Can You Trust Your Biological Age?
Biological age tests are common in longevity clinics, but what do epigenetic clocks actually measure — and does a lower score mean slower ageing?
– 2 min read
What Longevity Tests Actually Measure
A blood test at a longevity clinic may produce a result that sounds remarkably precise: chronological age 44, biological age 38. Repeat the test several months later and the number might fall again. But what, exactly, has changed?
There is no single universally accepted biological age. Many of the tests used in longevity programmes rely on epigenetic clocks, algorithms that analyse patterns of DNA methylation — chemical modifications attached to DNA that change with age and are associated with health and disease. Different clocks were built for different purposes. Some were trained primarily to predict chronological age; newer versions attempt to estimate mortality risk, health-related characteristics or the pace at which ageing-related changes are occurring.
A large analysis published in Nature Medicine in August 2026 helps clarify how differently these clocks can behave. Researchers harmonised data from 51 longitudinal human intervention studies, involving 3,128 samples, and calculated the same 16 major epigenetic clocks, together with 94 additional DNA-methylation biomarkers, across the datasets. This allowed them to compare how the measures responded to lifestyle changes, medications, supplements and medical procedures using a consistent methodology.
The results suggest that some newer clocks are more responsive than older ones. Measures designed around mortality risk or pace of ageing — particularly DunedinPACE and PCGrimAge — showed more consistent responses across interventions. Pharmacological and lifestyle interventions produced stronger changes in DNA-methylation biomarkers overall than supplements or the medical procedures included in the analysis.
That does not mean researchers have proved that these interventions make people live longer. Responsiveness is only one requirement for a biomarker to become a useful surrogate endpoint — a short-term measure capable of reliably predicting a meaningful long-term health outcome. The authors state that no minimal clinically important difference has yet been established for DNA-methylation clocks. A lower result therefore cannot currently be translated directly into additional years of life, reduced disease risk or improved physical function.
This distinction matters as biological-age testing becomes increasingly common in preventive-health and longevity programmes. A four-year fall in an epigenetic-age score may indicate that a particular biomarker has changed. It does not mean that four years of ageing have been reversed.
For anyone presented with a biological-age result, the useful questions are more specific than the number itself: Which clock was used? What was it designed to measure? Has it been associated with meaningful health outcomes? How reproducible is the test? And is the reported change larger than normal technical variation?
The research also comes with disclosures worth noting. Two authors are co-inventors of SystemsAge, several authors have consulting or employment relationships with TruDiagnostic, and private datasets from TruDiagnostic were included in the analysis. These relationships are disclosed by Nature Medicine and do not negate the findings, but they reinforce the need for independent validation as biological-age testing moves further into consumer healthcare.
For now, biological age is better understood as a developing biomarker than as a verdict on how long someone will live. The science is becoming more sophisticated; the number on the report still requires context.
