A phase HI comparative trial of tacrolimus- vs. cyclosporine-based graft-vs.-host disease (GVHD) prophylaxis for human leukocyte antigen (HLA)-identical sibling bone marrow transplantation showed less GVHD but poorer survival in the tacrolimus arm. To test the comparability of the two treatment arms with respect to baseline survival prognosis, a matched control study using exclusively cyclosporine-treated patients from the International Bone Marrow Transplant Registry (IBMTR) database was performed. Controls were matched (2:1) based on age (within 5 years), disease, and pretransplant disease status. Two-year survival for tacrolimus-treated clinical trial patients was similar to that of their cyclosporine-treated matched controls (27 and 24%, respectively), and 2-year survival of the cyclosporine-treated clinical trial patients was similar to that of their cyclosporine-treated matched IBMTR controls (42 and 45%, respectively). Consistent with the clinical trial results, the cyclosporine-treated IBMTR controls matched to the tacrolimus group had significantly poorer 2-year survival than the cyclosporine-treated IBMTR controls matched to the cyclosporine group (24 and 45%, respectively; p < 0.01). No significant difference was seen in GVHD between the cyclosporine-treated clinical trial patients and their matched controls; however, the tacrolimus-treated clinical trial patients had significantly less GVHD than their cyclosporine-treated IBMTR controls (p < 0.01). These results support the hypothesis that the survival difference in the phase III trial resulted from an imbalance in the underlying risk factors for death in the two groups rather than from the randomized immunosuppressive regimen.