midline targeted for specific types of shift in mm, and the presence or therapy absence of one or more surgical masses

Cordobés et al., 198621

Eisenberg et al., 199022

Marshall et al., 199115

aTraumatic

Study conclusions

coma was produced in 45 monkeys by accelerating the head without impact in one of three

directions.

986


GW Zhu, F Wang, WG Liu Computed tomographic imaging in traumatic brain injury fractures of the skull base. These two types can then be further subdivided: skull vault fractures may be linear or depressed; while skull base fractures may be in the anterior fossa, medial fossa or posterior fossa.29 The incidence of traumatic carotid cavernous fistula is higher in patients with medial fossa fractures than in those with anterior or posterior fossa fractures, especially when the fractures are transverse or oblique.30 Twoand three-dimensional CT can be used to identify and classify Le Fort type fractures; three-dimensional CT can provide valuable information concerning spatial relationships, enabling pre-operative planning of surgical treatment.31 With a rising prevalence of firearm injuries, it is increasingly common to find foreign bodies in the head. In addition to locating foreign objects and determining whether removal is necessary, CT imaging can also help track the path and subsequent movement of the foreign body, so the corresponding complications can be anticipated. Non-contrast CT remains the imaging modality of choice. Depending on their size and velocity, foreign bodies can cause damage by different mechanisms: direct laceration, shock-wave transmission (pulsations that emanate from the front of a projectile), and cavitation (caused by the suction force created along the path of the foreign body).32

Prediction of outcome in TBI Accurate prediction of long-term outcomes after emergency admission to hospital is useful as an aid to clinical decision-making in patients with severe head injury.33 The reliability of patient assessment using the GCS and pupil reactivity is high.34,35 Based on the International Mission on Prognosis and Analysis of Clinical Trials in TBI (IMPACT) database,36 a series of papers have

reported associations between the Glasgow Outcome Scale (GOS) and demographic characteristics,37 cause of injury,38 secondary insults,39 GCS and pupil response,40 blood pressure,41 CT scan features42 and laboratory parameters.43 The most powerful independent prognostic variables were age, GCS motor score, pupil response and CT characteristics, including the Marshall CT classification and traumatic SAH.44 Since the introduction of CT imaging, the treatment of patients with head injury has improved considerably and it has become a powerful prognostic tool for improving clinical care in TBI, significantly reducing both morbidity and mortality.8,9 In patients with severe to moderate TBI, the outcome is much better in the absence of intracranial abnormalities.45 The prognostic value of individual CT characteristics has been well documented, including the status of basal cisterns (Table 3),18,21,22,42,46 – 50 midline shift (Table 4),22,42,49 – 54 traumatic SAH (Table 5)17,18,22,25,27,28,42,50,53,55 – 57 and the presence and type of intracranial lesions (Table 6).18,42,49,53,58 – 60

THE MARSHALL CT CLASSIFICATION In 1991, Marshall et al.15 proposed a CT classification for grouping patients with TBI according to multiple CT characteristics. Currently, this CT classification is the most frequently used prognostic method that incorporates the anatomical nature of the injury in the determination of outcome after acute TBI. The Marshall CT classification uses the findings from CT scans on the status of the mesencephalic cisterns, the degree of midline shift and the presence or absence of local lesions to categorize patients into six different groups (Table 7). It allows the identification of patients at risk from deterioration from intracranial hypertension, which offers the possibility of

987



TABLE 3: Studies of the prediction of outcome in traumatic brain injury based on the status of the basal cisterns on computed tomography (CT) Reference van Dongen et al., 1983

Cordobés et al., 198621 Eisenberg et al., 199022 Selladurai et al., 199247 Liu et al., 199548

Kakarieka et al., 199549


Maas et al., 200518

Maas et al., 200742

Study conclusions 46

The state of the basal cisterns proved to be a very powerful prognostic indicator. The percentage of accurate predictions was markedly higher with a combination of clinical and CT features than with clinical or CT features alone The presence of ventriculocisternal collapse and an unfavourable outcome correlated with raised intracranial pressure Compression or obliteration of the mesencephalic cisterns was related to abnormal intracranial pressure and death Perimesencephalic cistern obliteration was associated with a poor outcome A grading system using changes in the brainstem and the perimesencephalic cistern was well correlated to dynamic changes in intracranial hypertension A system based on the status of the mesencephalic cisterns, the degree of midline shift and the presence or absence of mass lesions allowed high risk patients to be identified and outcome to be predicted Haematoma thickness, midline shift, status of the basal cisterns and presence of subarachnoid haemorrhage were related to outcome when identified on the initial (early) CT examination Prediction of outcome was further increased by adding intraventricular and traumatic subarachnoid haemorrhage and by more detailed differentiation of mass lesions and basal cisterns Partial obliteration of the basal cisterns, traumatic subarachnoid haemorrhage or midline shift were strongly related to a poorer outcome

