Lifestyle-associated blood metabolic pathways and functional performance in cognitive aging

Functional decline is a major clinical feature of Alzheimer's disease (AD), yet the blood metabolic pathways associated with lifestyle factors and multidimensional functional performance across cognitive stages remain incompletely characterized. We applied a pathway-level blood metabolomics framework to harmonized, de-identified data from aging and dementia-related cohort resources spanning cognitively normal aging (CN), mild cognitive impairment (MCI), and AD. Metabolites were mapped to curated pathways and summarized into pathway activity scores across five domains: energy metabolism, amino acid metabolism, lipid metabolism, inflammation/oxidative stress, and microbiome-linked metabolism. We evaluated associations among physical activity, diet quality, pathway activity scores, and functional outcomes, including activities of daily living, gait speed, grip strength, global cognition, composite function, and frailty. To summarize pathway patterns jointly associated with physical activity and diet quality, we derived a lifestyle-modulated metabolic pathway score (LMPS) using elastic net regression with cross-validation, out-of-fold score estimation, and bootstrap stability assessment. Lifestyle-associated pathway activity showed coordinated patterns across metabolic domains and was associated with functional performance across cognitive groups. Higher LMPS values were associated with better physical and cognitive function and lower frailty, with graded differences observed across CN, MCI, and AD. Internal robustness analyses indicated greater stability at the pathway-domain level than at the individual-pathway coefficient level. Sensitivity analyses adjusting for cognitive group attenuated but did not eliminate the directionally consistent associations between LMPS and major functional outcomes. Convergent pathway patterns involved mitochondrial energy metabolism, lipid remodeling, inflammatory regulation, and microbiome-related metabolism. Pathway-level blood metabolomics identified lifestyle-associated metabolic patterns related to multidimensional functional outcomes across the cognitive aging spectrum. LMPS provides a data-driven summary of lifestyle-associated pathway variation in this cohort and may help generate hypotheses about metabolic correlates of functional performance. Independent and longitudinal validation will be required to determine its reproducibility, temporal relevance, and translational utility.