We have analyzed the long term outcome of 197 patients who were treated for grade II to IV acute graft-versus-host disease (GVHD) following histocompatible allogeneic bone marrow transplantation (BMT). Of 469 recipients of sibling donor allografts performed at our center between January, 1979 and October, 1987, 197 patients (42%) developed greater than or equal to grade II acute GVHD at a median of 38 days (range 9 to 98 days) post-BMT. After treatment with corticosteroids (n = 160) or other immunosuppressive therapies (n = 37), 72 patients (41% +/- 8%; 95% confidence interval [CI]) achieved complete and continuing resolution of acute GVHD after a median of 21 days of therapy. Sixty-one patients required additional immunosuppressive therapy with high dose methylprednisolone, antithymocyte globulin (ATG)/steroids, or other therapies because of refractory or progressive symptoms of acute GVHD. Seven of these 61 patients eventually obtained complete and continuing remission after 13 to 57 days (median 50) of secondary treatment. The overall rate of chronic GVHD was 70% +/- 16%; 95% CI following grade II to IV acute GVHD. Twenty-five of the 197 patients never developed chronic GVHD, resulting in a Kaplan-Meier projection of 30% +/- 8% (95% CI) cure of moderate/severe acute GVHD. Analysis of clinical features associated with complete response (CR) to acute GVHD therapy identified more favorable responses to therapy in patients without either liver or skin involvement, patients with acute lymphoblastic leukemia, and donor/recipient pairs other than male patients with female donors. Older recipient age was not associated with more resistance to GVHD treatment. CR to GVHD treatment was associated with significantly better 5-year survival: 51% +/- 14% versus 32% +/- 11% for patients with therapy resistant acute GVHD (P = .004). GVHD was a major contributing cause of death in 49 of the 90 patients who died and was often complicated by infection or interstitial pneumonitis. Control of acute GVHD through immunosuppressive therapy did not affect the risk of leukemic relapse after transplantation.