Four patients with thrombotic thrombocytopenic purpura (TTP) were treated by plasma-exchange transfusion, three of whom recovered completely. Because previous reports in the literature describing exchange transfusion as treatment for TTP have demonstrated variable success rates, particular attention was given to "dose" and frequency of plasma exchange. Evans blue dye studies established a measure of "dose" under conditions of varying efficiency. Serum LDH activity was found to be diminished by plasma exchange, and the rate of return of serum LDH activity reflected residual disease activity. The magnitude of LDH activity reduction correlated with the adequacy of dose of plasma exchange and was an indicator for the need of repeated daily exchanges. Failure to obtain a spontaneous increment in platelet count also suggested the need for additional exchanges and/or larger dose of exchange. There is a need for a standard expression of dose of plasma exchange. Utilizing these markers (LDH, platelet count), it may be possible to improve the survival in TTP if adequate dose and frequency of plasma exchange are used.