Proficiency testing and standardization of prothrombin time: effect of thromboplastin, instrumentation, and plasma.

Prothrombin times accumulated from ten different proficiency testing surveys were analyzed in terms of a linear additive model described by Evatt et al (Clin Lab Haemat 1981; 3:331-342). Different types of lyophilized plasma samples were used, i.e., plasmas artificially depleted of coagulation factors by adsorption to aluminum hydroxide, and pooled plasmas of patients receiving coumarin drugs. For each plasma sample, both instruments and thromboplastins had a highly significant effect on the prothrombin time. For most instruments and thromboplastins, a good correlation was found between instrument effect or thromboplastin effect and the mean prothrombin time if various artificially depleted plasma samples from a single manufacturer were used. Artificially depleted plasmas from a second manufacturer gave different relationships between estimated effects and mean prothrombin time. Relationships based on lyophilized pooled patient plasmas were different from those of artificially depleted plasmas from either manufacturer. The potential use of the additive linear model for standardization of the prothrombin time in monitoring oral anticoagulant therapy is discussed. Additional studies are required to establish the suitability of this model to define a universal scale for prothrombin times of fresh plasma samples of anticoagulated patients. If suitable, the model can be linked to the International Normalized Ratio proposed by the World Health Organization.