Abstract
In adults, glial fibrillary acidic protein (GFAP) has been shown to out-perform S100β in detecting intracranial lesions on computed tomography (CT) in mild traumatic brain injury (TBI). This study examined the ability of GFAP and S100β to detect intracranial lesions on CT in children and youth involved in trauma. This prospective cohort study enrolled a convenience sample of children and youth at two pediatric and one adult Level 1 trauma centers following trauma, including both those with and without head trauma. Serum samples were obtained within 6 h of injury. The primary outcome was the presence of traumatic intracranial lesions on CT scan. There were 155 pediatric trauma patients enrolled, 114 (74%) had head trauma and 41 (26%) had no head trauma. Out of the 92 patients who had a head CT, eight (9%) had intracranial lesions. The area under the receiver operating characteristic curve (AUC) for distinguishing head trauma from no head trauma for GFAP was 0.84 (0.77-0.91) and for S100β was 0.64 (0.55-0.74; p<0.001). Similarly, the AUC for predicting intracranial lesions on CT for GFAP was 0.85 (0.72-0.98) versus 0.67 (0.50-0.85) for S100β (p=0.013). Additionally, we assessed the performance of GFAP and S100β in predicting intracranial lesions in children ages 10 years or younger and found the AUC for GFAP was 0.96 (95% confidence interval [CI] 0.86-1.00) and for S100β was 0.72 (0.36-1.00). In children younger than 5 years old, the AUC for GFAP was 1.00 (95% CI 0.99-1.00) and for S100β 0.62 (0.15-1.00). In this population with mild TBI, GFAP out-performed S100β in detecting head trauma and predicting intracranial lesions on head CT. This study is among the first published to date to prospectively compare these two biomarkers in children and youth with mild TBI.