Metrics of physiological network topology are novel biomarkers to capture functional disability and health

Background: Physiological networks are highly complex, integrating connections among multiple organ systems and their dynamic changes underlying human aging. It is unknown whether individual-level network could serve as robust biomarkers for health and aging.

Methods: We used personalized network analysis to construct single sample network and examine the associations between network properties and functional disability in the Rugao Longevity and Aging Study (RuLAS), the China Health and Retirement Longitudinal Study (CHARLS), the Chinese Longitudinal Healthy Longevity Survey (CLHLS), and the National Health and Nutrition Examination Survey (NHANES).

Results: We observed impairments in interconnected physiological systems among long-lived adults in RuLAS. Single sample network analysis was applied to reflect the co-occurrence of these multi-system impairments at the individual level. The ADL-disabled individuals' networks exhibited notably increased connectivity among various biomarkers. Significant associations were found between network topology and functional disability across RuLAS, CHARLS, CLHLS and NHANES. Additionally, network topology served as novel biomarkers to capture risks of incident ADL disability in CHARLS. Furthermore, these metrics of physiological network topology predicted mortality across four cohorts. Sensitivity analysis demonstrated that prediction performance of network topology remained robust, regardless of the chosen biomarkers and parameters.

Conclusion: These findings showed that metrics of network topology were sensitive and robust biomarkers to capture risks of functional disability and mortality, highlighting the role of single sample physiological networks as novel biomarker for health and aging.