Local Calibration of International Normalised Ratio Improves between Laboratory Agreement: Results from the UK National External Quality Assessment Scheme

Abstract

SummaryIn the present study we have performed local calibration of International Normalised Ratio (INR) measurement systems in a large series of laboratories. We assigned INRs to five lyophilised plasma calibrants, one prepared from normal plasma and four using plasma from war-farinised patients, using different International Reference Preparations for Thromboplastin. These five calibrants, and two lyophilised test plasmas were analysed by 349 centres using 60 different thromboplastin instrument combinations.Plasma calibrants were assigned INRs using the WHO reference thromboplastin RBT-90 or the European reference thromboplastin CRM 149R. Each participating centre determined PTs of the calibrants with their local system. These PTs were then used to construct a local calibration graph relating PT to INR. The PTs of test plasmas were converted directly into INR using the local calibration model and into INR using the conventional method. The overall medians of conventionally derived INRs of two test plasmas analysed in 349 centres were 2.50 and 3.10, compared to 2.47 and 3.04 after local calibration where RBT-90 was employed to assign INRs to calibrants. Use of CRM 149R to assign INRs to calibrants led to a significant (p <0.0001) increase in INR to 2.7 and 3.36 respectively. When results were grouped according to the thromboplastin employed, agreement between results with different reagents was improved by local calibration. There was a significant reduction (p <0.01) in the spread of results in different centres as indicated by a reduction in coefficient of variation.