OBJECTIVES: Minor head injury (MHI) can lead to deterioration, severe injury and death in a small number of cases. Using Computed Tomography (CT) scans on all those with MHI would result in large numbers receiving an unnecessary dose of radiation. Biochemical markers may be useful in reducing the number of scans. This study aimed to systematically identify and synthesize data estimating the diagnostic accuracy of biochemical markers for intracranial injury on CT in patients with MHI. METHODS: Key databases including MEDLINE, EMBASE & CINAHL were searched for potentially relevant literature. Studies reporting a cohort of more than 20 patients, with more than 50% having suffered a MHI (GCS 13-15), and which tested the diagnostic accuracy of a biochemical marker for intracranial or neurosurgical injury were included. Quality was assessed using the QUality Assessment of Diagnostic Accuracy Studies (QUADAS) checklist. Meta-analysis was used to estimate pooled sensitivity, specificity and likelihood ratios. RESULTS: Of the 12 included papers, nine provided diagnostic data on protein S100B only, one for Neuron-Specific Enolase (NSE) only, one for other markers and one study for both S100B and NSE levels. Data was only extracted and synthesized from S100B studies. Bayesian meta-analysis of these pooled data for 2442 adult subjects gave sensitivity of 96.8% (95% High Density Region (HDR), 93.8 to 98.6%) and specificity of 42.5% (95% HDR, 31.0 to 54.2%) with a negative likelihood ratio of 0.076 (95% HDR, 0.031 to 0.156). CONCLUSIONS: Evidence to support the addition of protein S100B as a triage tool for CT in MHI patients within three hours of injury is promising. Whilst the quality of studies is good, results are heterogeneous. S100B has the potential to be used in conjunction with a clinical decision rule. The marker therefore needs further testing as a component within such a diagnostic pathway.