Abstract: We performed a prospective pilot study on 12 patients to evaluate the efficacy of the anti-CD20 monoclonal antibody rituximab in relapsed idiopathic thrombocytopenic purpura (ITP). Inclusion criteria were relapse of ITP with a thrombocyte count <20 000 µL−1 and unsuccessful corticosteroid treatment. Eleven patients had a previous splenectomy, five patients had unsuccessful cytotoxic treatment, and six patients were refractory to intravenous immunoglobulins before rituximab therapy. Response criteria were as follows. Complete remission (CR): normalization of thrombocyte count for at least 30 d. Partial remission (PR): an increase of thrombocytes to above 30 000 µL−1 for at least 30 d. Minor response (MR): any increase above 30 000 µL−1 for less than 30 d but more than 10 d. No response (NR): failure to achieve any of the above responses. Treatment plan: We administered 375 mg m−2 of rituximab once weekly on up to four consecutive weeks, unless there was early CR. Five patients (41%) achieved CR, two patients (17%) PR, and two patients MR (overall response rate 75%, median follow-up of responders 320 d). Four CR patients are ongoing; one CR patient relapsed after 6 months. Adverse events included excessive thrombocytosis in one patient as well as minor infusion-related (grade I) toxicities in four patients. We conclude that rituximab is a promising agent in the treatment of relapsed ITP.