Results of treatment in immune thrombocytopenia.

One hundred and forty-eight adults with immune thrombocytopenia attending a single institution between 1971 and 1981 were analysed retrospectively and observed for a minimum of three years. Prednisone was administered to 134 patients (90 per cent) in whom a peak platelet count occurred at a median of eight days, at which time four distinct patterns of response became evident: 20 patients (15 per cent) achieved complete remission with platelet counts greater than 200 X 10(9)/I, and of these 13 relapsed; 46 (34 per cent) had incomplete remission with platelet counts between 100 and 200 X 10(9)/I, and of these 38 relapsed; 22 (17 per cent) had a partial remission with platelet counts between 50 and 100 X 10(9)/I, and 14 have relapsed; the remaining 46 (34 per cent) failed to respond to prednisone. The addition of cytotoxic drugs for patients who relapsed or failed to respond to this agent was uniformly ineffective with the spleen in situ. One hundred and two patients underwent splenectomy without mortality, and peak platelet counts were achieved at a median of two days after the operation. Forty-one patients (40 per cent) achieved complete remission and four have relapsed; 36 (35 per cent) achieved incomplete remission and five have relapsed; 18 (18 per cent) had a partial remission and two have relapsed; treatment failed in seven (7 per cent). Eleven of the post-splenectomy patients relapsed in a median of four months. Together with the seven who failed to respond (n = 18), a subgroup of 17 was treated with immunosuppressive regimens including prednisone, azathioprine, cyclophosphamide or vincristine. Ten achieved stable although incomplete remissions, with platelet counts between 50 and 200 X 10(9)/I. We suggest that the treatment of immune thrombocytopenia in the adult initially should be a course of prednisone, from which complete remission rate will be 4.7 per cent and further incomplete remissions of 5.5 per cent and partial remissions of 8.5 per cent can be anticipated; no correlation was demonstrable with the duration of symptoms and response to prednisone. Symptomatic patients and those requiring unacceptable doses of adrenocorticosteroids should then proceed to splenectomy, when an overall response rate of 93 per cent is possible. A relapse rate of 11 per cent may require a limited course of immunosuppressive treatment, which is likely to be effective in approximately 75 per cent of these individuals.