Association between genetic liability to physical health conditions and comorbidities in individuals with severe mental illness: an analysis of two cross-sectional observational studies in the UK

Background

Individuals with severe mental illness, including schizophrenia and bipolar disorder, have elevated rates of physical health conditions, contributing to increased morbidity and mortality. While environmental factors such as adverse effects from medication and lifestyle changes play a role, the contribution of genetic liability to physical health comorbidities remains underexplored. We investigated whether genetic risk for physical health conditions influences comorbidities in people with severe mental illness and compared these effects with those in the general population. Additionally, we explored the effects of psychiatric genetic liabilities and the occurrence of physical health problems in those with severe mental illness.

Methods

We analysed two UK cross-sectional cohorts of people with severe mental illness—the Cardiff Cognition in Schizophrenia study (CardiffCOGS) cohort and the National Centre for Mental Health (NCMH) cohort. Individuals were selected for analyses if they responded to a validated self-report questionnaire of physical health problems and if their genetic data passed quality control. These subsets of individuals were used to test associations between polygenic risk scores for six physical health conditions (high cholesterol, type 2 diabetes, hypertension, asthma, heart disease, and rheumatoid arthritis) and corresponding physical health conditions in this population. Models were further adjusted for demographic and clinical covariates (sex, age, smoking, and clozapine use). Effect sizes from these analyses were compared in magnitude to those reported in studies conducted in the general population. We also evaluated associations between psychiatric polygenic risk scores (schizophrenia, bipolar disorder, major depressive disorder, and ADHD) and physical comorbidities. People with lived experience were involved in the analysis planning and guided the choices of outcomes analysed.

Findings

Following exclusions due to missing phenotypic or genetic data (403 individuals in CardiffCOGS; 1704 individuals in NCMH), our analyses included 721 individuals from the CardiffCOGS cohort (mean age 43·7 years [SD 12·1], 267 [37·0%] females, 454 [63·0%] males, and 703 [97·5%] with self-reported White ethnicity) and 1011 from the NCMH cohort (mean age 47·6 years [SD 13·7], 553 [54·7%] females, 458 (45·3%) males, and 928 [91·8%] with self-reported White ethnicity). Polygenic risk scores for physical health conditions were associated with corresponding conditions in one or both of these cohorts, explaining between 1·4% and 6·5% of the variability in these comorbidities. Polygenic risk score effect sizes for at least one of the cohorts overlapped with the reported effects (within 84% CIs) in the general population. Adjustments for clinical and demographic factors had minimal impact on these associations. Psychiatric polygenic risk scores showed weaker and less consistent associations with physical comorbidities.

Interpretation

Our findings support the role of genetic risk in physical health comorbidities among individuals with severe mental illness. Genetic liability to physical health conditions was more strongly associated with comorbidities than psychiatric genetic liability, highlighting its additive contribution alongside environmental and clinical factors. These findings indicate that there would be value in incorporating genetic risk information into predictive algorithms for physical health comorbidities in those with severe mental illness, and that polygenic risk scores should be included in research studies developing and validating such algorithms.

Funding

EU Horizon 2020 programme.