Meta-analysis of the association between new hypoglycemic agents and digestive diseases

Abstract

Background: New hypoglycemic agents include sodium-glucose cotransporter-2 inhibitors (SGLT2is), glucagon-like peptide 1 receptor agonists (GLP1RAs), and dipeptidyl peptidase-4 inhibitors (DPP4is). The association between each class of these new hypoglycemic drugs and the risks of various digestive system diseases is unknown. We aimed to explore this relationship by performing a meta-analysis.

Methods: We included large randomized trials of SGLT2is, GLP1RAs, and DPP4is. Outcomes of interest were 91 kinds of digestive diseases including 75 kinds of gastrointestinal disorders and 16 kinds of hepatobiliary disorders. Meta-analysis was done to generate pooled risk ratio (RR) and 95% confidence interval (CI). Subgroup analysis was conducted according to 3 different drug classes.

Results: We included 21 large trials in this meta-analysis. Compared with placebo, GLP1RAs were associated with the higher risks of gastric ulcer hemorrhage (RR 2.68, 95% CI 1.07-6.68; Pdrug = .035; I2 = 0), pancreatitis (RR 1.48, 95% CI 1.02-2.15; Pdrug = .041; I2 = 0), cholangitis acute (RR 5.96, 95% CI 1.04-34.08; Pdrug = .045; I2 = 0), and cholecystitis acute (RR 1.52, 95% CI 1.08-2.15; Pdrug = .017; I2 = 1.5%), but were not significantly associated with the occurrences of the other 87 kinds of digestive diseases (Pdrug ranged from .064 to .999). SGLT2is versus placebo were not significantly associated with the occurrences of 91 kinds of digestive diseases (Pdrug ranged from .077 to .995). DPP4is versus placebo were not significantly associated with the occurrences of 91 kinds of digestive diseases (Pdrug ranged from .085 to .999).

Conclusions: Neither SGLT2is nor DPP4is are associated with the occurrences of various kinds of digestive diseases, whereas GLP1RAs are associated with the higher risks of 4 kinds of digestive diseases, namely, gastric ulcer hemorrhage, pancreatitis, cholangitis acute, and cholecystitis acute. These findings seem to suggest that GLP1RAs are not applicable for patients at high risk of 4 specific digestive diseases, whereas SGLT2is and DPP4is are safe for patients susceptible to digestive diseases. However, our findings require to be further verified by future studies with sufficient statistical power.