Abstract
This study utilized data from the National Health and Nutrition Examination Survey (NHANES), to train mortality prediction-based phenotypic ages (PhenoAge [systemic] and organ-specific ages [cardiovascular, kidney, liver, and musculoskeletal]) from NHANES-III, and applied it in the continuous NHANES. Weighted linear regression analyses revealed significant associations between five diet scores-Healthy Eating Index 2020, Alternate Healthy Eating Index, Dietary Approaches to Stop Hypertension, Alternate Mediterranean Diet Score, and Dietary Inflammatory Index-derived from 24-hour diet recalls and accelerations in biological ages, encompassing both phenotypic and epigenetic measures (GrimAge2 and DunedinPoAm). Reduced rank regression was used to derive five aging-related diet scores that considered food groups within each previously established score as predictors and phenotypic age accelerations as response. The strongest food predictors of favorable aging-related diet scores included dietary patterns high in vegetables, fruits and high-quality protein (dairy, fish and legumes), and low in added sugar, sugar-sweetened beverages and red/processed meat. Weighted Cox regression models revealed that aging-related diet scores were more strongly associated with mortality risk than their respective diet scores alone.