Rapamycin, Metformin, and the Gap Between Mice and Us
Two drugs dominate the longevity conversation. The animal data behind one is genuinely impressive. Neither has shown a lifespan effect in a human being.
Ask which pharmaceuticals might extend human healthspan and you will get the same two answers. They deserve different levels of confidence, and both deserve less than they usually receive.
Rapamycin: the strongest animal signal we have
Rapamycin inhibits mTOR, a nutrient-sensing pathway conserved across eukaryotes and central to nearly every mechanistic theory of ageing. It is a licensed immunosuppressant used in transplant medicine and in certain cancers.
The National Institute on Aging's Interventions Testing Program is the most rigorous longevity screening effort in existence: parallel testing at three independent sites, genetically heterogeneous mice, pre-specified protocols. Most candidates that arrive with excited press coverage fail there.
Rapamycin did not. It extended median and maximal lifespan in both sexes, replicated across sites, and — remarkably — worked when started at 20 months of age, the mouse equivalent of late middle age. That result is real and it is not a small deal.
Dog work through the Dog Aging Project is ongoing and represents a genuinely useful intermediate step: larger, longer-lived, sharing our environment.
The risks are not hypothetical. At transplant doses rapamycin causes immunosuppression, impaired wound healing, mouth ulcers, dyslipidaemia and glucose intolerance. The intermittent low-dose regimens used off-label are chosen specifically to reduce these effects, and there is no long-term safety data for those regimens in healthy people. That is not a reassurance; it is an absence.
Metformin: a weaker case that has been treated as a stronger one
Metformin is a first-line diabetes drug with an excellent safety record over sixty years. Enthusiasm traces largely to a 2014 observational analysis suggesting diabetic patients on metformin outlived non-diabetic controls.
That result is confounded in ways that are difficult to fix. Metformin is prescribed to healthier diabetic patients; sicker ones move to other agents. Comparing across those groups is comparing prognosis as much as treatment.
The animal data are also unimpressive relative to the reputation. Metformin performed poorly in the ITP — it did not extend lifespan on its own in the tested strain.
And there is a documented cost for anyone training: several trials show metformin blunts the adaptations to exercise, including attenuated improvements in aerobic capacity and insulin sensitivity. Given the strength of the fitness-mortality relationship, a drug that dampens training response is a strange choice for a healthy person.
The TAME trial, designed to test metformin against a composite of age-related diseases, has faced years of funding difficulty. Until it reports, this remains a hypothesis.
What I would say to a patient
For rapamycin: an important research compound with the best animal evidence in the field, no human efficacy data for this indication, real known risks, and unknown long-term risk at the doses being used off-label. Anyone taking it is a participant in an uncontrolled experiment that will not generate usable data because nobody is collecting it. If that appeals, do it with a physician and with monitoring.
For metformin: if you have type 2 diabetes or prediabetes, it is a good drug. If you do not, the longevity case is weak and there is a plausible cost to your training.
For both: the interventions with actual randomised human evidence remain exercise, sleep, not smoking, blood pressure control and lipid management. They are less interesting and they are what we have.
Dr. Raghunathan is a clinical pharmacologist writing in a personal capacity. This is general information about prescription medicines and is not medical advice.
