Aims Glucagon‐like peptide‐1 (GLP‐1) agonists and dipeptidyl peptidase‐4 (DPP‐4) inhibitors are both incretin‐based therapies for type 2 diabetes (T2DM) but have distinct efficacy and side effect profiles. We thus performed a systematic review and meta‐analysis to compare the effects of GLP‐1 agonists to DPP‐4 inhibitors on glycaemic control, weight and incidence of adverse events in adults with T2DM. We also sought to determine whether there was any additional effect in switching from DPP‐4 inhibitor to GLP‐1 agonist. Materials and methods We systematically searched PubMed, Embase and ClinicalTrials.gov for (1) randomized controlled trials (RCTs) comparing any GLP‐1 agonist to any DPP‐4 inhibitor and (2) interventional studies where a DPP‐4 inhibitor was switched to a GLP‐1 agonist. We assessed pooled data using random‐effects model (CRD42017057115). Results The pooled analysis of 13 RCTs (n = 4330) showed that, compared to DPP‐4 inhibitors, GLP‐1 agonists yielded a greater mean reduction in glycated haemoglobin (HbA1c) of −0.41% (95% CI −0.53 to −0.30) and in weight of −2.15 kg (−3.04 to −1.27). GLP‐1 agonists were associated with greater likelihood of gastrointestinal side effects with no increased risk of hypoglycaemia. In 5 interventional studies (n = 433), switching from DPP‐4 inhibitor to GLP‐1 agonist yielded further mean reduction in HbA1c of −0.69% (−1.03 to −0.35) and in weight of −2.25 kg (−3.12 to −1.38). Conclusions GLP‐1 agonists yield greater reduction in HbA1c and weight as compared to DPP‐4 inhibitors, with increased incidence of gastrointestinal symptoms but not hypoglycaemia. Replacing a DPP‐4 inhibitor with GLP‐1 agonist provides additional benefits in glycaemic control and weight loss.