This randomised controlled multicentre trial evaluated the effectiveness of recombinant human erythropoietin (rhEPO) in preventing anaemia and reducing the need for blood or erythrocyte transfusion in 122 ovarian cancer patients receiving platinum-based chemotherapy. The patients were randomly allocated to receive rhEPO 150 U/kg or 300 U/kg subcutaneously, three times a week, or open control. Patients also received up to 6 cycles of carboplatin or cisplatin, alone or in combination with other cytotoxic agents. Intention-to-treat analysis showed that 39.4% of patients in the control group received at least one blood transfusion, compared with 9.2% of patients treated with rhEPO. Patients treated with rhEPO experienced a significantly longer time to first erythrocyte transfusion than the control group and were less likely to experience nadir haemoglobin levels < 10 g/dl (P < 0.001 and < 0.05, respectively). A haemoglobin decrease < 1 g/dl during the first chemotherapy cycle, as well as a low baseline serum erythropoietin concentration, predicted a low transfusion need in rhEPO-treated patients but not in controls. During the study, 103 patients suffered at least one adverse event, but no serious, and only nine non-serious adverse events were considered possibly related to rhEPO therapy. These results indicate that treatment with rhEPO prevents anaemia, it reduces the need for blood or rhEPO erythrocyte transfusion in patients with ovarian cancer receiving platinum-based chemotherapy, and it is well tolerated. A starting dose of 150 U/kg of rhEPO, three times a week, may be recommended.