Abstract
AIMS: This study compares the therapeutic efficacy, gastrointestinal (GI) adverse event (AE) rates and the relationship between the therapeutic efficacy and GI AEs in randomized, placebo-controlled clinical trials (RCTs) of GLP-1 RAs and the dual GLP-1/GIP agonist tirzepatide. MATERIALS AND METHODS: A systematic PubMed search identified 38 phase 3 or 4 placebo-controlled RCTs of exenatide (b.i.d. and q.w.), lixisenatide, liraglutide, dulaglutide, albiglutide, semaglutide (s.c. and oral) and tirzepatide with a total of 16 660 individuals with type 2 diabetes (T2D) across 104 study arms. Changes in HbA1c, fasting plasma glucose and body weight and the proportion of GI AEs (nausea, vomiting or diarrhoea) were calculated by agent, preparation and dose. The correlation between odds ratios (ORs) for GI AEs and the magnitude of therapeutic efficacy was assessed in a linear regression analysis. RESULTS: Baseline characteristics were similar across studies: mean age 57 ± 10 years, diabetes duration 8 ± 6 years, body mass index (BMI) 31.9 ± 5.8 kg/m(2) and HbA1c 8.2% ± 0.9%. HbA1c reductions ranged from -0.63% ± 0.03% (lixisenatide, 20 μg q.d.) to -1.79% ± 0.09% (tirzepatide, 15 mg q.w.; p < 0.0001). Weight reductions ranged from -0.75 ± 0.10 kg to -9.65 ± 0.56 kg. Despite the high variability in therapeutic efficacy, ORs for GI AEs were similar across compounds/preparations. CONCLUSIONS: The magnitude of efficacy for intended therapeutic actions (HbA1c and body weight reduction) varied widely between incretin mimetic glucose-lowering agents. However, larger therapeutic efficacy was not systematically associated with higher GI AE or drug discontinuation rates, indicating better tolerability of the more effective agents/preparations.