Hepatic allograft rejection remains a major cause of morbidity related to the need for increased immunosuppression and continues to be a principal cause of late failure of the graft. Hepatic allograft rejection is defined on the basis of morphologic findings; cellular rejection is defined as portal or periportal hepatitis with nonsuppurative cholangitis or endotheliitis (or both), and ductopenic rejection is defined as loss of interlobular and septal bile ducts, typically in at least 50% of the portal tracts. The overall incidence of episodes of cellular rejection, which usually occur within the first 2 weeks after liver transplantation, varies from 50 to 100%. Ductopenic rejection occurs in approximately 8% of patients who undergo initial liver transplantation and is usually diagnosed between 6 weeks and 6 months after transplantation. Induction and maintenance immunosuppression with triple-drug (cyclosporine, prednisone, and azathioprine) therapy and other combinations that include antilymphocyte preparations, however, has decreased the incidence of both cellular and ductopenic rejection. In patients experiencing episodes of cellular rejection, high-dose intravenously administered corticosteroid therapy yields the best response and is associated with a lower incidence of ductopenic rejection than is low-dose and orally administered corticosteroid therapy. The correlation between degree of biochemical liver dysfunction and presence of histologic rejection is minimal early after transplantation. Histologic severity of rejection, however, suggests which patients will require more immunosuppression and which patients may need antilymphocyte therapy for controlling the rejection episode. With the availability of new immunosuppressive agents, distinguishing patients at high risk for rejection is important. The goals for use of new immunosuppressive agents and regimens are to improve graft and patient survival, to decrease the incidence of cellular and ductopenic rejection, and to minimize side effects and complications.