Are GLP-1s Longevity Drugs? What the Death and Outcome Data Show

GLP-1 drugs are not proven longevity drugs. What they have is narrower and still real: in people who already have heart disease, type 2 diabetes or kidney disease, randomized trials pooled together show about 12% fewer deaths from any cause over two to four years.^[8][10] No trial has tested them in healthy adults, none has measured lifespan, and the anti-aging claims that circulate online rest on biomarkers and mice. Dementia prevention failed its direct test.^[13] Cancer shows no clear effect in trials.^[14] The muscle cost is real and matters most in exactly the older people a longevity drug would target.

The verdict in four lines

• Deaths: lower, in sick people. Every large outcome trial enrolled people with established cardiovascular disease, diabetes at high risk, or chronic kidney disease. In them, all-cause death fell by roughly a fifth in the best single trials and about 12% across pooled trials.
• Healthy people: no data. No randomized trial has measured death in adults without obesity, diabetes or heart disease.
• Lifespan: mice only. One 2026 study extended median lifespan in old female mice.^[19] There is no human lifespan data and there will not be for decades.
• Aging biology: surrogates. Epigenetic clocks moved in one 32-week trial.^[20] A clock is an estimate, not an outcome.

All-cause mortality, trial by trial

The table below lists every death result we could confirm from the trial abstracts or the FDA label, with the status of each one. That status matters. A trial is designed around one primary question. Death is usually a secondary result, and secondary results carry less weight because a trial that checks many outcomes will find some by chance.

SELECT is the trial most cited in longevity articles, so its fine print deserves a direct reading. The Wegovy label prints the death result and then a footnote:

“Confirmatory secondary endpoint. Not statistically significant based on the prespecified testing hierarchy.”^[2]

— Wegovy prescribing information, Table 6 footnote 4 (DailyMed)

The reason is mechanical. Cardiovascular death came first in SELECT’s testing order, and its confidence interval, 0.71 to 1.01, touched no effect.^[2] Once that step failed, the trial could not formally claim the outcomes below it, all-cause death included. The 19% figure is a credible estimate. It is not a proven result under the rules the trial set for itself. A later analysis of SELECT’s 833 deaths found the gap spread across heart and non-heart causes, and that fewer deaths from infection, during the COVID-19 years, did much of the non-heart work.^[3] That is not the pattern you would expect from a drug slowing aging itself.

FLOW is the cleaner result. Death from any cause was a prespecified confirmatory outcome, tested in order, and it came out at 20% lower.^[4] FLOW was stopped early for benefit, which tends to inflate effect sizes somewhat, and it enrolled only people with diabetes and chronic kidney disease. The absolute terms are on our page about how many people need treating to help one.

Pooled across trials the signal holds: three separate meta-analyses land on a 12% reduction.^[8][9] The largest, at nearly 100,000 patients over a mean of 2.4 years, puts it at 121 people treated to avoid one death, and also counts the cost: gastrointestinal disorders up 63% and gallbladder disorders up 26%.^[10] A 2026 BMJ network meta-analysis of obesity drugs found that injected semaglutide was the only one linked to fewer deaths, and said that estimate came mostly from cardiovascular trials in high-risk people.^[11] Tirzepatide’s death data, from one trial against an older GLP-1, are on tirzepatide and longevity.

Who the trials studied, and who they did not

A longevity drug, as the term is used, is something a healthy middle-aged person takes to live longer. No GLP-1 trial has studied that person.

• Every mortality result above comes from people who were already ill. SELECT required established cardiovascular disease and a BMI of at least 27.^[1] FLOW, SOUL, LEADER and SUSTAIN-6 required type 2 diabetes plus heart or kidney disease or high risk.
• Higher risk makes benefit easier to see. A drug that cuts heart attacks will save more lives in people who are having heart attacks. The same relative effect in a lean, healthy 50-year-old would produce far fewer avoided deaths, against the same side effects.
• The weight-loss trials were not built to count deaths. Trials such as STEP and SURMOUNT measured body weight over a year or two. They were never sized to count deaths.
• The approval matches the evidence. Wegovy’s heart indication covers only adults who already have cardiovascular disease along with overweight or obesity.^[2] No GLP-1 label carries an aging or longevity indication.

