Four days after a study explained why epigenetic clocks disagree, a second one — in Nature Medicine today, described by Yale — asks whether they move. The team assembled TranslAGE, a harmonized database of 51 longitudinal intervention studies, and ran sixteen clocks and 94 other DNA-methylation markers across all of them. Nineteen interventions significantly lowered the aging measures, five raised them, and most did nothing detectable. The clocks built to predict mortality or pace of aging responded most; drugs and lifestyle changes drove the strongest responses; over-the-counter supplements had no measurable effect; and the markers moved more in people with diseases than in healthy volunteers. Then the authors wrote the sentence the industry will not quote: responsiveness alone does not establish a biomarker as a valid surrogate endpoint.
What was shown, and what wasn't. The study shows that some clocks change when people are treated, and that lifestyle and pharmacology change them more than pills from a shelf. It does not show that a changed clock means a longer or healthier life; that link, the authors say, still has to be made, and the time course and mechanisms are not understood. The disclosures matter here, and they are extensive: several authors hold patents on clocks licensed to a consumer testing company, some consult for it, and three are its employees, with the company among the data sources. The science is serious and public-funded. The commercial interest in the result is also serious, and the paper's own caution is the right one to carry.
Loree’s takeThe useful finding for my patients is the ranking, not the clocks: structured lifestyle change and prescription drugs moved the markers; supplements from a shelf did not. That is consistent with everything else we know. What I will not do is tell a patient that a lower clock reading means they are aging slower, because the people who built the database say that has not been shown. A biomarker that responds is a research tool. A biomarker that predicts is a clinical one. We have the first.
Ben’s takeRead the two August papers together. One says the clocks measure different things; this one says they respond to treatment but are not yet proof of benefit. Between them sits a consumer market selling a single number as both. A company whose employees co-authored the paper will sell the "responds" half and leave the "not a surrogate" half in the methods section. The honest version is on the page. Quote it back to them.
Why it matters to a recovering optimizer. If you are tracking a biological-age number, this is the state of the art: it may move when you change how you live, the strongest movers are the fundamentals and medicine, supplements did nothing, and nobody knows yet what the movement is worth. That is enough to keep sleeping, moving and eating well. It is not enough to buy a kit to prove it.
Let the fundamentals move your biology; skip the kit that claims to score it. Even the kit-makers' own study says the score isn't validated.