Nothing in this paper is a reason to buy a GLP-1, and it is going to be reported as one. What it is, is the first careful attempt to separate a drug’s effect on aging from the effect of simply eating less — and it is the outcome that the surrogate-marker literature keeps standing in for, measured directly, in animals.
What was done
Female mice of a single common laboratory strain were given daily semaglutide by injection starting at twenty months of age, which is late middle age for a mouse rather than the start of life. [1] Treatment continued for the rest of their lives.
Median lifespan came out at 742 days in the vehicle group and 834 days on semaglutide, compared by log-rank test. The gap of 92 days, about 12%, is this desk’s subtraction from those two published figures rather than a number the paper prints.
What it also found
Three months of treatment improved physiological function, attenuated recognized hallmarks of aging, and shifted nutrient-sensing pathways and conserved genetic regulators of aging in the direction calorie restriction shifts them.
The authors read the whole picture as semaglutide acting like a calorie-restriction mimetic, with some effects beyond what reduced intake explains — which is a mechanistic claim about mice, and a different kind of evidence from a biomarker moving in people.
Every reason to hold it loosely
Female mice only, one strain, one laboratory. Strain and sex both change lifespan results in this field routinely, and a finding in female C57BL/6 mice is a finding in female C57BL/6 mice until somebody repeats it elsewhere.
The schedule is not a human schedule either. These animals were injected daily; people take semaglutide weekly, at doses that do not map across species by any simple ratio. Drugs that extend rodent lifespan have a long history of not doing so in humans, and there is no registered trial asking this question in people.
What a reader should take from it
That the mechanism question is now genuinely open, and that it was mostly closed before this — the reasonable default was that any longevity effect of these drugs is a weight effect wearing a lab coat. This is real evidence against that default, in mice.
What it is not is a purchase. The drugs on this site are sold for weight, at prices set by sellers, to people who will take them for a year or two at most, and what happens outside a controlled study looks nothing like a mouse facility. A first-in-class result has to survive a long way before it means anything at a checkout — the same distance an early-phase weight figure has to travel.