Glucagon-like peptide-1 receptor agonist (GLP-1 RA) and hematologic cancer risk among older adults with type 2 diabetes.

6592 Background: Recent preclinical and clinical evidence suggests that glucagon-like peptide-1 receptor agonists (GLP-1 RA) may reduce cancer risk through inflammation modulation. However, comparative effectiveness research evaluating hematologic malignancy risks of GLP-1 RAs remain limited in real-world populations. Methods: Using 15% random samples of Medicare Part D beneficiaries from 2013 to 2022, we identified older adults with type 2 diabetes who initiated GLP-1 RA or DPP-4i. The primary outcomes included myelodysplastic syndromes (MDS), lymphoma, and leukemia. We performed 1:1 propensity score matching for confounding control. We estimated the hazard ratios (HR) and 95% confidence intervals (CI) with Cox proportional hazard model. We examined the heterogeneity in treatment effects by pre-specified patient demographic and baseline factors. Subgroup analyses were conducted for: a) age (≥75 vs. <75 years), b) sex (female vs. male), c) race/ethnicity (non-Hispanic Black vs. non-Hispanic White vs. Hispanic ethnicity), d) obesity status (overweight vs. obese vs. severe obesity), and e) specific GLP-1 RA class (e.g., liraglutide, dulaglutide, and semaglutide). In the sensitivity analysis, we used cause-specific Cox model to account for death as a competing risk and calculated E-values to examine the potential influence of unmeasured confounders. Results: After propensity score matching, there were 29,212 patients included in the cohort (14,606 GLP-1 RA users and 14,606 DPP-4i users). Of the cohort, mean (SD) age was 77.3 (7.4) years; 53.9% were female, 69.9% were non-Hispanic White, 11.1% were non-Hispanic Black, and 3.9% were Hispanic. Compared to DPP4i, GLP-1 RA showed reduced risks in MDS (HR: 0.70, 95% CI [0.53-0.94]), lymphoma (0.71 [0.57-0.89]), and leukemia (0.85 [0.74-0.99]). Overall, no statistically significant heterogeneity of treatment effect was detected in subgroup analyses. After accounting for death as a competing event, the hazard ratios are consistent with the conventional Cox models except monocytic leukemia (HR: 1.4 [0.45-4.56], p=0.48). E-values suggest moderate robustness to potential unmeasured confounding. Conclusions: GLP-1 RA initiation was associated with a reduced risk of hematologic malignancies compared with DPP-4i use among real-world patients with type 2 diabetes. These findings suggest a potential protective effect of GLP-1 RAs on hematologic cancer risk and warrant further mechanistic and prospective investigation.