Two hundred and twenty-two patients with advanced breast cancer were randomised in two separate trials of similar design to either concomitant combination treatment or sequential use of the same drugs given as single agents changed only at disease progression. Both trials used cyclophosphamide, methotrexate, 5-fluorouracil and prednisone; the WCSG using triiodothyronine and the SECSG using vincristine as the remaining agent. A common data base was generated for these trials and combined for analysis. Considering all patients, combination treatment was associated with a significantly increased response (46 versus 25%, P < 0.05) but not survival improvement. For the 141 patients without liver involvement, survival was closely comparable in both treatment arms. Combination therapy did result in significant survival benefit for patients with liver involvement (P <0.05). These studies demonstrate: (1) in the majority of breast cancer patients, sequential single agent therapy can result in survival comparable to combination treatment; and (2) sole consideration of response frequency does not represent the optimal criterion to compare therapeutic approaches in advanced breast cancer. © The Macmillan Press Ltd., 1989.