525 Background: Type 2 diabetes mellitus (T2DM) is associated with elevated risk of hepatocellular carcinoma. Insulin use may exacerbate hepatic carcinogenesis, whereas glucagon-like peptide-1 receptor agonists (GLP-1 RAs) have demonstrated favorable metabolic and hepatic effects. We investigated the association between GLP-1 receptor agonist initiation and the incidence of liver cancer. Methods: We conducted a retrospective, propensity score–matched cohort study using the TriNetX US Collaborative Network (72 healthcare organizations). Adults ≥18 years with T2DM who initiated a GLP-1 RA without insulin were compared with those initiating insulin without GLP-1 RAs between January 1, 2017, and September 5, 2025. Patients with prior GLP-1 RA, insulin exposure, or liver cancer were excluded. The outcome was incident liver cancer (ICD-10 C22), assessed beginning 181 days after treatment initiation. Matching balanced demographics, diabetes complications, substance use, liver disease, and healthcare utilization. Results: After 1:1 matching, 318,970 patients remained in each group. Following exclusion of those with prior liver cancer, 318,888 GLP-1 RA initiators and 318,396 insulin initiators were analyzed. Incident liver cancer occurred in 163 GLP-1 RA users (0.051%) and 773 insulin users (0.243%), yielding an absolute risk difference of –0.19% (≈1.9 fewer cases per 1,000), risk ratio 0.211 (95% CI 0.178–0.249), and odds ratio 0.210 (95% CI 0.177–0.249; p<0.001). Kaplan–Meier analysis showed survival probabilities at the end of follow-up of 99.75% (GLP-1 RA) versus 99.19% (insulin), with a hazard ratio of 0.306 (95% CI 0.258–0.363; log-rank p<0.001). Conclusions: Among adults with T2DM, initiation of GLP-1 RAs without insulin was associated with a 70–80% lower relative risk of incident liver cancer compared with insulin initiation, though the absolute risk reduction was small. These findings suggest a potential hepatoprotective effect of GLP-1 RAs, warranting confirmation in prospective studies.
Syed et al. (Sat,) studied this question.