Association Between Low Vitamin D Levels and Depression: A Systematic Review and Meta-Analysis

Depression is a globally prevalent, biologically heterogeneous condition for which standard pharmacotherapy fails approximately one-third of patients. Vitamin D receptors (VDRs) and the biosynthetic enzyme CYP27B1 are expressed in limbic cortical neurons, and calcitriol up-regulates tryptophan hydroxylase 2 (TPH2), the rate-limiting enzyme in central serotonin synthesis. These observations suggest a plausible biological mechanism linking serum 25-hydroxyvitamin D [25(OH)D] deficiency to depression risk. Prior meta-analyses have pooled cross-sectional and longitudinal designs indiscriminately, limiting causal inference. We restricted eligibility to longitudinal study designs and multivariable-adjusted randomised controlled trials (RCTs) to reduce reverse-causation bias. Four databases were searched systematically from inception to December 2024. Eligible designs were prospective cohort, retrospective cohort, and registry-based studies and RCTs reporting multivariable-adjusted associations between baseline serum 25(OH)D and subsequent depression. Fifteen studies met all eligibility criteria (N = 49,931 participants). Risk of bias was assessed using the Newcastle-Ottawa Scale (NOS) for 14 observational studies and the Cochrane Risk of Bias 2 (RoB 2) tool for the single included RCT. A DerSimonian-Laird random-effects meta-analysis was performed. The pooled odds ratio (OR) for depression among participants with 25(OH)D deficiency was 1.45 (95% confidence interval (CI): 1.36-1.56; Z = 10.61; p < 0.001), with no between-study heterogeneity (I² = 0%; τ² = 0.000; Q(df=14) = 11.13; p = 0.68). The association was consistent across study designs, geographic regions, deficiency thresholds, and depression assessment methods. European cohorts yielded the highest pooled estimate (OR 1.59; 95% CI: 1.40-1.82). Studies employing strict deficiency thresholds (<25 or <30 nmol/L) reported an OR of 1.66 (95% CI: 1.34-2.05). Sensitivity analysis restricted to high-quality studies (n = 13) yielded OR 1.49 (95% CI: 1.38-1.60). Egger’s regression test was marginally non-significant (t = 2.02; p = 0.065). Low serum 25(OH)D is associated with a 45% increase in depression odds across 15 studies (14 longitudinal cohorts and one RCT), with homogeneous findings (I² = 0%). Adequately powered randomised trials pre-selecting participants on confirmed deficiency are needed to establish whether repletion reduces incident depression.