Busulfan plus cyclophosphamide versus busulfan plus fludarabine as a preparative regimen for allogeneic haemopoietic stem-cell transplantation in patients with acute myeloid leukaemia: an open-label, multicentre, randomised, phase 3 trial

Summary Background The standard busulfan–cyclophosphamide myeloablative conditioning regimen is associated with substantial non-relapse mortality in patients older than 40 years with acute myeloid leukaemia who are undergoing allogeneic stem-cell transplantation. Because the combination of busulfan plus fludarabine has been proposed to reduce non-relapse mortality, we aimed to compare this treatment with busulfan plus cyclophosphamide as a preparative regimen in these patients. Methods We did an open-label, multicentre, randomised, phase 3 trial for patients with acute myeloid leukaemia at 25 hospital transplant centres in Italy and one in Israel. Eligible patients were aged 40–65 years, had an Eastern Cooperative Oncology Group performance status less than 3, and were in complete remission. Patients were randomly assigned 1:1 to receive intravenous busulfan plus cyclophosphamide or busulfan plus fludarabine. Treatment allocations were not masked to investigators or patients. Randomisation was done centrally via a dedicated web-based system using remote data entry, with patients stratified by donor type and complete remission status. Patients allocated to busulfan plus cyclophosphamide received intravenous busulfan 0·8 mg/kg four times per day during 2 h infusions for four consecutive days (16 doses from days −9 through −6; total dose 12·8 mg/kg) and cyclophosphamide at 60 mg/kg per day for two consecutive days (on days −4 and −3; total dose 120 mg/kg). Patients allocated to busulfan plus fludarabine received the same dose of intravenous busulfan (from days −6 through −3) and fludarabine at 40 mg/m 2 per day for four consecutive days (from days −6 through −3; total dose 160 mg/m 2 ). The primary endpoint was 1-year non-relapse mortality, which was assessed on an intention-to-treat basis; safety outcomes were assessed in the per-protocol population. This trial has been completed and is registered with ClinicalTrials.gov, number NCT01191957. Findings Between Jan 3, 2008, and Dec 20, 2012, we enrolled and randomly assigned 252 patients to receive busulfan plus cyclophosphamide (n=125) or busulfan plus fludarabine (n=127). Median follow-up was 27·5 months (IQR 9·8–44·3). 1-year non-relapse mortality was 17·2% (95% CI 11·6–25·4) in the busulfan plus cyclophosphamide group and 7·9% (4·3–14·3) in the busulfan plus fludarabine group (Gray's test p=0·026). The most frequently reported grade 3 or higher adverse events were gastrointestinal events (28 [23%] of 121 patients in the busulfan plus cyclophosphamide group and 26 [21%] of 124 patients in the busulfan plus fludarabine group) and infections (21 [17%] patients in the busulfan plus cyclophosphamide group and 13 [10%] patients in the busulfan plus fludarabine group had at least one such event). Interpretation In older patients with acute myeloid leukaemia, the myeloablative busulfan plus fludarabine conditioning regimen is associated with lower transplant-related mortality than busulfan plus cyclophosphamide, but retains potent antileukaemic activity. Accordingly, this regimen should be regarded as standard of care during the planning of allogeneic transplants for such patients. Funding Agenzia Italiana del Farmaco.