Abstract
Background. To determine the accuracy of transvaginal ultrasonography, sonohysterography and diagnostic hysteroscopy for the investigation of abnormal uterine bleeding in premenopausal women. Design. Systematic review of common diagnostic imaging tests. Data sources. Relevant papers were identified through electronic searching of MEDLINE (1980 to July 2001) and EMBASE (1980 to July 2001) and manual searching of a bibliography of primary and review articles. Review methods. Studies were selected if accuracy of transvaginal ultrasonography, sonohysterography and diagnostic hysteroscopy was compared with a reference standard and included data that could be abstracted into a two-by-two table in order to calculate sensitivity and specificity. Quality assessment and data extraction were performed by at least two independent reviewers. Diagnostic accuracy was determined by calculating positive and negative likelihood ratios for all intrauterine pathologies, submucous fibroids and endometrial hyperplasia. Results. Nineteen studies met the inclusion criteria. Statistically significant heterogenenity was present between the likelihood ratios for studies of transvaginal ultrasound. A positive test result with sonohysterography diagnosed submucous fibroids with a pooled likelihood ratio of 29.7 (17.8, 49.6). A positive test result with hysteroscopy diagnosed submucous fibroids with a pooled likelihood ratio of 29.4 (13.4, 65.3), and any intrauterine pathology with a pooled likelihood ratio of 7.7 (4.3, 13.7). A negative test result with hysteroscopy for diagnosing any intrauterine pathology had a pooled likelihood ratio of 0.07 (0.04, 0.15). Conclusion. Although there was considerable variability present between the studies, all three diagnostic tests were moderately accurate in detecting intrauterine pathology. However, sonohysterography and hysteroscopy performed better than transvaginal ultrasound in detecting submucous fibroids.