The review addressed a clear question with restrictive inclusion criteria and a search of just two databases for papers English. No attempts were reported to reduce the potential impact of language and publication biases and it seemed likely that relevant papers were missed. No details of the review processes were reported so we could not rule out reviewer error and bias.
The studies were partially described but with insufficient detail to permit assessment of their reliability. The narrative synthesis and tables did not provide data on any statistically significant differences. The reviewers highlighted a range of potential weaknesses in the primary studies.
Overall, the data as presented do not support the strong conclusions offered by the authors and these should be considered as unreliable.