Postmenopausal women with localised, early breast cancer (n = 285) were enrolled in a prospective subprotocol of the 'arimidex, tamoxifen, alone or in combination' (ATAC) trial to assess gynaecological abnormalities arising during treatment with anastrozole (1 mg/day) or tamoxifen (20 mg/day). After 6 years' follow-up, there appeared to be non-significantly fewer endometrial abnormalities with anastrozole than with tamoxifen (12.4% vs 20.2%, odds ratio 0.52; 95% confidence intervals 0.20, 1.32; p = 0.17). The time to first endometrial abnormality was non-significantly longer for patients receiving anastrozole compared with tamoxifen (hazard ratio 0.57; 95% confidence intervals 0.26, 1.22; p = 0.15), with most abnormalities occurring within the first year of treatment. Fewer patients treated with anastrozole appeared to require medical intervention for endometrial abnormalities, compared with patients on tamoxifen. This study showed that there was no significant difference in endometrial pathology between anastrozole and tamoxifen treatment groups.