Introduction: Mild traumatic brain injury (mTBI) is a major cause of morbidity but there are no validated tools to help clinicians predict post-concussion symptoms. This systematic review and meta-analysis aimed to determine the prognostic value of S-100B protein to predict post-concussion symptoms following a mTBI in adults. Methods: The protocol of this systematic review was registered with the PROSPERO database (CRD42016032578). A search strategy was performed on seven databases (CINAHL, Cochrane CENTRAL, EMBASE, MEDLINE, Web of Knowledge, PyscBITE, PsycINFO) from their inception to October 2016. Studies evaluating the association between S-100B protein level and post-concussion symptoms assessed at least seven days after the mTBI were eligible. Individual patient data were requested. Studies eligibility assessment, data extraction and risk of bias assessment were performed independently by two researchers. Analyses were done following the meta-analysis using individual participant data or summary aggregate data guidelines from the Cochrane Methodology Review Group. Results: Outcomes were dichotomised as persistent (≥3 months) or early (≥7 days <3 months). Our search strategy yielded 23,298 citations of which 29 studies presenting between seven and 223 patients (n=2505) were included. Post-concussion syndrome (PCS) (16 studies), neuropsychological symptoms (9 studies) and health-related quality of life (4 studies) were the most frequently presented outcomes. The S-100B protein serum level of patients with no PCS was similar to that of patients experiencing persistent PCS (mean difference 0.00 [-0.05, 0.04]) or early PCS (mean difference 0.03 [-0.02, 0.08]). The odds of having persistent PCS (OR 0.56 (95% CI: 0.29-1.10) or early PCS (OR 1.67 (95% CI: 0.98-2.85) in patients with an elevated S-100B protein serum level was not significantly different from that of patients with normal values. No meta-analysis was performed for other outcomes than PCS due to heterogeneity and small samples. Studies’ overall risk of bias was considered moderate. Conclusion: Results suggest that the prognostic value of S-100B protein serum level to predict persistent and early post-concussion symptoms is limited. Variability in injury to S-100B protein sample time and outcomes assessed could potentially explain the lack of association and needs further evaluation.