A genetically informed investigation into the relationship between childhood maltreatment and adult health

The underlying mechanisms linking childhood maltreatment (CM) to adverse health outcomes are unclear. Here, we estimated CM pleiotropy across human phenome, also assessing the role of socioeconomic factors. Specifically, we performed a phenome-wide association study (PheWAS) of CM-associated loci using electronic health records and genetic data from the All of Us Research Program (N = 98,551), FinnGen Project (N = 377,277), and Million Veteran Program (N = 449,042). To characterize CM pleiotropy, we also conducted genetic correlation (rg), latent causal variable analysis (LCV), and bidirectional Mendelian randomization (MR) analyses. Additionally, multi-trait conditional and joint analysis and multivariate MR (MVMR) were performed to estimate whether genetically inferred CM associations are partially explained by household income (HI, a proxy for socioeconomic factors). PheWAS revealed CM loci associations with 66 medical outcomes (p < 2.76 × 10 −6 ), including psychiatric, musculoskeletal, circulatory, neurologic, and respiratory conditions. Similarly, we observed 285 CM genetic correlations (p < 3.24 × 10 −5 ) with a range of medical domains. Among them, emotional disorders starting during childhood or adolescence showed the highest CM genetic overlap (rg=0.75). LCV analysis revealed CM genetic relationships with musculoskeletal, metabolic, respiratory, and eye-related outcomes. For these, MR analyses highlighted bidirectional and heterogeneous effects possibly due to dynastic confounders, and MVMR showed a reduced effect of CM genetic liability after accounting for HI. In conclusion, different pleiotropic mechanisms likely contribute to CM relationships with several health domains, which appear to be at least partially explained by socioeconomic factors.