We report the results of a controlled study in which BMT patients with moderate/severe acute graft-versus-host disease (GVHD) who responded to primary treatment with corticosteroids were prospectively randomized to short versus long taper of their steroid doses. Thirty patients with moderate/severe acute GVHD who responded by 14 days were eligible for random assignment of their steroid tapering schedule. Patients in the short taper group received a total PRED dose of 2275 mg/m2 over 86 days, whereas those in the long taper group received 6300 mg/m2 over 147 days. Patients in the long taper group achieved resolution of acute GVHD after a median of 30 days of therapy (range 6-30), whereas those receiving the short taper resolved after a median of 42 days (12-74) (P = 0.01). After 8 weeks of therapy, only 2 of 13 evaluable long taper and 3 of 13 short taper patients still had active GVHD. The median PRED dose required to achieve complete resolution of acute GVHD was not different between the two groups: 1300 mg/m2 for the long taper patients and 1800 mg/m2 for the short taper patients. Importantly, the incidence of chronic GVHD and survival at 6 months was similar in the 2 groups. The incidence of steroid-related complications was similar, as well. This study suggests that the rapid administration of high-dose PRED to a cumulative dose of 2000 mg/m2 might lead to complete and prompt resolution of acute GVHD in the majority of patients and that rapid PRED taper might provide a mechanism for minimizing steroid-related morbidity. Further investigation and formal studies of the dose-response relationships and kinetics of steroid administration may lead to improvement in the management of acute GVHD.