Metformin for Longevity: Why Peter Attia Reconsidered His Position

The Shift That Got People's Attention

For most of the 2010s, if you listened to longevity-focused physicians and researchers talk about geroprotective molecules, metformin was usually in the top three. Peter Attia talked about it on his podcast. David Sinclair wrote about it. Nir Barzilai built the TAME trial around it. Metformin was the longevity drug that had the longest human safety record, the largest supporting epidemiology, and the most tractable path to FDA-recognized anti-aging labeling.

And then, gradually and then all at once, the conversation shifted. By late 2022, Attia was publicly saying he had stopped recommending metformin to patients pursuing longevity goals, and that he had discontinued taking it himself. The stated reason was research showing that metformin appeared to dull some of the cellular changes in your muscles that exercise normally produces — the changes that build cardio fitness.^[1]

This wasn't a quiet walk-back. It was a public, specific, reasoned reversal from one of the most-followed voices in the longevity space. And because the reversal was grounded in a specific set of published papers rather than a vague change of heart, it spread through the biohacker community in a way that reshaped the default metformin position for a lot of informed readers.

This article is the full arc: the original rationale, the TAME trial context, the specific research that shifted the conversation, and where other longevity-focused clinicians have landed. Pepvio doesn't sell metformin — this is authority content, not a sales case.

The Original Longevity Case for Metformin

The longevity case for metformin rested on several pillars that had accumulated over decades:

Epidemiology in diabetic populations. Multiple retrospective cohort studies in patients with type 2 diabetes taking metformin showed lower all-cause mortality, lower cancer incidence, and lower rates of cardiovascular events compared to other diabetes medications. The strikingly counterintuitive finding was that diabetic patients on metformin sometimes appeared to have lower mortality than the non-diabetic comparison cohorts.^[1]

Preclinical mechanisms that lined up with aging biology. Metformin's primary mechanism — activation of AMPK, inhibition of mitochondrial complex I, downstream effects on mTOR — hit several of the hallmarks of aging frameworks that were coalescing in the 2010s. It was mechanistically plausible in a way that not many molecules were.

Favorable safety profile. Metformin has been in widespread use since the 1950s in Europe and since the 1990s in the US. The adverse event profile is well-characterized, serious adverse events are rare in otherwise-healthy adults, and the drug is extraordinarily cheap.

The TAME trial. Nir Barzilai and colleagues designed Targeting Aging with Metformin (TAME), a proposed multi-thousand-person trial that would test metformin for reduction in composite aging outcomes (cardiovascular events, cancer, cognitive decline, mortality) rather than any single disease. The design was important beyond metformin itself — it was attempting to establish a regulatory pathway where 'aging' is a treatable indication.^[2]

Given this stack — decent epidemiology, mechanistic plausibility, cheap and safe, ongoing trial design — metformin became the default 'hedge bet' longevity molecule for people who wanted something backed by more than mouse studies. Attia's position through most of this era was consistent with this framing.

The Konopka and Miller Research That Reframed the Conversation

The specific research that Attia has publicly cited as shifting his position is a body of work from Adam Konopka's lab (and independently from Ben Miller's group) examining how metformin interacts with the cellular adaptations to aerobic exercise.

The core finding: in a randomized controlled trial in healthy older adults undergoing a structured aerobic exercise training program, the metformin group's muscle cells responded less to the training, and they made smaller gains in cardio fitness (VO2max), than the placebo group doing the same workouts.^[1] The mechanism proposed is that exercise works partly because it briefly stresses your mitochondria, and they then adapt by getting stronger. Metformin softens that initial stress signal, so the adaptation that follows is smaller too.

This matters because cardiorespiratory fitness is one of the strongest predictors of all-cause mortality in observational epidemiology. Larger VO2max gains from exercise aren't a small effect — they're one of the biggest levers on longevity anyone has identified. If metformin blunts those gains, then for an individual who exercises seriously, the cost-benefit calculation of metformin looks meaningfully different than the cost-benefit in a sedentary diabetic population where metformin's effects were originally studied.

The argument Attia has made in subsequent podcasts and writing is roughly: the epidemiology that made metformin look good for longevity was in populations that weren't doing structured exercise training. For his actual patients — biohackers who train seriously — the exercise cost probably outweighs whatever longevity benefit metformin might offer.

