Use of warfarin is standard of care for stroke prevention in patients with atrial fibrillation (AF). However, AF patients experience high rates of warfarin discontinuation/interruption, resulting in increased health risks and health care costs. As such, it is important to study the rates and predictors of warfarin discontinuation/interruption in this population. To determine (a) rates of warfarin discontinuation and interruption and (b) demographic, clinical, and health care-related factors associated with discontinuation and interruption in patients with nonvalvular AF (NVAF) in the usual clinical practice settings in the United States. This retrospective cohort study used the MarketScan Database and included patients (aged ≥ 18 years) with NVAF who were initiated on warfarin. The study period was January 1, 2008, to June 30, 2012. To be included, patients were required to have at least 2 claims with AF diagnosis separated by ≥ 30 days and ≤ 12 months and at least 1 outpatient claim. Warfarin initiation had to occur within 30 days of the AF diagnosis. Patients also had to have continuous enrollment in prescription drug plans from 6 months prior to warfarin use to at least 12 months after warfarin initiation. Patients were followed for 1 year after warfarin initiation. Persistence was defined as warfarin therapy without a gap ≥ 45 days between the end date of the former prescription and the start date of the current prescription or with international normalized ratio (INR) monitoring at least every 42 days. Interruption was defined as a gap in warfarin therapy ≥ 45 days and ≤ 90 days between the end date of the former prescription and the start date of the current prescription and without INR monitoring at least every 42 days. Discontinuation was defined as greater than 90 days without warfarin therapy between the end date of the former prescription and the start date of the current prescription and without INR monitoring at least every 42 days. Chi-square tests were used to analyze categorical variables, and independent samples t-tests were used for continuous variables. Cox proportional hazards regression model was performed to determine factors associated with warfarin discontinuation/interruption, including demographic (e.g., age, gender); clinical (e.g., comorbidities, CHADS2 score); and health care-related (e.g., hospitalizations or emergency room visits) characteristics. Sensitivity analyses were conducted by varying prescription gaps by 7, 14, and 30 days. A total of 58,593 patients with NVAF were included. The mean age was approximately 71 years (SD = 12.00) and mean CHADS2 score was 1.66 (SD = 1.23). The majority of patients were male (60%). During 12 months after warfarin initiation, 45% of patients were persistent with warfarin; 12% had interruption without discontinuation; and 43% had discontinuation. The risk of warfarin interruption or discontinuation was significantly greater in patients who were younger than 65 years (HR = 1.22; 95% CI = 1.19-1.25), lived in the West (HR = 1.07; 95% CI = 1.03-1.11), had history of anemia (HR = 1.10; 95% CI = 1.06-1.14), had history of bleeding (HR = 1.10; 95% CI = 1.06-1.14), or had history of hospitalization or emergency room visits (HR = 1.11; 95% CI = 1.08-1.13). The significant factors associated with interruption and discontinuation were similar. In the sensitivity analyses, the significant factors associated with discontinuation/interruption were similar across different prescription gaps. In the U.S. clinical practice setting, more than 50% of NVAF patients discontinued or interrupted warfarin within 1 year after initiation. Aged less than 65 years, history of anemia, and history of hospitalization/emergency room visits were associated with increased risk of discontinuation/interruption. Given the high prevalence of warfarin discontinuation/interruption, health care providers should take a more active role in understanding and addressing the reasons behind patient discontinuation/interruption.