Chronic inflammation is a major cause of morbidity and mortality in end-stage renal disease. The associated anemia in these patients due to renal cortical atrophy and erythropoietin deficiency is treated with recombinant erythropoietin. Recent reports suggest a growing incidence of symptomatic venous thrombosis in cancer patients treated with recombinant erythropoietin. Several investigators have reported on different mechanisms of thrombosis in these patients. We hypothesize that thrombosis in patients with end-stage renal disease due to increased expression of C-reactive protein (CRP) as a result of chronic inflammation promotes the release of thrombin activatable fibrinolytic inhibitor causing fibrinolytic deficit and eventually thrombosis. Furthermore, because endothelial nitric oxide is responsible for the maintenance of the normal vascular function, the decreased levels of nitric oxide in chronic inflammation cause endothelial damage and result in thrombosis. To test this hypothesis, blood samples were collected from 106 patients (49 male and 57 female, aged 59.8±15.7 years) with end-stage renal disease undergoing hemodialysis and treated with recombinant erythropoietin at a mean dose of 201.8 U/kg/week. Blood samples were drawn in 5-mL tubes containing 3.2% sodium citrate just before the hemodialysis procedure. These blood samples were immediately centrifuged to obtain platelet-poor plasma, which was aliquoted and frozen at −70°C until further analysis. Erytropoietin antibodies were measured using an anti-EPO enzyme-linked immunosorbent assay (ELISA) method developed in our laboratory. Nitric oxide was measured using a NO analyzer (Sievers 280I, Ionics, Boulder, CO). Plasma CRP levels were measured with a highly sensitive ELISA method IMUNOCLONE CRP ELISA (American Diagnostica, Greenwich, CT). TAFI antigen levels in plasma were analyzed with an IMUCLONE TAFI ELISA kit (American Diagnostica, Greenwich, CT). TAFI functional activity was assayed with an ACTICHROME TAFI activity kit. The measured levels of nitric oxide, CRP, TAFI antigen, and TAFI functional were 37.36±36.8 (normal value, 37.49±18.96; range, 19.3-102 μM), 12.27±10.6 (normal value, < 1 μg/mL), 146.9±28.4% NHP (normal, 100% NHP), and 102.55±37% NHP (normal range, 22.3-165.7; mean, 89.5% NHP), respectively. The erythropoietin antibody was detected in 9.4% of the patient group. While 20% of the erythropoietin antibody-positive and 27.1% of the erythropoietin antibody-negative patients experienced chest pain, thrombotic events developed in 9.4% of the erythropoietin antibody-negative patients. These data provide the rationale for a novel mechanism of thrombosis through increased activity of CRP, nitric oxide, and TAFI, leading to fibrinolytic deficit and thrombosis in patients treated with erythropoietin.