Alternating v repeated postremission treatment in adult acute myelogenous leukemia: a randomized phase III study (AML6) of the EORTC Leukemia Cooperative Group.

The value of a postremission treatment in acute myelogenous leukemia (AML), with alternating combinations of non-cross-resistant drugs, has been prospectively assessed. Of 515 evaluable patients, 347 (67.4%) entered into complete remission (CR), following induction treatment with daunorubicin (DNR), vincristine (VCR), and cytosine arabinoside (ara-C). After one consolidation course, 248 patients were randomized for six courses of intensive maintenance: either repeated treatment with DNR-VCR-ara-C, or alternating treatment where amsacrine (AMSA) was combined with high dose ara-C on cycle 1,3, and 5 and with 5-azacytidine on cycle 2, 4, and 6. Ninety-nine patients were not randomized: 57 were introduced in a bone marrow transplantation (BMT) program, and 42 went off study, mainly for treatment toxicity or refusal. The main prognostic factors for achievement of CR were performance status, cytogenetics, and age, and for the disease-free survival (DFS): age and number of courses to CR. The rate of second remission was fairly high (64%) for patients relapsing off therapy. The DFS appeared identical (median, 53 weeks), in the two randomized arms, the alternating treatment not showing superiority to the repeated one, in spite of an increased toxicity. The median overall survival for patients achieving a CR was 90 weeks. The reason for the failure of alternating maintenance treatment to improve the DFS is probably related to an insufficient dose intensity: five patients who relapsed during maintenance arm B achieved a second CR with a more intensive combination of high-dose ara-C and AMSA. In addition, 60 patients underwent a BMT (43 allogeneic and 17 autologous). The DFS of patients treated with allogeneic BMT tended to be superior to the one obtained with the chemotherapy program. However the overall survival, as well as the event-free survival, seemed equivalent, including patients who relapsed before the planned BMT. Comparisons between allogeneic BMT, autologous BMT, and intensive consolidation during first CR deserve further prospective studies in AML.