Biological variation of International Normalized Ratio for prothrombin times, and consequences in monitoring oral anticoagulant therapy: computer simulation of serial measurements with goal-setting for analytical quality.

Oral anticoagulant therapy (OAT) has a well-established efficacy in prophylaxis and treatment of thromboembolic disorders. Because complications are related to intensity of OAT, optimal control of treatment is mandatory. In studies of OAT, as many as 30% of International Normalized Ratio (INR) measurements for prothrombin times fall outside the therapeutic interval. Preanalytical, analytical, and biological variation all contribute to this. Computer simulations of serial INR measurements were performed for various assumed in-treatment setpoints within the therapeutic interval INR 2.0-3.0 and for an "in-treatment within-subject variation" (CV) of 10.1%. Results are presented in difference plots with therapeutic intervals and critical differences. If the in-treatment setpoint is mid-interval (INR = 2.5), only 5% of simulated INR values fall outside the therapeutic interval. Setpoints deviating from the mid-interval and increases in the in-treatment within-subject variation considerably increase the number of observations outside the therapeutic interval and the critical differences. In conclusion, random variation, biological or analytical, and setpoints (targets) deviating from mid-interval explain a substantial number of the INR values outside therapeutic intervals observed in clinical studies. Analytical imprecision should be kept < 5% and analytical bias < +/- 0.2 INR.