Chelation therapy for coronary heart disease: an overview of all clinical investigations

CRD COMMENTARY: This review provided a short summary of the controlled and uncontrolled studies, and answered the question posed by the author at the outset. The inclusion of uncontrolled trials was decided after the search had been completed, but in the light of the lack of other data the inclusion appears reasonable. Some of the limitations of these studies were made clear, and it was sensible not to attempt to combine the results of controlled and uncontrolled studies. The aim of the study was to review effectiveness and efficacy, but the author made no distinction between these two measures throughout the paper. The search strategy included searching of the main scientific electronic databases and one database specific to complementary therapy, without any language restrictions. The choice of keywords used may have resulted in some studies being excluded; in particular, 'vascular disease' was the only method used to identify studies of CHD and 'controlled clinical trials' as the only way of identifying clinical trials. Furthermore, it was unclear if the search strategy was modified after the decision was made to include all evaluations, including uncontrolled studies and case reports. The keywords do not appear to have been linked to MeSH, and this may have further limited the studies found. Other ways of identifying relevant studies were appropriate. No attempt seems to have been made to identify unpublished studies. No data were given on the number of studies identified in the initial search but excluded from the review. There were no tests for publication bias. The details of individual studies were not adequately reported (e.g. treatment regimen, what counts as improved symptoms) nor considered in the narrative reports. A table referenced in the text was missing from the publication. It was unclear whether the single author made all decisions regarding the inclusion of studies. Validity assessments were absent. Many of the papers included were published by the same authors (22 studies but 14 authors), and the review author made no mention of whether these were reports of the same study or not. In the narrative report, the uncontrolled studies were not grouped by treatment regimen. Information on basic patient characteristics including age, gender and source of recruitment was missing. The author highlighted the studies using angiographic outcomes, but did not mention in the text that they each only included one patient. The two controlled studies were well described in the narrative. The author's conclusions are appropriate in view of the types of studies included in this review.