The Counter-Argument and Where Other Clinicians Have Landed

Not every longevity-focused clinician has followed Attia to the same conclusion. The counter-arguments worth taking seriously:

Effect size and population. The exercise-blunting effect in the Konopka work is real but modest in magnitude. For a patient who doesn't train hard — and the median US adult does not — the epidemiological metformin benefit may still outweigh the exercise cost.

Metformin's non-exercise benefits. Metformin has effects on glucose regulation, insulin sensitivity, and possibly on cancer risk that are independent of the mitochondrial-exercise pathway. Those benefits, at the population level, are still there.

Timing and dose questions. Some clinicians have suggested that timing metformin away from training windows (morning metformin, evening training; or cycling off during high-training periods) might preserve exercise adaptation while retaining other benefits. This is speculative — there isn't controlled trial data to confirm it — but it's the current thinking in some practices.

TAME still matters. Even clinicians who personally don't recommend metformin for longevity acknowledge that the regulatory precedent TAME is trying to set (establishing aging as a treatable indication) is important beyond metformin itself.

Barzilai's response to the Attia reversal has been roughly: the exercise-blunting effect is real and worth knowing, but metformin's effects at the population level remain interesting, and picking one mechanism out of many doesn't settle the debate. Other longevity-aligned physicians — including some associated with AgelessRx, which continues to prescribe metformin — take a similar position.

Bryan Johnson's Blueprint protocol also walked away from metformin around the same time, for similar reasons to Attia's. The informed-biohacker consensus, if one can be said to exist, has shifted toward 'not a default recommendation anymore for people who exercise seriously.'

What the Research Actually Does and Doesn't Say

A careful read of the state of evidence:

This is the reason the debate is interesting rather than settled. The epidemiology is compelling at the population level. The mechanism is plausible. The exercise-blunting finding is concerning specifically for the kind of active, health-optimizing patient most likely to be reading about longevity molecules. Depending on which of these weights most, reasonable clinicians land in different places.

The TAME Trial Status

Where TAME sits in 2026: the trial has faced funding, enrollment, and timeline challenges. Barzilai's group has continued working toward getting the trial fully funded and running, but it has not reported definitive results as of early 2026. Readers tracking the longevity research space will want to watch for TAME updates — the trial, if and when it completes, will be the most definitive prospective evidence on metformin-for-longevity in non-diabetic populations.

One thing worth understanding about TAME: even if it reports positive results, the effect size in a healthy non-diabetic population will probably be modest. To prove a small effect is real, you either need a big effect or a huge number of participants. TAME is designed around the latter assumption. The clinical takeaway from a positive TAME result would likely be 'metformin produces a modest but real reduction in aging-related outcomes in non-diabetic adults,' not 'metformin is a dramatic longevity intervention.'

Even a modestly positive TAME probably doesn't reverse the Attia-era reconsidered position for the active-biohacker population — because the exercise-blunting concern would still apply. It would, however, be a potentially meaningful result for sedentary older adults, for whom the population-level mortality benefit probably still outweighs the exercise-blunting cost (since they're not doing much exercise to blunt).

What This Means for a Reader Sorting Through Longevity Molecules

The practical takeaway for a reader trying to build a thoughtful longevity protocol in 2026: metformin is no longer the default recommendation it was five years ago. It's not contraindicated for longevity use — reasonable clinicians still prescribe it, particularly for patients with insulin resistance or metabolic syndrome components. But it's also no longer the 'everyone should consider this' molecule it was.

For an active patient whose exercise program is central to their health strategy, the current informed-biohacker default is leaning toward 'probably skip metformin, prioritize exercise, use other geroprotective levers.' For a sedentary or pre-diabetic patient, metformin may still be a reasonable consideration — and in that population, an endocrinologist or primary care physician's judgment should lead the conversation.

The broader lesson in the Attia reversal is probably worth more than the specific metformin conclusion. The lesson is: epidemiology in one population doesn't automatically translate to recommendations for a different population. Metformin's epidemiology is in diabetics. Most people considering metformin for longevity aren't diabetic. The translation required assumptions that the Konopka/Miller work has now partly undermined.

Pepvio's longevity catalog is rapamycin and NAD+ — molecules with different mechanisms that don't raise the exercise-adaptation concern. For a longer read on rapamycin specifically, see our rapamycin longevity guide. For NAD+, see our NAD+ protocol explanation.