Further assessment of the long-term risk of breast cancer associated with diethylstilbestrol (DES) during pregnancy.Follow-up continuation through June 1, 1989, of a historical cohort of DES-exposed and unexposed mothers ascertained by review of obstetric records.Totals of 3029 each of DES-exposed and unexposed mothers who had delivered live babies at four centers in the United States during 1940 through 1960. Questionnaires were returned for 92.6% of the DES-exposed and 88.8% of the unexposed women.Breast cancer incidence and mortality assessed from returned questionnaires and review of medical records and death certificates.The relative rate of breast cancer associated with DES exposure, after adjustment for demographic and reproductive variables, was 1.35 (95% confidence interval, 1.05 to 1.74). For 30 years or more following exposure, the relative rate was not appreciably higher (relative rate, 1.33; 95% confidence interval, 0.95 to 1.87) than that in earlier periods. Surveillance and increased detection seemed unlikely explanations for the increased risk, since DES-exposed women had excesses of both large and small breast cancers and the two cohorts reported similar breast cancer detection practices. A history of miscarriage before first term delivery was not associated with breast cancer occurrence.Exposure to DES during pregnancy is associated with a modest but statistically significant increased risk of breast cancer. Contrary to prior indications, the risk does not appear to increase greatly over time. The findings are sufficient to exclude the possibility of a doubling of risk for the period of 30 or more years following exposure.