The optimum preparative regimen for unrelated donor marrow transplantation in patients with severe aplastic anaemia remains to be established. We investigated whether the combination of fludarabine, anti-thymocyte globulin, and total body irradiation (TBI) would enable reduction of the cyclophosphamide dose to less than 200 mg/kg while maintaining engraftment and having a survival similar to or better than that with standard regimens using a cyclophosphamide dose of 200 mg/kg (known to be associated with significant organ toxicity) for unrelated donor transplantation for severe aplastic anaemia. We have previously shown that cyclophosphamide at 150 mg/kg resulted in excess toxicity and its omission (0 mg/kg) resulted in unacceptable graft failure (three of three patients had secondary graft failure). Here we report results for the 50 mg/kg and 100 mg/kg cohorts.

In a multicentre phase 1-2 study, patients (aged ≤65 years) with severe aplastic anaemia, adequate organ function, and an unrelated adult marrow donor HLA matched at the allele level for HLA A, B, C, and DRB1 or mismatched at a single HLA locus received bone marrow grafts from unrelated donors. All patients received anti-thymocyte globulin (rabbit derived 3 mg/kg per day, intravenously, on days -4 to -2, or equine derived 30 mg/kg per day, intravenously, on days -4 to -2), fludarabine (30 mg/m(2) per day, intravenously, on days -5 to -2), and TBI (2 Gy). Cyclophosphamide dosing started at 150 mg/kg and was de-escalated in steps of 50 mg/kg (to 100 mg/kg, 50 mg/kg, and 0 mg/kg). The primary endpoint was the selection of the optimum cyclophosphamide dose based on assessments of graft failure (primary or secondary), toxicity, and early death during 100 days of follow-up after the transplant; this is the planned final analysis for the primary endpoint. This trial is registered with ClinicalTrials.gov, number NCT00326417.

96 patients had bone marrow transplant. At day 100, 35 (92%) of 38 patients were engrafted and alive in the cyclophosphamide 50 mg/kg cohort and 35 (85%) of 41 in the 100 mg/kg cohort. Cyclophosphamide 50 mg/kg and 100 mg/kg resulted in posterior means for fatality without graft failure of 0·7% (credible interval 0-3·3) and 1·4% (0-4·9), respectively. Three patients (8%) had graft failure with cyclophosphamide 50 mg/kg and six (15%) with cyclophosphamide 100 mg/kg. Four (11%) patients had major regimen-related toxicity with cyclophosphamide 50 mg/kg and nine (22%) with cyclophosphamide 100 mg/kg. The most common organ toxicity was pulmonary (grade 3 or 4 dyspnoea or hypoxia including mechanical ventilation), and occurred in three (8%) and four (10%) patients given cyclophosphamide 50 mg/kg and 100 mg/kg, respectively.

Cyclophosphamide at 50 mg/kg and 100 mg/kg with TBI 2 Gy, fludarabine, and anti-thymocyte globulin results in effective conditioning and few early deaths after unrelated donor transplantation for severe aplastic anaemia. These doses of cyclophosphamide provide a framework for further regimen optimisation strategies.

US National Heart, Lung, and Blood Institute and National Cancer Institute.