AHA Scientific Sessions 2026 Abstract · Nov 2, 2026
Metabolomic Signatures of Fasting and Postprandial Glucose and Triglyceride Responses Identify Long-Term Cardiometabolic Risk
By Qian F, Aghi A, Mi M et al.
Background: Post-prandial glucose and triglyceride responses identify individuals at heightened cardiometabolic risk that is not readily apparent in the fasting state. We investigated the association between metabolomic signatures of post-prandial glucose and triglyceride responses with long-term cardiometabolic risk in community-dwelling adults. Methods: Participants from the Framingham Heart Study Generation 3, Omni 2, and New Offspring Spouses cohorts (N = 2,082), who were free of diabetes, underwent a standardized mixed meal tolerance test (MMTT, comprised of 48% carbohydrate [with 75 g glucose], 31% fat, and 21% protein). Blood samples were collected after a 10-hour fast and 2 hours post-MMTT. Untargeted metabolomic profiling of 446 circulating metabolites were performed on the fasting blood samples. We derived metabolomic signatures related to fasting and post-MMTT glucose and triglycerides using LASSO regressions adjusted for age and sex (post-MMTT signatures were also adjusted for fasting levels). Metabolite signatures were then associated with incident type 2 diabetes (T2D) and cardiovascular disease (CVD, composite of coronary artery disease, stroke, peripheral artery disease, and heart failure) in a separate sample of participants from the Framingham Offspring Study and Generation 3 cohorts at an earlier exam cycle (N = 3,024). Results: In the derivation sample, 54.3% were female, mean (SD) age was 59.5 (8.7) years, and mean (SD) BMI was 28.4 (5.4). Fasting glucose and triglycerides were 97.7 (9.4) mg/dL and 103.3 (58.0) mg/dL, respectively, whereas 2-hour post-MMTT glucose and triglyceride values were 108.1 (21.9) mg/dL and 136.0 (69.1) mg/dL. There were between 29 to 134 LASSO-derived coefficients for the four scores, with each metabolite signature weighted on metabolites with both known and novel associations with cardiometabolic traits ( Figure 1 ). The scores captured ~90% of the variance in fasting and post-MMTT triglycerides, with more modest explanatory variance for glucose (≈40% for fasting and post-MMTT values). Scores demonstrated modest correlation with each other (r≈0.4-0.6), except post-MMTT glucose and triglyceride scores (r=0.02) ( Figure 2 ). In multivariable-adjusted models, all metabolomic scores were associated with higher risk of T2D, with scores for fasting glucose, fasting triglycerides, and post-MMTT triglyceride excursion remaining significant after further adjustment for measured fasting glucose ( Table 1 ). For CVD, fasting triglyceride metabolomic score was related to a higher risk ( Table 2 ). Conclusion: Metabolomic signatures predicting higher fasting glucose or triglycerides and their post-prandial excursions were associated with T2D, independent of traditional risk factors and measured fasting glucose. Conversely, metabolomic signatures of post-prandial glucose and triglyceride excursion were not associated with CVD.