Autoimmune phenomena may be detected in patients with aplastic anemia. In the etiology of aplastic anemia, it is not known whether these processes are primary factors (causative), perpetuating factors (following other causative factors), or merely epiphenomena. Response to immunosuppressive therapy would suggest a causative or perpetuating role for autoimmunity in development of this disease. We have reviewed trials of immunosuppressive therapy, including the use of antilymphocyte globulin, cyclophosphamide, high-dose corticosteroids, cyclosporine, and apheresis, as well as combinations of these agents, for treatment of aplastic anemia. Responses occur in 40-70% of patients who receive various preparations of antilymphocyte globulin. Responses to other immunosuppressive regimens have been less frequent. Responses to all immunosuppressive regimens are usually incomplete, and late complications, such as relapse, paroxymal noctural hemoglobinuria, or leukemia, are being reported with increasing frequency. For all treatments, potential mechanisms of action other than immune suppression are possible. HLA-matched sibling-donor bone marrow transplantation is the treatment of choice for children with severe aplastic anemia. For children without a matched sibling, antilymphocyte globulin is the best current therapy. No treatment has been shown to alter the long-term course of mild aplastic anemia. Other immunosuppressive agents and transplant regimens should be considered experimental. The therapeutic value of these modalities can only be established by well-monitored trials performed at centers with special clinical and laboratory expertise.