Effect of Glucagon-Like-Peptide-1 Receptor Agonists (GLP-1 RA) on Neuropsychiatric Outcomes: A Systematic Review and...

Type 2 diabetes mellitus (T2D) is associated with an increased burden of neuropsychiatric disorders, including cognitive decline, affective disorders, and substance use disorders. Glucagon-like peptide-1 receptor agonists (GLP-1 RAs), developed for glycemic control and weight reduction, have been hypothesized to confer neuroprotective and psychotropic benefits via central and peripheral mechanisms. However, evidence across individual studies remains inconsistent.

We conducted a systematic review and meta-analysis (PROSPERO CRD420251025534) of randomized controlled trials (RCTs) and observational studies assessing the effect of GLP-1 RAs on neuropsychiatric outcomes. Searches were performed in MEDLINE, SCOPUS, CENTRAL, and Web of Science up to 30^th October 2025. Eligible comparators included placebo, standard care, or alternative glucose-lowering agents. Data were synthesized using random-effects models, and risk of bias was assessed using RoB2 and ROBINS-I. Evidence certainty was graded using GRADE methodology.

From 10,037 records, 82 studies were included. GLP-1 RAs were associated with a reduced risk of idiopathic Parkinson's disease (PD) (pooled HR 0.70, 95% CI: 0.53-0.92, I² = 38%) but no improvements in motor symptoms in established PD (MDS-UPDRS Part III off-medication: mean difference -3.29, 95% CI: -7.84, 1.26, I² = 80%). No consistent improvements in non-motor symptoms, quality of life, or dyskinesia scores were seen. GLP-1 RAs were associated with a reduced risk of cannabis use disorder (pooled HR 0.55, 95% CI: 0.39-0.77, I² = 71%) but showed no significant effect on opioid use disorder (pooled HR 0.61, 95% CI: 0.24-1.52, I² = 91%). Liraglutide reduced binge frequency in binge eating disorder (mean difference -1.28, 95% CI: -1.67, -0.89, I² = 53%) compared to placebo. In observational studies versus non-user controls, GLP-1 RA use was associated with a lower reported risk of suicidality (pooled OR 0.34, 95% CI: 0.18-0.63, I² = 99%), but no such association was seen with Semaglutide alone (pooled OR 0.71, 95% CI: 0.29-1.73, I² = 99%), active-comparator studies (pooled HR 0.96, 95% CI: 0.81-1.15, I²=0%) or RCTs (pooled RR 1.13, 95% CI: 0.53-2.41, I²=0%). No associations were observed for depression or anxiety. For dementia, findings were mixed: no benefit versus SGLT2 inhibitors (pooled RR 1.07, 95% CI: 0.73-1.56, I² = 78%), but there was a reduced risk of Alzheimer's disease when compared to unmatched or non-exposed controls (pooled RR 0.37, 95% CI: 0.14-0.96, I² = 98%). Certainty of evidence ranged from moderate to very low.

Current evidence, largely of low to very low certainty, suggests potential selective neuropsychiatric associations of GLP-1 RAs, particularly in Parkinson's disease risk reduction and binge eating disorder. Benefits for dementia risk reduction were observed primarily in comparisons with non-exposed or DPP4i, whereas no advantage was seen versus other active comparators, such as SGLT2 inhibitors. Our findings should be interpreted as hypothesis-generating due to significant heterogeneity and wide confidence intervals, indicating imprecision. Longer-term studies with neuropsychiatric outcomes as the primary endpoint are required.