The value of serum biomarkers in prediction models of outcome after mild traumatic brain injury

Abstract

Background: To determine, using a civilian model of mild traumatic brain injury (TBI), the added value of biomarker sampling upon prognostication of outcome at 1 week and 6 weeks postinjury.

Methods: The Galveston Orientation and Amnesia test was administered, and blood samples for serum protein S100B and neuron-specific enolase (NSE) were collected from 141 emergency department patients within 4 hours of a suspected mild TBI (mTBI). The Rivermead Post-Concussion Symptoms Questionnaire (RPQ) was administered via telephone 3 days postinjury. Patients were assessed by a physician at 1 week (n = 113; 80%) and 6 weeks (n = 95; 67%) postinjury. Neurocognitive and postural stability measures were also administered at these follow-ups.

Results: Levels of S100B and NSE were found to be abnormally elevated in 49% and 65% of patients with TBI, respectively. Sixty-eight percent and 38% of the patients were considered impaired at 1 week and 6 weeks postinjury, respectively. Stepwise logistic regression modeling identified admission Galveston Orientation and Amnesia test score, S100B level, and RPQ score at day 3 postinjury to be predictive of poor outcome at 1 week postinjury (c-statistic 0.877); female gender, loss of consciousness, NSE level, and RPQ score at day 3 postinjury were predictive of poor outcome at 6 weeks postinjury (c-statistic 0.895). The discriminative power of the biomarkers alone was limited.

Conclusions: Biomarkers, in conjunction with other readily available determinants of outcome assessed in the acute period after injury, add value in the early prognostication of patients with mTBI. Our findings are consistent with the notion that S100B and NSE point to biological mechanisms underlying poor outcome after mTBI.