PERIPHERAL BIOMARKER RELIABILITY IN MAJOR DEPRESSIVE DISORDER: A META-ANALYSIS OF S100B

S100B is a glial-derived calcium-binding protein that reflects astrocytic activity, neuroinflammatory signaling, and blood–brain barrier regulation. Although extensively investigated as a potential biomarker for major depressive disorder (MDD), findings have been highly inconsistent. This meta-analysis synthesized nineteen studies comparing serum, plasma, or cerebrospinal fluid (CSF) S100B concentrations between individuals with MDD (N = 914) and healthy controls (N = 890). Standardized mean differences (Hedges’ g) were pooled using random-effects models with restricted maximum likelihood estimation. The pooled analysis indicated higher S100B levels in MDD (Hedges’ g = 0.845, 95% CI 0.388– 1.301); however, heterogeneity was extremely high (I² = 90.6%), and the 95% prediction interval (−1.226 to 2.916) included zero, indicating substantial variability in study-specific true effects. Subgroup analyses by biospecimen type and univariate meta-regressions examining demographic, clinical, and methodological variables did not identify significant moderators. Evidence of small-study effects was detected. After adjustment for publication bias using trim-and-fill, the pooled effect was attenuated and no longer statistically significant. When publication bias adjustment and the pronounced between-study heterogeneity are considered, the evidence is insufficient to conclude that S100B levels are reliably elevated in MDD. The present findings highlight the methodological and biological complexities inherent in psychiatric biomarker research. These results emphasize the importance of rigorous validation standards for biomarker development in psychiatry, underscoring the need for more sophisticated analytic and integrative methodologies. Such approaches may be essential for advancing more reliable, biologically informed diagnostic and treatment frameworks in psychopharmacology.