Many women are now being maintained on tamoxifen therapy following treatment of breast cancer. The estrogenic effects of this drug can lead to proliferative endometrial changes that have the potential to lead to malignant change. Regular ultrasonic assessment of endometrial thickness shows promise as a screening tool in these women, but there is controversy as to the thickness at which sampling should be undertaken. Franchi and colleagues sought to determine the thickness that would best predict a need for endometrial sampling in order to avoid subjecting many women to unnecessary interventions.

The authors studied 163 postmenopausal women with breast cancer who had been receiving 20 mg per day of tamoxifen for at least six months. Women with previous endometrial hyperplasia or vaginal bleeding during the first six months of therapy were excluded from the study. All women had complete gynecologic evaluations, including sonographic assessment, before beginning tamoxifen therapy and then every 12 months during therapy. An additional sonogram was performed if vaginal bleeding occurred. No exogenous estrogen was permitted during the study. All women were examined by hysteroscopy, and samples were taken for histologic assessment.

Abnormal histologic findings occurred in 60 percent of women with an endometrial thickness of 9 mm or greater compared with only 6 percent of women with lesser endometrial thickness. Histologic abnormality was also significantly related to the duration of tamoxifen use: 46 percent of those who took the drug for 27 months or longer had positive findings, compared with 16 percent of those who took the drug for less time. The occurrence of vaginal bleeding was also predictive of histologic findings. Of the 163 women studied, 46 (28 percent) reported vaginal bleeding during the study. Only 43 percent of these women had atrophic endometrium compared with 84 percent of women without bleeding. Among the women who reported vaginal bleeding, 9 percent had simple hyperplasia, 4 percent had endometrial polyp and 4 percent had endometrial carcinoma. No cases of atypical hyperplasia, polyp or carcinoma were found in women who did not report vaginal bleeding.

The authors conclude that an endometrial thickness of 9 mm and a report of vaginal bleeding are independent predictors of endometrial pathology and could be used to select women for hysteroscopy and biopsy during tamoxifen therapy.