Are biological age tests actually accurate?

Epigenetic clocks read chemical markers on your DNA that change as you grow older. A clock that estimates calendar age scores an average deviation of 2.3 years. That sounds precise, but the real question is whether such a test also tells you something about how healthy you are for your age.

GrimAge is currently the best-supported test. It combines DNA markers with estimates of seven blood proteins and your smoking history. In large validation studies involving thousands of participants it showed a statistically strong association with risk of death, heart disease and cancer. These are, however, associations, not proven cause-and-effect relationships.

PhenoAge, based purely on blood samples, also outperforms earlier clocks in predicting mortality, various cancers and Alzheimer's disease. People whose PhenoAge is higher than their calendar age show more signs of chronic inflammation and less activity in cellular repair processes. Here too, we are dealing with associations.

DunedinPACE works differently: it does not measure an 'age number' but the pace of ageing, derived from decline across 19 organ measurements over twenty years. The test scores high on repeatability and adds predictive value on top of GrimAge. A cheek-swab variant (CheekAge) has a lower threshold for use but was validated exclusively by researchers employed by the company behind the test, which makes the independence of those findings uncertain.

The biggest problem is that there is still no consensus on what a biological age test should actually measure or how it should be used clinically. A broad expert consortium concludes that standards and validation protocols are largely absent. That means: GrimAge and DunedinPACE are scientifically promising as research instruments, but as consumer products they confront you with interpretation problems that have not yet been resolved.