BACKGROUND: The aim of this study is to evaluate the therapeutic effect of splenectomy in children with chronic idiopathic thrombocytopenic purpura(ITP) who do not respond to corticosteroids and intravenous immunoglobulin(IVIG). METHODS: Fifteen patients(7 boys and 8 girls) who were admitted to St. Mary's Hospital under the diagnosis of ITP from December 1990 to February 1997 were enrolled In this study. All had chronic forms(more than 6 months after diagnosis) and were refractory to corticosteroids(prednisolone 1~2mg/kg) and IVIG(400mg/kg for 5 days or 1 g/kg for 2 days). We you-finely employ pneumococcal vaccine in 13 patients before splenectomy slnce 1992. Antiplatelet antibody was examed in 10 patients. We measured platelet count at preoperative time, postoperative 1 day, 1 month, and 3 months or more. Hematologic results were analyzed according to Berchtold and McMillan's criteria (1) complete response(CR) >120 x 10(9)/l; (2) partial response(PR) >50 x 10(9)/l (3) no response(NR) < 50 x 10(9)/l. RESULTS: The median age at operation was 10 yr(6~16). The median time interval between diagnosis and splenectomy was 2 yr 2 mo(6 mo~5 yr 5 mo). Antiplatelet antibody was positive in 2 cases of 10 cases(20%). Of 15 cases, 12 cases were corticosteroid resistant and 3 cases were corticosteroid dependent. The postsplenectomy platelet count(median 407 x 10(9)/l of 1 day; 254 x 10(9)/l of 1 month; 227 x 10(9)/l of 3 months or more) was significantly higher than presplenectomy platelet count(median 9 x 10(9)/l)(P<0.05). All patients showed complete or parial response throughout the follow up period. Accessory spleen was found in 1 case and removed during operation. There were no serious complications following splenectomy except mild fever in 2 cases(14%). CONCLUSION: Splenectomy appears to be an effective and relatively safe treatment for patients with chronic ITP who have had inadequate response to conventional therapies.