NiaAge: a clinically interpretable measure of biological-age derived from long-term mortality-risk

Abstract

Biological-age models quantify the physiological aging process by relating biomarker profiles (e.g., blood biochemistry, DNA methylation) to all-cause mortality risk. These models outperform chronological age in predicting disease and mortality, making them useful metrics for preventative health. However, in existing biological-age models, biomarker contributions do not align with the non-linear associations biomarkers exhibit with long-term mortality risk, nor do they account for normative trajectories that occur in healthy aging, limiting their utility in a clinical setting. To address these limitations, we developed a biological-age framework (NiaAge) where biomarker contributions are derived directly from non-linear associations with long-term mortality risk and aligned with normative trajectories observed in healthy aging. As a result, biomarker contributions to NiaAge are consistent with known biomarker risk profiles and normative reference ranges. We trained NiaAge in the 1999-2000 cohort of the US National Health and Nutrition Examination Survey (NHANES; N=2028) on 59 biomarkers spanning multiple physiological domains (e.g., hematology, metabolism, inflammation), then evaluated it in the 2001-2002 cohort (N=2346). NiaAge predicted long-term mortality, physical-health, and cognitive-health significantly better than chronological age. It also outperformed several DNA-methylation age clocks on mortality and physical/cognitive health-span metrics, while performing comparably to leading physiological age clocks. These results position NiaAge as a valuable tool for preventative health.