Does Ozempic Slow Aging? What the Mouse Study Really Showed

On September 2, 2026, a team at UC Berkeley published a paper in Nature reporting that semaglutide — the molecule in Ozempic and Wegovy — extended lifespan in mice when started late in life. Within days it was everywhere: GLP‑1sGLP‑1 (glucagon-like peptide-1) A gut hormone released after eating that boosts insulin, slows the stomach, and signals fullness. Drugs like semaglutide mimic it. as longevity drugs, Ozempic beats dieting, the anti-aging shot.

The study is legitimate and the finding is interesting. It is also narrower than almost any coverage conveyed. Here is what was actually measured, how it compares to the other drugs in this category, and what it does and doesn’t change for anyone making a real decision.

What the Study Actually Found

The researchers took mice at 20 months of age — roughly the equivalent of a human in their sixties — and gave them semaglutide injections for the rest of their lives. The comparison animals got injections of plain salt water and ate freely.

The headline result: medianmedian The middle value when results are lined up from lowest to highest. Median lifespan is the age by which half the group has died. lifespan rose by about 12% — from 742 days to 834. The treated animals also did better on tests of muscle and cognitive function.

Looking at the animals’ tissue, the team found that some of the usual molecular signs of aging, like rising inflammation, were blunted. That matters. It suggests the drug was acting on aging itself, not just preventing one particular disease.

A separate experiment compared semaglutide against simply eating less. Over five months, the drug came out ahead on several measures: treated mice explored more, remembered better, and handled blood sugar better. They also held their metabolic rate steady, while the dieting mice saw theirs fall.

That last comparison generated most of the “better than dieting” headlines. It is also the part of the study most often misdescribed, which we’ll come back to.

Why “Female Mice” Is Not a Footnote

Read the paper’s actual title: Late-life semaglutide treatment slows ageing and extends lifespan in female mice.

The authors put the limitation in the title because they know what it means. No male mice were studied. No results in males were reported, because there are none.

Why One Lab Isn’t Enough

Mouse lifespan results are surprisingly fragile. Diet, room temperature, the gut bacteria in a particular facility, even which technician handles the animals can change how long mice live. Compounds that looked promising in one lab have repeatedly produced nothing when other labs tried to repeat them.

That’s why the NIH’s Interventions Testing Program (ITP) tests promising aging compounds in three independent labs at once. Rapamycin, the most rigorously tested anti-aging drug in animals, went through that process and worked in both sexes. It’s taken seriously because of that — not because its numbers were bigger. They were about the same as semaglutide’s.

Semaglutide hasn’t been through it yet. And even rapamycin, which passed, still isn’t a proven longevity drug in people — as our review explains.

The Calorie Restriction Comparison Didn’t Test Longevity

This is the most distorted part of the coverage, so it’s worth being precise.

The calorie-restriction arm ran five months and measured physiological outcomes: exploratory behavior, spatial memory, glucose handling, metabolic rate. It did not follow those animals to death. It was never a lifespan comparison.

That means the widely repeated claim that semaglutide “beats dieting” for longevity is not something this study measured. It measured that on several mid-life physiological markers over five months, the drug looked better than eating 24% less.

Nir Barzilai of Albert Einstein College of Medicine — one of the most prominent researchers in this field, and the architect of the metforminmetformin The standard first medication for type 2 diabetes. It lowers blood sugar mainly by reducing how much sugar the liver releases. longevity trial — flagged exactly this gap in the press coverage, saying of the missing head-to-head lifespan data: “The data is not there, and it bugs me.”

When a leading advocate for using drugs to slow aging is publicly pointing at what the study didn’t do, that’s a signal worth heeding.

The Question Nobody Can Answer Yet: Drug or Weight Loss?

Here’s the problem that determines whether any of this translates.

Semaglutide makes animals eat less and lose weight. Calorie restriction and lower body fat are, on their own, among the most reliably life-extending manipulations in rodents — that has been known for ninety years.

So when treated mice live 12% longer, there are two very different explanations:

1. Semaglutide has direct molecular effects on aging pathways, independent of weight.
2. Semaglutide causes mice to eat less and carry less fat, and that extends lifespan.

The study cannot cleanly separate them. The authors themselves say the mechanism isn’t established — asked why the drug outperformed dieting on those physiological measures, the team’s answer was essentially that they don’t yet know.

