Hepatic complications are a major cause of illness and death after bone marrow transplantation. To confirm the results of a pilot study that indicated that ursodiol prophylaxis could reduce the incidence of veno-occlusive disease of the liver. Randomized, double-blind, placebo-controlled study. Tertiary care teaching hospital. 67 consecutive patients undergoing transplantation with allogeneic bone marrow (donated by a relative) in whom busulfan plus cyclophosphamide was used as the preparative regimen and cyclosporine plus methotrexate was used to prevent graft-versus-host disease. Before the preparative regimen was started, patients were randomly assigned to receive ursodiol, 300 mg twice daily (or 300 mg in the morning and 600 mg in the evening if body weight was > 90 kg), or placebo. Patients were prospectively evaluated for the clinical diagnosis of veno-occlusive disease, the occurrence of acute graft-versus-host disease, and survival. The incidence of veno-occlusive disease was 40% (13 of 32 patients) in placebo recipients and 15% (5 of 34 patients) in ursodiol recipients (P = 0.03). Assignment to placebo was the only pretransplantation characteristic that predicted the development of veno-occlusive disease. The most significant predictor of 100-day mortality was the diagnosis of veno-occlusive disease. The difference in actuarial risk for hematologic relapse in patients with chronic myelogenous leukemia and nonhepatic toxicities between the two groups was not statistically significant (13% in the ursodiol group and 20% in the placebo group; P > 0.2). Ursodiol prophylaxis seemed to decrease the incidence of hepatic complications after allogeneic bone marrow transplantation in patients who received a preparative regimen with busulfan plus cyclophosphamide.