The review addressed a clear question supported by appropriate inclusion criteria (although no definition of chronic use was provided). Numerous databases were searched and, although search terms were not reported, the search was conducted with an expert reference librarian. It was unclear whether the authors took steps during the search to minimise the risk of language bias affecting the review. However, by screening, data extracting and quality assessing studies in-duplicate, the risk of reviewer error and bias was minimised. An assessment of study quality was made and was used in interpreting the results of the review.
Details of drug names and doses were not reported, which made it more difficult to appraise the included studies. Meta-analyses were conducted, but considerable clinical and statistical heterogeneity was seen but not then investigated (from the forest plot, one particular study seemed an obvious cause of the heterogeneity, but this was not explored or discussed).
Although this review had some shortcomings, the authors did acknowledge the limitations of the primary studies and made their conclusions appropriately cautious.