Recombinant human erythropoietin (r-HuEPO) corrects cancer-related anemia and, thereby, improves quality of life. The purpose of the present study was to measure the impact of erythropoietin on hemoglobin and mood state in patients with metastatic breast cancer and mild anemia (Hgb < 12.0 g/dL). Women were randomized to receive usual care (G1) or usual care plus r-HuEPO (G2). Usual care included transfusions as necessary and fatigue education. R-HuEPO was begun at 40,000U subcutaneously per week. At 4 weeks, the dose was increased to 60,000U if Hgb had not increased > or = 1.0 g/dL. The drug was discontinued at 8 weeks if hemoglobin improvement was < 1.0 g/dL. The study was terminated early (n = 27, G1 = 13, G2 = 14) when 4/14 (28.5%) subjects in G2 developed thrombotic events (deep vein thrombosis [DVT] in 1; DVT plus pulmonary embolism [PE] in 1; DVT plus PE 1 month after drug discontinuation in 1; and brachial vein thrombosis with infected Mediport in 1). In all four patients, Hgb levels were normal at the time of the event. No patient in G1 developed a thrombotic event. There were no significant differences in demographic characteristics or current chemotherapy regimen in G1 vs. G2. The decision to terminate the trial was made after considerable deliberation. The increased incidence of thrombotic events in the r-HuEPO (G2) arm of this study exceeds that in prior studies in this population and prior r-HuEPO trials. This may relate to the administration of r-HuEPO in this high-risk population, but the small sample size and possible predisposing risk factors preclude definitive conclusions.