Reduced-intensity conditioning versus standard conditioning before allogeneic haemopoietic cell transplantation in patients with acute myeloid leukaemia in first complete remission: a prospective, open-label randomised phase 3 trial

Background: Reduced-intensity conditioning regimens have been developed to minimise early toxic effects and deaths after allogeneic haemopoietic cell transplantation. However, the efficacy of these regimens before this procedure has not been investigated in a randomised trial. In this prospective, open-label randomised phase 3 trial we compared a reduced-intensity fludarabine-based conditioning regimen with a standard regimen in patients with acute myeloid leukaemia in first complete remission.

Methods: Patients were aged 18-60 years and had intermediate-risk or high-risk acute myeloid leukaemia (defined by cytogenetics) in first complete remission; an available HLA-matched sibling donor or an unrelated donor with at least nine of ten HLA alleles; and adequate renal, cardiac, pulmonary, and neurological function. Between Nov 15, 2004, and Dec 31, 2009, patients were randomly assigned (1:1, by a computer-based minimisation procedure that balanced patients for age, cytogenetic risk, induction therapy, and donor type) to receive either reduced-intensity conditioning of four doses of 2 Gy of total-body irradiation and 150 mg/m(2) fludarabine or standard conditioning of six doses of 2 Gy of total-body irradiation and 120 mg/kg cyclophosphamide. All patients were given ciclosporin and methotrexate as prophylaxis against graft-versus-host disease. Neither investigators nor patients were blinded to study treatment. Our primary endpoint was the incidence of non-relapse mortality, analysed in the intention-to-treat population. The trial is registered with ClinicalTrials.gov, number NCT00150878.

Findings: The trial was stopped early on Dec 31, 2009, because of slow accrual of patients. 99 patients were randomly assigned to receive reduced-intensity conditioning and 96 to receive standard conditioning. The incidence of non-relapse mortality did not differ between the reduced-intensity and standard conditioning groups (cumulative incidence at 3 years 13% [95% CI 6-21] vs 18% [10-26]; HR 0·62 [95% CI 0·30-1·31]). Relapse incidence (cumulative incidence 3 years 28% [95% CI 19-38] vs 26% [17-36]; HR 1·10 [95% CI 0·63-1·90]), disease-free survival (3 year disease-free survival 58% [95% CI 49-70] vs 56% [46-67]; HR 0·85 [95% CI 0·55-1·32]), and overall survival (3 year overall survival 61% [95% CI 50-74] vs 58% [47-70]; HR 0·77 [95% CI 0·48-1·25]) did not differ significantly between groups. Grade 3-4 of oral mucositis was less common in the reduced-intensity group than in the standard conditioning group (50 patients in the reduced-intensity conditioning group vs 73 patients in the standard conditioning group); the frequency of other side-effects such as graft-versus-host disease and increased concentrations of bilirubin and creatinine did not differ significantly between groups.

Interpretation: Reduced-intensity conditioning results in a similar incidence of non-relapse mortality and reduced toxic effects compared with standard conditioning without affecting survival outcomes, and thus could be preferentially used in patients younger than 60 years with acute myeloid leukaemia transplanted in first complete remission.