Discrepant sensitivity of thromboplastin reagents to clotting factor levels explored by the prothrombin time in patients on stable oral anticoagulant treatment: impact on the international normalized ratio system.

We tested the principle of local International Normalized Ratio (INR) calibration using INR calibrator plasmas (PT Calibration Plasma Kit, Behring), two thomboplastin reagents (Neoplastin plus, rabbit brain, Stago, and Recombiplastin, recombinant human tissue factor, Ortho Diagnostics) and the same coagulometer (STA, Stago) on 92 patients on stable oral anticoagulant treatment.A four-point calibration was obtained with each reagent by linear regression (sec/INR) on a log-log scale (r > or = 0.999). The bias between the two reagents (Recombiplastin - Neoplastin Plus) was reduced from 31.7% to 17.5% and 7.5% (p=0.001) when results were expressed, respectively, as PT ratio (using the mean normal prothrombin time as denominator term), INR (using instrument-specific ISI supplied by the manufacturers) and calibrated INR, but there was a consistently significant regression of the differences over the average values even after log transformation (r > or = 0.586). The bias between the reagents was reduced to 1% (p=ns) when assuming Recombiplastin as the reference thromboplastin and applying Tomenson's correction, but limits of agreements were as large as 20%. Factor VII, X, V and II activity was measured with the two thromboplastin reagents in all plasma samples using immunodepleted plasmas (Stago).Statistically significant biases were observed for all clotting factors with the two reagents (Recombiplastin Neoplastin Plus) and ranged from 3.5 % (FII) to 37.2% (FVII). In addition, for FVII and FV there was a significant regression of the difference over the average value (after log-transformation, r > or = 0.282). The patients were divided into 3 groups according to their degree of anticoagulation (INR <2.0; INR between 2.0 and 3.5; INR >3.5). Factor levels differed significantly with the two reagents throughout the 3 groups of patients. In addition, the relative distributions of the 3 vitamin K-dependent factors also differed in the 3 groups with the two thromboplastin reagents.The discrepant sensitivity to factor VII, X and V levels of the two thromboplastin reagents explored in this study prevents INR calibration with commercially available calibrator plasmas and is responsible for a significant variability in INR values even under optimal conditions of INR calibration.