Reducing head computed tomography after mild traumatic brain injury: Screening value of clinical findings and S100B protein levels

Primary objective: The present prospective study was performed to investigate whether primary clinical findings and serum S100B concentrations at 3 and 6 hours post-trauma can contribute to the selection of patients for an initial computed tomography (CT) scanning. Research design and methods: S100B was measured in serum samples obtained at 3 and 6 hours after the injury. Adjusted odds ratios (OR) and 95% confidence interval (CI) associated with demographics and clinical predictors of positive CT scan were calculated. Sensitivity, specificity, negative and positive predictive values were also calculated for S100B levels. Main outcomes and results: It was found that the presence of loss of consciousness (OR = 2.3; 95% CI = 1.00-4.01; p = 0.008) and post-traumatic vomiting >= 2 episodes (OR = 1.8; 95% CI = 1.08-3.29; p = 0.019) are factors associated with positive CT scan. In this study the best cut-off point of 0.115 mu g L-1 for 3-hour S100B has sensitivity of 94.9% (95% CI = 86.8-98.3) with specificity of 35.4% (95% CI = 25.2-47.0) to predict intracranial injury on CT scanning. The corresponding results for 6-hour S100B > 0.210 mu g L-1 were 98.7% (95% CI = 92.1-99.9) for sensitivity and 39.2% (95% CI = 28.6-50.8) for specificity. Conclusions: Serum S100B measurement along with clinical evaluation of patients with mild traumatic brain injury has promising screening value to support selection of patients for CT scanning.