Improved survival after bone marrow transplantation for early leukemia using busulfan-cyclophosphamide and individualized prophylaxis against graft-versus-host disease: a long-term follow-up.

To minimize immunosuppression, allow a graft-versus-leukemia (GVL) effect, and reduce relapse incidence, 73 leukemic recipients of human leukocyte antigens-identical sibling marrow were given graft-versus-host disease (GVHD) prophylaxis based on the estimated risk of GVHD development. Methotrexate (MTX) monotherapy was given to patients with an estimated low risk of developing GVHD, whereas MTX + cyclosporine (CsA) was given to 'high-risk' patients. After engraftment, CsA was discontinued, and weekly MTX was reinstituted and given until 3 months post-bone marrow transplant. Conditioning consisted of busulfan (BU) + cyclophosphamide (CY) (n = 35) or CY + total body irradiation (TBI) (n = 38). Retrospective controls were given CY + TBI and MTX + CsA (n = 39). The median observation time was 5 yr 11 months. Chronic GVHD increased to 53% in the individual BU + CY group and 46% in the individual CY + TBI group, compared to 25% in the control group (p = 0.05). This increase was restricted to the limited form. The actuarial relapse incidence decreased to 20% in the individual BU + CY group, compared to 52% in the control group, p = 0.03. In the individual CY + TBI group, the relapse incidence was 44% (n.s. versus controls, p = 0.04 versus individual BU + CY). The 5-yr relapse-free survival (RFS) in the individual BU + CY group was 66%, in the control group, 41% (p = 0.07), and in the individual CY + TBI group, 45% (p = 0.1 versus individual BU + CY). Patients with early leukemia in the individual BU + CY group had a RFS of 83%, compared to 44% in the control group (p = 0.02) and 42% in the individual CY + TBI group (p = 0.01). In the multivariate analysis, advanced leukemia beyond first complete remission and first chronic phase and conditioning with CY + TBI were correlated to poor RFS. In summary, the individualized prophylaxis itself did not reduce the relapse incidence. However, in patients with early leukemia conditioning with BU + CY, our method of individualizing the GVHD prophylaxis might be of value, since this group had the best RFS in this study.