Semaglutide, Multivitamins and Slower Aging: What the Evidence Can Support

A study can show that mice lived longer without telling us whether a treatment extends human life. It can also show a change in a biological-age estimate without establishing that anyone became healthier.

That distinction matters when reading the recent research on semaglutide and multivitamins. There are findings worth taking seriously here, including a demonstrated cardiovascular benefit and small improvements in cognitive performance. Calling them all “slower ageing” obscures what each study can tell you.

Semaglutide: a clinical benefit and a separate longevity question

The strongest human evidence discussed here comes from SELECT, a trial involving 17,604 adults with established cardiovascular disease and overweight or obesity, without diabetes.

Over an average follow-up of 39.8 months, cardiovascular death, non-fatal heart attack or non-fatal stroke occurred in 6.5% of participants assigned semaglutide, compared with 8.0% assigned placebo. The absolute difference was 1.5 percentage points. Adverse events led to permanent treatment discontinuation in 16.6% and 8.2%, respectively. Both the benefit and the difficulty some participants had continuing treatment matter. Lincoff and colleagues, SELECT investigators, *New England Journal of Medicine*, 2023.

SELECT established a reduction in these cardiovascular events in the population studied. It did not establish that semaglutide slows ageing in healthy middle-aged adults. If you are considering prescription treatment, its relevance depends on a clinical assessment of your circumstances, the treatment’s indications, and its benefits and risks.

What the mouse study adds

In September 2026, University of California, Berkeley researchers reported that semaglutide increased median lifespan in female mice from 742 to 834 days, approximately 12%. Treatment began at 20 months and continued throughout the survival experiment. This was a survival finding, although limited to one sex and one mouse strain. Feng and colleagues, *Nature*, 2026.

The researchers also reported a 24% reduction in food intake. In a separate experiment comparing semaglutide with matched calorie restriction, both treatments produced several similar functional benefits. Semaglutide performed better on some measures, including spatial memory. That comparison helps explore how the drug works; it does not establish that the lifespan extension was independent of reduced food intake. Feng and colleagues, University of California, Berkeley, 2026.

The human question remains open: could treatment delay disability or extend survival beyond the populations and conditions already studied clinically? The mouse finding supports investigating that question. Its percentage cannot be applied to your life expectancy.

Multivitamins: the outcome changes the answer

The Brigham and Women’s Hospital-led COSMOS trial enrolled 21,442 older adults. Its 2022 report found no statistically significant reduction in total invasive cancer, total cardiovascular disease or all-cause mortality over a median 3.6 years. That follow-up leaves questions about longer-term effects unresolved. Sesso and colleagues, COSMOS investigators, *American Journal of Clinical Nutrition*, 2022.

Cognition produced a different finding. A 2024 pooled analysis of COSMOS cognitive substudies found small benefits in global cognition and episodic memory. The difference in global cognition was 0.07 standard deviation units. This was a difference in measured cognitive performance, not evidence that multivitamins prevent dementia or extend life. Vyas and colleagues, COSMOS investigators, *American Journal of Clinical Nutrition*, 2024.

These results can coexist. The absence of a demonstrated mortality benefit does not erase the cognitive finding. Equally, better cognitive scores cannot answer the mortality question.

When biological clocks disagree

Two newer COSMOS analyses examined ageing-related measurements rather than diagnoses or survival.

The metabolomics study followed 399 participants over two years. It examined 168 blood metabolites, seven metabolomic ageing clocks and 24 disease-related metabolomic risk scores. DHA and several other measures moved in favourable directions, but the metabolite and disease-risk-score findings did not remain statistically significant after correction for multiple testing. All seven ageing clocks showed numerically favourable differences; none reached statistical significance. Li and colleagues, COSMOS investigators, *GeroScience*, 2026.

Multiple-testing correction matters because examining many outcomes creates more opportunities for an apparently positive finding to arise by chance. Several favourable estimates may warrant further study, but they do not become confirmations simply by pointing in the same direction. A lower modelled disease-risk score also does not mean that fewer participants developed that disease.

The other analysis examined DNA methylation, chemical marks on DNA, rather than circulating metabolites. Among 958 participants followed for two years, multivitamins modestly slowed the rate of increase in two of five epigenetic clocks. The clinical relevance of those changes remains unresolved. Li and colleagues, COSMOS investigators, *Nature Medicine*, 2026.

Expressing a result in biological-age years can make it sound like time gained. These studies do not establish that conversion. A change in an ageing-related measurement still needs to be connected to outcomes people care about, such as remaining independent or living longer in good health.

What would make a finding relevant to you?

If your concern is staying mobile, a study measuring mobility answers your question more directly than one measuring a blood-based age estimate. If you are concerned about dementia, changes in cognitive test scores and changes in the number of diagnoses are different outcomes.

Who took part matters just as much. Results in older adults do not automatically answer a question about a healthy 45-year-old. SELECT’s cardiovascular findings apply to a defined clinical population; the mouse survival finding remains animal evidence.

Once those distinctions are clear, consider the size and uncertainty of the result. The control group, absolute difference and confidence interval help you judge what the finding supports. Where researchers tested many outcomes, their handling of multiple comparisons matters too.

Personal observation has a more limited role. You might record how a familiar walk feels or whether afternoon fatigue interrupts your work. Those observations relate to daily function, but they are not validated estimates of lifespan.

If you choose to observe a familiar habit, such as a consistent bedtime, keeping other routines broadly stable can make the record easier to interpret. Note whether you followed the habit, any perceived benefit and any downside. Expectations and unrelated changes remain possible explanations, even if you feel better. Changing supplements, diet and training together makes attribution harder still.

Questions about medication or supplementation are best framed around a specific outcome and discussed with a clinician. On the evidence reviewed here, there is no established general reason to start semaglutide or a multivitamin solely to slow ageing. A finding can deserve your attention without requiring a change to your routine.

Sources