A significant number (some 20%) of patients with thrombotic thrombocytopenic purpura do not respond to standard therapy and die. We reasoned that early identification of those who are likely to fail standard therapy would allow the rational introduction of other treatment modalities. To this end we prospectively evaluated 27 consecutive patients, examining serum LDH levels and platelet counts as markers of disease activity and as predictors of outcome. All patients were treated, according to a written protocol, with high volume plasma exchange (35 ml/kg), prednisone, aspirin, and persantine. Twenty-one of the 27 patients (78%) are alive following therapy. Initial LDH and platelet values did not distinguish between the survivors and nonsurvivors. However, by day 3 of therapy both LDH levels and platelet counts differed significantly between the two groups. Mean day 3 LDH level for survivors was 364 units/L, and for nonsurvivors, 891 units/L (P < 0.005). Mean day 3 platelet count for survivors was 119,000/microL, and for nonsurvivors, 46,000/microL (P < 0.005). Receiver Operator Characteristic curves were constructed and assessed by calculating the area under the curve. This analysis showed that LDH is able to discriminate survivorship one day earlier than platelet count. Both LDH level and platelet count are excellent predictors of survival, under standard therapy, after 3 days of treatment. Early identification of those at greatest risk will facilitate the early institution and evaluation of alternative methods of treatment, such as splenectomy, intravenous Ig, or Vincristine.

A 30-year-old man with long-standing localization-related epilepsy and mental retardation had seizures that were partially controlled with valproate (VPA) 500 mg four times daily. Routine examination showed severe thrombocytopenia with mild leukopenia and chronic low-grade hemolytic anemia. Pertinent laboratory results included positive ANA, rheumatoid factor, anti-NIA, circulating immune complexes, and antihistone antibody. The patient was treated with high dosage prednisone with partial improvement, but continued to have exacerbations at lower dosages. Fourteen months later, VPA was discontinued, and rapid improvement ensued. Prednisone was subsequently discontinued, and the patient has now maintained normal platelet counts for 18 months.