early intervention. Since its introduction in 1991, this classification has been increasingly used for predicting outcome, including overall survival, GOS, elevated ICP and neuropsychological consequences. There is a very strong relationship between the Marshall CT classification, mortality and the frequency of elevated ICP.15 The presence of a midline shift of > 5 mm on the initial brain CT scan and a high or mixed density lesion > 25 cm3 in volume have both been correlated with early death.61 The relative risk of requiring a delayed operation has been shown to be related to the Marshall CT classification of initial CT scans (diffuse injury IV, 30.7%;

diffuse injury III, 30.5%; non-evacuated mass, 20.0%; evacuated mass, 20.2%; diffuse injury II, 12.1%; diffuse injury I, 8.6%).62 When studying outcomes at 6 months following trauma using the GOS, Ono et al.17 found that all diffuse brain injury I patients recovered well. In the diffuse brain injury II group, age, the GCS score and the presence of multiple parenchymal lesions on CT scans were significantly correlated with outcome. For the diffuse brain injury III and IV groups, the only significant prognostic factor was the GCS score. In patients with a mass lesion, the GCS score was the only significant prognostic factor in the epidural haematoma group, but the GCS score and the presence of SAH were

988



TABLE 4: Studies of the prediction of outcome in traumatic brain injury based on midline shift on computed tomography (CT) Reference Young et al., 1981


Eisenberg et al., 199022 Quattrocchi et al., 199152 Kakarieka et al., 199549


Azian et al., 200153

Pillai et al., 200354

Maas et al., 200742

GCS scores and shift data were highly accurate indicators of outcome Midline shift was related to abnormal intracranial pressure and death Quantification of midline shift was a predictor of poor outcome A system based on the status of the mesencephalic cisterns, the degree of midline shift and the presence or absence of mass lesions allowed high risk patients to be identified and outcome to be predicted Haematoma thickness, midline shift, status of the basal cisterns and presence of subarachnoid haemorrhage were related to outcome when identified on the initial (early) CT examination Predictors of outcome included intracerebral haemorrhage (ICH), extradural haemorrhage (EDH), intraventricular haemorrhage, subarachnoid haemorrhage, subdural haemorrhage, the site of the ICH, the volume of the EDH, and midline shift The most important predictors of poor outcome were the horizontal oculocephalic reflex, the motor score of the GCS, and midline shift on CT scan Partial obliteration of the basal cisterns, traumatic subarachnoid haemorrhage or midline shift were strongly related to a poorer outcome


predictive factors in the acute subdural haematoma group. Outcomes were unfavourable in the majority of patients with intracerebral haematoma. With regard to the frequency of elevated ICP, Hiler et al.63 reported that the Marshall CT classification correlated significantly but weakly with ICP measured during the first 24 h of monitoring but not with mean ICP over the total time spent in intensive care. In contrast, the 6 month GOS score correlated with the initial CT scan findings. These results all suggest that the Marshall CT scan classification provides accurate predictions regarding the likelihood of a fatal or non-fatal outcome. In an evaluation of the Traumatic Coma Data Bank (TCDB) classification by Levin et al.64 (based on the classification system devised by Marshall et

al.15), no group differences on neurobehavioural tests were found in a group of severe TBI patients. In contrast, Mataró et al.65 reported that there was a relationship between the acute intracranial lesion diagnosis according to the Marshall CT classification and neuropsychological results and ventricular dilatation indices at 6 months post-injury. Although the Marshall CT classification is an adjunct to clinical parameters and is easy to use, it does not take into account all the possible prognostic factors visible on CT and clearly has some limitations. An important limitation is that the classification is partially based on arbitrary assessments and is dependent on the accuracy of measured volumes of focal mass lesions. Furthermore, TCDB categories V (mass lesion surgically

