Cranial computed tomography (CT) scan is the most widely used tool to rule out intracranial lesions after pediatric traumatic brain injury (TBI). However, in pediatric population, the radiation exposure can lead to an increased risk of hematological and brain neoplasm. Defined in 2019 National Institute for Health and Care Excellence (NICE) guidelines as "troponins for the brain", serum biomarkers measurements, particularly S100B, have progressively emerged as a supplementary tool in the management of TBI thanks to their capacity to predict intracranial post-traumatic lesions. This systematic review was conducted following the PRISMA protocol (preferred reporting items for systematic reviews and meta-analyses). No chronological limits of study publications were included. Studies reporting data from children with TBI undergoing serum S100B measurement and computed tomography (CT) scans were included. Of 380 articles screened, 10 studies met the inclusion criteria. Patients admitted with mild-TBI in the Emergency Department (ED) were 1325 (80.25%). The overall pooled sensitivity and specificity were 98% (95% CI, 92-99%) and 45% (95% CI, 29-63%), respectively. The meta-analysis revealed a high negative predictive value (NVP) (99%; 95% CI, 94-100%) and a low positive predictive value (PPV) (41%; 95% CI, 16-79%). Area under the curve (AUC) was 76% (95% CI, 65-85%). The overall pooled negative predictive value (NPV) was 99% (95% CI, 99-100%). The measurement of serum S100B in the diagnostic workflow of mTBI could help informed decision-making in the ED setting, potentially safely reducing the use of CT scan in the pediatric population. The high sensitivity and excellent negative predictive values look promising and seem to be close to the values found in adults. Despite this, it must be pointed out the high heterogeneity (> 90%) found among studies. In order for S100B to be regularly introduced in the pediatric workflow for TBI, it is important to conduct further studies to obtain cut-off levels based on pediatric reference intervals.
Read full abstract