Effect of plasma pooling on the International Sensitivity Index of prothrombin time systems

Guidelines set by the World Health Organization (WHO) state that in order to calibrate a prothrombin time system for International Sensitivity Index (ISI), freshly prepared specimens from 20 normal individuals and 60 patients receiving coumarin are required. These numbers are required to obtain a precise value of the calibration line slope when there is considerable scatter of individual data about the regression line. The scatter can be reduced by pooling individual plasma samples. In the present study, four pooled plasmas were prepared, one from 20 normal individuals and three from three groups of 30 patients receiving treatment with long-term oral anticoagulants. Prothrombin times were determined with four thromboplastins, HepatoQuick (rabbit thromboplastin combined with adsorbed plasma), Recombiplastin (recombinant human thromboplastin), Thromborel-S (human placenta), and Thromboplastin-C Plus (rabbit brain). Calibration line slopes were calculated for the six possible combinations of thromboplastins using the set of all individual plasma samples and the set of four pooled plasmas. In most comparisons, the WHO calibration model was appropriate, i.e. the line calculated for the patients' samples passed through the mean of the normals. The calibration line slopes obtained with the set of four pooled plasmas did not differ by more than 5% from the corresponding slopes calculated with the original individual plasmas. For some combinations of thromboplastins non-linear relations were observed both with the individual plasmas and with the pooled plasmas. We conclude that pooling individual plasmas does not significantly change the calibration relation between prothrombin times determined with the original individual plasmas. Freshly pooled plasmas can be used to determine the ISI of prothrombin time systems with an acceptable degree of precision.