A diabetes pill that costs less than a dollar a day is quietly becoming one of the most studied drugs in modern medicine. Metformin, first approved in France in 1957 and now prescribed to roughly 150 million people worldwide for type 2 diabetes, is being tested against conditions its inventors never imagined: Long COVID, age-related frailty, knee osteoarthritis, and breast cancer. As of spring 2026, randomized trials and massive observational studies have produced results that range from genuinely promising to flatly negative, painting a complicated picture of what a cheap, familiar molecule can and cannot do.
Long COVID: the strongest case so far
The most compelling evidence for metformin beyond diabetes comes from the COVID-OUT trial, a multicentre, randomized, quadruple-blind, phase 3 study published in The Lancet Infectious Diseases. The trial enrolled more than 1,100 non-hospitalized adults who had tested positive for SARS-CoV-2 and started them on metformin or placebo within days of symptom onset. Over more than 10 months of follow-up, participants who received metformin were less likely to receive a provider diagnosis of Long COVID compared with those on placebo, with the trial reporting concrete effect sizes including hazard ratios and cumulative incidence figures.
A separate virologic analysis drawn from the same trial dataset offered a mechanistic explanation: metformin reduced SARS-CoV-2 viral load and cut the rate of viral rebound. If the drug helps the body clear the virus faster, fewer patients may develop the persistent inflammation and immune dysregulation associated with Long COVID.
An NIH-supported observational study through the RECOVER program reinforced those findings at scale. Using U.S. electronic health records covering a large cohort of adults with diabetes, researchers compared patients who were already taking metformin against those on other glucose-lowering medications. The RECOVER-linked cohort analysis found that metformin users had a meaningfully lower risk of Long COVID or death, with the association holding across multiple sensitivity analyses and time windows. An NIH news release confirmed the finding, though the agency stopped short of recommending the drug for that purpose. As of May 2026, no major health system or insurer has publicly announced coverage or protocol changes based on the Long COVID data, leaving the question of real-world implementation unanswered.
A peer-reviewed commentary published in Clinical Infectious Diseases in early 2026 went further, arguing that clinicians should consider offering metformin to outpatients during acute SARS-CoV-2 infection based on the accumulated trial and observational data. No regulatory body, however, has formally endorsed that recommendation.
Aging, arthritis, and the limits of repurposing
Beyond infectious disease, researchers have turned to metformin as a potential tool against the biology of aging itself. A double-blind, randomized, placebo-controlled trial conducted in England, known as MET-PREVENT, evaluated whether metformin could improve physical performance in older adults at risk of sarcopenia and frailty. The study measured gait speed and other functional outcomes, treating the trial as a direct test of whether a metabolic drug could slow physical decline. Results showed modest signals but fell short of the large, consistent effects that would prompt guideline changes.
A separate randomized clinical trial tested metformin for symptomatic knee osteoarthritis in patients with overweight or obesity, tracking pain scores, adverse events, and medication adherence. The logic was straightforward: metformin tamps down low-grade inflammation and improves metabolic health, both of which contribute to joint degeneration. Early results were encouraging enough to justify further study, but sample sizes remained small and follow-up periods limited.
The most sobering result came from oncology. The MA.32 trial, a large phase 3 randomized study, tested metformin as an add-on to standard treatment in non-diabetic women with early-stage breast cancer. The trial found no improvement in invasive disease-free survival. The negative result was a clear reminder that metformin is not a universal fix. Whatever anti-inflammatory or metabolic benefits the drug provides, they did not translate into better cancer outcomes in that population.
What remains uncertain
Even the Long COVID data carry important caveats. The RECOVER observational study, for all its scale, cannot prove causation. Adults who take metformin for diabetes differ from those on other medications in ways that electronic health records struggle to capture, including exercise habits, diet, and disease severity. Confounding remains a real possibility.
Long-term cardiovascular safety data for metformin in non-diabetic populations are thin. Decades of use in diabetes care have built a reassuring track record, but those patients carry different baseline risks. Extrapolating that safety profile to otherwise healthy people taking the drug for prevention or anti-aging purposes requires assumptions that have not been tested in large trials.
The most ambitious proposal in this space, the Targeting Aging with Metformin (TAME) trial, aims to test whether metformin can delay a composite of age-related diseases in older adults without diabetes. The trial’s design and rationale have been published, but as of spring 2026, TAME has faced persistent funding challenges and has not yet reported results. Until it does, claims that metformin “slows aging” remain speculative, however intriguing the biological rationale of improved insulin sensitivity, reduced inflammation, and effects on cellular energy sensors may be.
Timing is another open question. In the COVID-OUT trial, metformin was started within days of infection, a window when suppressing viral replication could plausibly prevent downstream complications. In chronic conditions like osteoarthritis or sarcopenia, the underlying pathology may already be entrenched by the time patients enroll in a study. Whether starting metformin earlier in life, or at preclinical stages of disease, would produce stronger benefits is unknown.
What this means for patients and doctors
For readers trying to make sense of the evidence, the key distinction is between randomized trials and observational studies. The COVID-OUT trial, the MET-PREVENT sarcopenia trial, the knee osteoarthritis trial, and the MA.32 breast cancer trial all randomly assigned patients to metformin or placebo and measured predefined outcomes. Their results are the most reliable guide to whether the drug works for a given condition. The RECOVER analysis, by contrast, drew on real-world prescribing patterns among a large patient population and detected a strong signal, but it cannot rule out the possibility that something other than metformin explains the lower Long COVID rates.
For someone currently managing a SARS-CoV-2 infection and weighing whether to ask a doctor about metformin, the published trial data show that a short course started early in illness reduced Long COVID diagnoses over extended follow-up. The drug is generic, widely available, and well-tolerated in most adults, though it can cause gastrointestinal side effects and is not appropriate for people with severe kidney impairment. Off-label use should be a conversation between patient and physician, balancing the randomized evidence against individual risk factors and the absence of formal regulatory endorsement.
For chronic conditions, the picture is more tentative. The aging and osteoarthritis trials hint at modest benefits in some groups, while the breast cancer data draw a clear boundary around the drug’s usefulness. Metformin should not be treated as a proven anti-aging therapy or a general-purpose disease modifier until larger, more diverse studies report.
Why a 69-year-old pill still drives new clinical trials
Metformin’s expanding research portfolio reflects a broader shift in how scientists think about old drugs. Rather than developing new molecules from scratch for every condition, researchers are revisiting inexpensive, well-characterized compounds to see if their biological effects reach further than originally understood. In metformin’s case, the answer so far is: sometimes yes, sometimes no, and often “we need more data.” The Long COVID prevention signal is real and supported by both randomized and observational evidence. The aging and joint-health signals are early and incomplete. The breast cancer result is a useful negative, narrowing the field of plausible applications. For a drug that still costs pennies per pill, that track record is remarkable enough to keep researchers enrolling patients for years to come.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
