Glucagon-like peptide-1 receptor agonists and risk of cancer: a systematic review and meta-analysis of randomized controlled trials

Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are established for glycaemic control and cardiovascular risk reduction in type 2 diabetes, with emerging use in obesity. However, evidence linking them to cancer risk remains inconclusive. PubMed, Embase, Web of Science, the Cochrane Library, and clinicaltrials.gov were searched (inception to 4 July 2025). Randomised controlled trials (RCTs) comparing glucagon-like peptide-1 receptor agonist therapy versus placebo or non-GLP-1RA drugs. 148 RCTs involving 168,875 patients (mean [SD] age, 56.4 (9.9) years; 95,558 [56.9%] men) were included. In pooled analysis, GLP-1RAs were not associated with overall cancer risk (RR, 0.99; 95% CI, 0.92–1.05). A nominal inverse association was observed for prostate cancer (RR, 0.79; 95% CI, 0.65–0.96; nominal P = 0.02; I² = 0%); however, this association did not survive correction for multiple testing across 25 cancer outcomes (FDR-adjusted P = 0.52). Subgroup analyse revealed that GLP-1 RA use was associated with lower risks of prostate cancer at participants with diabetes (95% CI 0.54, 0.87; P = 0.002, I 2 = 0%), higher doses (95% CI 0.46, 0.97; P = 0.03, I 2 = 0%), longer duration of use (95% CI 0.63, 0.95; P = 0.02, I 2 = 0%). Meta-regression analysis revealed that the type of control group was the only factor that significantly moderated the effect size ( P = 0.0295). Compared with active-controlled studies, GLP-1RAs were associated with a 56% lower risk of prostate cancer in placebo-controlled studies (RR = 0.44, 95% CI 0.21–0.92). Drug dose, trial duration, population, mean age, sample size, and baseline HbA1c level did not significantly moderate the association between GLP-1RAs and prostate cancer risk. This large-scale meta-analysis provides robust evidence that GLP-1 receptor agonists do not increase overall cancer risk. An exploratory signal of reduced prostate cancer incidence was observed in nominal analyses, but it was not confirmed after correction for multiplicity and should be regarded strictly as hypothesis-generating. These findings do not currently support altering clinical decision-making or drug selection for the purpose of cancer prevention. Prospero : CRD42024558497. Available from https://www.crd.york.ac.uk/PROSPERO/view/CRD42024558497 .