Clinical trials conducted in Western countries have shown that aromatase inhibitors are associated with better disease-free survival (DFS) than tamoxifen in postmenopausal early breast cancer. Because pharmacogenetic differences in drug-metabolizing genes may cause ethnic differences, assessment of the efficacy and tolerability of aromatase inhibitors in non-white women is warranted. This open-label, randomized clinical trial included 706 postmenopausal Japanese women with hormone-receptor-positive breast cancer, who had received tamoxifen for 1 to 4 years as adjuvant therapy. This study was closed early after entry of approximately 28% of the initially planned patients. They were randomly assigned to either switch to anastrozole or to continue tamoxifen for total treatment duration of 5 years. Primary endpoints were DFS and adverse events. At a median follow-up of 42 months, the unadjusted hazard ratio was 0.69 (95% confidence interval, 0.42-1.14; P = 0.14) for DFS and 0.54 (95% CI, 0.29-1.02; P = 0.06) for relapse-free survival (RFS), both in favor of anastrozole. The incidence of thromboembolic events in the tamoxifen group and bone fractures in the anastrozole group was not excessively high. Switching from tamoxifen to anastrozole was likely to decrease disease recurrence in postmenopausal Japanese breast cancer patients. Ethnic differences in major adverse events may be attributable to a low baseline risk of these events in Japanese.