BACKGROUND: Chronological age inadequately captures inter-individual variability in aging-related functional decline. Biological age metrics such as PhenoAge and PhenoAgeAccel, based on clinical biomarkers, have shown associations with frailty and mortality in clinical populations, but their utility in predicting physical performance decline in community-dwelling older adults remains uncertain.
METHODS: We used data from 979 participants aged ≥65 years in the InCHIANTI study, with complete baseline biomarker and physical performance data. Associations of standardized (z-scores) chronological age, PhenoAge, and PhenoAgeAccel with longitudinal changes in physical function (rescaled SPPB) and 10-year all-cause mortality were analyzed using linear mixed and Cox models respectively. A secondary analysis in 504 participants with normal baseline physical performance (SPPB ≥10) assessed the predictive value of each rescaled metric for the onset of compromised function (SPPB ≤9) at 6 years. Model performance was evaluated using AIC and AUC.
RESULTS: All three metrics showed statistically significant positive associations with physical function decline and mortality. Chronological age showed the strongest associations with rSPPB decline (β = -0.41, AIC = 3793) and mortality (HR = 2.78). PhenoAge (β = -0. 32, AIC 4338, HR = 2.57) and PhenoAgeAccel (β = -0.14, AIC 4642, HR = 1.71) showed weaker effects. Chronological age also outperformed PhenoAge and PhenoAgeAccel in predicting SPPB decline (AUC = 0.71 vs. 0.69 and 0.55, respectively).
CONCLUSION: While PhenoAge and PhenoAgeAccel are associated with adverse functional outcomes, they do not add predictive value over chronological age in a general older population.
