Adults with prediabetes who took metformin for more than two decades showed no reduction in heart disease, cancer, or death compared to those given a placebo, according to the longest follow-up yet from the National Institutes of Health’s Diabetes Prevention Program. The finding, drawn from data collected between 1996 and 2021, directly challenges the popular narrative that the inexpensive generic drug can slow aging and prevent chronic disease in people who do not already have diabetes. The result lands at a moment when millions of Americans take or consider metformin off-label, hoping it will do for them what earlier trials suggested it did for people with established type 2 diabetes.
Why metformin’s prediabetes results upend a decade of hype
The excitement around metformin as an anti-aging drug traces back to a single trial in a specific population. In the late 1990s, the UK Prospective Diabetes Study randomly assigned overweight patients newly diagnosed with type 2 diabetes to intensive glucose control with metformin or to conventional treatment. That trial, reported in a landmark Lancet analysis, found that metformin reduced myocardial infarction and all-cause mortality in those patients. A 10-year post-trial follow-up published in the New England Journal of Medicine later showed that the survival benefits persisted even after the formal intervention ended, a phenomenon researchers called the “legacy effect.”
Those results fueled a hypothesis: if metformin protects people with diabetes from heart attacks and early death, perhaps it works through mechanisms beyond blood sugar control. Maybe it slows biological aging itself. That idea gained traction in longevity research circles and led to off-label prescribing for people without diabetes. But the hypothesis rested on an untested assumption: that benefits observed in people with high blood sugar and active metabolic disease would carry over to people whose glucose levels were only mildly elevated or entirely normal.
The DPP/DPPOS data now available tests that assumption directly. Participants with impaired glucose tolerance were randomized between 1996 and 1999 to metformin, intensive lifestyle intervention, or placebo and followed through 2021. In this prediabetic population, the NIH reported that lifestyle intervention lowered the risk of developing multiple chronic conditions compared to placebo, while metformin showed no advantage on overall multimorbidity. The contrast is stark: structured diet and exercise changes worked in prediabetes, while the pill did not.
That divergence matters for how clinicians and patients interpret metformin’s reputation. In people with newly diagnosed type 2 diabetes, metformin clearly improves glycemic control and, in UKPDS, appeared to translate that into longer-term cardiovascular and survival gains. In people with prediabetes, the same drug over more than two decades did not prevent the accumulation of major chronic illnesses when compared with placebo. The implication is that metformin’s benefits are context-dependent, emerging in the setting of frank diabetes but not in earlier, milder metabolic disturbance.
UKPDS legacy data and the DPP’s 21-year cancer and mortality record
The strongest evidence for metformin’s benefits still comes from the UKPDS program, but even that evidence has limits. Post-trial monitoring extended to roughly 24 years in a later UKPDS follow-up, including the subgroup of overweight participants assigned to metformin. That long horizon reinforced the idea that early intensive glucose control can leave a durable imprint on cardiovascular risk, yet it did not fully disentangle how much of the effect was specific to metformin versus improved glycemia more broadly. Detailed, arm-specific breakdowns for cancer mortality across the entire extension have not been publicly reported, leaving uncertainty about any long-term anticancer signal in that cohort.
The DPP/DPPOS program, in contrast, has generated its own long-horizon data on both cancer and death in people who started with prediabetes. A 21-year cancer incidence analysis evaluated whether metformin or lifestyle intervention reduced adjudicated cancer outcomes in the randomized groups and did not find a protective effect of metformin compared with placebo. Separately, a peer-reviewed mortality analysis in Diabetes Care assessed all-cause, cardiovascular, and cancer mortality and similarly found no mortality reduction for metformin versus placebo over two decades of follow-up.
These clinical findings echo a broader evidence base. A large meta-analysis of randomized trials that restricted inclusion to studies with at least 500 participants and a minimum of one year of follow-up reported inconsistent signals for cancer incidence associated with metformin use. Some smaller or observational studies have hinted at anticancer effects, but when the lens is narrowed to adequately powered randomized data, the picture becomes far less convincing.
None of this erases what UKPDS observed in people with overt diabetes. It does, however, draw a clear boundary. Metformin’s cardiovascular and survival benefits appear tied to its glucose-lowering action in patients whose glycemia is already in the diabetic range, not to some broad anti-aging mechanism that operates regardless of metabolic status. The drug solved a specific problem in a specific population. Extending that success story to prediabetes or general longevity required evidence that has now, at least in this large trial, come back negative.
Open questions about metformin, aging biology, and what to watch next
Several gaps in the evidence prevent a final verdict on metformin’s role in aging science. Mechanistic work has suggested that metformin can influence cellular pathways implicated in longevity, including AMP-activated protein kinase signaling, mitochondrial function, and inflammatory cascades. The MILES trial, registered as NCT02432287, was designed to probe some of these questions directly by measuring metformin’s molecular effects on aging biology through gene expression and RNA-sequencing endpoints in human tissues. That study can help clarify whether metformin shifts cellular markers associated with aging, but it was not built or powered to detect reductions in heart attacks, cancer diagnoses, or deaths.
The DPP/DPPOS multimorbidity analysis also has limitations. While it shows no overall advantage for metformin over placebo in preventing the accumulation of major chronic conditions, it does not fully resolve whether particular subgroups-such as those who actually progressed to diabetes during follow-up, or those with the highest baseline fasting glucose-might have derived some targeted benefit. Post hoc subgroup analyses can be misleading, but they remain an area of active interest for researchers trying to reconcile the drug’s clear value in diabetes with its apparent futility in broader prediabetic populations.
Another open question is timing. The UKPDS experience suggests that initiating metformin at the time of diabetes diagnosis and maintaining good glycemic control early in the disease course can yield long-term cardiovascular dividends. The DPP/DPPOS results, by contrast, indicate that starting the drug even earlier-before diabetes has developed-does not further extend those benefits in terms of multimorbidity, cancer, or mortality. One possibility is that there is a threshold of metabolic dysfunction below which metformin simply has little to do; another is that lifestyle changes at the prediabetes stage are powerful enough to overshadow any subtle pharmacologic effect.
For clinicians and patients, the practical message is relatively straightforward. In people with type 2 diabetes, metformin remains a first-line therapy with a robust safety record and strong evidence for reducing microvascular complications, and in some settings, cardiovascular events. For people with prediabetes who are otherwise healthy, the DPP/DPPOS data argue that long-term metformin should not be prescribed with the expectation that it will broadly prevent heart disease, cancer, or early death. Lifestyle interventions-weight loss, physical activity, and dietary changes-have demonstrated benefits in this group that metformin has not matched.
The broader field of aging research is also taking note. Large-scale efforts such as the proposed TAME (Targeting Aging with Metformin) trial aim to test whether metformin can delay a composite of age-related diseases in older adults without diabetes. The DPP/DPPOS experience in prediabetes does not directly answer that question, since the populations and endpoints differ, but it does raise the evidentiary bar. Any claim that metformin is an anti-aging drug will have to contend with long-term randomized data showing no advantage on major clinical outcomes in a large, high-risk group.
Meanwhile, the concept of a “legacy effect” in glucose control continues to shape prevention strategies. The 10-year follow-up of the original DPP lifestyle arm, published in the New England Journal of Medicine as a long-term outcomes report, found that early, intensive lifestyle changes conferred durable reductions in the incidence of type 2 diabetes itself. That durability, now extended to multimorbidity and mortality advantages over decades, underscores a consistent theme across metabolic trials: when it comes to preventing chronic disease, behavior change initiated early can leave a lasting imprint that no pill has yet matched.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
