Inclusion criteria for the review were broadly defined. Several relevant databases were searched without language restrictions. There was the potential for publication bias as the authors did not state if unpublished studies were searched for; publication bias was not assessed due to the small number of identified studies. The authors did not state how many reviewers performed study selection, and quality assessment, but did state that the CHERG guidelines were adhered to, which should minimise error and bias in the review.
The quality assessment indicated the moderate to high quality of the included studies, which strengthened the review. Trials were combined using meta-analysis; heterogeneity was assessed, which was appropriate. However, some of the studies were conducted over 50 years ago, thus the relevance of the review findings to the current situation was unclear.
This issue, together with some of the methodological problems within the review, necessitates caution when interpreting the authors' conclusions.