LAUSR

This study (registered with PROSPERO, CRD42018085967) compares the efficacy (i.e. pro‐cognitive effects) and acceptability of antidiabetic agents for Alzheimer's disease (AD) and mild cognitive impairment (MCI). Cochrane Library (CENTRAL), PubMed/MEDLINE, EMBASE and PsycINFO were searched from inception to January 15, 2018 for randomized controlled trials comparing antidiabetic agents with placebo and/or another active antidiabetic agent for the treatment of AD or MCI. Nineteen eligible studies (n = 4855) evaluating the effects of 6 different antidiabetic drugs (i.e. intranasal insulin, pioglitazone, rosiglitazone, metformin, sitagliptin and liraglutide) were included. The results of 29 pairwise comparisons indicated that cognition was significantly improved in subjects treated with antidiabetic agents compared with placebo. Pioglitazone 15 to 30 mg demonstrated the greatest efficacy compared to placebo in network meta‐analysis. No significant differences in acceptability were identified when comparing agents with each other and with placebo. The current findings indicate a pro‐cognitive class effect of antidiabetic agents in AD/MCI. Other antidiabetic agents should also be investigated in future studies.