Objective: There is enormous research and clinical interest in blood-based biomarkers of mild traumatic brain injury (MTBI) sustained in sports, daily life, or military service. We examined the reliability of a commercially available assay for S100B used on the same samples by two different laboratories separated by 2 years in time.
Methods and procedures: A cohort of 163 adult patients (head CT-scanned, n = 110) with mild head injury were enrolled from the emergency department (ED). All had Glasgow Coma Scale scores of 14 or 15 in the ED (94.4% = 15). The mean time between injury and venous blood sampling was 2.9 h (SD = 1.4; Range = 0.5-6.0 h). Serum S100B was measured at two independent centers using the same high throughput clinical assay (Elecsys S100B®; Roche Diagnostics).
Results: The Spearman correlation between the two assays in the total sample (N = 163) was r = 0.93. A Wilcoxson Signed Ranks test indicated that the median scores for the values differed (Z = 2,082, p < .001, Cohen's d = 0.151, small effect size). The values obtained from the two laboratories were very similar for identifying traumatic intracranial abnormalities (sensitivity = 80.1% versus 85.7%).
Conclusions: The serum S100B results measured using the same assay in different laboratories yielded highly correlated and clinically similar, but clearly not identical, results.
Keywords: S100B protein; Traumatic brain injuries; computed tomography; emergency treatment; psychometrics.