Benefit and harm of anticoagulation in the prevention of thromboembolic stroke for non-valvular atrial fibrillation in haemodialysis patients: a Top End of Northern Australia study.

Abstract

Warfarin for the prevention of non-valvular atrial fibrillation (AF)-related thromboembolic stroke in patients on maintenance haemodialysis is controversial. Despite the exclusion of haemodialysis patients in randomised control trials, the American Heart Association/American College of Cardiology has recommended warfarin in high-risk AF patients. To retrospectively examine the utility of warfarin anticoagulation therapy in our prevalent haemodialysis patients over 10 years of follow up. Eligible patients were retrospectively identified and stratified to two cohorts based on whether warfarin was prescribed. The outcomes of interest were ischaemic stroke, haemorrhagic stroke and death from any cause. Rate ratio and Cox proportional hazard regression model were used to compare the differences in outcome between the two cohorts. The Kaplan-Meier method was used to analyse survival. Three ischaemic strokes and four haemorrhagic strokes occurred in the unexposed group of 166 patients over 484.44 patient-years of follow up. One ischaemic stroke and no cases of haemorrhagic stroke occurred in the exposed warfarin group of 16 patients over 39.32 patient-years of follow up. Eighty-seven percent of patients in both groups were indigenous. More than 90% of each cohort had a CHA2DS2-VaSc score ≥2. One hundred and one deaths, 90 in the unexposed group and 11 in the warfarin group, occurred in the follow-up period. A non-statistically significant trend towards increasing mortality was observed in the warfarin group (hazard ratio=1.63; P = 0.13). This retrospective study of prevalent haemodialysis patients with co-existing history of non-valvular AF failed to demonstrate sufficient evidence for the routine use of warfarin for prophylaxis of thromboembolic stroke.