We report the results of a randomized trial carried out by the Swiss Group for Clinical Cancer Research (SAKK) and in which 230 patients with advanced breast cancer receiving concurrently a hormonal treatment (oophorectomy for pre- and tamoxifen for postmenopausal women) were randomly allocated to three different regimens of combination chemotherapy. The therapeutic results registered with the two more intensive combinations (LMP/FVP and LMFP/ADM) were similar with regard to response rates, time to progression and survival. The patients receiving the low-dose chemotherapy lmfp showed a statistically significant lower response rate (32%, P less than 0.001) and a shorter survival (P = 0.03) than the results observed in patients treated with the two other regimens. This difference was particularly pronounced, at least regarding survival, in the following subgroups: postmenopausal women, patients with a poor performance status, dominant visceral lesions, two sites of disease and a disease-free interval longer than 12 months. Patients with bony metastases as dominant lesion fared similarly with all three regimens of chemotherapy. This latter subset of advanced breast cancer patients should probably be spared too intensive cytotoxic treatment. This is, to our knowledge, the first report of a randomized trial showing an evident correlation between response rate and survival in various subgroups of patients with advanced breast cancer treated with different chemotherapeutic regimens.