The review question and inclusion criteria were clear, although placebo was used in only one trial, despite being pre-specified in the review inclusion criteria. Only one database was searched, which meant relevant studies may have been missed. Search strategy details were not reported, so it was not possible to evaluate how comprehensive the PubMed search was. It was also possible that the review may have been subject to language and/or publication bias (relevant details were not reported). The authors did not report using methods which could minimise the risk of reviewer error and bias during the review process (such as independent duplicate study selection and data extraction).
Trial quality was not appraised, so it was not possible to fully evaluate the reliability of the trial evidence (although the lack of blinding of patients and care-providers may have resulted in performance bias affecting those trials). Appropriate methods were used to pool data and assess heterogeneity.
In light of the limitations in the conduct and reporting of the review, the authors' conclusions should not be considered as likely to be reliable.