The review question and inclusion criteria were clear. Relevant databases were searched. Attempts were made to search for unpublished studies. Appropriate steps were taken to minimise reviewer bias and error in study selection and data extraction, but was not reported for validity assessment.
Quality assessment appeared to have been performed using appropriate criteria. Trial results appear to have been pooled using appropriate methods, but reported results (text versus meta-analysis) were inconsistent, so it was difficult to confirm their reliability. Statistical heterogeneity was assessed, but the methods used were incompletely reported. The authors acknowledged a number of limitations including differences in the interventions, and a small number of trials with small sample sizes.
The reliability of the authors' conclusion is unclear given the potential for reviewer error and bias in the review methods, small number of included trials with small sample sizes, reporting errors, and incomplete reporting of review methods.