IDENTICAL-TWIN BONE-MARROW TRANSPLANTS FOR LEUKEMIA

Objective: To compare outcomes of identical-twin with HLA-identical sibling bone marrow transplants for leukemia. Design: Matched-pair analysis comparing relapse, treatment-related mortality, and leukemia-free survival in cohorts matched for disease and variables correlated with transplant outcome, with and without adjustment for graft-versus-host disease. Setting: 163 institutions worldwide between 1978 and 1990, reporting to the International Bone Marrow Transplant Registry. Participants: 103 identical-twin transplants: 24 for acute lymphoblastic leukemia (ALL) in first remission, 45 for acute myelogenous leukemia (AML) in first remission, and 34 for chronic myelogenous leukemia (CML) in first chronic phase. Results were compared with those in 1030 concurrent HLA-identical sibling transplants matched for prognostic factors. Results: Three-year probabilities of relapse after identical-twin compared with HLA-identical sibling transplants were as follows: ALL, 36% (95% CI, 17% to 55%) compared with 26% (CI, 20% to 32%); AML, 52% (CI, 37% to 67%) compared with 16% (CI, 12% to 20%); and CML, 40% (CI, 23% to 57%) compared with 7% (CI, 4% to 10%). Increased relapse risks in AML and CML persisted after adjusting for graft-versus-host disease (relative risk, 3.1 [CI, 1.9 to 5.1] and 5.5 [CI, 2.8 to 11.0], respectively). Although twins had less treatment-related mortality than HLA-identical siblings, leukemia-free survival was similar. Three-year leukemia-free survival probabilities after twin compared with HLA-identical sibling transplants were as follows: ALL, 57% (CI, 37% to 77%) compared with 58% (CI, 52% to 64%); AML, 42% (CI, 27% to 57%) compared with 55% (CI, 50% to 60%); and CML, 59% (CI, 42% to 76%) compared with 61% (CI, 56% to 66%). Conclusions: Identical-twin transplants in AML and CML are associated with increased relapse risk compared with HLA-identical sibling transplants. A similar trend was observed in ALL but was not statistically significant. Increased relapse in twin transplants is not explained by lack of graft-versus-host disease. Leukemia-free survival after twin and HLA-identical sibling transplants is similar because increased relapse in twins is offset by decreased treatment-related mortality.