Derivation of a bioclinical prognostic index in severe head injury

We investigated the problem of outcome prediction from seven risk factors in 40 severely head injured patients - 13 favorable and 27 unfavorable outcomes. By applying stepwise logistic discriminant analysis to the patients' data, we selected three significant risk variables: cerebrospinal fluid (CSF) CK-BB isoenzyme activity recorded on admission, severely raised intracranial pressure (more than 40 mmHg) and age, respectively. CSF CK-BB activity, which quantifies the initial neurological damage, proved to be the best prognostic factor. The presence of severe intracranial hypertension was always associated with a bad outcome, whereas its absence was not necessarily indicative of good prognosis. Finally, we combined the three selected variables into a single risk index, which allowed correct predictions in 92% of patients with favorable outcome and in 85% of patients with unfavorable outcome (total predictive efficiency 88%).