OBJECTIVES: To compare transvaginal ultrasonography with histological findings in endometrial evaluation of postmenopausal women using hormone replacement therapy and to evaluate endometrial safety of three hormone replacement therapy regimens. METHODS: In a randomized, comparative study in postmenopausal women, endometrial safety was evaluated using (1) no hormone replacement therapy, (2) oral micronized 17 beta-estradiol/oral sequential dydrogesterone, (3) transdermal 17 beta-estradiol/oral sequential dydrogesterone, or (4) oral tibolone. 85 Non-hysterectomised subjects underwent transvaginal ultrasonography immediately before Pipelle biopsy at baseline and subsequently after 12 and 24 months. Endometrial thickness and uterine dimensions were determined by transvaginal ultrasonography, and endometrial thickness (double-layer) was compared with biopsy results. RESULTS: Endometrial evaluation was conveniently performed by transvaginal ultrasonography, and endometrial thickness correlated well with biopsy findings. If endometrial thickness was < 5 mm, the endometrial biopsy sample was either inactive/atrophic or insufficient for histopathological diagnosis. Hyperplastic or malignant changes were not reported. After 24 months, endometrial thickness was increased both in the oral (P < 0.001) and transdermal (P < 0.001) 17 beta-estradiol/dydrogesterone groups, whereas with tibolone the change in endometrial thickness was not different from controls. CONCLUSION: transvaginal ultrasonography of the endometrium reliably predicts the histological picture in hormone replacement therapy users. Using 5 mm endometrial thickness as cut-off point, more than 75% of biopsies could be avoided. All three hormone replacement therapies were safe with respect to the endometrium. With sequential 17 beta-estradiol/dydrogesterone the expected progestogen-induced secretory pattern was observed, whereas endometrial histology under tibolone closely mimicked the natural atrophic postmenopausal state.