Allogeneic bone marrow transplantation is the treatment of choice for many childhood leukemias. The donor of choicean HLA matched sibling-is only available about 30% of the time. Unrelated donors are an alternative choice. In this report, we describe the results of unrelated donor bone marrow transplants (BMT) in 50 children with leukemia (25 acute lymphoblastic leukemia [ALL], 3 acute myeloid leukemia [AML], 3 juvenile chronic myelogenous leukemia [JCML], 10 chronic myeloid leukemia [CML]) or myelodysplastic syndrome (MDS; 9). The median age of the 31 male and 19 female patients was 9 years (range 2 to 18). Only 13 patients were serologically matched at HLA-A, B, DR, and DO. with their donors; 6 of these were reactive in mixed lymphocyte culture. The other 37 patients were mismatched for one (36 patients) or more (1 patient) HLA antigens. Pretransplant conditioning included cytosine arabinoside, cyclophosphamide, fractionated total body irradiation (TBI) (with lung, liver, and more recently, kidney shielding), and methylpred-nisolone. High-risk patients also received busulfan. Graftversus-host disease (GVHD) prophylaxis consisted of T-cell depletion with IgM monoclonal antibody T10B9 plus complement and posttransplant cyclosporine-A. Forty-nine patients (98%) engrafted. Median times to greater than 500 polymorphonuclear leukocytes (PMN)/μL and greater than 25,000 platelets/μL were 18 and 20 days, respectively. Acute GVHD ≥ grade II occurred in 16 patients (33%); 13 (81%) of these died. Chronic GVHD developed in 30 of 40 patients at risk, but was extensive in only 5. Event-free survival (EFS) for all patients was 44% ± 7% (median follow-up was 49 months), and overall survival was 50 ± 7%. Patients with low-risk disease (ALL or AML in first or second remission and CML in chronic phase) had a better EFS than children with high-risk disease (60% v 34%, P = .07). There was no significant difference in EFS between patients who were serologically matched with their donors (46%) and those who were partially mismatched (43%) (P = .97). These data compare favorably with published reports for children transplanted with HLA-matched sibling donors and should encourage earlier consideration of unrelated donor BMT in children with leukemia or myelodysplasia.