Bayesian network meta-analysis evaluates antidiabetic medications and dementia risk in 4,382,897 patients.

This Bayesian network meta-analysis synthesized evidence from 28 articles encompassing 4,382,897 patients with diabetes to evaluate the association between antidiabetic medications and dementia risk. The review examined insulin, metformin, pioglitazone, DPP4i, SGLT_2i, and GLP1 against placebo or other comparators. Primary outcomes focused on dementia incidence, with secondary outcomes including Alzheimer's dementia and vascular dementia. Follow-up duration was not reported in the source material.

Regarding dementia incidence versus placebo, insulin demonstrated a reduction with an odds ratio of 0.11 and a 95% CrI of 0.1 to 0.12. Metformin also reduced incidence, showing an odds ratio of 0.79 with a 95% CrI of 0.77 to 0.81. Pioglitazone similarly reduced incidence, yielding an odds ratio of 0.69 and a 95% CrI of 0.56 to 0.86. These findings suggest potential protective associations for these specific agents within the analyzed dataset.

In comparisons involving Alzheimer's dementia incidence versus SGLT_2i, DPP4i had higher incidence with an odds ratio of 1.78 and a 95% CrI of 1.66 to 1.91. For vascular dementia incidence versus SGLT_2i, DPP4i also had higher incidence, resulting in an odds ratio of 2.59 and a 95% CrI of 2.33 to 2.88. These relative increases highlight differences in risk profiles among the drug classes evaluated.

Safety data, including adverse events and discontinuations, were not reported in the abstract. The study is a meta-analysis, not a primary trial, which limits causal inference. Absolute numbers were not reported, preventing calculation of absolute risk reduction. Clinicians should interpret these pooled effect sizes cautiously given the lack of safety reporting and follow-up duration details.