Abstract 047: Plasma metabolic signatures of healthy diets and healthspan: an integrative analysis of metabolomics and genetics among 12,478 US adults with 32 years of follow-up

Background: Extending the healthspan – years of lived without disease – is a public health priority. Healthy diets have been linked to lower chronic disease risk, but the underlying mechanisms are largely unknown. We have previously identified plasma multi-metabolite signatures that measure intakes and metabolic responses to different dietary patterns. Hypothesis: Plasma metabolic signatures of heathy diets are associated with longer healthspan. Methods: We studied 12,478 participants from Nurses’ Health Studies and Health Professional Follow-up Study who were free of baseline chronic diseases. Diet was assessed by a food frequency questionnaire and indices for 3 recommendation-based, 3 plant-based, and 2 mechanism-based diets, were calculated. Plasma metabolome was profiled by liquid chromatography-mass spectrometry. Metabolic signatures for 8 dietary patterns, each based on 38-66 metabolites, were derived using elastic net model from our prior study. A polygenetic risk score (PRS) of shortened healthspan was calculated based on a published genome-wide association study using LDpred2. Cox regression was used to test associations with healthspan, which ended at the time of first onset of any of 11 major chronic diseases including cardiovascular disease. Results: During up to 32.4 years of follow-up, 6,021 individuals developed at least one chronic disease within an average of 23.8 years. In multivariable analyses, per standard deviation (SD) increase in metabolic signatures of all 5 examined healthful diets (e.g., Mediterranean diet) were associated with a lower risk of shortened healthspan (hazard ratio [HR]: 0.83-0.88; P < 6e-8), whereas signatures of 3 unhealthful diets (e.g., proinflammatory diet) were associated with a higher risk of shortened healthspan (HR: 1.08-1.26; P < 1e-3) ( Fig. A ). These associations persisted after adjusting for their corresponding dietary indices. While the derived PRS significantly predicted the genetic risk for any chronic diseases ( Fig. B ), the associations between metabolic signatures of diets and healthspan were not modified by such a genetic burden ( P-interaction >0.05; Fig. C ). Conclusions Plasma metabolic signatures that reflect intakes and metabolic responses to heathy diets are associated with longer healthspan, regardless of genetic burden for chronic diseases.