Acute hepatic allograft rejection occurs in approximately 50% to 60% of the patients undergoing liver transplantation. In this study, we compared the rate of acute rejection in liver transplant recipients randomized in a double-blind comparative study to treatment with mycophenolate mofetil (MMF) or azathioprine (AZA), both in combination with cyclosporine and corticosteroids. Five hundred sixty-five primary liver transplant recipients were randomly assigned to treatment with MMF, 1 g twice daily intravenously followed by 1.5 g twice daily orally (n = 278), or AZA, 1.0 to 2.0 mg/kg/d intravenously followed by oral administration (n = 287), in combination with cyclosporine and corticosteroids. Patients were followed up for at least 1 year, and efficacy analysis was based on intent-to-treat methods. Acute rejection was defined according to the Banff histological criteria. The two study groups were balanced for demographic and clinical baseline characteristics. The incidence of acute rejection or graft loss was 47.7% in the AZA patients and 38.5% in the MMF patients (P <.03). The incidence of biopsy-proven and treated rejection censoring for graft loss was 40.0% in the AZA group versus 31.0% in the MMF group (P <.06). Steroid-resistant rejection requiring treatment with either OKT3 or antithymocyte globulin occurred in 8.2% of AZA patients versus 3.8% in MMF patients (P <.02). Patient and graft survival rates at 1 year posttransplantation were 85.4% in the AZA group and 85.3% in the MMF group (P = not significant). MMF was superior to AZA in preventing acute rejection in the first 6 months posttransplantation. MMF and AZA were equivalent in preventing graft loss at 1 year, and the safety profiles between the two immunosuppressive agents were similar.