Twenty-five RCTs with 10,560 participants (range 49 to 1,091) were included. Thirteen trials assessed exenatide twice daily, eight assessed liraglutide, and four assessed exenatide once weekly. Three trials compared exenatide twice daily with liraglutide or exenatide once weekly. Randomisation and allocation concealment were considered adequate in all trials.
Glucagon-like peptide-1 receptor agonist treatment was associated with a statistically significant 2.9 kg improvement in weight loss compared with controls (weighted mean difference 2.90, 95% CI -3.59 to -2.22; 21 RCTs, 6,411 participants). Statistically significant heterogeneity was present, but Egger's test did not reveal significant bias or small study effects. Subgroup analyses showed a statistically significant effect on weight loss for patients with and without diabetes and for all three glucagon-like peptide-1 receptor agonist regimens. Differences between regimens were not significant. Body mass index at baseline and trial duration did not significantly predict treatment effect. Sequential analysis indicated that the evidence was sufficient to confirm the intervention effect after adjusting for multiple testing and random error.
Glucagon-like peptide-1 receptor agonists had beneficial effects on blood pressure, plasma cholesterol and measures of glycaemic control, but did not significantly affect plasma concentrations of liver enzymes. Compared with controls, glucagon-like peptide-1 receptor agonists were associated with significant increases in risk of nausea, vomiting and diarrhoea, but not with hypoglycaemia.