Abstract 18: Association of Serum Metabolic Profiles with Mortality and Longevity: The Trans-Omics for Precision Medicine (TOPMed) Study

Abstract

Introduction: All-cause and cause-specific mortality remain major measures of public health burden, despite the rising number of individuals achieving longevity (≥85 years). Although many mortality-related metabolites have been identified, metabolite predictors of long-term mortality and longevity across diverse populations remain understudied. We aim to identify novel metabolites associated with mortality and longevity. Methods: Circulating metabolite profiling was performed across seven cohorts in the Trans-Omics for Precision Medicine project. Cox models were used to examine the associations of 1,121 metabolites with all-cause, cardiovascular (CV), cancer, and respiratory mortality. Logistic regression was used to assess longevity, defined as living past 85 years at the end of follow-up, adjusting for clinical risk factors (CRF). Random-effects meta-analysis was used to estimate joint effects, and subgroup analyses were conducted by sex and race. Replication was performed using independent samples. Results: During an average follow-up of ten years among 26,091 participants (57% women, 41% Whites), there were 6,315 deaths, including 1,649 (26%), 1,387 (22%), and 314 (5%) from CV, cancer, and respiratory diseases, and 4,216 participants achieved longevity. A total of 183, 101, 12, and 23 metabolites were discovered and replicated (FDR < 0.05) for all-cause, CV, cancer, and respiratory mortality, with a range of 20% to 98% risk difference per SD increase of the metabolite. Nearly half of the metabolites were novel, and carnitines, glycerophospholipids, ceramides, and sphingolipids were leading pathways. A metabolite risk score derived from all-cause mortality-related metabolites improved the prediction of all-cause mortality by an average of 3.3%, using Harrell’s C, beyond CRF across participating cohorts. In the longevity analyses, 38 metabolites were discovered and replicated, and 31 were shared with all-cause mortality (correlation r = -0.97). Among the seven metabolites uniquely linked to longevity, taurocholate, glycocholate, and glycoursodeoxycholate suggested distinct bile acid metabolism, possibly driven by enterohepatic or microbiome-related processes. Subgroup analyses of all-cause mortality and longevity by sex and race revealed no significant heterogeneity across strata. Conclusions: We identified circulating metabolites associated with mortality and longevity, providing insight into slowing aging and the identification of at-risk populations.