The therapeutic efficacy of vincristine, vinblastine, and vindesine were evaluated in a prospective randomized trial in patients with metastatic breast carcinoma. All patients were refractory to doxorubicincontaining chemotherapy. Vincristine was administered at 0.4 mg/M2/I/day by continuous infusion (CI), vinblastine at 1.7 mg/M2/I/day by CI, and vindesine at 1.2 mg/M2/I/day by CI or intermittent bolus (IB) over 5 days. The courses were administered at 2‐week intervals for vincristine and 3‐week intervals for vinblastine and vindesine. Ninety‐nine patients were evaluable for response. The 15 patients treated with vincristine had no objective response. Of 23 patients treated with vinblastine, there were 2 complete responses (9%) and 5 partial responses (22%). Of 31 patients treated with CI vindesine, there were 6 partial responses (19%). Of 30 patients treated with IB vindesine, there were 5 partial responses (17%). The median duration of disease control was 13 weeks (range, 8–140 +) for vinblastine, 18 weeks (range, 8–34) for CI vindesine, and 20 weeks (range, 12–47) for IB vindesine. These data illustrate that vinblastine (CI) and vindesine (CI or IB) have significant antitumor activity in patients with refractory metastatic breast carcinoma and that vincristine has no antitumor activity in similar patients.