Semaglutide extended median lifespan in a new experiment—but the beneficiaries were older female mice, not people. That distinction is the starting point for understanding a study that is scientifically provocative precisely because it asks a narrower question than the headlines around longevity drugs often do.
Researchers at the University of California, Berkeley tested whether late-life activation of the GLP-1 receptor could do more than reduce appetite and body weight. Their results, published September 2 in Nature, connect the drug to longer survival, preserved physical and cognitive performance, and shifts across several molecular features associated with aging. None of that establishes that semaglutide slows aging or extends life in humans.
What changed in the mice
The lifespan experiment began with 20-month-old female C57BL/6 mice. Thirty-nine received saline and 40 received daily semaglutide injections until death. Median lifespan was 742 days in the control group and 834 days in the treated group—a 92-day difference, or roughly 12 percent relative to the control median.
Separate cohorts allowed the researchers to examine function and biology without waiting for the survival study to end. After three months, treated mice performed better on tests involving exploration, spatial memory, motor coordination, grip-related muscle function, treadmill endurance and glucose handling. Tissue analyses also found changes consistent with less inflammation and cellular senescence, improved mitochondrial function and protein maintenance, and altered blood and neural stem-cell features.
Those measurements span several systems, which makes the result more informative than a survival curve alone. But breadth is not the same as proof of a single anti-aging mechanism. Behavioral tests can be influenced by body weight, motivation and other drug effects; molecular markers can move without determining lifespan; and the experiments used one sex and one inbred mouse strain.
TENS analysis: the feeding control is the real hinge
The cleanest way to read this experiment is as a test of whether reduced food intake explains everything. Semaglutide cut food consumption by 24 percent, so longer survival could otherwise look like a familiar calorie-restriction effect wearing a pharmaceutical label.
The team therefore ran a separate five-month comparison among saline, semaglutide and a diet restricted by the same 24 percent. Ten mice were assigned to each group. Both interventions produced similar weight and fat loss and preserved several functions relative to controls. Yet the feeding patterns differed: restricted mice ate their allotment quickly and then fasted, while treated mice ate more gradually. Semaglutide also produced more favorable trajectories on exploratory behavior, spatial memory and glucose control.
That comparison weakens the simplest explanation that all observed benefits came from eating less. It does not isolate one alternative mechanism. GLP-1 receptor signaling changes appetite, glucose regulation and physiology across multiple tissues, while food timing itself can affect metabolism. The study creates a stronger mechanistic hypothesis, not a clean separation between “calorie restriction” and “drug effect.”
Why the human evidence is still different
Semaglutide already has substantial human evidence for approved metabolic indications and outcomes in defined patient groups. Longevity is a different claim. The mouse study did not test people, and its late-life design cannot tell us whether healthy older adults would receive the same balance of benefits and risks.
This is where the translation gap widens. A 2026 post-hoc analysis of the SLIM LIVER study examined three DNA-methylation aging measures in 41 people with HIV and metabolic liver disease who received semaglutide for 24 weeks. Group medians did not show consistent improvement across the three measures. Some individual changes tracked with liver-fat reduction or walking speed, but the study had no placebo group, was exploratory, and did not test survival or healthspan.
Those human findings neither confirm nor refute the mouse result. They measure different populations, durations and endpoints. They do show why a biomarker response cannot be treated as a substitute for a demonstrated extension of healthy life.
Evidence level and limitations
This is controlled preclinical evidence in aged female mice, supported by functional, molecular and survival endpoints. The survival groups were modest, the calorie-matched comparison used only ten animals per arm, and males were excluded to reduce injury-related confounding. Results may differ by sex, strain, dose, health status or starting age.
The researchers reported no treatment-attributable adverse effects among the endpoints monitored, but a mouse experiment cannot define long-term safety for healthy humans. Nor can a study built around daily injections in laboratory animals answer how treatment discontinuation, muscle loss, nutrition, competing illness or years of exposure would shape outcomes in older people.
What a genuine longevity test would require
The actionable next step is not off-label self-experimentation. It is a randomized human research program that prespecifies aging-related clinical outcomes, follows participants long enough to separate transient weight-loss effects from durable function, and measures harms as carefully as benefits. Biomarkers could help explain results, but they should not become the result.
The new study moves the field forward by showing that late-life GLP-1 receptor activation can be tested against matched food intake rather than assumed to be calorie restriction by another route. Its strongest contribution is a sharper experimental question. Whether that question leads to human healthspan gains remains entirely open.
Sources: Nature; National Institutes of Health; npj Aging.
TENS Magazine conceptual illustration
