Trilostane, which causes a perturbation of adrenal steroidogenesis, was studied in combination with hydrocortisone in 32 women with progressive metastatic breast cancer. Trilostane was administered orally at a dosage level of 240 mg four times daily after escalation over the first 10 days from 60 mg four times daily. Hydrocortisone was given orally at doses of 10 mg at 8 a.m. and 5 p.m. and 20 mg at bedtime. Patients must have been postmenopausal (81%) or previously castrated (19%), had a response to the hormonal treatment just prior to study (81%) or a positive estrogen receptor at time of entry on study (41%), and a measurable indicator lesion. The number of prior hormonal therapies was 1 in 19 patients (59%), 2 in 12 patients (38%), and 3 in 1 patient (3%), respectively. Twelve patients (38%) achieved an objective response, and a 95% confidence interval for this result is from 21 to 56%. The median time to disease progression was 140 days, median duration of response was 278 days, and median survival was 556 days. Common toxicities included lethargy, lightheadedness, diarrhea, and abdominal discomfort. Eleven patients required a dosage reduction, usually because of gastrointestinal side effects, and one additional patient had the trilostane discontinued because of leukopenia. We conclude that the combination of trilostane plus hydrocortisone appears to have definite antitumor activity in women with metastatic breast cancer who have characteristics favorable for response to hormonal therapy.