Importance of device evaluation for point-of-care prothrombin time international normalized ratio testing programs

OBJECTIVE: To determine the accuracy of 2 commercially available point-of-care devises relative to plasma international normalized ratio (INR) values. PATIENTS AND METHODS: Point-of-care INR testing was performed with the CoaguChek and ProTime 3 devices in consecutive patients attending an anticoagulation clinic between June 18, 2003, and August 6, 2003. Results were compared with plasma INRs using a sensitive thromboplastin (International Sensitivity Index, 1.0). RESULTS: Ninety-four patients agreed to participate in the study. Relative to the plasma INR, values were in agreement ±0.4 INR unit 82% and 39% of the time for the CoaguChek and ProTime 3 devices, respectively. The mean ± SD CoaguChek INRs were 0.2±0.31 unit lower, whereas ProTime 3 INRs were 0.8±0.68 unit higher than plasma INR values. Treatment decisions based on these data would have resulted in inappropriate dose adjustments 10% and 22% of the time for these 2 respective devices. Correlation with plasma was greater for the CoaguChek (r ²=0.90) compared with the ProTime 3 device (r²=0.73). CONCLUSIONS: Optimal warfarin treatment requires accurate measurement of the IMR. The choice of a point-of-care device for INR management depends on the reliability of INR data generated by the device.