Abstract A controlled multicenter trial was conducted in France over a period of 5 years in 40 patients with thrombotic thrombocytopenic purpura (TTP) to try to determine the best treatment of this disease. Patients presenting eligibility criteria were centrally randomized into two therapeutic groups: Group (A) daily plasma exchanges with a solution of fresh frozen plasma (15 mL/kg) in albumin (45 mL/kg) versus Group (B) daily transfusions of fresh frozen plasma (15 mL/kg), both accompanied by intravenous antiplatelet therapy. When the treatment was started early after the first symptoms of TTP, there was no significant difference among plasma exchange and plasma transfusion. But the later the treatment was started, the lesser was the initial therapeutic response and the poorer the prognosis. Furthermore the excessive use of plasma transfusions in initially unresponsive patients could be dangerous, allowing in some cases an irreversible development of the disease. The overall survival rate in group A was 85% with 80% complete remissions and 15% deaths opposed to a 57% survival rate with 52% complete remission and 43% deaths in group B. Thus plasma exchanges with fresh frozen plasma, if started early and performed daily in sufficient amounts, appear to be undoubtedly safer than plasma transfusions.