Stroke associated with discontinuation of warfarin therapy for atrial fibrillation.

ABSTRACT

OBJECTIVE: The objective of this study was to determine the association between warfarin discontinuation and stroke among patients with nonvalvular atrial fibrillation (NVAF). RESEARCH DESIGN AND METHODS: This was a retrospective, observational study of adult NVAF patients (≥ 18 years) who were on warfarin in the Truven MarketScan commercial claims and encounters and Medicare supplemental and coordination of benefits databases (1 January 2008 to 30 June 2012). Warfarin discontinuation was defined as a gap of ≥ 45 days in warfarin prescription within 1 year after initiation. Patients who did and did not discontinue warfarin were matched at a 11 ratio using a propensity score method. Matched patients were followed for up to 1 year to determine risks of ischemic stroke, transient ischemic attack (TIA), and hemorrhagic stroke. A multivariate Cox proportional hazards model was used to further adjust for the effects of potential confounders. RESULTS: A total of 27,000 patients were included. Patients who discontinued warfarin had higher rates of ischemic stroke compared to persistent patients (1.0 vs. 0.5 per 100 patient years, P < 0.01), but similar rates of TIA (1.2 vs. 0.9 per 100 patient years, respectively; P = 0.07) and hemorrhagic stroke (0.3 vs. 0.2 per 100 patient years, P = 0.31). After adjustment for potential confounders, warfarin discontinuation was significantly associated with increased risk of ischemic stroke (hazard ratio [HR] 2.04; 95% confidence interval [CI] 1.47-2.84), TIA (HR 1.36; 95% CI 1.04-1.78), and ischemic stroke or TIA (HR 1.50; 95% CI 1.20-1.87). CONCLUSIONS: Warfarin discontinuation is associated with increased risk of ischemic stroke and TIA. Health care providers may need to take a more active role in the management of warfarin discontinuation and clinical outcomes, e.g., by considering newer anticoagulants with favorable risk-benefit profiles. Key limitations of the study include unavailability of important clinical factors and measures in claims data.