This distinction is everything for humans. If the benefit runs through weight loss, then a person who is already lean has little to gain and a real amount to lose. If there’s a weight-independent effect on aging biology, that’s a genuinely new class of drug. Right now, nobody can tell you which is true — and the marketing, inevitably, assumes the flattering answer.

What We Actually Know in Humans

It’s worth separating the speculation from the evidence, because semaglutide’s human data is genuinely strong — just for something else.

The SELECT trial randomized more than 17,600 adults with obesity and heart disease, but no diabetes, to semaglutide or placeboplacebo A dummy treatment, such as a sugar pill or salt-water injection, that looks identical to the real one but contains no active drug. Comparing against it shows how much of an effect comes from the drug itself.. The result was a roughly 20% reduction in major cardiovascular events — real heart attacks and strokes prevented, in real people, in a properly powered trialpowered (statistical power) A properly powered trial enrolls enough people to reliably detect the effect it is looking for. An underpowered trial can miss a real effect or produce a fluke..

That is a far higher grade of evidence than anything in the mouse paper, and it’s the actual reason semaglutide is a serious drug. Preventing cardiovascular events is not the same claim as slowing aging, but it isn’t a trivial one either — cardiovascular disease is the leading cause of death, and a drug that reduces it meaningfully extends life expectancy at a population level whether or not it touches aging biology.

What does not exist in humans: any trial measuring lifespan, any trial measuring biological aging as a primary endpointprimary endpoint The main outcome a trial is designed and sized to measure, chosen before the study starts. Other results are treated as less reliable., and any data at all in people who are not overweight.

The Muscle Tension Worth Watching

One detail deserves a flag, because it points in an uncomfortable direction.

The mice on semaglutide showed improved muscle function. In humans, the well-documented pattern runs the other way: a substantial share of the weight lost on GLP‑1 therapy comes from lean tissue unless protein intake and resistance training actively counteract it — something we cover in detail in our article on GLP‑1s and muscle mass.

Those findings aren’t necessarily contradictory. Mouse muscle physiology, mouse diets, and mouse activity patterns differ from ours, and “muscle function” on a grip-strength or rotating-rod balance test isn’t the same measurement as lean masslean mass Body weight that isn’t fat — mostly muscle, bone, and water. on a DEXADEXA / DXA scan A quick, low-dose X-ray scan that measures body composition — how much of your weight is fat versus lean tissue — and bone density. The gold standard for tracking muscle and fat, rather than just scale weight. scan. But it’s a reminder that the animal result cannot simply be assumed to carry over — including the parts of it that sound most appealing. Sarcopeniasarcopenia The age- or illness-related loss of muscle mass and strength. It raises the risk of falls, frailty, and loss of independence. is one of the strongest predictors of death in older adults. A longevity drug that costs a 65-year-old lean mass is not obviously a longevity drug.

Where This Leaves Us

The reasonable position here is neither dismissal nor enthusiasm. GLP‑1 receptor agonistsreceptor agonist A drug that switches on a specific receptor, copying the body’s own signal — e.g. a GLP‑1 receptor agonist mimics the GLP‑1 hormone. have turned out to do far more than suppress appetite — cardiovascular protection, kidney protection, improvement in fatty liver disease. A direct effect on aging biology is a plausible extension of that pattern, and this paper is a legitimate reason to investigate it seriously.

It is not a reason to take the drug for that purpose. Here’s the evidence at a glance:

What can be said cleanly: this is the most interesting gerosciencegeroscience The field that studies the biology of aging itself, on the idea that slowing it could delay many age-related diseases at once. result of the year, and the mechanism question — drug versus weight loss — is now the most important open problem in the field. If the ITP runs semaglutide in both sexes across multiple labs and it holds, that changes the conversation substantially.

What cannot be said: that GLP‑1s slow human aging, that they beat dieting for longevity, or that any of this applies to a person at a healthy weight. As with metformin, the distance between a striking mouse result and a justified human decision is enormous, and the longevity market has a consistent habit of pretending otherwise.

If you are already on a GLP‑1 for obesity, metabolic disease, or cardiovascular risk, this paper is a pleasant bonus hypothesis and changes nothing about your regimen — keep prioritizing protein and resistance training. If you are healthy, lean, and now wondering whether to find a prescriber, the answer this month is the same as it was last month. The mice are promising. You are not a mouse.