A-373 Comparison of the CoaguChek ® and Coag-Sense ® PT2 Meter Point of Care INR Devices

Background Warfarin monitoring may be done at the point-of-care (POC) at home, in a clinic environment, or be performed using plasma specimens in a central laboratory. With a goal of standardizing International Normalization Ratio (INR) measurements and improving data integrity by enabling electronic transmission of results from any location, we evaluated two POC INR devices against an internal plasma INR reference method. Methods We conducted a multicenter study to test the accuracy of two commercially available point of care devices: Coag-Sense PT2 Meter (Coagusense, Inc., Fremont, CA) and CoaguChek XS Pro or XS Plus (Roche Diagnostics, Indianapolis, IN) compared with plasma INR values among warfarin treated patients. Bias and linear regression analysis were assessed for device-plasma INR comparisons. Warfarin dosing decisions were assessed for each device compared to an internal plasma INR reference method (Instrumentation Laboratories ACLTOP 500, 550, 750 analyzers (Werfen Diagnostics, Bedford MA) in one of three CLIA-certified laboratories. The thromboplastin used at our institution for INR measures is HemosIL RecombiPlasTin 2G (R2G) reagent (Werfen Diagnostics, Bedford MA) which has an International Sensitivity Index (ISI) of 1.0. Results Two hundred ninety-nine warfarin patients treated across three Mayo Clinic sites agreed to participate in the study. Common anticoagulant indications were atrial fibrillation (63.9%), venous thromboembolism (21.7%) and heart valve prosthesis (15.4%) with a 2.0–3.0 INR target range for 93.6% of patients. For the CoaguChek devices, 81.3% of values fell within 0.2 INR units of plasma INR referent and 95.3% within 0.4 units. The correlation between CoaguChek whole blood and plasma INR was excellent (R2 = 0.93). For the Coag-Sense device, 34.1% of values fell within 0.2 INR units and 60.2% within 0.4 INR units of plasma INR values, with a slightly weaker correlation between Coag-Sense whole blood and plasma INR (R2 = 0.83; P < 0.0001). Mean (± standard deviation) deviation from plasma INR was 0.04 ± 0.2 for CoaguChek compared to 0.4 ± 0.3 for Coag-Sense (P < 0.0001). The difference between POC devices was largely attributed to systematic negative bias between Coag-Sense and ACL TOP plasma INR values. Using the plasma INR as the reference method for dosing decision making, appropriate dosing recommendations would have occurred for 97.7% of CoaguChek and 81.6% of Coag-Sense results. Conclusion Compared to our institution's plasma reference method, INR values obtained from the CoaguChek devices (XS Pro and XS Plus) exhibited less systematic bias compared to the Coag-Sense values. This translates to greater percentage of concordant management decisions between POC and laboratory plasma INR methods when the CoaguChek device is used.