Sailing Between Scylla and Charybdis: Oral Long-term Anticoagulation in Dialysis Patients

Thilo Krüger; Vincent Brandenburg; Georg Schlieper; Nikolaus Marx; Jürgen Floege


Nephrol Dial Transplant. 2013;28(3):534-541. 

In This Article


Physicians caring for patients with advanced CKD or ESRD and atrial fibrillation cannot rely on evidence-based guidelines in their decisions about OAT. Observational studies in dialysis patients with atrial fibrillation provide very little evidence to encourage the use of coumarins in such patients. It must be stressed that recommendations developed for the non-CKD population cannot be simply transferred to HD patients. Moreover, avoiding treatment with coumarins might reduce the burden of vascular calcification and the risk of calciphylaxis in ESRD––a population cohort at special risk of these cardiovascular complications. Considering the current absence of suitable alternatives to coumarins in advanced CKD, we still see some indications for their use in this population (Figure 2), but we discourage using coumarins for the prevention of thromboembolic events in ESRD patients with atrial fibrillation.

Figure 2.

Suggested coumarin indications in patients with atrial fibrillation and ESRD. Green boxes denote situations where coumarins are indicated or may provide benefit. Red boxes denote situations, where coumarins increase the risks or are contraindicated.

Note added in press

Very recently a large scale Danish observational study assessed the incidence of stroke and bleedings in patients with end stage renal disease (ESRD) and atrial fibrillation. In 901 such patients dismissed from hospitals between 1998 and 2008 the use of warfarin significantly reduced stroke incidence but increased bleeding episodes. However the design and data analysis have significant limitations. There was no information on the reason why warfarin was given and no information on INR control. The ESRD group consisted of patients on hemodialysis (with thrice weekly full dose heparin use), peritoneal patients (without anticoagulation), and kidney transplant patients (with some residual renal function). Also the authors adjusted for age via the CHA2DS2VASc score. This is a problem as it only distinguishes 3 broad categories, i.e. <65, 65-74, and >75 years of age. Given that the "renal replacement" patients were on average nearly 10 years younger than the other groups, such a categorical adjustment may blunt important differences. Finally, an unexplained observation is that in the ESRD patients the apparent benefit in terms of stroke reduction derived from warfarin completely disappeared, when warfarin was combined with ASS.[47] Given these significant limitations, the conclusions of our article do not change.