A cohort study reveals shared and distinct serum metabolic biomarkers for major adverse cardiovascular events in middle-aged and older adults

Abstract

We assessed the association of serum metabolites with the occurrence of major adverse cardiovascular events (MACE) in middle-aged and elderly individuals, explored the value of metabolomics in predicting MACE, and compared the distinctions in MACE risk-related metabolic biomarkers between middle-aged and elderly groups. Among the participants of the UK Biobank who underwent baseline assessment through nuclear magnetic resonance (NMR)-based metabolomic profiling of 168 serum metabolites and had complete covariates and clinical lipid parameters, we included those without a previous diagnosis of ischemic heart disease, cerebrovascular disease, heart failure, or cardiac arrest and not on lipid-lowering medications. Relevant covariates included sociodemographic characteristics, lifestyle factors, clinical information, and fasting time. Cox regression gave adjusted hazard ratios for metabolites, including the concentrations of various lipoprotein particles, compositional profiles of different lipoproteins, ketone bodies, amino acids, fatty acids, and additional low-molecular-weight metabolic biomarkers. The least absolute shrinkage and selection operator (LASSO) regression was applied to these metabolites to screen characteristic metabolic variables. Selected feature metabolic biomarkers were added to the established model for predicting MACE risk; risk differentiation (C-statistic) and reclassification (continuous net reclassification improvement [NRI], integrated differentiation index [IDI]) were evaluated. This study included 54,561 UK Biobank participants (34,797 middle-aged adults and 19,764 elderly adults) and was followed for a median of more than 12 years. Of these, there are 1799 middle-aged individuals and 2527 elderly individuals incident of MACE (ischemic heart disease, stroke, and cardiovascular deaths). After adjusting for relevant covariates, Cox regression yielded metabolic biomarkers associated with the occurrence of MACE in the population (false discovery rate controlled P < 0.05). In the elderly, the metabolites associated with increased MACE risk were notably diminished compared to the middle-aged; and the elderly group underscored the protective function of medium and small HDL and their constituents, docosahexaenoic acid, and glycine. The more comprehensive model, which additionally includes the feature metabolic biomarkers, demonstrated enhanced discriminatory power and predictive accuracy for MACE occurrence among middle-aged individuals, evidenced by improved C-statistics (from 0.711 [95% CI 0.699–0.722] to 0.723 [0.711–0.734]), a continuous NRI of 0.247 [0.207–0.315], and an absolute IDI of 0.005 [0.004–0.008]. Its evaluation value is superior to that in the elderly. Our study explored the association of circulating metabolites with MACE risk in middle-aged and elderly adults and made comparisons. Metabolomic insights have revealed biomarkers associated with new-onset MACE in different age populations, highlighting the value of protective metabolites in the elderly. This provides instrumental information to possibly implement precision medicine for preventing MACE.