Thrombocytopenic purpura is a common hematologic abnormality occurring in individuals infected with the human immunodeficiency virus, HIV-1. Of the nearly two million people infected with HIV-1, approximately 11% have platelet counts of less than 100,000/mm3. If no etiology other than HIV-1 infection can be found for the thrombocytopenia, the syndrome is referred to as HIV-1-associated thrombocytopenia (HAT). Steroids lead to an improvement in the platelet count in 60% to 80% of effected individuals but the majority of those who respond cannot maintain a normal platelet count when steroids are withdrawn. Furthermore concern over chronic steroid therapy in HIV-1-infected individuals has led to the investigation of other forms of treatment for this syndrome. This report describes the experience at Duke University Medical Center with eight patients who developed HAT and subsequently underwent splenectomy. In this group there was 1 complete response, 5 partial responses, and 2 patients who did not respond. There were no perioperative deaths and minimal perioperative morbidity. No evidence for the progression of HIV-1 infection in asymptomatic patients after splenectomy to AIDS related complex (ARC) or to the acquired immune deficiency syndrome (AIDS) was seen. In addition no increase in the susceptibility to infections by encapsulated organisms as a result of splenectomy was observed after a mean follow-up of 13.25 months. A review of 79 other cases reported in the literature suggests a higher response rate than that observed in our patients. Reasons for this discrepancy are discussed and an algorithm defining the role of splenectomy in the management of HAT is presented.