Meta-analysis confirms neuroprotective benefit of metformin

Key takeaways

Metformin may be associated with a lower risk of neurodegenerative disease, in particular when the drug has been prescribed for at least 4 years, according to a systematic review and meta-analysis of longitudinal data.

However, the heterogeneity between the available studies and the potential heterogeneity of diagnostic criteria may mean that validation studies are needed.

Why is this important?

Data suggest that metformin, the most commonly prescribed antidiabetic drug, may be neuroprotective, while diabetes is associated with an excess risk of neurodegenerative disease. Results of studies conducted specifically to investigate the benefit of the antidiabetic drug on cognitive prognosis have been unclear. A meta-analysis was published in 2020, but it included cross-sectional and case-control studies. Given the long observation period needed to measure such an outcome, only cohort studies conducted over several years can provide reliable results. This new meta-analysis attempts to circumvent this limitation.

Methods

The meta-analysis was conducted using studies published up to March 2021 that met the inclusion criteria (population-based cohort studies published in English in which the administration of metformin and associated risk of exposure were reported).

Main results

Twelve studies were included in this analysis, of which eight were retrospective and 11 were considered to be of good methodologic quality. In total, 194,792 patients were included.

Pooled data showed that the relative risk associated with onset of neurodegenerative disease was 0.77 (95% CI, 0.67-0.88) for patients with diabetes taking metformin versus those not taking metformin. However, heterogeneity between studies was high (I2; 78.8%; P < .001).

The effect was greater with longer metformin use, with an RR of 0.29 (95% CI, 0.13-0.44) for those who took metformin for 4 years or more. Similarly, the studies conducted in Asian countries versus other locations suggested an added benefit for this population (RR, 0.69; 95% CI, 0.64-0.74).

Sensitivity analyses confirmed these results, and subtype analyses showed no difference according to the nature of the neurodegenerative disease.

A version of this article first appeared on Univadis.