Lifespan: the mouse study and what it can carry

The strongest lifespan result is a 2026 Nature paper. Female mice started on semaglutide at 20 months old, already late in a mouse’s life, lived to a median of 834 days against 742 for untreated mice.^[19] The dose cut their food intake by 24%, and the authors frame the drug as a mimic of calorie restriction, which also lengthens mouse lifespan. Only females were studied; the authors chose them to avoid fighting injuries among males. The university behind the work has filed a patent on GLP-1 drugs for healthy aging.^[19]

That is a careful study and a real finding in mice. It does not tell anyone how long a person will live on the drug. Calorie restriction itself extends life in laboratory mice, and no study has shown it does the same for human lifespan. The same caution applies to the human biomarker work: one 32-week trial in people with HIV-related fat changes reported slower epigenetic clocks, a post hoc result that our page on whether semaglutide slows aging takes apart line by line.^[20]

Dementia: a strong signal that failed its test

For years the brain was the most exciting part of the longevity case. A meta-analysis of 26 randomized trials found GLP-1 drugs linked to 45% lower odds of dementia, though in trials that were built to study the heart and diabetes, not dementia.^[12] Large health-record studies pointed the same way. Our dementia risk page lays out why those analyses disagree.

Then the question was asked directly. EVOKE and EVOKE+ randomized 3,808 people in the early, symptomatic stage of Alzheimer’s to daily semaglutide tablets or placebo for two years. Clinical decline was the same in both groups, and both trials were discontinued for negative outcome.^[13] Prevention in people without symptoms has not been tested, so the door is not fully closed. But the one direct trial failed. Detail is on our EVOKE result page.

Cancer: no clear effect in trials

Cohort studies suggest fewer obesity-related cancers. One US study of 919,609 people with type 2 diabetes and excess weight found a 13% lower rate in GLP-1 users than in users of other diabetes drugs.^[15] Cohorts like this cannot rule out that healthier or better-insured people are more likely to get the drug. The randomized data, 48 trials and 94,245 people, found that GLP-1 drugs “may have little or no effect” on obesity-related cancers either way, with the caveat that none of those trials was designed or long enough to measure cancer.^[14] Both sides are on what the cohorts show and the randomized answer.

Muscle and frailty: the trade-off a longevity user takes on

Weight loss on any method removes some lean mass. A meta-analysis of 20 trials put the lean share at 35.2% of weight lost on semaglutide and 25.4% on tirzepatide, against 26.2% with lifestyle programs, and found resistance training brought it down to 17.5%.^[16] That comparison is on lean mass, drug versus dieting.

The concern sharpens with age. A 2026 review noted that longer-term and retrospective studies in older people with diabetes reported falling grip strength on semaglutide, even though shorter trials in mixed-age groups did not.^[17] There is a counterweight from SELECT: in a post hoc analysis, the drug’s benefit on heart events and deaths did not shrink in the frailest participants, and their frailty scores were more likely to improve than worsen.^[18] Those people had heart disease and obesity. For a lean older adult taking a GLP-1 to live longer, there is no benefit data to set against the muscle cost. See GLP-1s after 65.

Kidney: the hardest outcome after death

FLOW is the one trial whose main question was organ failure. Among its 3,533 participants, all with diabetes and kidney damage, semaglutide cut the combined risk of kidney failure, a 50% loss of kidney function, or death from kidney or heart causes by 24% (HR 0.76, 0.66 to 0.88), and cardiovascular death by 29%.^[4] It is strong evidence of organ protection in people whose kidneys are already failing. It says nothing about healthy kidneys. The short-term kidney warnings on the label are covered in GLP-1s and your kidneys.

How the other longevity candidates compare

Measured by hard outcomes in randomized trials, GLP-1 drugs are ahead of the usual longevity supplements, because those have none. NAD+ boosters raise blood NAD+ and have no outcome data at all; see NAD and longevity. Metformin, the older longevity favorite, is compared directly in metformin vs GLP-1 for longevity. Being ahead of supplements is a low bar. It does not make a GLP-1 a sensible prescription for someone who is healthy and wants to live longer.

Any benefit also depends on staying on the drug, which turns longevity use into an open-ended commitment. What happens to heart risk after people stop is on the cardiovascular cost of stopping.