The purpose of this single arm phase II study was to test a modified version of the three drug combination vincristine, adriamycin and dexamethasone (m-VAD), in which intravenous vincristine (0.4 mg/d) and adriamycin (9 mg/m2 per day) infusions are administered for only 2 h on days 1-4 of each 28 d cycle, in patients with refractory multiple myeloma. In addition, only two 4 d courses of dexamethasone 40 mg/d was given during each cycle. The entry criteria for 44 patients included plasma cell myeloma and a measurable monoclonal peak, either refractory to initial treatment with melphalan and prednisone, or resistant to melphalan and prednisone after initially responding (resistant relapsed disease, 27 patients). Patients treated previously with chemotherapy other than melphalan and prednisone were excluded. There were no complete responses. Of the 41 evaluable patients who completed at least one course of therapy 11 had a partial response (27%, 95% C.I. 14-40%). The response rates were 19% for primary refractory disease patients, and 32% for those with resistant relapsed disease. The median duration of response was 4 months. The median survival for all 44 patients was 7.6 months (5.5 months for primary refractory patients, and 10 months for relapsed resistant disease patients). Episodes of documented bacterial infection occurred in 12 patients, and 10 patients had minor viral infection. The dexamethasone dose was reduced in 12 patients. The median neutrophil nadir was 1.2 x 10(9)/l, and median platelet nadir was 147 x 10(9)/l. Five deaths were judged as treatment related and occurred during marrow cytopenia. The results of this modified form of VAD are inferior to that reported previously for 4 d continuous infusions of vincristine and doxorubicin. This could be related to either patient selection factors, or to a reduction of the efficacy of the drug combination produced by either the shortened intravenous infusions and/or omission of one 4 d course of dexamethasone.