Poor comparability of prothrombin fragment 1+2 values measured by two commercial ELISA methods: influence of different anticoagulants and standards

We compared F 1 + 2 results obtained with two commercial ELISA methods (Behring and Baxter) assaying the same plasma samples. There was little correlation between the results of the two methods, as shown by the low correlation coefficient (r = 0.50) and by low percentage of concordant classification (normal or abnormal) of the samples (24%). Such poor correlation is probably due to the different anticoagulants suggested, because correlation improved when both methods were carried out in plasmas collected with the same anticoagulant. However, the Baxter method still gave significantly lower F 1 + 2 values than the Behring method. Assuming that this difference is due to the use of standards with different F 1 + 2 concentrations, the standards from Behring and Baxter were evaluated by both methods. Parallel dose-response curves were obtained when the standards were run by the Behring method but not by the Baxter method, indicating that the two standards are qualitatively different. This study demonstrates that the two F I + 2 methods give different values for the same samples and that these values are poorly correlated. Standardization of the F 1 + 2 assays cannot be achieved easily simply by using a common standard and the use of different anticoagulants appears to be the main reason for poor standardization.