Abstract 11308: Genome-Wide Polygenic Risk Scores Incoronary Artery Disease Risk Prediction: Results From the Asian Indian Diabetic Heart Study/Sikh Diabetes Study

Abstract

Introduction: Polygenic risk score (PRS) has been shown to be highly effective in predicting coronary artery disease (CAD) risk in Western European populations. However, such studies on South Asians are scarce despite the fact that they account for 50% of the global burden of CAD. Hypothesis: In this study, we evaluated the predictive efficacy of PRS derived from the Asian Indian (AI) ancestry-specific score (PRS AI ) and European-derived PRS (PRS EU ). Also, we compared these with the clinical risk score (CRS). Methods: The study used 4602 participants (791 CAD cases and 3790 controls) from the Asian Indian Diabetic Heart Study/Sikh Diabetes Study. Weighted PRS was constructed using 100 significant SNPs from our Punjabi/Sikh CAD GWAS and 75 SNPs identified from the European GWAS catalog. The CRS was derived using the clinical risk factors described for the Framingham risk score. Results: Ancestry-specific PRS AI showed an enhanced efficacy of over 30% in estimating the relative risk for CAD over PRS EU . In sensitivity analysis, the area under the curve (AUC) for PRS AI and PRS EU were 0.84 and 0.72, respectively, while the AUC remained unchanged on combining the PRS (AI+EU) , i.e., 0.84. PRS AI also predicted the risk for increased waist to hip ratio (β=0.11, p=4.1x10 -13 ) in both gender, while fasting glucose levels (β=0.08, 3.0x10 -3 ) were confined to females. Conclusions: The results highlight evidence for the utility of PRS for identifying genetically predisposed high-risk individuals and attest to its broader clinical value. Also, sex differences may play a role in determining the risk for CAD, and increased fasting glucose and type 2 diabetes (T2D) are known to increase the risk of heart disease in women more than men, especially after menopause.