Polygenic risks and cardiovascular treatment effects in severe mental illness.

OBJECTIVE: Patients with severe mental illness (SMI) experience increased cardiovascular risks, leading to reduced life expectancy. Polygenic risk scores (PRS) prediction is promising for assessing cardiovascular risks. This study evaluated the predictive utility of cardiovascular PRS and the impact from risk-reducing interventions among patients with SMI. METHODS: Using samples from the PRIMROSE programme, involving longitudinal cardiovascular interventions within primary care, we calculated seven cardiovascular and two psychiatric (bipolar/schizophrenia) PRS to predict seven corresponding cardiovascular measures [total cholesterol/high-density lipoprotein cholesterol/low-density lipoprotein cholesterol (LDL)/triglyceride/systolic blood pressure/diastolic blood pressure/BMI] assessed at baseline and 12-month follow-up. We applied multiple linear regression models at the two time points and explored the interactions between cardiovascular and psychiatric PRS on these treatment outcomes. RESULTS: At baseline, most cardiovascular PRS were associated with the respective measures, except LDL. At follow-up, the participants showed significant improvements in total cholesterol and systolic blood pressure measures; however, these two PRS's prediction effects attenuated toward the null. LDL measures became negatively associated with bipolar PRS posttreatment, though no significant interaction effects were found. Participants in the highest bipolar PRS quartile group had 0.58 mmol/L lower LDL measures than the lowest quartile group at follow-up. These results were robust to potential power reduction, participants' age, sex, prescribed medications, smoking habits, alcohol consumption, and physical activity. CONCLUSION: Our findings underscore the dynamic interplay between genetic risks and treatment effects on cardiovascular outcomes in SMI and warrant careful PRS assessment timing. While the clinical utility of PRS is still evolving, future research should explore different disorders' subtype-specific genetic interactions with interventions.