Abstract
BACKGROUND: This study evaluated the potential of plasma metabolites as novel biomarkers for secondary major adverse cardiovascular event (MACE) development and risk prediction in patients with coronary artery disease. METHODS: We analyzed data from 10 175 UKB (UK Biobank) participants with coronary artery disease and metabolites (n=249) measured through nuclear magnetic resonance at baseline. Cross-validated elastic net regression models were used in the training cohort (n=7122) to select metabolites to be included in the metabolomic risk score (MRS). Hazard ratios (HRs) were estimated to examine the association between MRS and MACEs (composite of nonfatal myocardial infarction, ischemic stroke, and cardiovascular death) in the testing cohort (n=3053). Predictive performance of MRS was evaluated with C-index and calibration plots. RESULTS: Over a median follow-up of 10 years, 1624 (16%) MACEs occurred. Elastic net regression identified 26 metabolites for construction of the MRS. Participants in the highest 20% of the MRS had significantly increased risk of MACEs (HR, 2.19 [95% CI, 1.61-2.99]) and cardiovascular death (HR, 3.69 [95% CI, 2.35-5.78]), compared with the lowest 20%, after adjusting for traditional risk factors. Adding MRS to a traditional risk factor model modestly improved the 10-year MACE prediction (ΔC-index: 0.012 [95% CI, -0.003 to 0.025]); calibration slopes (0.91 [95% CI, 0.79-1.03] versus 0.88 [95% CI, 0.74-1.02]). CONCLUSIONS: Metabolomic biomarkers are independently associated with and predict MACEs, particularly cardiovascular death, in middle-aged individuals with coronary artery disease, providing valuable insights into the biological pathways underlying MACEs in high-risk populations.