Summary
Revista Brasileira de Ginecologia e Obstetrícia. 2003;25(9):667-672
DOI 10.1590/S0100-72032003000900008
PURPOSE: to compare the results of hysterosonography with those of hysteroscopy and the histopathologic study in postmenopausal women. METHODS: hysterosonography, hysteroscopy and endometrial biopsy were performed in 59 women who had an endometrial echo over 4 mm, age above 40 years and amenorrhea over one year, and whose follicle-stimulating hormone levels were over 35 mIU/mL. Patients using hormones were excluded, as well those in whom it was impossible to perform histerosonography, histeroscopy or endometrial biopsy. The statistical analysis was performed using the nonparametric "G"-Cochran and McNemar tests. In addition, sensitivity and specificity, as well as positive and negative predictive values were determined. The value of 0.05 or 5% for rejection level of the null hypothesis was applied. RESULTS: the agreement rates of hysterosonographic results compared to hysteroscopy and histopatology were 94.8 ande 77.6%, respectively. Sensitivity and specificity of hysterosonographic evaluation of the abnormal endometrial cavity were 98 and 75%, respectively, when compared to hysteroscopy. In addittion, positive and negative predictive values of hysterosonography were 96 and 86%, respectively. When the histopathologic study was used as the gold standard, sensitivity and specificity were 98 and 33%, with positive predictive value of 76% and negative predictive value of 86%, for the detection of the endometrial cavitary changes. One great concern were the histopathologic results of two patients with uterine synechia who showed endometrial hyperplasia. Also, one patient was diagnosed as normal using histerosonography and the histopatological result showed simple hyperplasia. CONCLUSIONS: our data suggest that hysterosonography presented good sensitivity as compared with hysteroscopy. However, uterine synechia is the great limitation of this method as compared with histopathology.