Potential link between long-term metformin use and reduced dementia risk emerges from new study

Researchers tracking participants of the Diabetes Prevention Program have identified a possible association between prolonged metformin use and a lower incidence of dementia, though findings remain preliminary and inconclusive.

Researchers following participants in the Diabetes Prevention Program have found a possible link between long-term metformin use and a lower chance of dementia, although the authors say the overall cognitive picture remains largely unchanged. The preprint, posted on medRxiv on 7 August, examined adults who entered the late 1990s trial at high risk of type 2 diabetes and then tracked them for more than two decades. By 2022-24, when the surviving participants had a median age of 74, a masked expert panel had assigned 1,483 people to cognitive categories after neuropsychological testing, informant interviews, video neurological examination and blood markers associated with Alzheimer’s disease.

Across the three original trial groups, rates of cognitive impairment were broadly similar. But dementia itself appeared less common in the metformin group: 1.9% compared with 3.5% in the placebo arm and 4.2% in the intensive lifestyle arm. In relative terms, that translated into adjusted odds ratios of about 0.40 and 0.38, although the authors note that the finding rests on just 48 dementia cases overall, including nine in the metformin group. Before adjustment, the metformin-versus-placebo comparison was not statistically significant.

The same paper found only a modest signal on memory performance. Participants originally assigned to metformin scored slightly higher on a word-list recall test over five rounds of assessment, but the difference was small enough that the authors described it as minimal. Rates of mild cognitive impairment were not meaningfully different, and declines in memory and executive function were similar across all groups over time. That means the study does not suggest a broad slowing of brain ageing, only a narrow and uncertain hint of benefit for one outcome.

The result fits with earlier observational research that has repeatedly raised the possibility that metformin may be linked with lower dementia risk, particularly with longer exposure. A meta-analysis of 14 cohort studies covering more than 396,000 people reported a lower overall risk of dementia among metformin users, with the strongest association in those treated for at least four years. Other cohort studies in Taiwan, South Korea and elsewhere have reported similar dose-related patterns in older adults with diabetes. Even so, those studies cannot rule out confounding, and the new trial-based analysis is still too small to settle the question. The authors themselves frame the findings as a signal worth watching rather than a reason to start the drug for brain protection, especially since the trial involved people who were overweight and pre-diabetic rather than metabolically healthy.