Prediction of incident atherosclerotic cardiovascular disease with polygenic risk of metabolic disease: Analysis of 3 prospective cohort studies in Korea

Abstract

Background and aimsStudies have demonstrated that the risk of atherosclerotic cardiovascular disease (ASCVD) can be assessed by polygenic risk score (PRS) using common genetic variants. Because metabolic syndrome is a well-known, robust risk factor of ASCVD, we established PRS of metabolic disease and analyzed whether this PRS could predict incident ASCVD. MethodsWe constructed PRSs for eight quantifiable metabolic phenotypes—systolic/diastolic blood pressure, body mass index (BMI), four blood lipid components, and fasting blood glucose—by genome-wide association studies of two prospective Korean cohorts (n = 37,285). We conducted a grid search of combinations of metabolic PRSs to identify the most optimal weighted score for incident ASCVD (PRSMetS-ASCVD). The utility of PRSMetS-ASCVD was validated in an independent prospective cohort (n = 4333). ResultsThe individuals in the highest PRS quintile demonstrated a 1.4–2.0-fold increased risk of incident hypertension, obesity, hyperlipidemia, and diabetes. Using the PRSMetS-ASCVD, we identified 6.7% of the population as a high risk group demonstrating a 3.3-fold (95% confidence interval 1.7–6.1, p < 0.001) higher risk for incident ASCVD. The model combining the PRSMetS-ASCVD demonstrated a better performance for predicting ASCVD than that consisting of only conventional risk factors, such as age, sex, BMI, smoking, hypertension, diabetes and hyperlipidemia. The population with high PRSMetS-ASCVD minimally overlapped with that of high Framingham risk score, thus suggesting the additive independent benefits beyond the Framingham risk score, especially in younger individuals. ConclusionsThe polygenic risk of metabolic disease independently predicts those at an increased risk of ASCVD, identifying those at a genetically high risk of incident ASCVD.