Effect of genetic risk score on coronary heart disease prediction

Abstract

Background: Coronary heart disease (CHD) is among the most health issues worldwide, and most people have multiple risk factors for this condition. The Framingham Risk Score is the representative CHD prediction model considering with risk factors simultaneously. However, it may over-estimate the risk of CHD in low-risk populations, and it did not consider genetic information. The purpose of this study was to examine the effect of genetic factors, in addition to traditional risk factors, on CHD risk prediction. Methods: Data from the Korean Cancer Prevention Study II (KCPS II) were used. A total of 415 CHD incident cases and 2,237 controls were included for developing and evaluating the CHD risk prediction model. This model was developed using multiple logistic regression analysis with 3, 15, and 17 single-nucleotide polymorphisms (SNPs) drawn from 48 CVD-related SNPs and evaluated using methods including comparison of the are under the receiver-operation characteristic curve (AUC), net reclassification index (NRI), and integrated discrimination improvement (IDI).Results: Various types of genetic risk score (GRS) improved predictive power when added to the traditional risk factors for CHD. When the weighted GRS using 15 SNPs was divided into tertiles, the degree of predictive power improvement was the highest (AUC improvement, 0.054; NRI, 20.85%; IDI, 4.23%).Conclusions: These results suggest that genetic information that is expressed as a GRS improves the power of CHD risk prediction in Koreans. Additional efforts to define CHD-related SNPs in Koreans, and to validate how these genetic information improve CHD risk predictive power through various Korean populations, should be supported.