Current issues of ACP Journal Club are published in Annals of Internal Medicine


Plasma exchange was more effective than plasma infusion for thrombotic thrombocytopenic purpura

ACP J Club. 1991 Nov-Dec;115:81. doi:10.7326/ACPJC-1991-115-3-081

Source Citation

Rock GA, Shumak KH, Buskard NA, et al. Comparison of plasma exchange with plasma infusion in the treatment of thrombotic thrombocytopenic purpura. N Engl J Med. 1991 Aug 8;325:393-7.



To determine whether thrombotic thrombocytopenic purpura is more effectively treated with fresh-frozen plasma exchange or infusion.


Randomized, controlled trial, with assessment after 9 days and 6 months.


16 medical centers in Canada.


All patients diagnosed with thrombotic thrombocytopenic purpura who had a platelet count of < 100 × 109/L and microangiopathic hemolytic anemia were considered. Those with contraindications to plasma infusion or identified causes of the condition (e.g., disseminated intravascular coagulation) were excluded. 102 patients with an average platelet count of 23.3 ± 16.9 × 109/L (SD) were randomized. All patients were included in intention-to-treat analysis.


51 patients were assigned to plasma exchange for ≥ 7 procedures over the first 9 hospital days, with 1.5 times the predicted volume of plasma exchanged on the first 3 days and 1.0 times the predicted volume exchanged thereafter. 51 patients were assigned to plasma infusion of 30 mL/kg body weight in the first 24 hours and 15 mL/kg each day thereafter; diuretics were used if required. Patients could cross over if initial treatment failed. All patients received oral dipyridamole (400 mg/d) and aspirin (325 mg/d) for a minimum of 2 weeks.

Main outcome measures

Mortality and treatment response after initial treatment and at 6 months; complete response to treatment was defined as a platelet count of > 150 × 109/L for 2 consecutive days with no deterioration in neurologic status after 9 days.

Main results

After 1 cycle of either plasma exchange or infusion, complete responses were noted in 24 of 51 (47%) and 13 of 51 (25%) patients, respectively (P = 0.025). {This absolute risk improvement (ARI) of 22% means that 5 patients would need to be treated (NNT) with plasma exchange (rather than infusion) for 1 additional complete response to occur, 95% CI 3 to 34; the relative risk improvement (RRI) was 85%, CI 8% to 224%.}* 2 patients (4%) in the exchange group died compared with 8 patients (16%) in the infusion group { P = 0.05}.* {This absolute risk reduction (ARR) of 12% means that 8 patients would need to be treated with plasma exchange for 1 additional patient to survive, CI 4 to 386; the relative risk reduction [RRR] was 75%, CI 2% to 94%.}* 31 of 51 plasma infusion patients and 0 of 51 plasma exchange patients were then crossed over to the other treatment. After 6 months, 40 of 51 patients (78%) in the exchange group compared with 25 of 51 patients (49%) initially assigned to infusionwere treatment successes (P = 0.002); {ARI 29%, NNT 3, CI 2 to 9; RRI 60%, CI 19% to 124%}*. 11 patients assigned to exchange died compared with 19 assigned to infusion (22% vs 37%, { P = 0.08; ARR 15%, CI 3% to 33%}*).


Plasma exchange was more effective than plasma infusion as the initial treatment for thrombotic thrombocytopenic purpura.

Source of funding: National Health Research and Development Program.

Address for article reprint: Dr. G.A. Rock, Canadian Apheresis Study Group, 206-435 St. Laurent Boulevard, Ottawa, Ontario KlK 2Z8, Canada.

*Numbers calculated from data in article.


The study reported by Rock and colleagues represents the first randomized, controlled clinical trial comparing therapeutic approaches for thrombotic thrombocytopenic purpura. The authors compared two treatments that have been reported to be effective: plasma exchange with fresh-frozen plasma and plasma infusion with fresh-frozen plasma. Both approaches have been reported to be associated with high response rates, but previous reports have been either anecdotal or small, preliminary trials. This study is important because it included most patients with thrombotic thrombocytopenic purpura seen in Canada during the 7-year period of the trial. The comparison of the two treatments clearly shows that plasma exchange is superior both in complete response rate following 1 cycle of therapy and in overall survival. The results also confirm reports indicating the long-term outcome of plasma exchange is excellent. A subsequent report has been published showing that > one third of the patients in this trial who survived an acute episode of thrombotic thrombocytopenic purpura had at least one relapse during 10 years of follow-up (1).

Only patients with less severe forms of thrombotic thrombocytopenic purpura were eligible, and those with findings such as anuria were excluded. Although this would probably not affect the differences between the outcomes of the treatments, it may suggest a more favorable outcome to the disease than would occur in practice.

Unfortunately, the results of the study do not shed any light on the cause of thrombotic thrombocytopenic purpura. Plasma exchange has been advocated in order to remove possible toxic substances from the plasma. In this study, plasma exchange may also have delivered possible helpful substances better than plasma infusion, because during the first treatment cycle, exchange delivered a mean of 21.5 liters of fresh-frozen plasma, compared with a mean of only 6.7 liters by infusion. It is now clear that the first line of therapy for thrombotic thrombocytopenic purpura should be plasma exchange using fresh-frozen plasma. The minimal toxicity of this therapy and the improvement in long-term survival are important benefits. The authors are to be congratulated for their contribution to an area that has remained problematic and filled with anecdotal reports.

Ronald M. Bukowski, MD
Cleveland Clinic Cancer CenterCleveland, Ohio, USA.


1. Shumak KH, Rock GA, Nair RC, and the Canadian Apheresis Group. Late relapses in patients successfully treated for thrombotic thrombocytopenic purpura. Ann Intern Med. 1995;122:569-72.