AIBRIEF
Twenty years after a 1990s trial randomized 1,483 people at risk for type 2 diabetes, the metformin group had 1.9% dementia incidence, against 4.2% in a diet and exercise program and 3.5% on placebo.
A 20-year follow-up of a 1990s diabetes-prevention trial found the lowest dementia rate in the arm given metformin, the standard type 2 diabetes drug taken daily by roughly 150 million people worldwide.
Among 1,483 participants originally randomized to metformin, a lifestyle program, or placebo, 1.9% of the metformin group had developed dementia by the end of follow-up, against 4.2% in the lifestyle arm and 3.5% on placebo. Only 48 dementia cases accrued across the whole cohort, a ceiling on how much weight the gap can carry. The metformin arm also scored better on memory testing.
Lead investigator José Luchsinger, of Columbia University, called the result "the available best evidence, short of a long-term randomised clinical trial, that metformin may decrease the risk of dementia". He flagged the mechanism as unresolved, and some prior studies have gone the other way, suggesting possible harm via metformin-linked vitamin B12 deficiency. A 2022 meta-analysis found lower dementia risk on average, and a 2023 commentary in Nature Reviews Neurology treated metformin and dementia as an open question.
The trial that would actually settle causation, the NIH-funded MAP study (NCT04098666), has not yet reported results.