Abstract 41: Comprehensive Plasma Metabolomic Risk Scores to Predict Incident Coronary Heart Disease Events

Introduction: Integrating multiple metabolomics signatures related to coronary heart disease (CHD) into a metabolomic risk score (MRS) may contribute to improving the risk prediction beyond traditional risk factors, although applications of the MRS remain underdetermined, especially in people at usual risk. Hypothesis: We analyzed associations of plasma metabolites with the risk of CHD to develop MRS for predicting risk of incident CHD by incorporating comprehensive metabolic pathways among women at usual risk from the Nurses’ Health Study (NHS). We also tested whether changes in CHD-related plasma metabolites were associated with significant reductions in atherosclerotic cardiovascular disease (ASCVD) risk scores among participants in the POUNDS Lost trial. Methods: Untargeted metabolomics profiling was performed to measure plasma metabolites at baseline in a prospective nested case-control of 762 incident cases of CHD (fatal CHD and nonfatal myocardial infarction) and 762 matched controls identified over 10-14 years of follow-up time in NHS. An MRS of CHD risk was created based on metabolites that were selected through conditional logistic regression with the elastic net procedure. The POUNDS Lost trial included adults with overweight/obesity who consumed weight-loss diets with different macronutrient intakes. We measured untargeted plasma metabolomics signatures at baseline, 6 months, and 2 years to calculate changes in response to the dietary interventions. The primary outcome in the POUNDS Lost trial was the reduction of 10-year ASCVD risk scores calculated by the validated pooled cohort equations. Results: Various plasma metabolites (such as gut-microbiota-related and others in the pathways of ceramides, phospholipids, and polyamine metabolism) were associated with the risk of incident CHD ( P FDR <0.05). We developed MRS for CHD combining 122 metabolites and found that the highest quartile (Q4) was associated with 11.3 times (95% CI: 6.88, 18.6) higher risk of CHD, as compared to the lowest quartile (Q1), even after adjusting for covariates including common CHD risk factors (such as a family history of MI, lifestyle factors, obesity, diabetes, hypertension, and dyslipidemia). Every 1 SD increment of MRS was associated with an adjusted RR of 3.23 (95% CI 2.62, 3.97) for CHD events. In the POUNDS Lost, diet-intervention-induced changes in CHD-related metabolites, particularly those in ceramide pathways, were significantly associated with reductions in the ASCVD risk score at 6 months and 2 years. Conclusions: Comprehensive metabolomic scores were strongly related to the risk of incident CHD among women without prior CHD. Dietary interventions may modify the unfavorable relations between CHD-risk-related metabolites and ASCVD risk. Further studies in other populations are warranted to validate the performance of MRS in predicting ASCVD events.