Which DNA clock is best: GrimAge, DunedinPACE or PhenoAge?
GrimAge2 has the strongest evidence base for predicting mortality and healthy life years. If you want to know which clock tells you the most about your biological ageing, GrimAge2 is currently the strongest candidate, but none of the clocks is yet ready for use as a personal health monitor.
GrimAge2 scores highest of all biological clocks studied when it comes to predicting mortality and how long a person lives in good health. In a comparison of 39 ageing biomarkers in more than 20,000 people, GrimAge2 had a hazard ratio of 2.57 for mortality and 2.00 for a longer healthy lifespan. Those are the strongest figures of all 39 biomarkers compared.
GrimAge is also the only clock that, across multiple studies with more than 28,000 participants combined, is significantly associated with physical frailty both at a single point in time and over time. PhenoAge and DunedinPACE did not show that long-term association. DunedinPACE also showed substantial variation between studies, which limits its reliability.
In a large study with nearly 19,000 participants, second- and third-generation clocks (GrimAge, DunedinPACE, PhenoAge) all three proved clearly superior to older clocks in predicting 174 diseases and mortality over ten years. For 27 specific diseases, including lung cancer and diabetes, the predictive value was even stronger than for mortality in general.
Worth noting: how accurately a clock predicts chronological age says very little about how well it predicts disease or mortality. The two are barely related. In short, a high score on chronological accuracy, as with the older Horvath clock, is no advantage in clinical practice.
There are also limitations. The groups studied consist largely of people of European descent. In a small study of older Black South African men, GrimAge performed better than PhenoAge, but this is partly related to smoking-related damage that GrimAge specifically measures. The clocks have not yet been validated across all population groups. Furthermore, all comparisons concern associations, not demonstrated causes: a higher biological age on the clock does not necessarily mean that a person is actually declining faster. How much the clocks can be influenced through lifestyle or medications such as metformin is plausible for now but not yet firmly proven.
This answer draws on several large studies, including a systematic review of 24 studies with more than 28,000 participants combined and a comparison study with nearly 19,000 people. For the comparison of 39 biomarkers, more than 20,000 people were followed. The findings on GrimAge are therefore reasonably robust. For populations outside Europe the evidence is thin: only a small study with 120 participants. The researchers emphasise that the clocks measure associations, not causes, and that clinical trials to test whether they can be improved through lifestyle or medication are barely available yet.