We administered combination chemotherapy with cyclophosphamide, doxorubicin, and cisplatin to 25 previously untreated patients with metastatic prostate cancer in order to assess the efficacy of chemotherapy before any hormonal manipulation. Hormonal therapy was administered only after progression of disease to chemotherapy. All 25 patients were followed until time of death and all were able to receive hormonal therapy. We did not find substantially improved response rates when combination chemotherapy was applied before endocrine treatment since the 33% objective response rate to chemotherapy was only minimally higher than the response in our patients who had failed hormonal therapy and then received identical or similar chemotherapy. Furthermore, the introduction of intensive combination chemotherapy before hormonal therapy in our study did not result in any striking improvement in overall survival compared with patients who received initial hormonal therapy in many other studies. Responses to chemotherapy were not attributable to suppression of serum testosterone since all 12 patients with partial response (PR) or stable disease (SD) and four of seven patients with no response (NR) had normal testosterone levels at the time of response assessment. The initial use of chemotherapy did not adversely affect the expected high percentage of objective responses (68%) to subsequent hormonal manipulation. The frequency, duration, and quality of responses to hormonal therapy exceeded the responses to chemotherapy. The disappointing responses to chemotherapy reflect the very modest efficacy of even aggressively delivered cytotoxic agents.