Abstract
BACKGROUND: Postpartum depression (PPD) significantly affects maternal health and infant development, yet its underlying neurochemical mechanisms remain poorly understood. While observational studies have linked cerebrospinal fluid (CSF) metabolites with mood disorders, causal relationships are difficult to establish due to confounding environmental factors. METHODS: We applied bidirectional two-sample Mendelian randomization (MR) to explore causal associations between 338 CSF metabolites and PPD. Genetic instruments were obtained from a metabolome-wide GWAS (n=291), and PPD summary statistics were derived from the FinnGen study (7604 PPD cases; 59,601 controls). Forward MR assessed how CSF metabolites affect PPD risk, while reverse MR evaluated the impact of PPD on metabolite levels. The primary analysis used inverse-variance weighting (IVW), supported by MR-Egger, weighted median/mode, and sensitivity tests including Cochran's Q and MR-PRESSO. RESULTS: Forward MR identified 10 CSF metabolites with causal effects on PPD. Protective metabolites included guanosine, argininosuccinate, maleate, N6-methyllysine, and homocarnosine. In contrast, increased levels of N-acetyl-isoputreanine, glucose, benzoate, gluconate, and arachidonate were associated with higher PPD risk. Reverse MR showed that PPD may causally lower CSF maleate levels, suggesting mitochondrial involvement in disease progression. All findings were robust to sensitivity analyses, with no evidence of pleiotropy. CONCLUSION: This study is the first to use bidirectional MR to investigate the causal role of CSF metabolites in PPD. Our results highlight potential metabolic drivers and feedback mechanisms in PPD, especially the reciprocal link with maleate. These findings offer new insights into PPD's neurobiology and suggest novel targets for early detection and may inform novel therapeutic strategies in maternal mental health.