Endometrial sampling for postmenopausal bleeding. Should we put the sampling tools away?

critical appraisal ❖ évaluation critique Endovaginal ultrasound to exclude endometrial cancer and other endometrial abnormalities. Should we still be screening patients with post-menopausal bleeding using endometrial sampling now that endovaginal ultrasound (EVUS) is available? Type of article and design Meta-analysis by two reviewers. Family physicians often see patients with post-menopausal bleeding. Researchers estimate that 3500 women in Canada 1 and 36 100 women in the United States 2 will be newly diagnosed with endometrial cancer this year. Until endometrial biopsy became available and was tested in the 1980s, diagnosis was made by uterine dila-tion and curettage (D & C). Endometrial biopsy was found to be less accurate than D & C in 20% of cases if the lesion was focal 3 and was associated with pain. 4 The false-negative rate for endometrial biopsy was 2% to 6%. Many patients star ted on hormone replacement therapy (HRT) will have abnormal bleeding patterns and must be screened for endometrial cancer. 5 Endovaginal ultrasound, a newer method of screening these patients, causes less pain than biopsy and will often allow you to reassure patients without obtaining tissue by sampling or D & C. 6 This article reviews the sensitivity of EVUS for screening these patients. The study reviewed the literature from 1966 to 1996. Only prospective studies that evaluated EVUS before endometrial tissue was obtained were selected independently and reviewed by two reviewers. The reviewers excluded studies that were retrospective, that pooled premenopausal and postmenopausal patients, and that measured endometrial tissue thickness after biopsy. Thirty-five studies met the inclusion criteria; 14 were non-English language. Three outcomes were considered in the pooled data: cancer, benign endometrial abnormalities (atypical and complex hyperplasia and polyps), and normal. The HRT status of subjects was recorded. Cases with cervi-cal disease were excluded from data review. The authors abstracted and recorded the number of true-positive, false-positive, true-negative, and false-negative cases using reported thickness measurements of 3 to 10 mm. For each study reviewed, sensitivity, specificity, and exact 95% confidence intervals (CI) were calculated for all EVUS measurements. The authors calculated mean weighted pooled estimates of sensitivity and specificity for each threshold, for any endometrial disease, and for cancer alone. Results Endovaginal ultrasound was better at detecting cancer than it was at detecting polyps or hyperplasia. Mean endometrial thickness was 4 mm for women with normal histology, 10 mm for women with endometrial polyps, 14 mm for women with hyperplasia, …