Severe renal impairment and risk of bleeding during anticoagulation for venous thromboembolism

Background Detection of severe renal impairment in patients with venous thromboembolism (VTE) is mandatory both for selecting anticoagulant therapy and for evaluating major bleeding risk, increased by severe renal impairment. Objectives To determine whether the Cockcroft and Gault (CG) and Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formulas identify severe renal impairment in the same VTE patients presenting the same risk of major bleeding. Patients/Methods We compared clinical characteristics and outcomes during the first 3 months of anticoagulation between VTE patients in the RIETE registry with severe renal impairment according to the CG and/or CKD-EPI formula (estimated glomerular filtration rate <30 mL/min and <30 mL/min/1.73 m(2), respectively). The primary outcome was major bleeding. Results Up to October 2017, 41 796 patients were included in RIETE. Among the 4676 patients with severe renal impairment according to at least one of the formulas, this was not confirmed by the other formula in 1904 (40.7%). Major bleeding risk was increased in every patient subgroup with severe renal impairment vs patients without this condition (CG or CKD-EPI < 30: odds ratio [OR] = 2.26, 95% confidence interval [CI 2.01-2.53], only CG < 30: OR = 1.72, 95% CI [1.37-2.13], only CKD-EPI < 30: OR = 2.34, 95% CI [1.77-3.05], CG+CKD-EPI < 30: OR = 2.47, 95% CI [2.16-2.83], all vs CG+CKD-EPI > 30). Conclusion The CG and CKD-EPI formulas identify different subgroups of patients with severe renal impairment, leading to discordant results in 40.7% of these patients. Irrespective of the formula used for their identification, patients with severe renal impairment have a higher risk of major bleeding under anticoagulant therapy.