Abstract P131: Importance of Algorithm-Consistent Warfarin Dosing in the Quality of Anticoagulation Control in Atrial Fibrillation: A Multilevel Analysis

Background: The quality of anticoagulation control in AF patients on warfarin varies between countries, but the reasons are unclear. We determined the effect of algorithm-based warfarin dosing on anticoagulation control after adjusting for patient, center, and country characteristics. Patients: RE-LY trial patients with AF on warfarin. Centers with >5 patients on warfarin were included. Design: We tracked INR measurements and warfarin doses in patients. Dose adjustments were considered algorithm-consistent if they were within 5% of the recommendation made in the RE-LY warfarin nomogram. We developed a multilevel linear regression model with patients (1st level) nested in centers (2nd level), and centers nested in countries (3rd level), and examined the effect of algorithm-consistent dosing on Time in Therapeutic Range (TTR) of the INR. Results: A total of 4577 patients (62% male) from 402 centers and 40 countries were included. Considerable regional variation in mean TTR was found, ranging from 54 ± 22% in East Asian countries to 73 ± 15% in North European countries. The degree of algorithm-consistent warfarin dosing correlated with mean country TTR (r 2 =0.6) (Fig.). After adjusting for patient, center, and country variables, algorithm-consistent warfarin dosing was found to strongly predict TTR; a 1% increase in consistency with algorithm-based dosing increased mean TTR by 0.67% [0.61-0.73%, p<0.001]. Conclusion: Algorithm-consistent warfarin dosing is predictive of anticoagulation control after adjusting for patient, center, and country factors. There is potential to improve the quality of anticoagulation control by simple algorithm-based warfarin dosing.