Joint consideration of LDL-C and polygenic risk for incident coronary heart disease in a multi-ethnic cohort of 48,881 individuals

Elevated concentrations of LDL-cholesterol (LDL-C) are a major independent modifiable risk factor for coronary heart disease (CHD). However, the degree to which background polygenic risk can modulate the effect of high LDL-C remains unclear.

We utilized data from a subset of the Genetic Epidemiology Resource in Adult Health and Aging (GERA) multi-ethnic cohort (n=48,881; mean ± SD age = 58.1 ± 10.3 years; 63% female; 18% non-European ancestry) who were non-diabetic and not taking cholesterol lowering medication at baseline in 2003-2007. We stratified the cohort into three groups of CHD genetic risk (low=quintile 1; intermediate=quintiles 2, 3 and 4; and high=quintile 5) using a 12-SNP Polygenic Risk Score (PRS) for CHD (CARDIO inCode-Score, GENinCode Plc). Incident CHD consisted of primary in-patient codes for angina pectoris, myocardial infarction, revascularization procedures or CHD death through 12/31/2022 (n=1,725); mean follow-up was 14.3 years. Age-adjusted CHD rates per 10,000 person-years were estimated using Poisson regression (accounting for death and health plan disenrollment) according to polygenic risk and LDL-C level. Hazard ratios and 95% confidence intervals for 1 standard deviation (SD) increment of LDL-C in each PRS group were obtain using Cox regression adjusting for 10 principal components of ancestry plus traditional risk factors. We tested for formal interaction LDL-C*PRS as continuous variables in the fully adjusted model. As a complimentary approach we tested interaction between high PRS (quintile 5 vs. lower) and LDL ≥190 (vs lower). Finally, we calculated number needed treat (NNT) to prevent 1 CHD event in each PRS group.

After multivariate adjustment, each 1 SD increment of LDL-C was associated with 1.07 (95% CI, 0.95-1.21; p=0.24), 1.22 (95% CI, 1.15-1.29; p=0<0.0001) and 1.14 (95% CI, 1.03-1.26; p=0.009) increased hazard of CHD among low, intermediate, and high PRS groups, respectively. There was no interaction between linear effects of LDL-C and PRS (p=0.34). On the other hand, we found evidence of interaction (p=0.002) between high PRS and LDL ≥ 190. The NNT was lowest in the high PRS group and highest in the low PRS group (see Figure below).

Our results are consistent with a multiplicative model of high polygenic risk and high LDL on CHD risk. While all patients with elevated LDL-C ought to be treated, our data suggest that an earlier and more aggressive lipid lowering therapy may be directed towards those with high polygenic risk.

Figure.