Metformin, a drug prescribed to tens of millions of people with type 2 diabetes, is producing measurable shifts in molecular markers of aging across multiple independent human studies. A randomized, placebo-controlled pilot trial in Madrid tracked 11 epigenetic clocks over 96 weeks and found that metformin slowed biological aging in non-diabetic participants. Separately, an analysis of nearly 13,910 veterans linked metformin exposure to lower epigenetic age acceleration. The findings are converging around a single, charged question: can a cheap, generic pill slow the aging process itself?
Why metformin’s anti-aging signal demands scrutiny in 2026
The tension behind these results is straightforward. Epigenetic clocks, which estimate biological age by reading chemical tags on DNA, have become the most widely used surrogate markers in aging research. But no one has yet shown that turning back those clocks translates into fewer heart attacks, cancers, or deaths. The studies now accumulating around metformin are the closest anyone has come to bridging that gap in humans, and they are doing so with a drug that costs a few cents per pill and already has decades of safety data in diabetic populations.
A testable explanation for how metformin achieves these clock changes centers on its effects inside muscle tissue. The drug appears to alter gene networks tied to mitochondrial function and inflammation, two biological systems that deteriorate with age. If those network-level shifts are what drive the epigenetic signal, then lower doses or intermittent dosing schedules might reproduce the effect without meaningfully changing fasting glucose, a possibility that would reshape how clinicians think about prescribing metformin to people who do not have diabetes.
Three trials building the case from DNA to tissue
The strongest new evidence comes from the METFORAGING trial, a double-blind, randomized, placebo-controlled pilot study conducted in Madrid. Researchers enrolled non-diabetic people living with HIV aged 50 and older, a population that experiences accelerated biological aging. They measured 11 epigenetic biomarkers and clocks at baseline and again at week 96. The trial’s design, with its placebo arm and nearly two-year follow-up, represents one of the most rigorous tests of metformin’s effect on biological age conducted so far in humans.
Running in parallel, the Million Veteran Program analysis drew on VA electronic health records and whole-blood DNA methylation data from approximately 13,910 participants with a mean age of about 66 years. That cohort is predominantly male and drawn from the VA system, which limits its generalizability. Still, the scale of the dataset, roughly ten times larger than most epigenetic aging studies, gives it statistical weight that smaller trials cannot match. The analysis linked metformin exposure to reduced epigenetic age acceleration across multiple clock measures.
Mechanistic detail comes from the Metformin in Longevity Study, or MILES, a randomized, double-blind, placebo-controlled crossover trial that enrolled older adults with impaired glucose tolerance. Participants completed six-week periods on metformin and placebo, and researchers collected skeletal muscle and adipose tissue biopsies. The resulting transcriptomic data, drawn from participants around 70 years old, showed that metformin shifted gene expression in both metabolic and non-metabolic pathways in muscle and fat. Those findings, published in peer-reviewed form and deposited in public gene expression databases, offer the clearest window yet into what metformin actually does at the tissue level in aging humans.
A separate trial, TRIIM, combined growth hormone, DHEA, and metformin and reported changes in epigenetic clock-based biological age alongside immune system markers. Because TRIIM used a cocktail rather than metformin alone, it cannot isolate the drug’s individual contribution. But the trial added to a growing body of evidence that epigenetic age can be shifted by pharmacological intervention in living people.
Gaps that separate clock changes from real health gains
The most significant limitation across all of these studies is the absence of hard clinical endpoints from a metformin-only trial in non-diabetic, non-HIV populations. No completed study has yet shown that metformin, given to otherwise healthy older adults, reduces the incidence of cancer, cardiovascular disease, or dementia. The Targeting Aging with Metformin trial, known as TAME, was designed to answer exactly that question using a composite endpoint of age-related diseases, but its results have not yet been reported.
The MeMeMe randomized controlled trial tested 1,700 mg per day of metformin in adults aged 50 to 79 with metabolic syndrome, aiming to prevent type 2 diabetes, cardiovascular disease, and cancer. That trial provides clinical endpoint data in a metabolically at-risk population, but its participants already had metabolic syndrome, making it difficult to extrapolate results to healthy aging.
Even in the METFORAGING and Million Veteran Program datasets, the primary readouts are methylation-based clocks and related biomarkers, not heart attacks or survival curves. That leaves open the central question of whether a reduction in epigenetic age of, say, one to three years corresponds to a clinically meaningful drop in disease risk. It is possible that these clocks capture only part of the aging process, or that their calibration in younger and middle-aged cohorts does not hold in older adults with multiple comorbidities.
There are also unresolved safety questions when metformin is used outside of diabetes. The drug’s long-term safety record in people with type 2 diabetes is reassuring, but those patients often have higher baseline risks that may obscure rare adverse effects. In relatively healthy older adults, chronic gastrointestinal side effects, vitamin B12 deficiency, or subtle impacts on muscle performance could matter more. Until large, event-driven trials in non-diabetic populations are completed, those trade-offs will remain speculative.
What the metformin story means for aging research
Despite those caveats, the convergence of epigenetic, transcriptomic, and clinical data around metformin is reshaping the field. For one, it strengthens the argument that aging itself can be targeted pharmacologically, even if the magnitude of benefit is still uncertain. The fact that a decades-old generic drug can nudge molecular aging markers in multiple tissues raises the bar for newer, more expensive interventions that claim to modify aging.
The metformin data are also forcing a more nuanced view of epigenetic clocks. Rather than treating them as definitive readouts of biological age, researchers are beginning to use them as one component in a broader panel that includes proteomic signatures, inflammatory markers, and functional measures like grip strength and gait speed. In this framework, metformin’s effect on methylation becomes a hypothesis-generating signal that must be tested against concrete health outcomes.
For clinicians and patients, the message in 2026 is one of cautious interest rather than immediate action. Off-label use of metformin for longevity is already widespread, driven by early epidemiological hints of reduced cancer and cardiovascular risk in diabetics taking the drug. The new epigenetic and tissue-level findings will likely intensify that demand. Yet without randomized evidence of reduced disease incidence in non-diabetic older adults, prescribing metformin solely for anti-aging remains a gamble on surrogate markers.
Regulators and trial designers, meanwhile, are watching metformin as a test case for how to evaluate geroprotective drugs. If TAME or similar studies can show that a single agent reduces the aggregate burden of age-related disease, it would validate the use of composite endpoints and potentially accelerate the approval pathway for future therapies. If they fail, it will prompt a rethinking of which biomarkers genuinely track the aging process and which are merely passengers.
In that sense, metformin’s most important contribution to longevity science may not be the extra months or years it might eventually add to human life, but the rigorous experimental framework it is helping to build. By tying together epigenetic clocks, tissue-level gene expression, and, eventually, hard clinical outcomes, the current wave of trials is creating a template for how to test aging interventions in humans. Whether or not metformin itself becomes a standard longevity drug, the scrutiny it is receiving in 2026 is likely to shape how the next generation of anti-aging therapies is developed, evaluated, and prescribed.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
