Patterns of international normalized ratio values among new warfarin patients with nonvalvular atrial fibrillation

Limited information exists regarding the relationship between international normalized ratio (INR) control/stability and the discontinuation of warfarin therapy among patients with nonvalvular atrial fibrillation (NVAF). This study evaluated the association between INR stabilization and warfarin discontinuation and assessed INR patterns before and after INR stabilization among patients (> 18 years) with NVAF who newly initiated warfarin (Veterans Health Administration datasets; October 1, 2007 through September 30, 2012). Achievement of INR stabilization (> 3 consecutive in-range therapeutic INR measurements >7 days apart) was examined from warfarin initiation through the end of warfarin exposure. Proportion of time in therapeutic range during warfarin exposure was calculated (Rosendaal method) and categorized as at least 60% or less than 60%. Among 34 346 patients, 49.4% achieved INR stabilization (mean time to stabilization, 98 days). Approximately 40% of INR values were out-of-range, even after achieving stabilization. During 30 days following an INR 4.0 or higher, patients had more INR testing than the overall mean (2.51 vs. 1.67 tests). Warfarin discontinuation was 4.2 times more likely among patients without INR stabilization versus those with INR stabilization (P < 0.00001). Patients with poor INR control (time in therapeutic range >60%) were 1.76 times more likely to discontinue warfarin within 1 year (P < 0.0001). INR stabilization is a better predictor of warfarin discontinuation than poor INR control. Improved approaches are necessary to maintain appropriate anticoagulation levels among patients with NVAF.