INR Quality Measure and Bleeding Risk Scores

Study Questions:

Do bleeding risk scores for anticoagulation therapy improve when you include a measure of international normalized ratio (INR) quality?


The authors used the combined warfarin-treated arms of the SPORTIF (Stroke Prevention using an Oral Thrombin Inhibitor in patients with atrial Fibrillation) III and V studies, where warfarin was compared to ximelagatran for stroke prevention in nonvalvular atrial fibrillation (NVAF). Four contemporary bleeding risk scores (HAS-BLED, ORBIT, ATRIA, and HEMORR2HAGES) were calculated for all patients with available clinical data. Major bleeding events were analyzed both as reported by the local site investigators and based on centrally adjudicated events. Poor quality INR management, as measured by the Rosendaal time in the therapeutic range (TTR) linear interpolation method, was defined as a TTR <65%. Bleeding risk scores with and without the inclusion of poor TTR were compared using receiver operating characteristics (ROC) and net reclassification index (NRI) methods.


Of the 3,665 patients randomized to the warfarin treatment arm, 162 major bleeding events were identified by the local investigators, of which 127 were validated by the central adjudication committee. Unadjusted analyses demonstrated an association between all bleeding risk scores and major bleeding risk, but only the HAS-BLED and ORBIT score associations persisted after adjustment for gender and type of NVAF. The HEMORR2HAGES score performed worse than ORBIT and ATRIA based on ROC curve analysis. Addition of TTR <65% to the ORBIT, ATRIA, and HEMORR2HAGES scores improved all three scores based on NRI analysis.


The authors concluded that different bleeding risk scores provide different discriminatory capacity for major bleeding in NVAF patients treated with warfarin. They also concluded that adding labile INR (TTR <65%) to existing bleeding risk scores improves their predictive performance for major bleeding.


This study highlights the importance of re-assessing bleeding risk in patients after initiating warfarin therapy. While most clinicians assess stroke risk at the time of anticoagulation initiation, they less often formally assess bleeding risk. And rarely do clinicians formally assess bleeding risk after a patient has been maintained on anticoagulant therapy for 3+ months. However, this study highlights the important role that warfarin control (as measured by the TTR) has on bleeding risk. While most clinicians do not currently have ready access to a ‘real-time’ TTR measurement in their electronic health records, a more straightforward assessment of the percent of in-range INRs may be just as useful (Chan PH, et al., Can J Cardiol 2015;Nov 6:[Epub ahead of print]). Continually assessing our patients’ bleeding risk (preferably using a formal risk score) and discussing strategies to reduce that risk is an important ongoing issue.

< Back to Listings