Healthy Eating Index, Epigenetic Age Acceleration and Mortality Risk in US Adults.

Abstract

This study examined associations between diet quality, epigenetic age acceleration (EAA), and mortality in two U.S. cohorts: NHANES (n = 2158) and HRS (n = 1752), while accounting for demographic and socioeconomic (SES) determinants. Diet was evaluated as a potentially modifiable exposure within broader social and biological pathways. Participants were linked to the National Death Index. Cox proportional hazards models, additive Bayesian networks (ABN), generalized structural equation models (GSEM), and four-way decomposition were used to estimate direct, indirect, and interaction effects, with SES specified as an upstream determinant of diet and aging processes. Higher diet quality (HEI-2015) was associated with lower EAA and reduced mortality. GrimAgeEAA was the strongest mortality predictor (NHANES: HR = 1.61; HRS: HR = 1.76, both p < 0.001 per SD). ABN and GSEM identified SES as a driver of both diet and biological aging pathways. In HRS, approximately 44% of the inverse association between diet and mortality (HR = 0.85, p < 0.05) was explained by GrimAgeEAA (pure indirect effect), whereas mediation was not evident in NHANES. In sensitivity analyses adjusting for lifestyle factors and energy intake, the total effect in HRS was attenuated to non-significance, with physical activity emerging as a key confounder. Diet quality also modified the association between PhenoAge and mortality (~22% attributable to interaction). Poor diet quality is associated with accelerated epigenetic aging and increased mortality risk, with patterns consistent with a modest mediating role of epigenetic mechanisms. However, these associations appear partly confounded by lifestyle factors, particularly physical activity, highlighting the importance of integrated behavioral and biological pathways.