Aldo Clerico1, Martina Zaninotto2, Andrea Padoan2, Rudina Ndreu3, Veronica Musetti4, Silvia Masotti4, Concetta Prontera4, Claudio Passino4, Gian Carlo Zucchelli3, Mario Plebani2, Marco Migliardi5. 1. Fondazione CNR-Regione Toscana, G. Monasterio, and Scuola Superiore Sant'Anna, Pisa, Italy. Electronic address: clerico@ftgm.it. 2. Dipartimento di Medicina di Laboratorio, Azienda Ospedaliera Universitaria di Padova, and Dipartimento di Medicina, Università di Padova, Padova, Italy. 3. CNR Institute of Clinical Physiology and CNR spin-off QualiMedLab, Pisa, Italy. 4. Fondazione CNR-Regione Toscana, G. Monasterio, and Scuola Superiore Sant'Anna, Pisa, Italy. 5. S.C. Laboratorio Analisi, A.O. Ordine Mauriziano di Torino, Torino, Italy.
Abstract
BACKGROUND: The study aim was to evaluate whether is possible to harmonize the results of two hs-cTnI methods using a recalibration procedure based on linear regression models and measured values of external quality assessment (EQA) and clinical samples. METHODS: A two-step experimental approach was used. The preliminary step was performed using 16 EQA samples. The harmonization procedure was then validated by using 2530 heparinized plasma samples collected by 12 Italian University and Regional Hospitals from apparently healthy volunteers and patients admitted to emergency department with cardiac diseases. Two hs-cTnI methods were tested: Architect Stat High Sensitive Troponin-I, and the Access hs-cTnI using DxI platform. Linear regression models based on mean values measured with the two hs-cTni methods were used. RESULTS: A significant reduction in the measurement bias between the two methods was found after recalibration procedure. The agreement between-methods improved of about 2.5 folds after recalibration, as assessed by reduction in mean CV values from 38.4% (SD 25.9%) before recalibration to 15.0% (SD 10.6%) after recalibration for hs-cTnI values ≤ 500 ng/L (n = 13, P = 0.0111 by non-parametric test for paired data). CONCLUSIONS: A recalibration procedure based on means of measured concentrations with hs-cTnI methods, which use monoclonal antibodies with similar binding characteristics, can be used to significantly reduce systematic bias and so to improve harmonization between methods. The results of this study can aid laboratorians and clinicians to better compare the concentrations respectively measured with the Architect and Access hs-cTnI methods.
BACKGROUND: The study aim was to evaluate whether is possible to harmonize the results of two hs-cTnI methods using a recalibration procedure based on linear regression models and measured values of external quality assessment (EQA) and clinical samples. METHODS: A two-step experimental approach was used. The preliminary step was performed using 16 EQA samples. The harmonization procedure was then validated by using 2530 heparinized plasma samples collected by 12 Italian University and Regional Hospitals from apparently healthy volunteers and patients admitted to emergency department with cardiac diseases. Two hs-cTnI methods were tested: Architect Stat High Sensitive Troponin-I, and the Access hs-cTnI using DxI platform. Linear regression models based on mean values measured with the two hs-cTni methods were used. RESULTS: A significant reduction in the measurement bias between the two methods was found after recalibration procedure. The agreement between-methods improved of about 2.5 folds after recalibration, as assessed by reduction in mean CV values from 38.4% (SD 25.9%) before recalibration to 15.0% (SD 10.6%) after recalibration for hs-cTnI values ≤ 500 ng/L (n = 13, P = 0.0111 by non-parametric test for paired data). CONCLUSIONS: A recalibration procedure based on means of measured concentrations with hs-cTnI methods, which use monoclonal antibodies with similar binding characteristics, can be used to significantly reduce systematic bias and so to improve harmonization between methods. The results of this study can aid laboratorians and clinicians to better compare the concentrations respectively measured with the Architect and Access hs-cTnI methods.