Search

Hematopoietic cell transplantation (HCT) and prolonged chemotherapy are standard postremission strategies for adult acute lymphoblastic leukemia in first complete remission, but the optimal strategy remains controversial. There are no randomized trials of allogeneic HCT. In the present study, updated individual patient data were collected and analyzed from studies with information on availability of matched sibling donor (used to mimic randomization) and from randomized trials of autograft versus chemotherapy. Data from 13 studies including 2962 patients, excluding Philadelphia chromosome–positive patients, showed a survival benefit for having a matched sibling donor for patients < 35 years of age (OR = 0.79; 95% CI, 0.70-0.90, P = .0003) but not for those ≥ 35 years of age (OR = 1.01; 95% CI, 0.85-1.19, P = .9; heterogeneity P = .03) because of the higher absolute risk of nonrelapse mortality for older patients. No differences were seen by risk group. There was a trend toward inferior survival for autograft versus chemotherapy (OR = 1.18; 95% CI, 0.99-1.41; P = .06). No beneficial effect of autografting was seen compared with chemotherapy in this analysis. We conclude that matched sibling donor myeloablative HCT improves survival only for younger patients, with an absolute benefit of approximately 10% at 5 years. Improved chemotherapy outcomes and reduced nonrelapse mortality associated with allogeneic HCT may change the relative effects of these treatments in the future.

Identifying risk factors that lead to graft failure may reduce morbidity and mortality after bone marrow transplantation (BMT) for hematologic malignancies. We evaluated engraftment of all patients with acute myelogenous leukemia (AML), chronic myelogenous leukemia (CML) and myelodysplastic syndrome (MDS) receiving an unmanipulated marrow allogeneic BMT at the Detroit Medical Center from 1987 to 1992 using a busulfan, cyclophosphamide +/- cytarabine preparative regimen. Three of 118 patients had graft failure (2.5%; (95% confidence interval (CI) 0.7%, 6.4%). Graft failure was high in patientsor = 15 years with 3 of 12 patients with failure (25.0%) compared with 0 of 106 patients15 years (p = 0.002). Failure to engraft was not seen in HLA-identical (related or unrelated) donor transplants (0 of 103) whereas 3 of 15 HLA-mismatched donors failed (p = 0.003). Patient diagnosis, locus of HLA-mismatch, cytarabine in the preparative regimen, marrow cell dose and the relative reactive index (RRI) were not significant factors. Altered busulfan kinetics secondary to young age was probably not a major factor since 8 of 8 HLA-identical donor transplants engrafted in children. These findings demonstrate that patients receiving an unmanipulated marrow graft using busulfan-containing regimens were at a high risk for graft failure only if they wereor = 15 years of age and had an HLA-mismatched donor. More immunosuppressive preparative regimens, possibly including total body irradiation, should be considered to prevent potential graft failure in children.