Blood Metabolism Maps How Lifestyle Tracks Function in Cognitive Aging
Key Insights
A pathway-level analysis of blood metabolomics links diet quality and physical activity to functional outcomes including mobility, cognition, and frailty across the cognitive aging spectrum.
The study, drawing from ADNI, ROSMAP, and AIBL cohorts, included 360 participants classified as cognitively normal, mild cognitive impairment (search), or Alzheimer's disease (search).
A lifestyle-modulated metabolic pathway score (LMPS) was derived and found to be associated with better activities of daily living, global cognition, and gait speed, with stronger associations in cognitively normal individuals.
A new study published in Scientific Reports has identified coordinated blood metabolic patterns that connect diet quality and physical activity to functional performance across the cognitive aging continuum. The research, leveraging data from the Alzheimer's Disease (search) Neuroimaging Initiative (ADNI), the Religious Orders Study and Rush Memory and Aging Project (ROSMAP), and the Australian Imaging, Biomarkers, and Lifestyle Study of Aging (AIBL), demonstrates that lifestyle-associated metabolic signals are organized across multiple biological domains and are not confined to a single pathway class.
The analysis included 360 participants: 150 classified as cognitively normal (CN), 120 with mild cognitive impairment (search) (MCI), and 90 with Alzheimer's disease (search) (AD). Individuals with AD were older and had lower levels of education than CN individuals, and the prevalence of apolipoprotein E (search) (APOE) ε4 carriers increased progressively from CN through MCI to AD groups.
Pathway-Level Metabolomics Approach
Rather than analyzing metabolites in isolation—an approach that has yielded inconsistent findings in prior AD research—the investigators mapped fasting blood metabolomics data to curated metabolic pathways and estimated pathway activity scores by aggregating metabolite signals. These pathways were categorized into five broad domains: amino acid metabolism, energy metabolism, microbiome-related metabolism, lipid metabolism, and oxidative stress and inflammation-related metabolism.
Linear regression models assessed associations between lifestyle exposures (diet quality measured by Food Frequency Questionnaire and physical activity quantified by the International Physical Activity Questionnaire), functional outcomes, and pathway activity scores. Models were adjusted for body mass index, age, sex, education, hypertension, statin use, diabetes, and APOE ε4 carrier status.
The Lifestyle-Modulated Metabolic Pathway Score
A key innovation of the study was the derivation of a lifestyle-modulated metabolic pathway score (LMPS) using elastic net regression. This score jointly summarizes diet-quality- and physical-activity-associated pathway patterns. Cross-validation, out-of-fold score estimation, and bootstrap stability assessment were used to evaluate internal robustness.
LMPS differed significantly across cognitive groups, with higher scores observed in CN individuals and lower scores in MCI and AD groups. Higher LMPS values were associated with better activities of daily living (ADL), global cognition, composite function, and gait speed, as well as lower frailty. Notably, these associations were stronger in CN participants than in those with MCI or AD.
Cross-Domain Metabolic Patterns
Multi-layer convergence mapping revealed that pathway-level signals were distributed across metabolic domains rather than concentrated in a single class. Lipid metabolism pathways demonstrated greater heterogeneity between CN and AD groups, while oxidative stress- and inflammation-related pathways showed greater dispersion and separation in the AD group. Microbiome-related pathways exhibited mixed patterns.
A clustered dot heatmap identified several pathway clusters with aligned association patterns across lifestyle exposures and functional outcomes. Pathways positively associated with lifestyle exposures were also positively associated with ADL, cognition, and gait speed and inversely associated with frailty. The researchers emphasized that these visualizations were intended to summarize association patterns and prioritize hypotheses rather than independently validate biological mechanisms.
Limitations and Future Directions
The study's cross-sectional design and lack of an independent external validation cohort represent important limitations. The researchers explicitly caution that LMPS should be interpreted as a hypothesis-generating research measure rather than a validated biomarker, mediator, or clinical tool. Additionally, the untargeted metabolomics data reflected relative signal intensities rather than absolute metabolite concentrations.
"Further validation is needed to evaluate its reproducibility, relevance, and translational utility," the authors note, underscoring that while the lifestyle-metabolism-function relationships span multiple metabolic domains, their clinical applicability remains to be established through prospective studies with independent cohorts.