Glucagon-like peptide-1 receptor agonists (GLP-1RAs) are widely used for type 2 diabetes and obesity, but their long-term cancer risks remain controversial in observational studies. We conducted a drug-target Mendelian randomization (MR) analysis using genetic variants in the GLP1R gene region as instruments to proxy GLP-1RA exposure. Summary-level data for 21 cancer types were obtained from public GWAS consortia. Inverse-variance weighted (IVW) MR was used as the primary method, supplemented by sensitivity analyses, Cox proportional hazards regression analyses, and mediation testing for BMI and HbA1c. Genetically proxied GLP-1RA use was associated with a reduced risk of breast cancer [the log of the odds ratio (log{OR}) = -0.343, P = 6.46 × 10-4] and prostate cancer [log(OR) = -0.004, P = 2.25 × 10-4], but an increased risk of head and neck cancer [log(OR) = 0.350, P = 1.17 × 10-4] and thyroid cancer [log(OR) = 1.016, P = 9.12 × 10-4]. The directions of these effects were validated by the Cox proportional hazards regression model using UK Biobank data and weighted polygenic scores for GLP1R expression, though the association did not reach significance (P > 0.05). Mediation analyses indicated that BMI explained about 3% of the causal effects on endometrial cancer, and HbA1c level explained 3% on head and neck cancer and 4% on thyroid cancer. This MR study provides evidence that GLP-1RA use may causally influence the risk of several cancers. These findings support the need for cancer-specific monitoring in patients using GLP-1RAs and further mechanistic research.
