Abstract
BACKGROUND: Polygenic risk scores derived from coronary artery disease genome-wide association studies are associated with statin relative risk reduction. OBJECTIVE: Examine the relationship between coronary artery disease polygenic risk scores that include variants below thresholds of genome-wide significance and statin primary prevention of major adverse cardiovascular events. METHODS: We generated coronary artery disease polygenic risk scores for participants with no past evidence of myocardial infarction from three electronic health record-linked genetic biobanks: All of Us Research Program, Genetic Epidemiology Research on Adult Health and Aging, and the Million Veteran Program. We scored each participant using three different polygenic risk scores: two with both genome-wide and sub-genome-wide significant variants (metaGRS and PRS2022) and one with only variants meeting genome-wide significance (164SNP). We used covariate-adjusted Cox regression models to compare risk of major adverse cardiovascular events between statin users and nonusers matched on age, sex, smoking status, type 2 diabetes mellitus, and hypertension within strata defined by polygenic risk. For the primary analysis, we performed a meta-analysis across the three cohorts for the delta hazard ratio of statin effectiveness between high and low polygenic risk. RESULTS: Across all cohorts, statin use was more strongly associated with reduced risk of major adverse cardiovascular events among participants with no myocardial infarction at index in the highest versus lowest polygenic risk score group for the PRS2022 (interaction beta 0.19, standard error 0.07, interaction P=2.3E-3) and metaGRS (interaction beta 0.14, standard error 0.07, interaction P=.02) scores. However, the association was not statistically significant (interaction beta 0.09, standard error 0.07, interaction P=.08) for the 164SNP risk score. CONCLUSIONS: We demonstrated that the association between coronary artery disease polygenic risk scores and statin relative risk reduction can by enhanced with the inclusion of sub-genome-wide variants, paving the way for more research to establish clinical utility.