Tamoxifen and endometrial pathologies: a prospective study

Abstract

Objective: The purpose of this study was to prospectively follow a group of women with breast cancer, on tamoxifen, for the development of endometrial pathologies.

Materials and methods: Eighty women with breast cancer, on tamoxifen, were prospectively followed every 6 months with pelvic examination, Pap smear, vaginal ultrasound, and endometrial biopsy.

Results: Nine women were lost to follow-up prior to initiation of treatment and 4 refused biopsies, leaving 67 patients for evaluation. Fifty (74.6%) of the 67 patients were already on tamoxifen for a mean duration of 15.8 +/- 16.6 months and had a baseline benign, unremarkable endometrium at the time of entry into the study. The total duration of treatment was 32.5 +/- 19.6 months (median 30 months). The mean age of the patients was 51.7 +/- 9.9 years (median 52 years). Of the patients, 56.7% were postmenopausal. Sixty-three patients had a benign endometrium (mean age 51.8 +/- 10.1 years, mean duration 33.1 +/- 19.6 months). Two patients had simple hyperplasia (mean age 43.5 years, duration 28.5 +/- 33.2 months), 1 patient had complex hyperplasia with atypia (age 57 years, duration 13 months), and another patient developed adenocarcinoma (grade 3) after 22 months. These 4 patients had abnormal vaginal bleeding. Seven patients developed endometrial polyps (mean age 54.0 +/- 8.5 years, duration 36 +/- 24.2 months). The mean endometrial thickness for patients with histologically unremarkable and abnormal endometrium was not significantly different (7.6 +/- 3.9 vs 8.8 +/- 5.0 mm, respectively) (median 7.0 mm for both groups). No endometrial thickness cutoff point reached statistical significance. The patient who developed endometrial cancer had a thickness of only 3 mm.

Conclusion: All patients who developed an abnormal endometrium had abnormal vaginal bleeding. There was no correlation between endometrial thickness and endometrial pathology; thus the value of routine screening remains controversial.