The relationship between the risk of endometrial cancer and the use of noncontraceptive estrogens by Japanese postmenopausal women was investigated in a hospital-based case-control study of 1025 women with endometrial cancer and 1267 with other conditions. The overall odds ratio (OR) for estrogen use with or without progestins, compared with never use of any type of estrogens, was 0.917 (95% confidence interval (CI) 0.622-1.353), suggesting that hormone replacement therapy is not a causative agent for endometrial cancer in Japanese women, and that a recent increase in the incidence of endometrial cancer in Japanese women may be related to changes in their life-style. However, although not statistically significant, women who used estrogen without progestin for 12 or more months had an OR of 2.552 (CI 0.231-28.192), while those who used estrogen with progestin for 12 or more months had an OR of 0.425 (CI 0.086-2.113). These results indicate that the addition of a progestin should be considered for reducing the risk of endometrial cancer in Japanese women.