Objective: To analyze the time course of serum protein S-100b in patients with traumatic brain injury deteriorating to brain death and to investigate the predictive value of initial S-100b levels in relation to clinical and radiologic measures of injury severity with regard to brain death.
Methods: Forty-seven patients who sustained severe head injury were studied. Blood samples for measurement of S-100b were drawn on admission in the intensive care unit and every 24 hours thereafter for a maximum of 6 consecutive days or until brain death occurred. Variables related to outcome were recorded, including age, sex, Glasgow Coma Scale (GCS), and brain CT findings on admission. Outcome was defined as deterioration to brain death or not.
Results: Of the 47 patients studied, 17 deteriorated to brain death and 30 did not. On admission, patients who became brain dead had higher median serum S-100b levels compared with those who did not (2.32 microg/L vs 1.04 micro g/L, p = 0.0028). Logistic regression analysis showed that initial S-100b was an independent predictor of brain death (p = 0.041), in the presence of advanced age (p = 0.043) and low GCS score (p = 0.013). The odds ratio of 2.09 (95% CI, 1.03 to 4.25) indicates a more than doubling of the probability of deteriorating to brain death per 1- micro g/L increase in S-100b on admission. At clinical brain death, median S-100b was higher in patients with brain death compared with the peak S-100b value obtained over a 6-day period in those who did not become brain dead (6.58 microg/L vs 1.49 microg/L, p < 0.0001).
Conclusions: Prediction of brain death after severe head injury can be improved by combining clinical and S-100b data; thus, serum S-100b determination deserves to be included in the neuromonitoring of patients with severe traumatic brain injury.