The correction of coagulometer effects on international normalized ratios: a multicentre evaluation.

One hundred and twenty-nine centres in the U.K. participated in a study to test the reliability of the three methods of correction for coagulometer effects on international normalized ratios (INR). Results from 37 centres which tested three warfarinized plasmas by the manual method were taken as the 'true' INR for the assessment of coagulometers. 63 centres (11 manual and 52 using coagulometers) determined their local International Sensitivity Index (ISI) in a calibration exercise. This was performed with a set of 20 lyophilized plasma calibrants with certified manual prothrombin times for the thromboplastin used in the study. The following methods of INR derivation were compared by assessing the percentage deviation from the three INR values established by 37 manual users: I. No coagulometer correction, i.e. (local PT/reference manual normal PT)manual ISI II. Coagulometer ratio correction, i.e. (local coagulometer PT/local coagulometer MNPT)manual ISI III. Local system ISI, i.e. (local coagulometer PT/local coagulometer MNPT)local system ISI IV. System ISI, i.e. (local coagulometer PT/local coagulometer MNPT)system ISI The local system ISI with the plasma calibrants (method III) gave the most reliable correction (mean deviation from 'true' INR 4.87%). The method which gave the least was with the coagulometer ratio correction, i.e. the manual ISI and local coagulometer MNPT (mean 11.25%). The system ISI tested with ACL coagulometers gave less correction than the local ISI calibration. The local system calibration with lyophilized plasmas also avoids some of the constraints on conventional thromboplastin calibrations.