The relation between diazepam use and breast cancer was explored in two case-control studies. The first (1981-1987) was a hospital-based study in the United States of 3,078 cases of breast cancer, 1,259 controls with other malignancies, and 672 controls with nonmalignant conditions. The relative risk estimates for regular diazepam use (at least 4 days per week for at least 6 months) that antedated the diagnosis of breast cancer by at least 12 months were 1.0 (95% confidence interval (CI) 0.6-1.7) using the cancer controls and 0.8 (95% CI 0.4-1.8) using the noncancer controls. Risk factors for breast cancer were taken into account in the estimates. There was no association for regular use lasting at least 5 years or for regular use that took place exclusively in the recent or more distant past. The second study (1982-1986), conducted in Toronto, Ontario, Canada, was community-based, and included 607 cases of breast cancer and 1,214 neighbor controls selected from municipal voting and census records. After control of confounding, the relative risk estimate for regular diazepam use was 0.8 (95% CI 0.5-1.3). Again, there was no association for long-term, past, or recent regular use. The results of both studies suggest that regular diazepam use does not increase the risk of breast cancer. The findings are strengthened by the similarity of the results using three different control groups--women with cancer, women with nonmalignant conditions, and neighbors.