Abstract
BACKGROUND: Both glucagon-like peptide 1 receptor agonist (GLP1RA) and pioglitazone are associated with cardiovascular and hepatic benefits in patients with type 2 diabetes (T2D). However, studies that directly compare their cardio-hepatic effects are lacking. We emulated a target trial to compare their effects on adverse liver and cardiovascular outcomes in T2D patients. METHODS: We adopted an "active comparator, new user" design involving T2D patients newly prescribed GLP1RA or pioglitazone between January 2008 and December 2022. The primary outcomes were major adverse liver outcomes (MALO) and cardiovascular events (MACE), with their individual outcomes and heart failure (HF) as secondary outcomes. Cox proportional hazards models were used to estimate hazard ratios (HRs) via intention-to-treat (ITT) and per-protocol (PP) analyses. RESULTS: A total of 8922 patients (N = 4461 each group) were included. Compared to pioglitazone users, GLP1RA users had comparable risks of incident MALO (ITT: HR 0.94, 95% CI 0.66-1.34; PP: HR 1.13, 95% CI 0.60-2.15) and MACE (ITT: HR 0.99, 95% CI 0.80-1.22; PP: HR 1.10, 95% CI 0.77-1.58). The risk of HF was significantly lower in GLP1RA users in ITT analysis (HR 0.65, 95% CI 0.51-0.83). The results were consistent across most subgroup and sensitivity analyses. CONCLUSIONS: The effects on incident major adverse liver and cardiovascular outcomes were comparable between GLP1RA and pioglitazone, although the risk of incident HF was lower with GLP1RA. Treatment decisions for T2D patients at risk of adverse cardio-hepatic events should be individualized, considering HF risk and the need for weight loss, which if present, GLP1RA may be preferred.