Biological Age Estimation

That naive approach, sometimes called the multiple linear regression method, suffers from a well-known problem: it pulls estimates toward the sample mean, so it tends to make old people look younger and young people look older, and it underuses the aging information in the biomarkers. Klemera and Doubal showed that treating biological age as a latent quantity estimated from all biomarkers jointly, with each marker weighted by how strongly and cleanly it tracks age, yields a more accurate and less biased estimate. Their formula is the reason KDM became the preferred statistical method.

Biological age acceleration is the difference between estimated biological age and chronological age, often refined by residualizing biological age on chronological age so it is uncorrelated with age by construction. A positive value means a person's biomarkers look older than their years. It matters because it isolates the between-individual variation in aging and predicts mortality and age-related disease beyond what chronological age explains. This makes acceleration the key quantity for studying determinants of aging and for evaluating whether exposures or interventions speed up or slow down the aging process.

Both aim to measure biological age and both rest on regression against chronological age, but they use different inputs. KDM combines clinical and physiologic biomarkers across organ systems, whereas epigenetic clocks predict age from DNA methylation levels at selected CpG sites. KDM is cheaper and built from routine measures, while epigenetic clocks probe a molecular layer of aging. The two are correlated but not interchangeable and often capture distinct facets of aging, so many studies compute both and compare their associations with outcomes.