989



TABLE 5: Studies of the prediction of outcome in traumatic brain injury based on the presence of traumatic subarachnoid haemorrhage (SAH) on computed tomography (CT) Reference

Study conclusions

Eisenberg et al., 199022

The presence of subarachnoid blood was related to abnormal intracranial pressure and death The outcome of patients with traumatic SAH was significantly worse than that of patients whose first CT scan did not show subarachnoid blood Patients with traumatic SAH associated with a non-penetrating head injury had a worse outcome than similar patients without traumatic SAH Haematoma thickness, midline shift, status of the basal cisterns and presence of SAH were related to outcome when identified on the initial (early) CT examination. Patients with SAH on early CT were at highest risk for associated evolving contusions The GCS score and the presence of SAH were predictive factors in the acute subdural haematoma group Predictors of outcome included intracerebral haemorrhage (ICH), extradural haemorrhage (EDH), intraventricular haemorrhage, subarachnoid haemorrhage, subdural haemorrhage, the site of the ICH, the volume of the EDH, and midline shift Traumatic SAH on admission CT scans was an independent prognostic factor. Death among patients with traumatic SAH was related to the severity of the initial mechanical damage rather than to the effects of delayed vasospasm and secondary ischaemic brain damage Traumatic SAH is associated with more severe CT findings and a worse patient outcome Prediction of outcome was further increased by adding intraventricular and traumatic subarachnoid haemorrhage and by more detailed differentiation of mass lesions and basal cisterns The prognosis was poor in patients with poor GCS scores on admission, cysternal or fissural haemorrhage, traumatic SAH with cerebral contusion, or acute subdural haematoma Baseline predictors of outcome studied were age, motor score, pupillary reactivity, CT classification, traumatic SAH, hypoxia, hypotension, glycaemia and haemoglobin. Covariate adjustment for strong predictors should be incorporated in the analysis of future trials of traumatic brain injury Partial obliteration of the basal cisterns, traumatic SAH or midline shift were strongly related to a poorer outcome


Greene et al., 199655


Ono et al., 200117 Azian et al., 200153


Mattioli et al., 200328 Maas et al., 200518

Okten et al., 200625

Hernández et al., 200657

Maas et al., 200742 GCS, Glasgow Coma Scale.

evacuated) and VI (mass lesion not operated) are, in part, retrospective in nature as they depend on the decision to operate or not.66 Vos et al.66 evaluated the TCDB CT

classification in severe head-injury patients and found that it had high inter-observer and intra-observer reliability, but suggested that inter-observer agreement could be

990



TABLE 6: Studies of the prediction of outcome in traumatic brain injury based on the presence and type of intracranial lesions on computed tomography (CT) Reference Narayan et al., 1981



Azian et al., 200153

Caroli et al., 200159

Bahloul et al., 200460

Maas et al., 200518

Maas et al., 200742

A combination of clinical data predicted outcome significantly, including age, GCS score, pupillary response, presence of surgical mass lesions, extra-ocular motility and motor posturing A system based on the status of the mesencephalic cisterns, the degree of midline shift and the presence or absence of mass lesions allowed high risk patients to be identified and outcome to be predicted Predictors of outcome included intracerebral haemorrhage (ICH), extradural haemorrhage (EDH), intraventricular haemorrhage, subarachnoid haemorrhage, subdural haemorrhage, the site of the ICH, the volume of the EDH, and midline shift The presence of a subdural and intracerebral haematomapredominant lesion was a statistically significant predictor of a bad outcome (death or vegetative state) On multivariate analysis, age > 40 years, a simplified acute physiology score > 40, a GCS score < 7, an intracranial mass lesion, cerebral herniation, diabetes insipidus, and blood sugar level > 10 mmol/l all correlated with a poor prognosis Prediction of outcome was further increased by adding intraventricular and traumatic subarachnoid haemorrhage and by more detailed differentiation of mass lesions and basal cisterns The prognosis in patients with mass lesions was better for patients with an epidural haematoma and poorer for acute subdural haematoma


further improved by considering diffuse injury groups III and IV together and groups V and VI together.

