The review addressed a clear question and was supported by appropriate inclusion criteria. Attempts to identify relevant studies were undertaken by searching electronic databases and checking references, but the restriction to searching only for studies published in English meant some relevant studies may have been missed. Although it appeared generally that suitable methods were used to minimise the risk of reviewer error and bias throughout the review, the description of the processes used for quality assessment and data extraction could have been clearer.
Study quality was assessed and the assessment used in interpreting the results of the review, but lack of details about the criteria being assessed made interpretation of quality scores difficult; all studies scored highly, but the sample sizes of all included studies appeared small. In light of the clinical heterogeneity between studies the authors appropriately conducted a narrative synthesis. The risk ratios and standard mean differences calculated by the authors were not referred to in the synthesis and instead presented in tables without confidence intervals or p values, which meant it was difficult to interpret their purpose and the significance of the results.
Considering the review limitations and the small sample sizes of included studies, the authors' conclusions should be interpreted with a degree of caution.