We compared 15 patients with systemic lupus erythematosus (SLE) treated with splenectomy for thrombocytopenic purpura and/or hemolytic anemia to 15 similar SLE patients treated only medically. There was no significant difference between the splenectomized and the nonsplenectomized patients when their entire course, as well as the presplenectomy and postsplenectomy or their equivalent control periods, were compared by means of an overall severity index. Splenectomized patients, however, had a significantly higher frequency of cutaneous vasculitis after splenectomy than in their own presplenectomy period and a significantly higher frequency of cutaneous vasculitis than the nonsplenectomized patients. Serious infections were more frequent in the postsplenectomy period than in an equivalent period in the nonsplenectomized patients. Splenectomy produced only short-term benefit in the management of hemocytopenic episodes in SLE and seems only warranted as an emergency procedure in patients unresponsive to medical treatment.