AIMS To compare the efficacy and safety of the addition of a dipeptidyl peptidase-4 (DPP-4) inhibitor or a placebo in patients with type 2 diabetes inadequately controlled with insulin.

METHODS We searched randomised controlled trials (RCTs) from MEDLINE, EMBASE, LILACS, the Cochrane Central Register of Controlled Trials and the ClinicalTrials.gov online registry. Studies of at least 12week treatment duration were eligible if they were RCTs in patients with type 2 diabetes comparing addition of a DPP-4 inhibitor to insulin therapy (INS/DPP4i) with addition of a placebo to insulin therapy (INS/PCB) and contained information on the change in glycated haemoglobin (HbA1c) from baseline.

RESULTS Of 3105 potentially relevant published articles and 206 registered trials, 9 studies were included for meta-analysis. Compared to INS/PCB, INS/DPP4i exhibited a greater reduction in HbA1c (weighted mean difference [WMD] -0.58%; 95% CI -0.70, -0.46) and fasting plasma glucose (WMD -0.59mmol/L; 95% CI -0.79, -0.40) with less daily insulin doses (WMD -1.86IU; 95% CI -3.27, -0.45) and with no difference in weight gain (WMD -0.04kg; 95% CI -0.25, 0.16). The risk of hypoglycaemia was similar between INS/DPP-4i and INS/PCB (the RR in favour of INS/PCB was 0.94; 95% CI 0.84, 1.05).

CONCLUSIONS Compared to placebo, DPP-4 inhibitors exhibit a better glycaemic control without further increasing the risk of weight gain and hypoglycaemia in patients with type 2 diabetes inadequately controlled with insulin.