Usefulness of lyophilized calibration plasmas for International Normalized Ratio determination with the bovine combined thromboplastin (Thrombotest): results of a collaborative study

Abstract

The logical solution to account for the influence of coagulometers on the International Sensitivity Index (ISI) is local calibration with freeze-dried plasmas. However, because of their unpredictable behavior these plasmas must be validated before large-scale implementation. We report on a collaborative exercise designed to evaluate the suitability of a set of such plasmas used with Thrombotest in combination with a coagulometer provided by the manufacturer to be used with that reagent. This was a two-step study. First, one lot of reagent was calibrated against the international standard OBT/79 in two expert laboratories. The calibrated lot was then used as an intermediate standard to calibrate two additional lots of the same reagent in four field laboratories where the ISI was determined for both plasma and native blood. The International Normalized Ratio (INR) for the patient plasmas tested in each laboratory were calculated using two algorithms: the World Health Organization-recommended ISI mode (gold standard), and the simplified calibration plasma mode. In the latter, the INR was derived from the local calibration curve constructed by plotting the certified INR versus local coagulation times obtained with calibration plasmas. The between-algorithm INR differences indicate that this set of calibration plasmas may be employed for local INR calibration of the investigated reagent/instrument combination, especially when plasma is used for INR determination where the average INR (range) difference is 5% (3-13%) or 2% (3-8%) according to whether the INRs to calibration plasmas were assigned by the manufacturer or by the two expert laboratories. A slight but measurable difference of the INR may be predicted [9% (6-20%) or 6% (8-15%)] if this set of calibration plasmas is used for local calibration when native blood is employed for INR determination. Whether this bias is of practical significance is to be determined.