In addition, the Marshall CT classification does not take the presence of traumatic SAH into account. The incidence of CT-

TABLE 7: The Marshall computed tomography (CT) classification of traumatic brain injury15 Category

Definition

Diffuse injury I Diffuse injury II

No visible intracranial pathology seen on CT scan Cisterns present with midline shift of 0 – 5 mm and/or lesion densities present; no high or mixed density lesion > 25 cm3; may include bone fragments and foreign bodies Diffuse injury III (swelling) Cisterns compressed or absent with midline shift of 0 – 5 mm; no high or mixed density lesion > 25 cm3 Diffuse injury IV (shift) Midline shift > 5 mm; no high or mixed density lesion > 25 cm3 Evacuated mass lesion (V) Any lesion surgically evacuated Non-evacuated mass lesion (VI) High or mixed density lesion > 25 cm3 not surgically evacuated

991


GW Zhu, F Wang, WG Liu Computed tomographic imaging in traumatic brain injury documented traumatic SAH in patients with severe TBI is 23 – 63%.23 It is most common in patients with subdural haematoma or haemorrhage contusion and is associated with a worse prognosis.23 – 25 Adding SAH to any CT classification system may enhance its predictive power for outcome and may have consequences for treatment and outcome issues, and in clinical trials.23,24

OTHER PROGNOSTIC MODELS BASED ON CT IMAGING A number of other CT classifications exist,11,53,67 – 71 but none of these has been as extensively evaluated as the Marshall CT classification. International guidelines on prognosis include CT classification as a major predictor of outcome based on class I evidence.72 In 2005, a study by Maas et al.18 compared alternative CT models with the Marshall CT classification and found it was preferable to use combinations of individual CT predictors rather than the Marshall CT classification for prognostic purposes in TBI. Later, in 2007, Maas et al.42 again evaluated

the prognostic value of CT scan characteristics in TBI and found that individual CT characteristics added substantially to the prognostic value of the CT classification alone. They found that making greater use of individual CT characteristics allowed them to improve on the already sizeable predictive value of the original Marshall CT classification scheme. These findings need to be corroborated through further prospective investigations.

Conclusion The Marshall CT classification was developed for descriptive purposes, but also yields important prognostic information in TBI. Future studies should address the complicated issue of how optimally to combine CT characteristics for prognostic purposes and how to improve on currently used CT classifications to predict outcome more accurately.

Conflicts of interest The authors had no conflicts of interest to declare in relation to this article.

• Received for publication 2 March 2009 • Accepted subject to revision 5 March 2009 • Revised accepted 27 July 2009 Copyright © 2009 Field House Publishing LLP References 1 Teasdale GM, Jennet B: Assessment of coma and impaired consciousness. Lancet 1974; ii: 81 – 84. 2 Malec JF, Brown AW, Leibson CL, et al: The Mayo Classification System for Traumatic Brain Injury Severity. J Neurotrauma 2007; 24: 1417 – 1424. 3 Balestreri M, Czosnyka M, Chatfield DA, et al: Predictive value of Glasgow Coma Scale after brain trauma: change in trend over the past ten years. J Neurol Neurosurg Psychiatry 2004; 75: 161 – 162. 4 Buechler CM, Blostein PA, Koestner A, et al: Variation among trauma centers’ calculation of Glasgow Coma Scale score: results of a national survey. J Trauma 1998; 45: 429 – 432. 5 Moskopp D, Stähle C, Wassmann H: Problems of the Glasgow Coma Scale with early

intubated patients. Neurosurg Rev 1995; 18: 253 – 257. 6 Firsching R, Woischneck D, Klein S, et al: Classification of severe head injury based on magnetic resonance imaging. Acta Neurochir (Wien) 2001; 143: 263 – 271. 7 Uchino Y, Okimura Y, Tanaka M, et al: Computed tomography and magnetic resonance imaging of mild head injury – is it appropriate to classify patients with Glasgow Coma Scale Score of 13 to 15 as ‘mild injury’? Acta Neurochir (Wien) 2001; 143: 1031 – 1037. 8 Teasdale G, Galbraith S, Murray L, et al: Management of traumatic intracranial haematoma. Br Med J 1982; 285: 1695 – 1697. 9 Siegel S, Castellan NJ: Nonparametric Statistics for the Behavioral Sciences, 2nd edn. New York: McGraw-Hill, 1988. 10 Lee B, Newberg A: Neuroimaging in traumatic

