The authors concluded that incretin-based treatments were effective in glycaemic control in patients with type 2 diabetes; they also offered other advantages (like weight loss with exenatide and liraglutide). Aspects of this review were well conducted, but it is difficult to assess the reliability of the findings without further information on the quality of the included trials.

To assess the clinical efficacy and safety of incretin-based medications in patients with type 2 diabetes.

EMBASE and the Cochrane Library were searched for English language publications from January 2000 to July 2009. Search terms were reported. Reference lists of relevant publications and abstracts from relevant conferences from August 2008 to July 2009 were also checked.

Randomised controlled trials (RCTs; phase II, phase III, parallel design) that compared exenatide, liraglutide, vildagliptin or sitagliptin with placebo in patients with type 2 diabetes were eligible for inclusion. Trials had to have a minimum of 100 participants (aged 18 years or over) and had to report a dispersion measure (standard deviation, standard error of the mean, or confidence interval) for both treatment arms. Trials were required to report on at least one of the following outcomes at baseline and post-treatment: change from baseline in glycated haemoglobin (HbA_Ic); change from baseline in weight; and/or body mass index. Patient-reported hypoglycaemic events were also evaluated.

In the included trials, the mean age of participants ranged from 50.9 to 62.9 years; the proportion of men ranged from 42 to 67%. The duration of diabetes diagnosis ranged from 1.7 to 9.0 years. Mean baseline glycated haemoglobin ranged from 6.7 to 9.4 across trials; mean baseline weight ranged from 61.5 to 104kg across trials (where reported). The trials ranged in duration from four to 52 weeks. Various doses for each treatment were reported.

Three reviewers were involved in the study selection, with any discrepancies resolved by consensus.

The quality of the trials was assessed using criteria on method of randomisation, intention-to-treat analysis, and blinding.

Three reviewers were involved in the assessment of study quality.

In each trial, mean and standard deviation were extracted to enable calculation of mean differences (MDs) and 95% confidence intervals (CIs).

Data were extracted by one reviewer and checked by a second reviewer.

A meta-analysis examined pooled weighted mean differences (WMDs) or relative risk (RR) and 95% CIs were performed using a fixed-effect model. When there was significant heterogeneity, a random-effects model was used. Heterogeneity was assessed using the X² test and I².

Meta-regression was conducted to assess the impact of major adjustment variables (age, study duration, duration of diabetes gender and treatment as monotherapy or in combination) on the pooled analyses.

Thirty-eight RCTs (n=16,502 participants) were included in the review. Thirty-four RCTs were double-blind.

Glycaemic control: All treatments were shown to produce a statistically significant reduction in glycated haemoglobin compared with placebo: vildagliptin (unadjusted WMD -0.67, 95% CI -0.83 to -0.52; 11 RCTs); sitagliptin (unadjusted WMD -0.79, 95% CI -0.93 to -0.65; 12 RCTs); exenatide (unadjusted WMD -0.75, 95% CI -0.83 to -0.67; eight RCTs); and liraglutide (unadjusted WMD -1.03, 95% CI -1.16 to -0.90; seven RCTs).

Weight: Vildagliptin (unadjusted WMD 0.56, 95% CI 0.27 to 0.84) and sitagliptin (unadjusted WMD 0.60, 95% CI 0.33 to 0.87) were both associated with a significant increase in weight when compared with placebo. Exenatide treatment was associated with significantly decreased weight when compared with placebo (unadjusted WMD -1.10, 95% CI -1.32 to -0.88). There was no significant difference between weight in liraglutide and placebo treated participants.

Hypoglycaemia: There was no significant difference between vildagliptin and placebo in the number of participants who reported hypoglycaemic episodes. Participants treated with sitagliptin (unadjusted RR 2.56, 95% CI 1.23 to 5.33), exenatide (unadjusted RR 2.40, 95% CI 1.39 to 4.11) and liraglutide (unadjusted RR 1.69, 95% CI 1.00 to 2.86) were more likely to experience hypoglycaemia compared with participants randomised to placebo.

There were a few instances where the covariates evaluated in meta-regression had a statistically significant impact on the outcomes.

Incretin-based treatments were effective in glycaemic control and offered other advantages (like weight loss with exenatide and liraglutide) which may have an important impact on patient adherence to medication.

The review addressed a clear question and was supported by appropriate inclusion criteria. Attempts to identify relevant studies were undertaken by searching relevant databases and other sources, including conference abstracts, which minimised the potential for publication bias. The search was limited to English language publications, which introduced the potential for language bias. Three reviewers were involved in the review process, which reduced the potential for reviewer bias and error.

Aspects of trial quality were assessed, but they were not reported or considered in the analyses. Trial details, participant characteristics and outcomes were reported in detail. Appropriate methods were used to pool the results and to investigate heterogeneity.

Aspects of this review were well conducted, but it is difficult to assess the reliability of the findings without further information on the quality of the included trials.

The authors did not state any implications for practice or research.

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.