Thirty-seven men with symptomatic bone metastases from prostate cancer that had progressed following earlier treatment with estrogens and/or orchidectomy were treated with low-dose prednisone (7.5 to 10 mg daily). The rationale for this treatment was that some patients might still have hormone-sensitive disease that was stimulated by weak androgens of adrenal origin, and that these androgens could be suppressed by prednisone through its negative feedback on secretion of adrenocorticotrophic hormone (ACTH). Response to treatment was assessed by requirement for analgesics, by the McGill-Melzack pain questionnaire, and by a series of 17 linear analog self-assessment (LASA) scales relating to pain and to various aspects of quality of life. Fourteen patients (38%) had improvement in indices used to assess pain at 1 month after starting prednisone, and seven patients (19%) maintained this improvement for 3 to 30 months (median, 4 months). Reduction in pain was associated with improvement in other dimensions of quality of life, and in the scale for overall well-being. Prednisone treatment led to a decrease in the concentration of serum testosterone in seven of nine patients where it was not initially suppressed below 2 nmol/L, and caused a decrease in serum levels of androstenedione and dehydroepiandrosterone sulfate in more than 50% of patients. Symptomatic response was associated with a decrease in serum concentration of adrenal androgens. We conclude that (1) low-dose prednisone may cause useful relief of pain in some patients with advanced prostatic cancer; (2) relief of pain was associated with suppression of adrenal androgens; and (3) measures of pain and quality of life can be used to assess possible benefits of systemic therapy in patients with metastatic prostate cancer.