The review question was clear in terms of the intervention and patients, and several relevant databases were searched. However, the exclusion of non-English papers might have led to some important data being omitted and unpublished studies were not specifically sought, therefore both language and publication bias may be present. Reporting of the review process was poor: the methods used to select the studies and extract the data were not reported, therefore it is unclear whether attempts were made to reduce error and bias. In addition, there was no assessment of the methodological quality of the included studies. The narrative synthesis was appropriate given the clinical heterogeneity between the studies. The authors were only able to undertake a minimal synthesis, although the differences between studies that prevented a more formal, and extensive, analysis were discussed extensively. Given the lack of good-quality studies, the potential for missing studies and the lack of reporting of the review process, the reliability of the results of the review may be compromised. However, in light of these shortcomings, the conclusion was suitably cautious.