Data from a population-based case-control study were used to evaluate the risk of endometrial cancer among women who have used an intra-uterine device (IUD). Incident cases were identified between 1985 and 1991 among women aged 45-74 years who were residents of one of 3 counties in Washington State. Controls were selected by random digit dialing, and both groups of subjects received an in-person detailed interview. In this study population, women who had ever used an IUD were estimated to have a risk of endometrial cancer that was 0.61 times that of other women (95% CI 0.41-0.89). The reduction in cancer risk was not found to be dependent on duration of IUD use. There was a suggestion that women who had used intra-uterine contraception relatively late in reproductive life experienced a greater reduction in risk than those whose use was more distant or at a younger age. The relative risk among the small number of women who were currently using an IUD was 0.49 (95% CI 0.12-2.80). These results apply to the use of inert and copper IUDs as there was no use of progestin-releasing IUDs among women in the study population. The data from this and several other studies of the question support the hypothesis that use of an IUD has a favorable effect on the subsequent risk of endometrial cancer. The reason(s) for such a reduced risk is unclear.