Treatment-related mortality ranged from 5% to 42% in the donor groups compared with 3% to 27% in the no-donor group. Of 18 trials reporting RFS across all cytogenetic risk groups, the combined hazard ratio (HR) for overall RFS benefit with allogeneic transplant was 0.80, indicating a statistically significant reduction in death or relapse in CR1. Of the 15 trials reporting OS across all cytogenetic risk groups, the combined HR for OS was 0.90, again indicating a statistically significant reduction in death or relapse in CR1.
In subgroup analysis according to cytogenetic risk category, there was no RFS or OS benefit of allogeneic transplant for patients with good-risk AML (RFS: HR = 1.07; 95% confidence interval [CI], 0.83-1.38; P = .59; OS: HR = 1.06; 95% CI, 0.64-1.76; P = .81). However, a transplant benefit was seen for patients with intermediate (RFS: HR = 0.83; 95% CI, 0.74-0.93; P < .01; OS: HR = 0.84; 95% CI, 0.71-0.99; P = .03) or poor-risk cytogenetics (RFS: HR = 0.73; 95% CI, 0.59-0.90; P < .01; OS: HR = 0.60; 95% CI, 0.40-0.90; P = .01). The conclusion from this meta-analysis was that allogeneic transplant from a sibling donor in CR1 is justified on the basis of improved RFS and OS for patients with intermediate- or poor-risk, but not good-risk, cytogenetics.[Level of evidence: 2A]
An important caveat to this analysis is that induction and postremission strategies for AML among studies included in the meta-analysis were not uniform; nor were definitions of cytogenetic risk groups uniform. This may have resulted in inferior survival rates among chemotherapy-only treated patients. Most U.S. leukemia physicians agree that transplantation should be offered to AML patients in CR1 in the setting of poor-risk cytogenetics and should not be offered to patients in CR1 with good-risk cytogenetics.
The use of matched, unrelated donors for allogeneic BMT is being evaluated at many centers but has a very substantial rate of treatment-related mortality, with DFS rates less than 35%. Retrospective analysis of data from the International Bone Marrow Transplant Registry suggests that postremission chemotherapy does not lead to an improvement in DFS or OS for patients in first remission undergoing allogeneic BMT from an HLA-identical sibling.[Level of evidence: 3iiiA]
Autologous BMT yielded DFS rates between 35% and 50% in patients with AML in first remission. Autologous BMT has also cured a smaller proportion of patients in second remission.[17,18,19,20,21,22,23] Treatment-related mortality rates of patients who have had autologous peripheral blood or marrow transplantation range from 10% to 20%. Ongoing controversies include the optimum timing of autologous stem cell transplantation, whether it should be preceded by postremission chemotherapy, and the role of ex vivo treatment of the graft with chemotherapy, such as 4-hydroperoxycyclophosphamide (4-HC)  or mafosphamide, or monoclonal antibodies, such as anti-CD33. Purged marrows have demonstrated delayed hematopoietic recovery; however, most studies that use unpurged marrow grafts have included several cycles of postremission chemotherapy and may have included patients who were already cured of their leukemia.