The relation between breast cancer and selected nonestrogenic drugs was evaluated in the Group Health Cooperative of Puget Sound, Seattle, Washington, a prepaid health care organization with computerized information on diagnoses and outpatient drug use. No important positive associations with breast cancer were found in a follow-up study of 302 women aged 35-74 years. These women were newly diagnosed with breast cancer in 1977-1980 and were studied in relation to exposure in the six months prior to diagnosis to one or more of the following drugs: diazepam, digitalis glycosides, medroxyprogesterone acetate, methyldopa, metronidazole, phenothiazines, tricyclic antidepressants, thiazides, thyroid/levothyroxine sodium, or spironolactone. A modest association between recent reserpine use and breast cancer was present (risk ratio = 1.7, 90% confidence interval 0.9-3.3).