Utility of genetic and non-genetic risk factors in predicting coronary heart disease in Singaporean Chinese

Abstract

Background

Although numerous phenotype based equations for predicting risk of ‘hard’ coronary heart disease are available, data on the utility of genetic information for such risk prediction is lacking in Chinese populations.

Design

Case–control study nested within the Singapore Chinese Health Study.

Methods

A total of 1306 subjects comprising 836 men (267 incident cases and 569 controls) and 470 women (128 incident cases and 342 controls) were included. A Genetic Risk Score comprising 156 single nucleotide polymorphisms that have been robustly associated with coronary heart disease or its risk factors (p < 5 × 10⁻⁸) in at least two independent cohorts of genome-wide association studies was built. For each gender, three base models were used: recalibrated Adult Treatment Panel III (ATPIII) Model (M₁); ATP III model fitted using Singapore Chinese Health Study data (M₂) and M₃: M₂ + C-reactive protein + creatinine.

Results

The Genetic Risk Score was significantly associated with incident ‘hard’ coronary heart disease (p for men: 1.70 × 10⁻¹⁰–1.73 × 10⁻⁹; p for women: 0.001). The inclusion of the Genetic Risk Score in the prediction models improved discrimination in both genders (c-statistics: 0.706–0.722 vs. 0.663–0.695 from base models for men; 0.788–0.790 vs. 0.765–0.773 for women). In addition, the inclusion of the Genetic Risk Score also improved risk classification with a net gain of cases being reclassified to higher risk categories (men: 12.4%–16.5%; women: 10.2% (M₃)), while not significantly reducing the classification accuracy in controls.

Conclusions

The Genetic Risk Score is an independent predictor for incident ‘hard’ coronary heart disease in our ethnic Chinese population. Inclusion of genetic factors into coronary heart disease prediction models could significantly improve risk prediction performance.