Difference between Biological Age and Chronological Age Predicts Mortality and Hospitalization in a Longitudinal...

Biological age (BA) reflects overall health status and may deviate from chronological age (CA). Whether the magnitude of this difference (Age_diff) independently predicts hospitalization and mortality remains unknown. Retrospective longitudinal cohort of 2,597 adults (27.5% women), aged 21-92 years, participating in an executive screening program at Sheba Medical Center, Israel, contributing 6,772 clinical encounters between 2006 and 2020, mean follow-up 9.2 (2.2) years. BA was estimated to use a previously established deep-learning model trained on routine clinical laboratory tests. Bayesian time-dependent Cox models evaluated associations between Age_diff and all-cause mortality, and negative-binomial models were applied to hospitalizations, adjusting for age, sex, BMI, smoking, and baseline comorbidities. The Age_diff associated with a hazard ratio (HR) of 1.5 for mortality was determined for each chronological age, and Age_diff ×age interaction was tested. Each 1-year increase in Age_diff was associated with a 15% higher mortality hazard (HR 1.15, 95% CI 1.10-1.21) and a 6% higher hospitalization rate (incidence rate ratio 1.06, 95% CI 1.03-1.09). The Age_diff ×age interaction was not significant (p = 0.56). The Age_diff required to reach HR = 1.5 decreased with chronological age, from approximately 5 years at age 40 to 3 years at age 80. Age_diff ≥3 years was associated with a clinically significant increase in mortality. In this cohort, a BA estimated from routine laboratory results exceeding CA by 3 or more years was associated with increased hospitalization and mortality risk, suggesting potential utility for risk stratification in clinical settings.