The review question was clear and appeared to be supported by well-defined selection criteria. A relatively limited number of bibliographic sources were consulted and language restrictions were applied to the searches so some studies may have been missed. Additional cases of perforations that were not deemed to be sufficiently well reported to be included in the review were presented briefly. Contacting the authors of these studies may have been appropriate to confirm ineligibility. It was unclear whether data were extracted in duplicate to reduce reviewer error and bias during data extraction.
The quality of the included studies was not assessed formally. All data came from generally small retrospective case reports which suggested that the quality of the evidence base was suboptimal. The choice of a narrative synthesis was appropriate in view of the limited number of events and types of studies included.
The conclusions reflected the evidence presented. However, concerns about several aspects of the review process and the poor quality of the evidence mean that their reliability is unclear.