Homeostatic dysregulation and clocks: a distance-based perspective.

Aging can be viewed as a progressive departure from physiological homeostasis accompanied by increasing morbidity and mortality risk. This perspective motivates two complementary but non-equivalent approaches. The first is homeostatic dysregulation (HD), which summarizes multivariate deviation from a prespecified reference distribution. The second comprises supervised aging measures, which are target-dependent model outputs trained to predict outcomes such as chronological age, mortality, morbidity, functional status, or pace of aging. Their outputs are often only modestly correlated and should not be assumed to represent the same latent biological age or a literal geometric distance. We propose that their joint value should be tested for outcome-specific risk stratification, longitudinal monitoring, and gerotherapeutic trials. Such applications require prospective validation, calibration, and evidence that each measure adds information beyond chronological age and established clinical predictors.