High quality of warfarin treatment is important to prevent recurrence of venous thromboembolism (VTE) without bleeding complications. The aim of this study was to examine the effect of individual time in therapeutic range (iTTR) and International Normalised Ratio (INR) variability on bleeding risk and mortality in a large cohort of well-managed patients with warfarin due to VTE. A cohort of 16612 patients corresponding to 19502 treatment periods with warfarin due to VTE between January 1, 2006 and December 31, 2011 was retrieved from the Swedish national quality register AuriculA and matched with the Swedish National Patient Register for bleeding complications and background characteristics and the Cause of death register for occurrence and date of death. The rate of bleeding was 1.79 (confidence interval (CI) 95 % 1.66-1.93) per 100 treatment years among all patients. Those with poor warfarin treatment quality had a higher rate of clinically relevant bleeding, both when measured as iTTR below 70 %, 2.91 (CI 95 % 2.61-3.21) or as INR variability over the mean value 0.85, 2.61 (CI 95 % 2.36-2.86). Among those with both high INR variability and low iTTR the risk of clinically relevant bleeding was clearly increased hazard ratio (HR) 3.47 (CI 95 % 2.89-4.17). A similar result was found for all-cause mortality with a HR of 3.67 (CI 95 % 3.02-4.47). Both a low iTTR and a high INR variability increase the risk of bleeding complications or mortality. When combining the two treatment quality indicators patients at particular high risk of bleeding or death can be identified.
Keywords: INR variability; TTR; all-cause mortality; bleeding; venous thromboembolism; warfarin.