Warfarin May Increase Risk of Bleeding Without Decreasing Risk of Stroke in Elderly Patients With Atrial Fibrillation Receiving Dialysis
EBM Focus - Volume 9, Issue 6
Warfarin is widely used for thromboembolic prophylaxis in patients with atrial fibrillation, including patients with chronic kidney disease. A recent randomized trial has shown that adjusted-dose warfarin reduces the risk of ischemic stroke or systemic embolism compared to low-dose warfarin or aspirin in patients with moderate kidney disease (Clin J Am Soc Nephrol 2011 Nov;6(11):2599 full-text). However, the role of warfarin in patients with more advanced kidney disease remains unclear. No randomized trials evaluating warfarin for prevention of cardiovascular outcomes in patients receiving dialysis have been conducted, and observational studies in this population have been conflicting. Now, a new cohort study has evaluated the use of warfarin among elderly patients on dialysis who developed atrial fibrillation.
A total of 1,626 patients aged 65 years or older receiving hemodialysis or peritoneal dialysis prior to hospitalization for atrial fibrillation were retrospectively evaluated for an association between warfarin use and risk of bleeding or stroke. Bleeding events included intracerebral, gastrointestinal, or intraocular bleeding, hematuria, or bleeding at an unspecified location. Stroke events included ischemic stroke, transient ischemic attack, or a retinal infarct. About 46% of patients received warfarin within 30 days of hospital discharge. Comparing the baseline risks for patients receiving warfarin vs. those not receiving warfarin, 84% vs. 86% had a high risk of bleeding (HAS-BLED score ≥ 3) and 77% vs. 69% had a high risk of stroke (CHADS2 score ≥ 2), but p values for these differences were not reported.
In an unadjusted analysis, the rate of bleeding was 10.9 per 100 person-years with warfarin vs. 7.3 per 100 person-years with no warfarin (p < 0.001), with no significant difference in the rate of stroke (3.4 per 100 person-years with warfarin vs. 2.9 per 100 person-years with no warfarin). A separate analysis adjusted for propensity to receive warfarin (with propensity scores based on multiple clinical and demographic factors including baseline risk) had results consistent with those of the unadjusted analysis.
Oral anticoagulation in patients with advanced kidney disease has recently been called into question based on results from several observational studies. Indeed, the routine use of oral anticoagulants in patients with chronic kidney disease requiring dialysis is no longer recommended in guidelines from either Kidney Disease: Improving Global Outcomes (Kidney Int 2011 Sep;80(6):572) or the Canadian Cardiovascular Society (Can J Cardiol 2012 Mar-Apr;28(2):125). The findings from this latest observational study support these updated guidelines, and suggest that warfarin may increase the risk of bleeding with no benefit in primary stroke prevention among patients with atrial fibrillation who require dialysis. Although this study has attempted to account for potential confounders in their statistical analyses, it is nonetheless limited by its observational design, and a randomized trial evaluating oral anticoagulation for primary prevention of stroke in patients with atrial fibrillation who require dialysis is warranted. In the meantime, use of oral anticoagulants in this patient population should be approached with caution.