992



brain imaging. NeuroRx 2005; 2: 372 – 383. 11 Gennarelli TA, Spielman GM, Langfitt TW, et al: Influence of the type of intracranial lesion on outcome from severe head injury. J Neurosurg 1982; 56: 26 – 32. 12 Lobato RD, Cordobes F, Rivas JJ, et al: Outcome from severe head injury related to the type of intracranial lesion. A computerized tomography study. J Neurosurg 1983; 59: 762 – 774. 13 Lipper MH, Kishore PR, Enas GG, et al: Computed tomography in the prediction of outcome in head injury. Am J Roentgenol 1985; 144: 483 – 486. 14 Macpherson P, Teasdale E, Dhaker S, et al: The significance of traumatic haematoma in the region of the basal ganglia. J Neurol Neurosurg Psychiatry 1986; 49: 29 – 34. 15 Marshall LF, Marshall SB, Klauber MR, et al: A new classification of head injury based on computerized tomography. J Neurosurg 1991; 75: S14 – S20. 16 Servadei F, Nanni A, Nasi MT, et al: Evolving brain lesions in the first 12 hours after head injury: analysis of 37 comatose patients. Neurosurgery 1995; 37: 899 – 906. 17 Ono J, Yamaura A, Kubota M, et al: Outcome prediction in severe head injury: analyses of clinical prognostic factors. J Clin Neurosci 2001; 8: 120 – 123. 18 Maas AI, Hukkelhoven CW, Marshall LF, et al: Prediction of outcome in traumatic brain injury with computed tomographic characteristics: a comparison between the computed tomographic classification and combinations of computed tomographic predictors. Neurosurgery 2005; 57: 1173 – 1182. 19 Becker DP, Miller JD, Ward JD, et al: The outcome from severe head injury with early diagnosis and intensive management. J Neurosurg 1977; 47: 491 – 502. 20 Gennarelli TA, Thibault LE, Adams JH, et al: Diffuse axonal injury and traumatic coma in the primate. Ann Neurol 1982; 12: 564 – 574. 21 Cordobés F, Lobato RD, Rivas JJ, et al: Posttraumatic diffuse axonal brain injury. Analysis of 78 patients studied with computed tomography. Acta Neurochir (Wien) 1986; 81: 27 – 35. 22 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 1990; 73: 688 – 698. 23 Kakarieka A: Review on traumatic subarachnoid hemorrhage. Neurol Res 1997; 19: 230 – 232. 24 Greene KA, Marciano FF, Johnson BA, et al: Impact of traumatic subarachnoid hemorrhage on outcome in nonpenetrating head injury. Part I: a proposed computerized tomography grading scale. J Neurosurg 1995; 83: 445 – 452.

25 Okten AI, Gezercan Y, Ergün R: Traumatic subarachnoid hemorrhage: a prospective study of 58 cases. Ulus Travma Acil Cerrahi Derg 2006; 12: 107 – 114 [in Turkish]. 26 Atzema C, Mower WR, Hoffman JR, et al: Prevalence and prognosis of traumatic intraventricular hemorrhage in patients with blunt head trauma. J Trauma 2006; 60: 1010 – 1017. 27 Kakarieka A, Braakman R, Schakel EH: Clinical significance of the finding of subarachnoid blood on CT scan after head injury. Acta Neurochir (Wien) 1994; 129: 1 – 5. 28 Mattioli C, Beretta L, Gerevini S, et al: Traumatic subarachnoid hemorrhage on the computerized tomography scan obtained at admission: a multicenter assessment of the accuracy of diagnosis and the potential impact on patient outcome. J Neurosurg 2003; 98: 37 – 42. 29 Wang ZC (ed): Neurosurgery. Wuhan: Hubei Science and Technology Press, 2005; pp 414 – 417. 30 Liang W, Xiaofeng Y, Weiguo L, et al: Traumatic carotid cavernous fistula accompanying basilar skull fracture: a study on the incidence of traumatic carotid cavernous fistula in the patients with basilar skull fracture and the prognostic analysis about traumatic carotid cavernous fistula. J Trauma 2007; 63: 1014 – 1020. 31 Chen WJ, Yang YJ, Fang YM, et al: Identification and classification in Le Fort type fractures by using 2D and 3D computed tomography. Chin J Traumatol 2006; 9: 59 – 64. 32 Kim PE, Go JL, Zee CS: Radiographic assessment of cranial gunshot wounds. Neuroimaging Clin N Am 2002; 12: 229 – 248. 33 Signorini DF, Andrews PJ, Jones PA, et al: Predicting survival using simple clinical variables: a case study in traumatic brain injury. J Neurol Neurosurg Psychiatry 1999; 66: 20 – 25. 34 Braakman R, Avezaat CJ, Maas AI, et al: Inter observer agreement in the assessment of the motor response of the Glasgow ‘coma’ scale. Clin Neurol Neurosurg 1977; 80: 100 – 106. 35 Teasdale GM, Knill-Jones R, van der Sande J: Observer variability in assessing impaired consciousness and coma. J Neurol Neurosurg Psychiatry 1978; 41: 603 – 610. 36 Marmarou A, Lu J, Butcher I, et al: IMPACT database of traumatic brain injury: design and description. J Neurotrauma 2007; 24: 239 – 250. 37 Mushkudiani NA, Engel DC, Steyerberg EW, et al: Prognostic value of demographic characteristics in traumatic brain injury: results from the IMPACT study. J Neurotrauma 2007; 24: 259 – 269. 38 Butcher I, McHugh GS, Lu J, et al: Prognostic value of cause of injury in traumatic brain injury: results from the IMPACT study. J

