The records of 133 patients treated with megestrol acetate as primary hormonal therapy for advanced breast cancer were reviewed retrospectively, using International Union Against Cancer (UICC) criteria for response. The median age was 65 years, 121 patients were over age 50, and the age range was 39 to 94 years. Response rates (complete response [CR] + partial response [PR]/total) by qualitative receptor level, with levels of 10 fmol/mg of protein considered positive, were as follows (ER = estrogen receptor, PgR = progesterone receptor): For ER + PgR+, 13 + 15/56 (50%); for those with one positive receptor, 0 + 12/47 (26%); for ER - PgR-, 0 + 0/12 (0%); and for receptor-unknown cases, 2 + 3/18 (14%). Response for ER less than 30 fmol/mg was 2 + 6/39 (21%); for ER 30 to 50, 1 + 5/16 (40%); and for ER greater than 50 fmol/mg, 11 + 15/56 (46%). For PgR less than 30, response was 0 + 6/37 (16%); for PgR 30 to 50 fmol/mg, 1 + 4/14 (36%); and for PgR greater than 50 fmol/mg, 12 + 13/54 (46%). For the 75 patients with a disease-free interval (DFI) of 2 years or less, the response rate was 5 + 1/75 (8%), and for the 58 patients with DFI greater than 2 years, 10 + 12/60 (37%). For patients with prior chemotherapy, 3 + 8/49 (22%) had an objective response. For those with no prior chemotherapy, 12 + 19/84 (37%) responded. Response by dominant site of disease was as follows: soft tissue 12 + 9/43 (49%), bone 2 + 13/49 (31%), viscera 2 + 5/41 (17%). Of these seven patients with visceral dominant disease who responded, all had PgR levels greater than 50 fmol/mg, all but one had an ER level over 100 fmol/mg, all but one were over age 65, and all but two received no prior chemotherapy. We conclude that megestrol acetate is effective initial hormonal therapy for patients with advanced breast cancer. It may have some role to play in the treatment of carefully selected cases with visceral disease.