Evidence for overlapping genetic architecture between lifestyle factors and severe mental disorders with different patterns across diagnoses

BACKGROUND: Severe mental disorders (SMDs) are associated with unhealthy lifestyle, contributing to increased risk of comorbid cardiovascular disease. Genetic factors influence both SMDs and lifestyle behaviours, but their genetic relationships remain unclear. Here, we aimed to unravel the shared genetic architecture of SMDs and lifestyle factors. Additionally, we assessed if genetic propensity to SMDs predicts body mass index (BMI) and lipids through lifestyle factors. METHODS: We analysed genome-wide data on major depression (MD) (N = 480,359), schizophrenia (SCZ) (N = 130,644), bipolar disorder (BIP) (N = 353,889) and self-reported lifestyle factors (N = 266,048-606,820), including food intake, physical activity, sedentary behaviours, and accelerometer-assessed activity from All of Us (N = 30,132) and UK Biobank (N = 91,105) to obtain objective measures for sensitivity analysis. We estimated the shared genetic architecture using bivariate MiXeR. Shared genetic loci were identified using conjunctional false discovery rate and mapped to genes, which were subject to enrichment analyses. We applied structural equation modelling (SEM) to assess if lifestyle mediates the relationship between polygenic risk score for SMDs and BMI and lipids. People with lived experience were involved in the research. FINDINGS: There was extensive genetic overlap between lifestyle factors and SMDs, with different patterns of effects. MD was genetically correlated with less physical activity and more sedentary behaviour. SCZ and BIP displayed opposite patterns with genetic associations with less sedentary behaviour, more physical activity and healthier food intake. This divergent pattern across SMDs was largely consistent using accelerometer-assessed activity. We identified 551 shared loci, implicating biological processes related to neurodevelopment and synaptic and neuronal properties. Further analyses indicated that lifestyle factors partly mediate the relationship between genetic risk for SMDs and BMI and lipids. INTERPRETATION: The results show a genetic propensity towards unhealthier lifestyle behaviours in MD, while SCZ and BIP displayed a divergent pattern. The genetic correlations reflecting mixed effect directions may also imply subgroups with different genetic propensity, which can form the basis for risk stratification and more tailored lifestyle interventions and personalised treatment. FUNDING: Research Council of Norway (grants, 273291, 273446, 300309, 324252, and 326813), South-East Norway Regional Health Authority (grants 2023-031 and 2022-073), NordForsk (University Cooperation Grant 164218, PreciMENT), European Union's Horizon 2020 Research and Innovation Programme (grant 847776, CoMorMent; grant 964874, RealMent), and the National Institutes of Health (grant R01MH125938).