993



Neurotrauma 2007; 24: 281 – 286. 39 McHugh GS, Engel DC, Butcher I, et al: Prognostic value of secondary insults in traumatic brain injury: results from the IMPACT study. J Neurotrauma 2007; 24: 287 – 293. 40 Marmarou A, Lu J, Butcher I, et al: Prognostic value of the Glasgow Coma Scale and pupil reactivity in traumatic brain injury assessed pre-hospital and on enrollment: an IMPACT analysis. J Neurotrauma 2007; 24: 270 – 280. 41 Butcher I, Maas AI, Lu J, et al: Prognostic value of admission blood pressure in traumatic brain injury: results from the IMPACT study. J Neurotrauma 2007; 24: 294 – 302. 42 Maas AI, Steyerberg EW, Butcher I, et al: Prognostic value of computerized tomography scan characteristics in traumatic brain injury: results from the IMPACT study. J Neurotrauma 2007; 24: 303 – 314. 43 Van Beek JG, Mushkudiani NA, Steyerberg EW, et al: Prognostic value of admission laboratory parameters in traumatic brain injury: results from the IMPACT study. J Neurotrauma 2007; 24: 315 – 328. 44 Murray GD, Butcher I, McHugh GS, et al: Multivariable prognostic analysis in traumatic brain injury: results from the IMPACT study. J Neurotrauma 2007; 24: 329 – 337. 45 Chesnut RM: Evolving models of neurotrauma critical care: an analysis and call to action. Clin Neurosurg 2000; 46: 185 – 195. 46 van Dongen KJ, Braakman R, Gelpke GJ: The prognostic value of computerized tomography in comatose head-injured patients. J Neurosurg 1983; 59: 951 – 957. 47 Selladurai BM, Jayakumar R, Tan YY, et al: Outcome prediction in early management of severe head injury: an experience in Malaysia. Br J Neurosurg 1992; 6: 549 – 557. 48 Liu HM, Tu YK, Su CT: Changes of brainstem and perimesencephalic cistern: dynamic predictor of outcome in severe head injury. J Trauma 1995; 38: 330 – 333. 49 Kakarieka A, Braakman R, Schakel EH: Classification of head injuries based on computerized tomography: prognostic value. Neurologia 1995; 10: 159 – 161 [in Spanish]. 50 Servadei F, Nasi MT, Giuliani G, et al: CT prognostic factors in acute subdural haematomas: the value of the ‘worst’ CT scan. Br J Neurosurg 2000; 14: 110 – 116. 51 Young B, Rapp RP, Norton JA, et al: Early prediction of outcome in head-injured patients. J Neurosurg 1981; 54: 300 – 303. 52 Quattrocchi KB, Prasad P, Willits NH, et al: Quantification of midline shift as a predictor of poor outcome following head injury. Surg Neurol 1991; 35: 183 – 188. 53 Azian AA, Nurulazman AA, Shuaib L, et al: Computed tomography of the brain in predicting outcome of traumatic intracranial

