Results of a collaborative study for fibrinogen measurement. Evidence that the use of a common calibrator improves interlaboratory agreement

Abstract

One hundred and sixty-five laboratories were asked in two different surveys to perform duplicate fibrinogen measurements on the same lyophilized plasmas by their own methods and reagents and two different methods of calibration. The between-laboratory variability, expressed as coefficient of variation (CV), and the extent of between-method and between-reagent differences were taken as a measure of agreement between measurements. It was found that the overall between-laboratory range of CV values was 9-16% and that it could be improved if a common calibrator was used. The improvement was more pronounced for the thrombin time-derived (Clauss) method (average CV reduction 43%) than for the prothrombin time (PT)-derived method (average CV reduction 12%). A satisfactory agreement was found between the PT-derived and the Clauss method for the normal-PT plasma, whereas the PT-derived was slightly higher than the Clauss method for the abnormal-PT plasmas, irrespective of calibration. There were statistically significant between-reagent differences within the Clauss method, which were reduced if the common calibrator was used. In conclusion, the results shows that a system of common calibration for fibrinogen determination improves interlaboratory agreement by reducing the CV and minimizing between-reagent differences, particularly for the Clauss method. The authors recommend manufacturers of reagents to link their calibrators to the International Standard for fibrinogen.