A longitudinal Health Atlas of age-associated multisystem physiology from population health examinations in China

Routine health examinations capture longitudinal physiology but are seldom used to characterize aging. We constructed a Health Atlas from 94,055,326 observations of 99 biomarkers in 189,095 participants from Zhejiang, organized into ten prespecified domains. Biomarkers and domains followed heterogeneous age trajectories, with predefined deviation thresholds crossed at sex-specific ages. Greater mean absolute domain burden was associated with mortality (HR per SD, 1.058; 95% CI: 1.045–1.071). Exploratory analyses linked longitudinal physiological instability to mortality, whereas cross-sectional signed dispersion was not independently associated after adjustment for coverage and burden. Evaluation in an independent Chinese cohort preserved domain ranking despite scale differences. DynLifeNet integrated raw longitudinal measurements and atlas features to predict seven future clinical abnormalities in held-out participants (macro AUROC, 0.802). External molecular data supported immune, metabolic, and renal axes. Routine examinations therefore provide a scalable framework for studying heterogeneous multisystem aging.