Page 134 - The diagnostic work-up of women with postmenopausal bleeding
P. 134
Chapter 7
sampling. We included observational studies on the evaluation of the diagnostic accuracy of endometrial sampling in women with PMB. References cited in the selected articles were checked for further relevant articles not identified by the electronic searches.The search strategy can be found in the Appendix.
Selection criteria
This review focused on diagnostic studies in which the histology results of endometrial sampling were compared with the results of a reference standard.The articles had to study women with postmenopausal uterine bleeding, the diagnostic test of interest was endometrial sampling (histology), the reference standard had to be endometrial histological findings from (blind) D&C, diagnostic hysteroscopy with histology by targeted biopsy or D&C or hysterectomy.
Identified articles were merged into a common file,duplicates were deleted,and results were divided between two reviewers (NvH and MMP) who independently examined the assigned articles and classified each as“exclude”,“include”,or“unsure.” Initial screening began with a title screen. Subsequently, abstracts were retrieved and screened to determine eligibility. Finally, full text articles were retrieved and screened for inclusion. A third reviewer (MB) settled discrepancies. For ar ticles, which included both pre- and postmenopausal women, but did not report separately on the postmenopausal group, we sent an email to the corresponding author to ask for the data on postmenopausal women. For articles which were published before 1997 and therefore no email address of the corresponding author was mentioned, we searched the internet (Google, PubMed) for an email address to contact the corresponding author. We calculated the agreement on the selection of studies between the reviewers.
Quality assessment
Two reviewers (NvH and MMP) independently assessed the methodological quality of each selected paper using the QUADAS-2 tool for diagnostic studies, modified to conform to this review.11 Disagreements were resolved via consensus and if necessary via a third reviewer (MB).
We decided a priori the criteria of each study for low risk of bias in each of the four main domains of the Quadas-2 tool: patient selection, index test, reference standard, and flow and timing11. For patient selection, the in- and exclusion criteria had to be clearly stated, and the patient sample had to be consecutive. For the index test, the independent assessment of the pathologist for endometrial sampling
132