부산대학교 도서관

Introduction: A target INR of 2-3 has been recommended for patients with atrial fibrillation (AF) and risk factors for thromboembolism. This recommendation is largely based on evidence from observational studies a decade ago. Our study utilized collective data from modern trials with warfarin controls to examine the longitudinal relationship between warfarin anticoagulation, as assessed by INR, on the clinical outcome events of interest.Methods: We obtained subject-level data for warfarin-treated patients from three large clinical studies supporting the approval of dabigatran (Pradax®), apixaban (Eliquis®) and edoxaban (SAVAYSA®). Ischemic stroke, intracranial hemorrhagic (ICH) and all-cause mortality were selected as the outcome events of interest. We performed multivariate Cox regression modeling to examine the association between time-dependent INR and each outcome event. Benefit-Risk analysis was assessed by summing estimated annual event rate for both ischemic stroke and ICH.Results: A total of 21,883 patients consisting of 322 ischemic strokes, 288 ICHs and 657 all-cause deaths were included in our analysis. Our models suggest that the risk of ischemic stroke is greatly reduced when INR exceeds 2; in contrast, the risk of ICH increases monotonically as INR increases. When combining ischemic stroke and ICH events, the lowest estimated annual event rate was observed between INR of 2 and 2.5; the risk only slightly increases (<20%) between INR of 1.8 and 3.0. Similarly, a U-shaped relationship between INR and the risk of all-cause death was found; both low and high INRs were associated with an increased risk of all-cause death with the lowest risk occurring between INR of 2 and 2.5.Conclusions: Our study using collective warfarin data from recent large prospective trials indicate that INR between 2-2.5 provides the best balance between ischemic stroke and ICH as well as optimal protection against death in patients with AF.