Chemotherapy given to patients with metastatic carcinoma of the breast is palliative in intent. Longer regimens would be justified if there was a proven prolongation of symptom response or survival. We conducted a randomised trial to assess the survival of patients receiving up to six extra courses of chemotherapy compared with our conventional regimen of six courses. The patients received either VAC, VEC (vincristine, doxorubicin or epirubicin and cyclophosphamide) or MMM (mitozantrone, methotrexate and mitomycin C) therapy. Patients who had stable disease or were responding after six courses of chemotherapy were randomised to either stop or continue treatment for another six courses. Those patients receiving maintenance therapy had a significantly longer duration of response (P < 0.02) and a significantly longer progression-free survival (P < 0.01). However, there was no survival difference between the two groups. Furthermore, treatment toxicity, which was similar in the two groups, persisted for longer in the maintenance group. These results indicate no clinical advantage for giving maintenance chemotherapy in order to prolong survival of patients with metastatic breast cancer.