My colleagues and I were the first to point out that as more patients survived their initial episode of thrombotic thrombocytopenic purpura, relapses were not rare but rather might afflict more than one third of patients [1]. Our experience impels me to question on two counts the value of splenectomy purported by Crowther and colleagues [2].

First, we have seen several patients with a pattern of relapses over a few years who spontaneously cease experiencing relapses. Thus, any intervention (or nonintervention) in such patients, who were preselected for frequent relapses, would seem to result in a lower rate of subsequent relapses. Pertinent examples are a middle-aged woman and a 60-year-old man who each had four episodes of thrombotic thrombocytopenic purpura over 6 years. Neither patient had any further episodes over several years. If splenectomy had been done for the first patient, it would seem to have been highly effective in preventing relapses. The second patient actually underwent splenectomy during the first episode!

Second, I must reemphasize the generally more benign course and outcome of relapsed thrombotic thrombocytopenic purpura [1]. We have treated more than 15 patients with relapse through more than 30 relapse episodes. In contrast to the 20% mortality rate we observed with first-episode thrombotic thrombocytopenic purpura, in 20 years we have witnessed no deaths or any long-lasting residue resulting from relapse. In fact, episodes of relapse almost always respond to treatment more quickly than do first episodes. One explanation is that patients having relapse are attuned to symptoms of recurrence and thus seek medical attention earlier. Further, by virtue of surviving their first episode, these patients have demonstrated responsiveness to therapy.

Data supporting the efficacy of splenectomy in preventing relapses are unconvincing. The procedure entails its own morbidity and risks, and I cannot recommend it for relapses of thrombotic thrombocytopenic purpura.