Lower risks of cirrhosis and hepatocellular carcinoma with GLP‐1RAs in type 2 diabetes: A nationwide cohort study using target trial emulation framework

Abstract Background To assess the association of cirrhosis and hepatocellular carcinoma (HCC) with the use of glucagon‐like peptide‐1 receptor agonists (GLP‐1RAs) versus long‐acting insulins (LAIs), which are the two commonly prescribed injectable glucose‐lowering agents (GLAs) for patients with type 2 diabetes (T2D) after the failure of multiple oral GLAs. Methods We emulated a target trial using the nationwide data of a Taiwanese cohort with T2D. Incident new users of GLP‐1RAs and LAIs during 2013–2018 were identified, and propensity score (PS) matching was applied to ensure between‐group comparability in baseline patient characteristics. The primary outcome was the composite liver disease including cirrhosis or HCC. Each patient was followed until the occurrence of a study outcome, death, or the end of 2019, whichever came first. Subdistribution hazard models were employed to assess the treatment‐outcome association. Sensitivity (e.g., stabilized inverse probability of treatment weighting analysis, time‐dependent analysis), E ‐value, and negative control outcome analyses were performed to examine the robustness of study findings. Results We included 7171 PS‐matched pairs of GLP‐1RA and LAI users with no significant between‐group differences at baseline. Compared with LAIs, the use of GLP‐1RAs was associated with significantly reduced risks of composite liver disease (subdistribution hazard ratio [95% confidence interval]: 0.56 [0.42–0.76]), cirrhosis (0.59 [0.43–0.81]), and HCC (0.47 [0.24–0.93]). Results were consistent across sensitivity analyses and among patients with different baseline characteristics. Conclusion Among T2D patients who require injectable GLAs, the use of GLP‐1RAs versus LAIs was associated with lower risks of cirrhosis and HCC.