Background : Rituximab, a monoclonal anti-CD20 antibody, is a promising therapeutic agent for adults with immune thrombocytopenic purpura (ITP). Pooling results of the published literature suggest that rituximab achieves an initial response in about 50 % of adults with chronic ITP, with a 25–40% rate of sustained response. However, heterogeneity of patient pre-treatment status, short follow up and bias due to retrospective studies limit the interpretation of the data.Objective : To assess the safety and efficacy of rituximab in adults with chronic ITP (duration ≥ 6 months) and platelet count ≤ 30x109/L candidate to splenectomy.Methods : a multicenter prospective open-label single arm phase 2 trial designed according to Fleming's single stage procedure was conducted. After informed consent, non splenectomized adults received 4 weekly intravenous infusions of rituximab at a dose of 375 mg/m2. All other ITP treatments were stopped. Treatment success was defined as a platelet count ≥ 50 x109/L with at least a 2 fold increase of the initial value at one year after the first rituximab infusion. Patients who received another treatment during follow up were considered as non responders. A sample size of at least 56 patients was calculated by Fleming's single-stage design to ensure 90 % power for proving lack of efficacy if the true complete response rate was below 25%.Results : Sixty consecutive patients (40F/20M) were included over a 21 months period. Mean age was 48 years (range 18–84). Mean ITP duration was 4.8 years. Mean platelet count at inclusion was 16±10x109/L. All but one patient, in whom reversible serum sickness disease was diagnosed after 2 infusions, received 4 infusions of rituximab. Fifteen other patients experienced transient side effects that did not lead to treatment discontinuation. No patient was lost to follow up. Success was achieved in 40 % (24/60 patients) (95 % confidence interval 28% to 52%) and was found significantly different from 25% (p=0.007). Among the 24 long term responders, platelet count at one year was ≥ 150x109/L in 18 and between 50 and 150 x109/L in 6. Two other patients (3%) had and incomplete response defined as a platelet count at one year between 30 and 50 x109/L but at least 2 fold the initial value. Thirty four (57%) patients failed to respond and amongst them, 18 have already undergone splenectomy.Conclusion: rituximab appears to be a safe and good splenectomy-sparing strategy in adults with chronic ITP leading to a significant and durable response in 40% of the cases.