Metformin and healthy aging: what the evidence shows

Metformin is a familiar medicine in diabetes care and an active subject in aging research. Those are different questions. Its established use is to help improve blood glucose control in people with type 2 diabetes; whether it slows aging, prevents age-related illness, or helps people live longer is still being studied.

One medicine, two different questions

The current U.S. label lists metformin, alongside diet and exercise, to improve glycemic control in adults and children aged 10 years and older with type 2 diabetes. The DailyMed label also describes digestive side effects and a rare but serious risk of lactic acidosis. That established use does not mean metformin is approved for aging or longevity. These facts provide context for the research; they are not a personal risk assessment.

Why aging researchers study it

Metformin can affect how cells sense and use energy. Researchers have studied pathways that include AMPK, insulin signaling, and other nutrient-sensing systems because they are involved in metabolism and stress responses. In some animal experiments, metformin has changed health measures and lifespan. For example, one mouse study reported longer healthspan and lifespan in male mice at one tested exposure, while a higher exposure was toxic. A result in one mouse model is a reason to ask a question in people, not evidence that a human benefit follows.

A biological pathway is an explanation for how an effect might occur. It is not itself a health outcome. A change in a pathway, a gene-expression pattern, or a laboratory marker does not show that people become less frail, avoid disease, or live longer.

What observational studies can tell us

Health records can reveal patterns among people who use metformin, but those people differ from comparison groups in ways that are hard to fully account for: diabetes severity, kidney function, other conditions, and the reasons a particular medicine was chosen. Cohort comparisons of metformin-treated people with type 2 diabetes and people without diabetes have reached different conclusions about survival in one UK study and a later Danish analysis. These studies can generate questions, but they cannot isolate a metformin effect or establish a longevity benefit.

What small studies and biomarkers show

The MILES study was a randomized crossover study in 14 adults aged 60 and older with impaired glucose tolerance. Participants had six weeks on metformin and six weeks on placebo. Researchers found changes in blood-sugar measures and in gene-expression patterns from muscle and fat tissue. The study was designed to examine biology over a short period, not physical independence, disease prevention, or lifespan.

Some newer studies measure biological age with DNA methylation “clocks.” These are statistical summaries of chemical tags on DNA. They can be useful research measures, but a change in a clock does not automatically mean a person will feel better, avoid illness, or live longer. Different clocks can give different results, and a biomarker is not a substitute for a meaningful clinical outcome.

What recent human trials have measured

In MET-PREVENT, 72 adults in England, with a mean age of about 80 and probable sarcopenia with prefrailty or frailty, were assigned metformin or placebo for four months. Metformin did not improve the primary outcome, four-meter walking speed. This was a small, short study of older adults already experiencing functional limitations; it was not a lifespan trial.

In the six-month DEMFOS trial, 114 adults aged 65 to 85 with obesity took part in one of three groups. Adding metformin to an intensive diet and exercise program did not improve physical performance more than the same intensive program with placebo. Nearly nine in ten participants were men. The result answers a specific question in older adults with obesity and does not establish what would happen in midlife women or in other groups.

METFORAGING studied 40 older adults living with HIV who did not have diabetes. After 96 weeks, the trial found no significant difference between groups on its primary epigenetic-age measure. This small pilot assessed a biomarker in a specific population, not changes in illness, function, or lifespan.

A 2026 medRxiv preprint reports a two-year randomized study of 145 non-frail or prefrail older adults with glucose intolerance. It reports improvement in a 95-deficit frailty index and in two epigenetic clocks. It did not find a difference in the Fried frailty phenotype after removing its weight-loss item. The finding is preliminary: the record identifies it as a preprint that has not been peer reviewed, and none of these measures shows that participants lived longer.

What the evidence does—and does not—establish

The studies ask different questions in different populations, over periods ranging from weeks to two years. Some measure pathways or biomarkers; others test a specific functional outcome. Results from mice, people with diabetes, frail older adults, or people living with HIV cannot be assumed to apply to midlife women generally. No cited study establishes that metformin slows human aging or extends lifespan.

The careful summary is that metformin has an established role in blood-sugar control for type 2 diabetes and a plausible biological rationale for further aging research. Human results so far are mixed and limited by study size, duration, population, and outcome. Healthy aging is broader than any one drug, pathway, score, or clock, and the longevity claim remains investigational.