Introduction: Diabetes mellitus (DM) is a common comorbidity of cirrhosis, but there is limited research on the impact of dual diabetes therapy on mortality and hepatic decompensation in cirrhosis, particularly in non-alcoholic fatty liver disease (NAFLD).
Methods: We did propensity score-matched analyses of DM patients with cirrhosis comparing those on metformin with those on both metformin and an GLP1-RA. We queried for people with both Type 2 DM (DM2) and cirrhosis using ICD-10 codes on the TriNetX network. We collected patient demographics—age, sex, race, among others—and our primary outcome was mortality in three years. Our secondary outcome was a composite of hepatic decompensation events over three years.
Results: We identified 841 patients with cirrhosis and DM2 who were on both metformin and an SGLT2-i. This cohort included 547 Whites (65.0%), 373 women (44.4%), had a mean age of 63.8 years, and was matched with a baseline group of patients who were on metformin alone (Table 1). The monotherapy group had a greater mortality risk (RR 2.0, 95%CI 1.4-2.8, p< 0.0001) with survival probability 82.6% at 3 years compared to the dual therapy group (91.1%, p< 0.0001). We also identified a subset of 116 patients within the dual therapy group that was confirmed to have cirrhosis from non-alcoholic steatohepatitis (NASH) and that included 85 Whites (73.7%), 63 women (54.2%), and had a mean age of 60.8 years. When comparing the NASH cohorts, they had equivalent mortality risk (RR 1.0, 95%CI 0.4-2.3, p=1), though survival at 3 years was shorter for the monotherapy group (90.9% vs. 97.1%, p=0.04). The composite risk for hepatic decompensation was equivalent between both therapy groups (Figure 1).
Discussion: We found a potential mortality benefit in cirrhosis patients on dual DM2 therapy with metformin and SGLT2-i compared to metformin alone that was durable over a 3-years and within a subset of NASH patients. This finding was consistent for men, women, White, and non-White patients in our matched cohort analyses. There was no clear effect on the composite risk for hepatic decompensation, so improved mortality maybe be due to beneficial effects on the patients’ comorbidities rather than on their liver issues. There may be underreporting of NASH patients in the database since our analysis relied in part on patients’ having ICD-10 codes (K76.0, K75.8). Further prospective studies in patients with biopsy-confirmed NASH and from underrepresented populations are needed to investigate our outcomes.
