Busulfan dose intensity and outcomes in reduced-intensity allogeneic peripheral blood stem cell transplantation for myelodysplastic syndrome or acute myeloid leukemia

Abstract

Comparisons of myeloablative conditioning versus reduced-intensity conditioning (RIC) have demonstrated a tradeoff between relapse and toxicity. Dose intensity across RIC regimens vary and may affect treatment outcomes. In this retrospective analysis, we investigated the effect of i.v. busulfan dosing (total dose 3.2 mg/kg versus 6.4 mg/kg) in RIC regimens that combined fludarabine and busulfan on outcomes in patients who were undergoing hematopoietic stem cell transplantation (HSCT) for myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML). A total of 217 consecutive patients with MDS or AML underwent first busulfan and fludarabine RIC peripheral blood stem cell transplantation from well-matched related or unrelated donors at our institutions between 2004 and 2009. Of the 217 patients, 135 patients received Bu1 (3.2 mg/kg of busulfan) and 82 patients received Bu2 (6.4 mg/kg of busulfan), both with daily fludarabine (30 mg/m(2)/day for 4 days). The choice of RIC regimen was based on temporal institutional standard, enrollment on protocols, and physician choice. Patients had similar characteristics with a few notable differences: Patients who received Bu1 were younger (median age 61 versus 64 years, P < . 001), received more single-antigen mismatched unrelated grafts (14.1% versus 1.2%, P < . 001), received more sirolimus-based graft-versus-host disease (GVHD) prophylaxis regimens (63% versus 45%, P < .0001), received less antithymocyte globulin for GVHD prophylaxis (0% versus 22%, P < .001), and had less enrollment on a clinical trial that used prophylactic rituximab for the prevention of chronic GVHD (2.2% versus 11.0%, P = .011). Clinical disease status was similar between the groups. Median follow-up for survivors was 4.4 years for Bu1 and 3.2 years for Bu2. Because of the differences in characteristics, the 2 groups were compared with the adjustment of a propensity score that predicted Bu2 to account for measured differences. The day +200 cumulative incidence rates of grades II to IV acute GVHD (Bu1, 17%, versus Bu2, 8.5%; hazard ratio [HR], .56; 95% confidence interval [CI], .22 to 1.41; P = .22) or grades III to IV acute GVHD (Bu1, 6.7%, versus Bu2, 4.9%) were not different. The 2-year cumulative incidence of chronic GVHD was not significantly different between Bu1 and Bu2 (41.5% versus 28%, respectively; HR, .70; CI, .42 to 1.17; P = .09). Two-year nonrelapse mortality rates were similar for Bu1 and Bu2 (8.9% versus 9.8%, respectively; HR, .80; CI, .29 to 2.21; P = .67). Two-year progression-free survival and overall survival were also similar between Bu1 and Bu2 (progression-free survival: 40.6% versus 39.3%, respectively; HR, .82; CI, .57 to 1.30; P = .33; and overall survival: 47.4% versus 48.8%, respectively; HR, .96; CI, .64 to 1.44; P = .85). Subset analysis defined by clinical disease and cytogenetic risk with the propensity risk score applied suggest that in patients with high clinical disease risk and nonadverse cytogenetics, the higher dose busulfan RIC regimen may be of marginal benefit (2-year progression-free survival: HR, .54; CI, .29 to 1.03; P = .062). For the majority of patients with MDS or AML undergoing busulfan and fludarabine RIC peripheral blood stem cell transplantation, however, the dose of busulfan (3.2 mg/kg versus 6.4 mg/kg) is not associated with significant differences in overall outcomes.