The risk for endometrial cancer among women with breast cancer might increase following use of tamoxifen, recently classified as a carcinogen of the human endometrium. However, the strength of the association remains uncertain and it is unknown whether use of this drug--widely prescribed in Sweden since the mid-1980s--has had any measurable effect at the population level. We analyzed all cases of breast cancer (n = 131,614) detected in the nationwide Swedish Cancer Registry in 1958-93. Incident cases of endometrial cancer during follow-up were identified also through the Cancer Registry. Standardized incidence ratios (SIR) and their 95 percent confidence intervals (CI) were computed by use of nationwide rates of endometrial cancer, adjusted for age and calendar year. During follow-up of up to 35 years of the breast cancer cohort, 803 incident endometrial cancers were identified, yielding an overall SIR of 1.58 (CI = 1.47-1.70). In univariate analyses, there was no increase in SIR in recent years. However, the excess risk increased linearly with increasing age at breast cancer diagnosis (P trend < 0.0001) and decreased markedly with longer follow-up (P trend < 0.0001). A multivariate regression analysis, with adjustment for age and year of follow-up, revealed no increased risk for subsequent endometrial cancer among breast cancer cases diagnosed more recently (P trend = 0.19); compared with the period 1958-63, the risk estimate in the last time period (1989-93) was 0.72 (CI = 0.51-1.01). We conclude that the risk for endometrial cancer following breast cancer has not increased over time in Sweden.