Nine RCTs of GLP-1 analogues (n=3,184) and 20 RCTs of DPP4 inhibitors (n=9,812) were included.
Allocation concealment was clearly described in three studies, and small baseline differences between groups were noted in three studies. Most studies did not use a true intention-to-treat analysis. Participant withdrawal was approximately 19% in GLP-1 analogue studies and 18% in DPP4 inhibitor studies.
Both GLP-1 analogues and DPP4 inhibitors reduced HbA1C compared with placebo: the WMD was -0.97% (95% CI: -1.13, -0.81) for GLP-1 analogues and -0.74% (95% CI: -0.85, -0.62) for DPP4 inhibitors. GLP-1 analogues and insulin did not differ significantly for this outcome (based on two studies), while the comparison between DPP4 inhibitors and hypoglycaemic agents showed a small but statistically significant difference favouring the latter (WMD 0.21%, 95% CI: 0.02, 0.39; based on four studies). GLP-1 analogues were associated with weight loss compared with control interventions (WMD -2.37 kg, 95% CI: -3.95, -0.78), although heterogeneity was high for this outcome (I-squared 98%). DPP4 inhibitors were associated with a small but statistically significant increase in weight relative to placebo (WMD 0.48 kg, 95% CI: 0.30, 0.66). GLP-1 analogues were associated with increased risk of gastrointestinal adverse events relative to comparators (RR 2.92, 95% CI: 2.02, 4.24 for nausea; RR 3.32, 95% CI: 2.51, 4.41 for vomiting; RR 2.23, 95% CI: 1.72, 2.89 for diarrhoea), while DPP4 inhibitors were associated with increased risk of urinary tract infections (RR 1.52, 95% CI: 1.04, 2.21) and headache (RR 1.38, 95% CI: 1.10, 1.72). Other outcomes were reported.