Ought dialysis patients with atrial fibrillation be treated with oral anticoagulants?

Background: Dialysis patients with atrial fibrillation have an increased thrombolic risk. Dicoumarin anticoagulant therapy is often considered contra-indicated in chronic renal insufficiency in which the risk of haemorrhage, though not defined, is perceived to be high. We assessed haemorrhage complications in dialysis patients receiving dicoumarin anticoagulant therapy to establish whether the haemorrhage risk justifies the contra-indication of anticoagulant therapy in patients with atrial fibrillation. Patients and methods:, Over a period of a decade in our dialysis centre, 29 patients receiving anticoagulant therapy over a protracted period presented haemorrhage complications. These were classified with respect to severity and location and compared with 211 patients not receiving anticoagulant therapy. The relative risk of haemorrhage was calculated and was compared to risk of thrombo-embolism in dialysis patients with atrial fibrillation. Results: Of the 29 patients, nine had 13 episodes of haemorrhage complications (26 episodes/100 patient-years). None was fatal, nor intra-cranial nor with serious clinical sequelae. In the group without anticoagulants, 29 patients had 39 haemorrhage complications (11 episodes/100 patient-years); four (10.2%) intra-cranial and all fatal. The relative risk of bleeding with anticoagulant therapy was 2.36 (95% confidence interval=1.19-4.27). Conclusions: (1) Dialysis patients with anticoagulant therapy presented with a higher risk of haemorrhage; (2) the relative risk of bleeding was double that of the dialysis population without anticoagulant therapy; (3) despite the high risk of haemorrhage that we observed, the high risk of thrombo-embolism and the attendant serious sequelae to which dialysis patients with atrial fibrillation are predisposed indicates that oral anticoagulation therapy ought not to be considered automatically contra-indicated in this patient group but that an exhaustive evaluation of the risk-benefit needs to be conducted on an individual patient basis. (C) 2002 Elsevier Science Ireland Ltd. All rights reserved.