Abstract P105: Weighted Multi-marker Genetic Risk Scores for Incident CHD Among Individuals of African, Latino or East-Asian Ancestry

Abstract

Background: GWAS have identified genetic variants associated with coronary heart disease (CHD). One of the potential uses of these genetic biomarkers is to improve the predictive capacity of existing risk functions. We studied the clinical utility of 4 multi-locus genetic risk scores (GRS’s), previously validated in European subjects, among persons of African (AFR), Latino (LAT) and East-Asian (EA) ancestry. Methods: We used data from the GERA cohort of Kaiser Permanente in Northern California (30-74 years old, 69 to 73% female) that included 2,079 AFR, 4,329 LAT and 4,801 EA. We generated 4 GRSs based on 8, 12, 36 and 51 SNPs, respectively, associated with CHD weighted by the magnitude of the association reported by the CardiogramplusC4D Consortium. We used the Framingham Risk Score (FRS) to estimate 10-year CHD risk (< 10%=low, 10 to 20%=intermediate, > 20%=high). Results: After a mean (± SD) follow-up of 5.9 (± 1.5) years, 77, 109 and 101 incident CHD events (AMI, angina pectoris, revascularization procedures and/or CHD death) were documented in AFR, LAT and EA, respectively. In models adjusted for individual FRS risk factors and principal components, GRS_8 and GRS_51 were significantly associated with CHD among LAT while GRS_36 and GRS_51 were significantly associated with CHD among EA. In fixed effects meta-analysis there was no evidence of heterogeneity (all p>0.53). The inclusion of the GRS on top of the FRF did not improve the Harrel’s C-statistics in any of the ethnic subgroups nor in the meta-analysis. The bias-corrected NRI in the intermediate FRS group (c-NRI) was statistically significant for GRS_8 and GRS_12 in EA and in the meta-analyses. Conclusions: All 4 GRS’s were linearly and directly associated with an increased risk of CHD events among minority subjects in GERA. Reclassification was overall better for GRS_8 and GRS_12 than for GRS_39 or GRS_51. These results support the consideration of the inclusion of genetic information in classical functions for risk assessment among subjects of minority background.