Mendelian randomization reveals the associations between modifiable risk factors and intermediate phenotypes and spontaneous abortion.

backgroundSpontaneous abortion (SA) is a natural process of abortion that endangers public health, and early identification of its risk factors plays a significant role in prevention. This study employed Mendelian randomization to explore the influencing factors of SA, with the aim of informing early monitoring and intervention strategies.

methodsWe performed a two-sample Mendelian randomization (MR) to evaluate the causal effects between 42 modifiable risk factors and intermediate phenotypes and spontaneous abortion. Single nucleotide polymorphisms associated with widely recognized confounders were removed based on current research and ensured sufficient F-statistics with the remaining SNPs. The primary statistical model was inverse-variance-weighted (IVW), and we performed sensitivity analyses for pleiotropy and heterogeneity. Outliers were detected and removed using MR-PRESSO. We found no statistical evidence of directional pleiotropy (based on MR-Egger intercepts and MR-PRESSO global tests), and traits showing heterogeneity were analyzed using the IVW random-effects model. Finally, to assess the robustness of the primary MR analysis findings, we employed an additional Genome-Wide Association Study (GWAS) dataset to validate the observed statistically significant results.

resultsThe results from the IVW model were reported primarily. Our analysis revealed that HLA-DR expression on monocytes (including the proportion of HLA-DR + + monocytes in leukocytes and HLA-DR on CD14 + monocytes) and the habit of cycling were positively corelated with an increased risk of SA (both with [OR] > 1 and p-adjust < 0.05). Conversely, the absolute count of CD14 + CD16- monocytes, age at first sexual intercourse, never smoking, and education level were negatively associated with SA (all with [OR] < 1 and p-adjust < 0.05). Following a validation analysis, our research has ultimately revealed statistically significant associations between SA and the following factors: age at first sexual intercourse and never smoking. Given dense LD across the extended MHC, these associations are best interpreted as locus-level (MHC) signals rather than HLA-DR-specific effects. We therefore avoid mechanistic attribution to HLA-DR per se.

conclusionsThe preliminary two-sample MR findings suggest immune signals (e.g., monocyte-related phenotypes) and highlight lifestyle correlates as potential priorities. Genetic instruments within the extended MHC region were associated with spontaneous abortion; however, given dense linkage disequilibrium across the locus, these signals are best interpreted as MHC-region associations rather than HLA-DR-specific effects. These results also require triangulation with complementary approaches-including polygenic risk score (PRS) analyses and prospective epidemiologic studies-before any causal interpretation or prevention guidance can be established.