Related Versus Unrelated Donor Transplantation for High Risk (HiRi) Acute Myeloid Leukemia (AML) in First Complete Remission (CR1).

Allogeneic hematopoietic stem cell transplantation (allo-SCT) from an HLA-identical sibling donor (SIB) is considered the preferred postremission therapy for younger patients with HiRi AML in CR1. The role of allo-SCT from volunteer unrelated donors (VUDs) is less clear and randomized controlled trials addressing this issue do not exist. We performed an observational landmark analysis on parallel cohorts of patients aged <60 years with AML in CR1 and HiRi cytogenetics who had been enrolled into the AML Cooperative Group (AMLCG) 1999 trial. 2347 patients were evaluable for the present update. 243/2347 patients were <60 years of age and had unfavorable cytogenetics [complex karyotype, -5/5q-, -7/7q, abnormal 3q21/3q26, or abnormal 11q23]. 111 of the 243 patients (45.7%) achieved CR1 after double induction with either TAD-HAM (123) or HAM-HAM (120). Of those 111 patients, 21 received an allo-SCT from a SIB while in first or higher CR (16 in CR1) and 34 were transplanted from a matched VUD in first CR (25; ≥6/10 HLA match: n=8, ≥8/10 HLA match: n=17) or higher CR (9). Of the 56 CR1 patients receiving conventional postremission therapy, 37 with relapse-free survival (RFS) ≥90 days were eligible for the control cohort. The 90-day landmark was set in order to account for the median time from entry into CR1 to transplant in the two allo-SCT groups. Median age in the SIB allo-SCT group was 46 years (range: 25–58), in the VUD cohort 45 (21–58), and in the control cohort 49 (22–59). Overall survival (OS) at 2 years for the SIB cohort was 68% (95% CI: 43–92%), for the VUD cohort 59% (95% CI: 38–81%), and for the control group 28% (95% CI: 13–44%). Projected RFS and OS at 4 years for the SIB allo-SCT cohort were 23% and 45%, for the VUD allo-SCT cohort 39% and 44%, and for the control group 14% and 25%, respectively (RFS: p=0.02; OS: p=0.03; log-rank test for differences between groups). The 4 year OS after any allo-SCT (SIB or VUD) is 48% versus 25% in the control group (p=.009). In a Cox regression model, survival end points in HiRi AML patients in CR1 were influenced by patient age and type of postremission therapy only. Hazard ratios (HRs) for RFS were 0.6 (95% CI: 0.3–1.3: p=.19) for SIB allo-SCT and 0.4 (0.2–0.8; p<.01) for VUD allo-SCT, respectively, versus the control group in a model including patient age. HRs in OS were 0.6 (0.2–1.3; p=.16) for SIB allo-SCT and 0.4 (0.2–0.9; p=.03) for VUD allo-SCT. Despite the limitations inherent to this type of analysis and the limited numbers of patients, we consider the data as best available controlled evidence that patients <60 years of age with HiRi AML in CR1 benefit from allo-SCT as postremission consolidation with the results of VUD allo-SCT in this setting being at least comparable, if not superior to those achieved with SIB allo-SCT.