The indication for splenectomy in chronic idiopathic thrombocytopenic purpura (ITP) remains a controversial subject. The mortality rate of persistent thrombocytopenia is very low, except in severe cases. Conversely, the risks of splenectomy are significant (in the present series, morbidity: 4.1% mortality: 1.4%), with a success rate of only 60-75%. It is therefore useful to define a parameter able to predict the efficacy or failure of splenectomy. An analysis of 578 cases of chronic ITP, where the site of platelet destruction has been determined, is presented. 268 of these cases had been splenectomized. When platelet destruction was splenic, 96% of subjects aged 5-30 years and 91% of cases over the age of 30 years obtained a remission. Conversely, when platelet destruction was hepatic or diffuse, failure or incomplete results were observed in 92% of cases. The site of platelet destruction therefore constitutes a parameter which can help the clinician to make the decision to perform splenectomy.