Causal associations between vitamin D and postpartum depression: A bidirectional mendelian randomization study

Previous epidemiological studies have reported associations between vitamin D and postpartum depression (PPD); however, the findings are inconsistent. This study employs bidirectional Mendelian Randomization (MR) to investigate the causal link between serum 25-hydroxyvitamin D [25(OH)D] levels and PPD. By utilizing genetic data from cohorts, this research aims to provide a more robust understanding of the potential relationship between vitamin D and PPD, addressing a critical gap in the current literature.

A bidirectional MR analysis was conducted to investigate the genetic association between serum 25(OH)D and PPD using summary statistics extracted from GWAS datasets. The study included data from 15,668 patients with PPD and 376,755 healthy controls of European ancestry. The GWAS data for 25(OH)D were obtained from two studies within the UK Biobank, encompassing 496,946 and 79,366 participants. The primary analysis employed the inverse-variance weighted (IVW) method, while supplementary MR estimates were derived through the MR-Egger and weighted median (WME) methods. Furthermore, sensitivity analyses were implemented to ensure robustness and reliability, including Cochran's Q test, MR-PRESSO, MR-Egger intercept test, and the leave-one-out test.

The MR study revealed no substantial genetic correlation between serum 25(OH)D levels and PPD (OR = 1.065, 95%CI = 0.878-1.293, P = 0.522 for set A; OR = 0.978, 95 % CI = 0.669-1.430, P = 0.910 for set B). Additionally, in the reverse analysis, we did not observe a significant causal impact of PPD on serum 25(OH)D (OR = 1.001, 95%CI = 0.974-1.028, P = 0.951 for set A; OR = 1.011, 95%CI = 0.992-1.031, P = 0.261 for set B). The results obtained from MR-Egger and WME analyses concord with those derived from the IVW method. Conducting leave-one-out tests did not identify any single nucleotide polymorphism that might have influenced the MR results, confirming the robustness and reliability of the findings.

The results suggest the absence of a causal link between vitamin D concentrations and PPD. Inconsistent observations in previous observational studies may be attributed to residual confounding.