No winners or losers: clinical chemistry-based biological aging metrics perform similarly across cohorts and health outcomes.

Abstract

Aging is the leading risk factor for most chronic disease. However, disease risk varies substantially between individuals of the same age. Biological aging measures attempt to quantify this difference using biomarkers; such measures have amassed substantial evidence as reliable correlates of morbidity and mortality. Although many have been developed throughout the years, there is no clear consensus as to which one is the best, if any. This study evaluates four methods for measuring biological aging: Klemera and Doubal's method for biological age (KDM BA), phenotypic age (PA), homeostatic dysregulation (DM), and Pace of Aging (Pace). Using five cohort studies from four different countries (InCHIANTI from Italy, WHAS I and II from the US, NuAge from Canada, and the UK Biobank), we assessed the relationship of these metrics with six health outcomes. The metrics were calculated using a consistent set of biomarkers to facilitate comparison. The biological aging measures correlated only weakly with each other (r > 0.5 for six of 21 correlations). The meta-analyses performed on the results from each dataset revealed that all biological age measures were significantly associated with at least one health outcome; however, no single metric consistently outperformed the others, with strength of association strikingly similar across metrics. This study is the first to combine an international multi-cohort analysis using a consistent set of biomarkers across biological age metrics. While there are no net winners or losers, effect sizes are heterogeneous across cohorts, highlighting the importance of replicating findings in different contexts and with different metrics.