The efficacy of corticosteroids in childhood acute idiopathic thrombocytopenic purpura (ITP) is controversial and has infrequently been evaluated in a controlled randomized fashion. We administered prednisone (2 mg/kg/day for 14 days with subsequent tapering and discontinuation by day 21) or placebo to 27 children, aged 10 years or less, with newly diagnosed ITP. Platelet count, bleeding time (a test of the integrity of the platelet-microvasculature interaction), and clinical bleeding score (based on a 0-4 scale) were determined before (day 0) and six times following initiation of drug therapy (days 1-2, 3-5, 7, 14, 21, and 28). There were no statistically significant (p less than 0.05) differences between the two treatment groups in any of the three study parameters except on day 7 of therapy when children receiving prednisone had higher platelet counts and lower bleeding scores and bleeding times than those taking placebo. Bleeding time correlated inversely with the platelet count in both treatment groups. Prednisone did not appear to influence bleeding time independent of its effect on platelet count. This treatment regimen of prednisone did not clearly improve hemostasis in childhood acute ITP except transiently at the end of 1 week of treatment.