Idiopathic thrombocytopenic purpura: predictors of chronic disease.

We studied the extent to which patient characteristics influenced outcome in childhood idiopathic thrombocytopenic purpura in a historical cohort of 289 children over a 20 year period (1968-87). Outcome was classified as acute or chronic depending on whether the platelet count had returned to normal (150 X 10(9)/l) by six months after diagnosis. Fifty three cases (18%) had chronic idiopathic thrombocytopenic purpura. The likelihood of chronic disease was determined by logistic regression analysis of five patient variables: age, sex, season of onset of symptoms, history of recent viral illness, and duration of symptoms at presentation. A history of symptoms of greater than 14 days at presentation, adjusted for the other variables, was strongly predictive of chronic idiopathic thrombocytopenic purpura; the other variables did not significantly affect outcome. At 28 days after diagnosis 138 (47%) of the study cohort had normal platelet counts. Children whose platelet counts were less than 150 X 10(9)/l had a threefold risk of progressing to chronic idiopathic thrombocytopenic purpura, which increased to fivefold if counts were less than 50 X 10(9)/l. Two thirds of patients in the chronic group, irrespective of treatment, remained thrombocytopenic two years after diagnosis. We conclude that a history of symptoms for greater than two weeks at presentation is strongly predictive of chronic idiopathic thrombocytopenic purpura. If platelet counts are subnormal 28 days after diagnosis the risk of chronic idiopathic thrombocytopenic purpura is increased with prolonged thrombocytopenia being very likely if platelet counts remain low three months after diagnosis.