Use of the CRASH study prognosis calculator in patients with severe traumatic brain injury treated with an intracranial pressure-targeted therapy

Based on the Corticosteroid Randomisation after Significant Head Injury (CRASH) trial database, a prognosis calculator has been developed for the prediction of outcome in an individual patient with a head injury. In 47 patients with severe traumatic brain injury (sTBI) prospectively treated using an intracranial pressure (ICP) targeted therapy, the individual prognosis for mortality at 14 days and unfavourable outcome at 6 months was calculated and compared with the actual outcome. An overestimation of the risk of mortality and unfavourable outcome was found. The mean risk for mortality and unfavourable outcome were estimated to be 44.6 +/- 32.5% (95% confidence interval [CI], 35.1-54.2%) and 69.3 +/- 23.7% (95% CI, 62.3-76.2%). The actual outcome was 4.3% and 42.6% respectively. The absolute risk reduction (ARR) for mortality was 33.1% and for unfavourable outcome 29.8%. A logistic fit for outcome at 6 months shows a statistically significant difference (p < 0.01). A receiver operating characteristic (ROC) curve analysis shows an area under the curve (AUC) of 0.691. The CRASH prognosis calculator overestimates the risk of mortality and unfavourable outcome in patients with sTBI treated with an ICP-targeted therapy based on the Lund concept. We do not advocate the use of the calculator for treatment decisions in individual patients. We further conclude that patients with blunt sTBI admitted within 8 hours of trauma should be treated regardless of their clinical status as long as the initial cerebral perfusion pressure is > 10 mmHg. (C) 2012 Elsevier Ltd. All rights reserved.