The review question and inclusion criteria were clear. Relevant databases were searched, but the authors did not appear to have made attempts to locate unpublished data. The authors used limited search terms which may have omitted potentially relevant studies. No language restrictions were applied, which reduced potential for language bias. Two authors were involved in study selection, data extraction and quality assessment, which minimised potential for error and bias. It was unclear why studies with a sample size less than 40 were excluded, but this would not necessarily alter the conclusions of the review. Inclusion of only full-length published trials risked publication bias.
Quality assessment using the Cochrane Risk of Bias tool indicated the generally good quality of the included trials. Trials were pooled by meta-analysis. Statistical heterogeneity was assessed and the authors acknowledged subjective outcome assessment as an issue that could have contributed to heterogeneity in the meta-analyses.
Although there were some limitations (limited search terms and unnecessary exclusion of potentially relevant studies), the authors' conclusions reflected the evidence presented and are likely to be reliable.