Metformin (longevity claim)

The verdict

Our read at a glance.

Whether a drug that treats diabetes slows ageing in people who do not have diabetes.

The signal

For type 2 diabetes the evidence is as good as it gets: decades of use, cardiovascular outcome data, and an approved first-line role. The longevity case is a different standard of evidence. It leans on observational analyses reporting that diabetic patients on metformin had mortality comparable to or better than matched non-diabetic controls, on model organism data, and on plausible mechanisms involving AMPK, mitochondrial complex I and nutrient sensing. Those comparisons carry well-recognised confounding. The purpose-built trial - designed to test whether metformin delays the onset of age-related disease in non-diabetic older adults - has been discussed for years and has not delivered a result. Meanwhile a separate literature has found that metformin blunts some of the adaptations to aerobic and resistance exercise training in older adults, which is a direct trade-off for this exact population.

The unknown

Whether any longevity benefit exists in non-diabetic people, and whether it is offset by the blunting of training adaptation - which matters because exercise is the intervention with the strongest evidence in this space.

Our read

The most defensible longevity candidate and still not close to demonstrated. An excellent diabetes drug with a mechanism worth testing, an observational literature that cannot settle the question, no completed trial in healthy people, and one replicated finding that it interferes with the benefits of training. Taking it for ageing is a bet, not a conclusion.