Acute subdural hematoma--prediction of outcome with a linear discriminant function

We devised a linear discriminant function to predict the outcome for patients with acute subdural hematoma (ASDH) based on a consecutive series of 170 ASDH cases with mild to severe head injury [Glasgow Coma Scale (GCS) 3-15]. Functional recovery was achieved in 50.0% of patients and the mortality was 36.5%. The relationship between initial clinical and radiological signs and the outcome 3 months after admission was studied retrospectively by Mann-Whitney's U-test and Pearson's chi-squared test. Fourteen factors (GCS, pupillary response, motor paresis, age, hematoma volume and thickness, midline shift, association with cerebral contusion and subarachnoid hemorrhage, obliteration of the basal, ambient, or quadrigeminal cistern on computed tomography, fibrin-fibrinogen degradation product level, and intracranial pressure) were found to correlate significantly (p < 0.01) with outcome. Linear discriminant functions were formulated by multivariate analysis to investigate the relationship between these factors and recovery or poor prognosis. The following formula was obtained: Z = -0.110 + 0.013 (Age) - 0.108 (GCS) + 0.397 (Eye) + 0.003 (Shift) + 0.268 (Ambient). Functional recovery could be predicted by a negative Z value, with an accuracy of 90.59%. This simple discriminant function is useful for predicting the outcome of ASDH.