Polygenic risk score and coronary artery disease: A meta-analysis of 979,286 participant data.

Abstract Background and aims Coronary artery disease (CAD) is a complex disease with a strong genetic basis. While previous studies have combined common single-nucleotide polymorphisms (SNPs) into a polygenic risk score (PRS) to predict CAD risk, this association is poorly characterised. We performed a meta-analysis to estimate the effect of PRS on the risk of CAD. Methods Online databases were searched for studies reporting PRS and CAD. PRS computation was based on log-odds (PRSLN), pruning or clumping and thresholding (PRSP/C + T), Lassosum regression (PRSLassosum), LDpred (PRSLDpred), or metaGRS (PRSmetaGRS). The reported odds ratio (OR), hazard ratio (HR), C-indexes and their corresponding 95% confidence interval (95% CI) were pooled in a random-effects meta-analysis. Results Forty-nine studies were included (979,286 individuals). There was a significant association between 1-standard deviation [SD] increment in PRS and adjusted risks of both incident and prevalent CAD (OR [95% CI]: 1.67 [1.57–1.77] for PRSmetaGRS, 1.46 [1.26–1.68] for PRSLDpred). The risk of incident CAD was highest for PRSP/C + T (HR [95% CI]: 1.49 [1.26–1.78]), PRSmetaGRS (1.37 [1.27–1.47]), and PRSLDpred (1.36 [1.31–1.42]). Analysis of model performance demonstrated that PRS predicted incident CAD with C-index of up to 0.71. Importantly, addition of PRS to clinical risk scores resulted in modest but statistically significant improvements in CAD risk prediction, with 1.5% observed for PRSP/C + T (p  Conclusions Polygenic risk score is strongly associated with increased risks of CAD. Future prospective studies should explore the usefulness of polygenic risk scores for identifying individuals at a high risk of developing CAD.