Does Semaglutide Slow Ageing? What the New Nature Study Actually Shows

Female mice given semaglutide from 20 months of age outlived their controls by 92 days at the median, 834 against 742, in a study published in Nature on 2 September 2026 that also reports attenuation of several hallmarks of ageing.¹ The design includes the control that most such studies omit: a calorie-restricted group matched to the food intake reduction the drug produced. Semaglutide reproduced many of the benefits of calorie restriction and beat it on a few measures. This is one laboratory, one strain, one sex, and no humans. Condor Research supplies reference materials for laboratory research use only, and nothing below describes or implies human use of any compound.

What did the study actually do?

Feng and colleagues at the University of California, Berkeley treated 20-month-old female C57BL/6 mice with daily subcutaneous semaglutide at 10 nmol per kg, or an equal volume of saline.¹ Twenty months is roughly the equivalent of late middle age in this strain, which is the point of the design: the question was not whether a lifelong intervention helps, but whether starting one late still does. For the lifespan arm, 39 mice received saline and 40 received semaglutide, and treatment continued for the rest of their lives. A separate cohort was treated for three months for physiological, molecular and cellular work, and a third cohort of ten animals per group compared saline, semaglutide and calorie restriction over five months.

How much longer did the treated mice live?

Median lifespan was 742 days in the control group and 834 days in the semaglutide group, an increase of 92 days or about 12%, assessed by log-rank test.¹ The paper notes that the control group’s lifespan fell within the range reported for C57BL/6 mice in the literature, which matters: an apparent extension can be an artefact of unusually short-lived controls, and here it does not appear to be.

+92 days of median lifespan in treated female mice, from a starting point of 20 months of age, in a single-site study of 79 animals.

Was the effect simply eating less?

This is the question that decides how interesting the result is, and the authors ran the control that answers it. Semaglutide reduced food intake by 24%, so they included a calorie-restricted group restricted by that same amount.¹ The comparison found substantial overlap: semaglutide recapitulated many of the functional benefits of calorie restriction, and liver transcriptomics showed both interventions suppressing inflammation and lipid metabolism genes while inducing adaptive immune response, glucose and insulin response, and proteostasis pathways. Where they diverged, semaglutide showed more favourable trajectories in exploratory drive, spatial memory and glycaemic control.

One methodological detail deserves attention. The degree of restriction in the calorie-restricted group was set from the mean reduction in food intake that semaglutide produced, and was not adjusted over time.¹ If the drug’s effect on intake changes across months, as anorectic effects often do, then the two groups are matched at the start and drift apart afterwards. That does not undo the comparison, but it makes it approximate rather than exact, and a reader should hold the divergences a little more loosely than the overlaps.

The interesting claim is not that a drug extended lifespan. It is that it did so partly through mechanisms that eating less does not explain.

The broader concern is real. An analysis of the DrugAge database found significant correlations between drug-induced weight loss and lifespan extension in mice, particularly in the studies run by the Interventions Testing Program, which is a warning that compounds reducing intake can look like geroprotectors for uninteresting reasons.² Running the matched restriction arm is exactly the right response to that problem.

Which hallmarks of ageing were measured?

The paper examines the established set: maintenance of the haematopoietic system as a proxy for stem cell attrition, inflammation, cellular senescence, mitochondrial function, proteostasis and genomic instability, alongside physiological measures including open-field behaviour, body composition and glucose handling.¹ The authors also report effects on nutrient sensors and conserved genetic regulators of ageing, which is the mechanistic bridge to the calorie restriction framing. Their conclusion is that late-life GLP-1 receptor activation behaves as a calorie restriction mimetic with additional effects not attributable to reduced intake alone.

Why does it matter that the mice were female?

Because sex-specific results are the norm rather than the exception in mouse longevity research, not a footnote. Several interventions that extend lifespan in one sex do nothing or less in the other, and the field has enough examples that a single-sex result is treated as provisional by default.²⁴ The title of this paper is careful about it, and so should any summary be: this is a result in female C57BL/6 mice. Whether male mice respond the same way is not merely unknown, it is one of the two or three most likely places for the finding to fail.

What would make this a robust longevity result?

There is an established answer to that question. The National Institute on Aging’s Interventions Testing Program exists precisely because single-site, single-strain lifespan studies have a poor replication record; it tests compounds in genetically heterogeneous mice at three independent sites, with prespecified protocols, for exactly this reason.³ The rigorous version of this finding is an ITP-style replication in both sexes in heterogeneous animals. That is a multi-year undertaking, and until it exists the appropriate description of the present result is a well-controlled single-laboratory finding that merits replication, which is a real thing to be and not a dismissal.

All entries describe a rodent study. Doses in animal research are not convertible to human doses and are reported here only as experimental parameters. Nothing in this table supports any human use.

Does this say anything about people?

Not about human lifespan, no. There is no human evidence that any GLP-1 receptor agonist extends life, and this study does not provide any. What it does is give a mechanistic reason to take the observation seriously that these drugs produce benefits which are hard to explain by weight loss alone, an observation that has accumulated from cardiovascular and renal outcome trials in people. Connecting those two levels is a hypothesis, not a finding. The authors declare a University of California patent application covering GLP-1 receptor agonists for healthy ageing, which is ordinary practice and worth knowing when reading the framing.¹

For the wider evidence picture in this area, see our scorecard of human evidence for longevity compounds, our review of senolytics ranked by evidence, and our discussion of what exercise mimetics actually are, where the same lifespan-versus-healthspan distinction keeps recurring.

What is established, and what is not

Established: in this experiment, semaglutide begun at 20 months extended median lifespan in female C57BL/6 mice by 92 days, reduced food intake by 24%, attenuated several measured hallmarks of ageing, and outperformed matched calorie restriction on exploratory drive, spatial memory and glycaemic control.

Not established: any effect in male mice, in other strains, in genetically heterogeneous animals, or in any species other than mice. Not established: maximum lifespan effects, durability after stopping treatment, or the safety of indefinite late-life administration. Unresolved: how much of the benefit is anorectic and how much is receptor signalling, since the matched restriction arm was fixed rather than dynamic. What would change the assessment: independent replication in both sexes at multiple sites, which is the standard this field already has and should apply here.

How this was checked. The full text was read from PubMed Central on 13 September 2026 rather than from press coverage, and every figure quoted here, including the median lifespan values, the group sizes, the dose, the 24% intake reduction and the design of the calorie-restricted arm, was taken from the paper itself. The paper is open access under a CC BY licence. No p value is quoted because the relevant comparison is reported as significant by log-rank without a figure we verified. Version 1.0, first published 13 September 2026.

Condor Research supplies characterised reference materials for laboratory research use only: not for human or veterinary use, not for diagnostic or therapeutic application, and not for any food or cosmetic purpose. We do not supply semaglutide. This article is commentary on published preclinical research. Nothing in it is a therapeutic claim or a recommendation to use any compound in humans, and animal doses reported here must not be interpreted as applicable to people.

Condor Research · Scientific desk
Atrio Sciences s.r.o., IČO 57 669 651, Nitra (SK) · info@condorresearch.com