Abstract 2188: Bleeding Events and Anticoagulant Discontinuation in Patients with Atrial Fibrillation

Abstract

BACKGROUND: Annually, about 2.3% of patients receiving warfarin prophylaxis experience major bleeding. Discontinuation of warfarin due to hemorrhage can leave patients unprotected against potential stroke. OBJECTIVE: To assess the rate of warfarin discontinuation and identify factors associated with warfarin discontinuation among patients with atrial fibrillation (AF). METHODS: Retrospective cohort study was conducted using the Market Scan Database between1/12005 and 6/30/2008. Study patients were ≥18 years old with ≥2 AF diagnoses where at least one diagnosis was prior to the first warfarin prescription (index). An additional warfarin prescription within 6-months after index date was required to ensure continuous treatment. Patients were excluded if they had previous warfarin treatment or a cardioversion during one year history period. Continuous health benefit coverage was required during study period. Patients were followed until a hemorrhagic event, warfarin discontinuation, or end of study period. Discontinuation was defined by ≥90 days without a refill after depletion of the drug supply from the prior prescription claim. Recent hemorrhage was defined as hemorrhage within 90 days before discontinuation. Multiple regression analyses were conducted to identify the factors associated with warfarin discontinuation. RESULTS: 7,971 patients met study eligibility criteria. Mean (±SD) patient age was 67.8 (±12.7), with majority ≥65 years old (59.3%). 41.2% were female. During follow-up, 51.7% of patients discontinued warfarin (p<.001). A higher proportion of patients who had a recent hemorrhage (61.8%) discontinued warfarin compared to those without (51.3%). After adjustment, patients with a recent hemorrhage were 35% more likely to discontinue warfarin, compared to those without (RR=1.35; CI:1.16-1.58). Age of ≥85 (reference: <65), and history of ≥1 hospitalization or emergency room visit were associated with increased likelihood of discontinuation (all p<.001). Female, daily doses>5mg (reference: ≤5mg), concomitant use of gastroprotective agents, and CHADS2 scores ≥1 were associated with decreased likelihood of discontinuation (all p<.05). CONCLUSIONS: In this retrospective cohort, half of AF patients receiving warfarin discontinued their treatment. A recent hemorrhage was associated with an increase in discontinuation. A novel therapy with a predictable profile may help minimize bleeding risk and treatment discontinuation.