haemorrhage in Malaysian patients. Acta Neurochir (Wien) 2001; 143: 711 – 720. 54 Pillai SV, Kolluri VR, Praharaj SS: Outcome prediction model for severe diffuse brain injuries: development and evaluation. Neurol India 2003; 51: 345 – 349. 55 Greene KA, Jacobowitz R, Marciano FF, et al: Impact of traumatic subarachnoid hemorrhage on outcome in nonpenetrating head injury. Part II: relationship to clinical course and outcome variables during acute hospitalization. J Trauma 1996; 41: 964 – 971. 56 Servadei F, Murray GD, Teasdale GM, et al: Traumatic subarachnoid hemorrhage: demographic and clinical study of 750 patients from the European brain injury consortium survey of head injuries. Neurosurgery 2002; 50: 261 – 267. 57 Hernández AV, Steyerberg EW, Butcher I, et al: Adjustment for strong predictors of outcome in traumatic brain injury trials: 25% reduction in sample size requirements in the IMPACT study. J Neurotrauma 2006; 23: 1295 – 1303. 58 Narayan RK, Greenberg RP, Miller JD, et al: Improved confidence of outcome prediction in severe head injury. A comparative analysis of the clinical examination, multimodality evoked potentials, CT scanning, and intracranial pressure. J Neurosurg 1981; 54: 751 – 762. 59 Caroli M, Locatelli M, Campanella R, et al: Multiple intracranial lesions in head injury: clinical considerations, prognostic factors, management, and results in 95 patients. Surg Neurol 2001; 56: 82 – 88. 60 Bahloul M, Chelly H, Ben Hmida M, et al: Prognosis of traumatic head injury in South Tunisia: a multivariate analysis of 437 cases. J Trauma 2004; 57: 255 – 261. 61 Rapenne T, Lenfant F, N’Guyen Kim L, et al: Predictive factors of short-term mortality in patients with severe head injury. Presse Med 1997; 26: 1661 – 1665 [in French]. 62 Lobato RD, Gomez PA, Alday R, et al: Sequential computerized tomography changes and related final outcome in severe head injury patients. Acta Neurochir (Wien) 1997; 139: 385 – 391. 63 Hiler M, Czosnyka M, Hutchinson P, et al: Predictive value of initial computerized tomography scan, intracranial pressure, and state of autoregulation in patients with traumatic brain injury. J Neurosurg 2006; 104: 731 – 737. 64 Levin HS, Gary HE Jr, Eisenberg HM, et al: Neurobehavioral outcome 1 year after severe head injury. Experience of the Traumatic Coma Data Bank. J Neurosurg 1990; 73: 699 – 709. 65 Mataró M, Poca MA, Sahuquillo J, et al: Neuropsychological outcome in relation to the Traumatic Coma Data Bank classification of computed tomography imaging. J Neurotrauma

994



2001; 18: 869 – 879. 66 Vos PE, van Voskuilen AC, Beems T, et al: Evaluation of the traumatic coma data bank computed tomography classification for severe head injury. J Neurotrauma 2001; 18: 649 – 655. 67 Teasdale G, Teasdale E, Hadley D: Computed tomographic and magnetic resonance imaging classification of head injury. J Neurotrauma 1992; 9(suppl 1): S249 – S257. 68 Hukkelhoven CW, Steyerberg EW, Habbema JD, et al: Predicting outcome after traumatic brain injury: development and validation of a prognostic score based on admission characteristics. J Neurotrauma 2005; 22: 1025 – 1039. 69 Lannoo E, Van Rietvelde F, Colardyn F, et al:

Early predictors of mortality and morbidity after severe closed head injury. J Neurotrauma 2000; 17: 403 – 414. 70 Nissen JJ, Jones PA, Signorini DF, et al: Glasgow head injury outcome prediction program: an independent assessment. J Neurol Neurosurg Psychiatry 1999; 67: 796 – 799. 71 Schreiber MA, Aoki N, Scott BG, et al: Determinants of mortality in patients with severe blunt head injury. Arch Surg 2002; 137: 285 – 290. 72 Chesnut RM, Ghajar J, Maas AR: Guidelines for the management and prognosis of severe traumatic brain injury. Part II: early indicators of prognosis in severe traumatic brain injury. J Neurotrauma 2000; 17: 557 – 627.

Author’s address for correspondence Dr WG Liu Institute of Brain Medicine, Zhejiang University, 88 Jiefang Road, Hangzhou City 310009, Zhejiang Province, China. E-mail: [email protected]

995
