CSF biomarkers in leak detection: A systematic review & meta-analysis of diagnostic test accuracy (DTA) studies

Background: Cerebrospinal fluid (CSF) leaks result from dural tears due to traumatic and non-traumatic causes, potentially leading to meningitis, intracranial infections, subdural hematomas, and even death. Neuroimaging is still the standard diagnostic tool but presents economic and logistical challenges. Non-invasive biochemical methods offer valuable first-line diagnostics. This systematic review and meta-analysis evaluate the accuracy of CSF biomarkers in detecting leaks. Objective: To assess the diagnostic accuracy of CSF biomarkers in suspected CSF leak cases. Methods: A comprehensive search was conducted across electronic databases and citation mining from the date of inception till February 2025. The study, registered on PROSPERO (CRD42024601252), followed PRISMA 2020 guidelines. Of 455 retrieved studies, eighteen met inclusion criteria; seventeen were used for meta-analysis, involving 1,914 patients. Results: Pooled CSF leak prevalence was 31 %. Pooled sensitivity and specificity were 0.96 (95 % CI: 0.869–0.988) and 0.946 (0.746–0.991) for Beta-2-transferrin (B2T), and 0.949 (0.909–0.972) and 0.999 (0.945–1) for Beta-trace protein (BTP). Diagnostic odds ratios (DOR) were 418.09 (48.582–3597.985) for B2T and 14,566.52 (387.368–547,756.3) for BTP. Conclusion: CSF biomarkers show high diagnostic accuracy, offering a promising alternative to neuroimaging. Standardizing biomarker thresholds and assay methodologies are crucial for reliability. Further research should address confounding factors and variability to enhance diagnostic effectiveness. Prospero registration: The above systematic review has been registered in PROSPERO Registration Id: CRD42024601252.