Population Heterogeneity and Selection of Coronary Artery Disease Polygenic Scores

Background/Objectives: The identification of coronary artery disease (CAD) high-risk individuals is a major clinical need for timely diagnosis and intervention. Many different polygenic scores (PGSs) for CAD risk are available today to estimate the genetic risk. It is necessary to carefully choose the score to use, in particular for studies on populations, which are not adequately represented in the large datasets of European biobanks, such as the Italian one. This work aimed to analyze which PGS had the best performance within the Italian population. Methods: We used two Italian independent cohorts: the EPICOR case–control study (576 individuals) and the Atherosclerosis, Thrombosis, and Vascular Biology (ATVB) Italian study (3359 individuals). We evaluated 266 PGS for cardiovascular disease risk from the PGS Catalog, selecting 51 for CAD. Results: Distributions between patients and controls were significantly different for 49 scores (p-value < 0.01). Only five PGS have been trained and tested for the European population specifically. PGS003727 demonstrated to be the most accurate when evaluated independently (EPICOR AUC = 0.68; ATVB AUC = 0.80). Taking into account the conventional CAD risk factors further enhanced the performance of the model, particularly in the ATVB study (p-value = 0.0003). Conclusions: European CAD PGS could have different risk estimates in peculiar populations, such as the Italian one, as well as in various geographical macro areas. Therefore, further evaluation is recommended for clinical applicability.