We tested the hypothesis that oral beclomethasone dipropionate (BDP) would control gastrointestinal graft-versus-host disease (GVHD) in patients with anorexia, vomiting, and diarrhea. Patients were randomized to prednisone for 10 days and either oral BDP 8 mg/d (n = 62) or placebo (n = 67) tablets for 50 days. At study day 10, prednisone was rapidly tapered while continuing study drug. On an intent-to-treat basis, the risk of GVHD-treatment failure was reduced for the BDP group at study day 50 (hazard ratio [HR] 0.63, 95% confidence interval [CI] 0.35-1.13) and at 30 days follow-up (HR 0.55, 95% CI 0.32-0.93). Among patients eligible for prednisone taper at study day 10, the risk of GVHD-treatment failure was significantly reduced at both study days 50 and 80 (HR 0.39 and 0.38, respectively). By day 200 after transplantation, 5 patients randomized to BDP had died compared with 16 deaths on placebo, a 67% reduction in the hazard of mortality (HR 0.33, P = .03). In 47 recipients of unrelated and HLA-mismatched stem cells, mortality at transplantation day 200 was reduced by 91% in the BDP group compared with placebo (HR 0.09, P = .02). The survival benefit was durable to 1 year after randomization. Oral BDP prevents relapses of gastrointestinal GVHD following tapering of prednisone; survival is statistically significantly better among patients receiving BDP.