A patient sends me their DunedinPACE score the way they used to send me a cholesterol panel. The number is precise, the report looks clinical, and the implicit question is always the same: am I aging faster or slower than my birthday says. In 2026, for the first time, there is a real body of randomised trial data that speaks to whether interventions move these clocks at all. The honest answer is: some do, modestly, and we still do not know what that means for how long someone lives or how well.
The clearest new signal came from a post hoc analysis of a 32 week randomised, double blind, placebo controlled phase 2b trial of semaglutide, published in Nature Communications in 2026. The trial was originally designed to study HIV associated lipohypertrophy, not aging, which matters: the epigenetic aging endpoint was exploratory, added after the fact to blood samples that already existed. Within that constraint, the analysis reported that semaglutide slowed epigenetic aging relative to placebo over the study period. That is a real, peer reviewed, randomised finding. It is also a single exploratory readout in a metabolically distinct population, not a general claim about GLP-1 drugs and lifespan.
What a moving clock does and does not tell you
Epigenetic clocks estimate biological age from patterns of DNA methylation at defined sites across the genome. Second generation and "next generation" clocks, including DunedinPACE and various PhenoAge derivatives, are built to correlate not just with chronological age but with mortality risk and disease burden in cohort studies. That correlation is genuinely useful for population level research. The open question, laid out carefully in a Nature Medicine analysis published in August 2026 on the responsiveness of these biomarkers to longevity interventions, is whether a clock that moves in response to a drug or a lifestyle change is actually tracking a change in the rate of biological aging, or whether it is tracking something more proximate, like inflammation, weight loss, or a change in cell type composition of the blood sample.
That distinction is not academic. Weight loss itself changes blood cell composition and circulating inflammatory markers, both of which feed into methylation based estimates. A drug that produces meaningful weight loss could, in principle, move a clock without doing anything specific to the biology of aging beyond what would be expected from the metabolic improvement alone. The Nature Medicine analysis is explicit that most current clock validation work has not adequately separated these effects, and it calls for trials that report clock changes alongside conventional biomarkers so the two can be disentangled.
The smaller, quieter data points
A pilot trial in Madrid called METFORAGING, published in EClinicalMedicine in 2026, tested metformin against placebo for its effect on epigenetic age in older, non-diabetic adults living with HIV who were already well controlled on antiretroviral therapy. It was double blind, randomised, and explicitly framed as a pilot, meaning it was sized to test feasibility and generate an effect estimate, not to settle the question. That kind of honesty about a trial's own limits is worth noting, because it is rarer than it should be in this field.
Meanwhile the trial most people ask me about, TAME, the Targeting Aging with Metformin study designed by the American Federation for Aging Research to test whether metformin delays the onset of age related disease in a large general population cohort, has still not produced efficacy results as of September 2026. It has faced a long funding path, in part because metformin is off patent and does not attract the same commercial sponsorship that a novel molecule would. In its place, a government funded trial referred to as VITAL-H has begun testing a combination approach, spreading risk across several agents rather than betting on one. This is a sensible reallocation of scarce trial capacity, not a sign that the geroscience hypothesis has been abandoned.
A separate strand of 2026 literature, including a Frontiers in Genetics review cataloguing every human study reporting a clock change from a pharmaceutical intervention, shows the field is now large enough to need systematic tracking. That is progress. It also means clinicians and patients need a habit of asking, for any single reported result, whether it was a primary or exploratory endpoint, whether it was placebo controlled, and in what population.
What this means for clinics selling the number
Longevity clinics that sell serial epigenetic age testing as a standalone product are, in effect, selling a research grade biomarker as a consumer diagnostic before the field has agreed what changes in that biomarker predict. That does not make the test worthless. Tracked over time in the same person, using the same assay, a clock can plausibly reflect a real change in health trajectory, particularly around large changes in weight, fitness or smoking status. What it cannot yet do is tell an individual patient, with any precision, how many years of healthy life a given intervention bought them. Selling it as if it can is the gap between a real research tool and a marketing instrument, and it is worth being direct about that gap with patients who ask.
The practical posture I recommend to colleagues is to treat clock results the way we treated early biomarkers in cardiology, as promising surrogates that need outcome trials before they carry clinical weight. Use them for hypothesis generation and to watch trend lines in patients undergoing major lifestyle change. Do not let a single number reset a treatment plan, and do not let a vendor's white paper substitute for a peer reviewed, placebo controlled trial.
Key Signals
Two randomised trials in 2026, one on semaglutide and one a metformin pilot, reported statistically detectable effects on epigenetic aging clocks, but both were small or exploratory and neither was designed to prove a lifespan benefit. A major Nature Medicine analysis this August formally flagged that clock movements can be confounded by weight loss and inflammation rather than a distinct aging process, which should temper how these results are marketed. TAME, the flagship trial meant to settle the metformin question for the general population, remains without efficacy data as of September 2026, while government funders have shifted resources toward a combination trial called VITAL-H. For clinicians, the sensible near term use of these clocks is longitudinal tracking of the same patient over time, not one off consumer testing sold as a verdict on how fast someone is aging.
