Phenoage and phenoageaccel do not outperform chronological age in predicting physical function decline and mortality in community-dwelling older adults

Abstract

Aim: Chronological age is a strong predictor of physical decline and mortality but fails to capture inter-individual variability in health and aging. Biological age metrics, such as PhenoAge and PhenoAgeAccel, integrate clinical biomarkers (e.g., inflammatory markers) to better reflect physiological aging. While these measures may predict well inflammation-driven outcomes, their utility for broader functional outcomes in general older populations remains unclear. This study compares their predictive power for physical function decline and all-cause mortality to chronological age in community-dwelling older adults.
Methods: Data from the InCHIANTI study were analyzed3. Participants aged ≥ 65 years with complete biomarker data for PhenoAge calculation (creatinine, albumin, glucose, C-reactive protein, red cell distribution width, mean corpuscular volume, lymphocyte percentage, white blood cell count, and alkaline phosphatase) and baseline Short Physical Performance Battery (SPPB) assessments were included (N=979; median age 73 years; 56% women). PhenoAgeAccel was calculated as the difference berween PhenoAge and chronological age. Physical function was assessed using the Short Physical Performance Battery (SPPB). Linear mixed models and Cox regression assessed associations with longitudinal changes in a continuous rescaled score of SPPB (rSPPB) and 10-year all-cause mortality. Logistic regression examined, in a subset of participants with normal physical function at baseline (N=504) the associations of these metrics with the onset of compromised physical function, defined as a drop in SPPB score from normal (≥10) at baseline to impaired (<10) at 6-year follow-up.
Results: Chronological age showed the strongest association with rSPPB decline (-0.50 points/10 years, p<0.001) and all-cause mortality (HR 1.15, p<0.001). PhenoAge and PhenoAgeAccel were associated with physical function decline (-0.32 and -0.15 points/10 years, respectively; p<0.001) and mortality (HR 1.10 and 1.09, respectively; p<0.001) but did not outperform chronological age. For the onset of compromised physical performance, chronological age demonstrated the strongest association and the highest predictive accuracy (OR 1.17, AUC=0.71) compared to PhenoAge (OR 1.10, AUC=0.69) and PhenoAgeAccel (OR 1.05, AUC=0.55).
Conclusions: While significantly associated with physical function decline and mortality, PhenoAge and PhenoAgeAccel do not surpass chronological age as predictive tools for these outcomes in general older populations. These outcomes are likely influenced by a complex interplay of factors – including musculoskeletal, psychosocial, and environmental determinants – that extend beyond those captured by these measures biomarker panels. These findings highlight the need for more comprehensive biological aging metrics to improve risk stratification and intervention planning in geriatric populations.