The staphylococcal cell-wall protein known as protein A has been explored as a therapeutic modality in the treatment of cancer and allied diseases. Protein A binds the Fc fragment of IgG 1, 2 and 4, and preferentially binds to IgG incorporated into immune complexes. Early investigators focused on the immune-suppressive effects of immune complexes in cancer and, based on in vitro experiments, postulated that clearance of immune complexes in vivo would permit effective immune clearance of cancer cells. A large clinical trial of the perfusion of cancer patient plasma over protein A was subsequently undertaken. Results were generally disappointing, with no complete remissions and overall response rates of 22%. Response rates for Kaposi's sarcoma (39%) and breast adenocarcinoma (26%) were somewhat encouraging, and further clinical trials in these disorders are ongoing. More impressive have been the responses to protein A perfusion in immune thrombocytopenia and hemolytic-uremic syndrome. Using a protein A-silica device, Snyder et al. reported responses in 42% of immune thrombocytopenia patients, with mean increases in platelet count from 27 x 10(9)/l to 120 x 10(9)/l. On the basis of these results, the protein A-silica column was approved by the United States Food and Drug Administration for treatment of immune thrombocytopenia. Equally encouraging are reports of an overall 59% response rate in cancer chemotherapy-related hemolytic-uremic syndrome. Reported toxicities include fever, chills, hypotension, dyspnea and musculoskeletal pain. With rare exceptions, these reactions are easily treated and do not result in cessation of therapy. Unfortunately, the mechanism of action of plasma perfusion over protein A is very unclear. The best available evidence would point to an immunomodulatory role, manifested by stimulation of an anti-idiotype response in immune thrombocytopenia. A better understanding of how protein A immunoadsorption alters the immune response will be necessary to permit optimum use of this therapy.