Purpose: To analyze the French experience of chemotherapeutic preparation before human leukocyte antigen (HLA)-identical bone marrow transplantation (BMT) in children with acute myeloblastic leukemia (AML) in first complete remission (CR).
Patients and methods: The data base used for this study was a French BMT registry for childhood AML. Twenty-three children were conditioned with busulfan and 120 mg/kg cyclophosphamide (Bu-Cy 120 group). Nineteen received busulfan and 200 mg/kg cyclophosphamide (Bu-Cy200 group). During the same time period, 32 patients were prepared with total-body irradiation (TBI group) most often in combination with 120 mg/kg of cyclophosphamide.
Results: The probability of relapse was 54%, 13%, and 10% for the Bu-Cy120, Bu-Cy200, and TBI groups, respectively (P < .05 in the univariate analysis, log-rank test, 2 df). In the multivariate analysis, a conditioning regimen with Bu-Cy120 was significantly associated with a higher risk of relapse (P = .02; relative risk, 3.62). The probability of transplant-related mortality (TRM) was 0% for Bu-Cy120, 5% for Bu-Cy200, and 10% for TBI. Kaplan-Meier estimations of event-free survival (EFS) were 46% +/- 24%, 82% +/- 18%, and 80% +/- 14%, respectively, for the three groups, with median follow-up durations of 28 months (range, 3 to 78), 31 months (4 to 68), and 48 months (2 to 73). In the multivariate analysis, two factors adversely affected EFS: a conditioning regimen with Bu-Cy120 (P = .07) and a long interval from diagnosis to BMT (> or = 120 days, P = .08).
Conclusion: Bu-Cy120 is a well-tolerated preparation, but results in a high risk of relapse for children with AML in first CR. This high risk of relapse is not observed when the dose of cyclophosphamide is increased to 200 mg/kg.