Biological age acceleration and multidomain functional impairment: Mortality associations and incremental value in NHANES and CHARLS.

Biological age acceleration and multidomain functional impairment may capture complementary dimensions of late-life vulnerability, but their outcome associations and incremental predictive value are uncertain. We analyzed adults aged 60 years or older in the National Health and Nutrition Examination Survey (NHANES) 2011-2014 and the China Health and Retirement Longitudinal Study (CHARLS) 2011-2020. Biological age acceleration (zB) used the same 12-biomarker Klemera-Doubal panel in both cohorts. Mortality models used a four-domain functional impairment score (zF4: locomotion, vitality, psychological capacity, and cognition); disability models used a non-overlapping three-domain score (zF3) excluding locomotion. Scores were age- and sex-residualized, standardized, and entered jointly with covariates. In NHANES (N = 2348; 406 deaths), zB was associated with mortality (hazard ratio [HR] 1.206, 95% confidence interval [CI] 1.030-1.411), as was zF4 (HR 1.798, 1.603-2.016). In CHARLS, zB was not associated with incident ADL limitation (N = 3791; 431 events; odds ratio [OR] 1.012, 0.913-1.122) or IADL limitation (N = 3193; 486 events; OR 0.997, 0.899-1.105), whereas zF3 was associated with both. zB was also associated with CHARLS mortality (N = 4570; 397 deaths; HR 1.109, 1.010-1.217), although expanded covariate adjustment attenuated this estimate (HR 1.085, 0.984-1.196). A descriptive two-cohort synthesis yielded an HR of 1.133 (1.046-1.228). Adding zB to functional impairment and covariates did not consistently improve out-of-sample discrimination or Brier performance. Biological age acceleration therefore carried a modest cross-cohort mortality association but no disability association, and it added little predictive information beyond multidomain functional impairment.