Warfarin use in hemodialysis patients: what is the risk?

BACKGROUND There is a paucity of data concerning the risks associated with warfarin in hemodialysis (HD) patients. We compared major bleeding episodes in this group with HD patients not receiving warfarin and with a cohort of non-HD patients receiving warfarin. METHODS A retrospective review of 141 HD patients on warfarin (HDW), 704 HD patients not on warfarin (HDNW) and 3,266 non-dialysis warfarin patients (NDW) was performed. Hospital admissions for hemorrhagic events and ischemic strokes were examined as was hospital length of stay and blood product use. INR variability was also assessed. RESULTS The incidence rates for major hemorrhage per 100 patient years was 10.8 in the HDW group as compared to 8.0 in the HDNW (p = 0.593) and 2.1 in the NDW (p < 0.001) groups. Mean units of red blood cell transfusions required was higher in patients on dialysis with no significant difference between HDW and HDNW groups. The risk of ischemic stroke per 100 patient years was 1.7 in the HDW group as compared to 0.7 in the HDNW groups (p = 0.636) and 0.4 in the NDW (p = 0.003). The HDW group had higher inter-measurement INR variability compared to the NDW group (p = 0.034). In patients with atrial fibrillation, HDW group had a higher incidence of ischemic stroke than the NDW group (2.2 versus 0.4 events per 100 patient years; p = 0.024). CONCLUSIONS This study confirms the higher bleeding risk associated with HD/ESRD but suggests that warfarin use in these patients may not add significantly to this risk. We also demonstrated high rates of ischemic stroke in HD patients despite warfarin use. SUMMARY Our study compares the frequency of major hemorrhage and secondarily, ischemic stroke in HD patients receiving or not receiving warfarin, with non-HD patients receiving warfarin. The major finding was that frequency of hemorrhage was higher in HD patients receiving warfarin than in non-HD patients receiving warfarin, but not different in HD patients with or without warfarin. A secondary finding was that INR variability was significantly higher in HD patients than non-HD patients on warfarin.