Background/Objectives: Type 2 diabetes, a widely recognized risk factor of Alzheimer's disease (AD), is mainly due to the role played by insulin resistance on neuronal degeneration. Mild Cognitive Impairment (MCI) also significantly increases the risk of developing AD by approximately 40%. Epidemiological studies have demonstrated correlations between glucagon-like peptide-1 agonists (GLP-1RAs) on reducing the risk of AD. We synthesize current data from randomized controlled trials (RCTs) to summarize the effects of GLP-1RA treatment in adults with Type 2 diabetes or with MCI on cognitive performance. Methods: We searched databases from inception to 1 January 2026 investigating effects of GLP-1RA(s) on cognition. PubMed, Web of Science and Scopus databases were explored following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). We selected only RCTs in adults investigating GLP-1RA(s) on cognitive performance. Results: Overall mean age was 64.5 yrs with an average follow-up period of 28.4 months. A total of 13,711 participants from five RCTs were included. Following a source-data audit, four randomized controlled trials contributed analytically compatible data to at least one cognitive domain. Cognitive domains were analyzed separately using Hedges' g. Random-effects models used restricted maximum likelihood (REML) estimation with Knapp-Hartung inference when at least two compatible studies contributed. For global cognition, three studies yielded g = 0.014 (95% CI: -0.061 to 0.088; p = 0.515; I2 = 0%). For attention, two studies yielded g = 0.072 (95% CI: -2.164 to 2.308; p = 0.753; I2 = 31.1%). For delayed memory, executive function, and verbal function, only one compatible study was available per domain; these were therefore reported as single-study estimates rather than pooled effects. A sensitivity analysis excluding the dominant Cukierman trial yielded an imprecise and non-significant effect estimate for global cognition (g = -0.205, 95% CI: -2.678 to 2.269) indicating that the robustness of the pooled result could not be established. Conclusions: Current randomized evidence is insufficient to establish a consistent beneficial effect of GLP-1RA(s) on cognitive function in adults at risk of AD. The available findings should be considered exploratory and hypothesis-generating because few analytically compatible trials were available within each cognitive domain.
