RESUMEN
This study aimed to assess analytical characteristics and diagnostic accuracy in management of venous thromboembolism (VTE) in the Emergency Department (ED) of the Abbott D-dimer assay applied on the Alinity c clinical chemistry analyzer (Abbott Laboratories, Chicago, IL) compared to the INNOVANCE D-dimer assay (Siemens Healthineers, Marburg, Germany). Precision was determined at three concentration levels following the CLSI EP15-A3 protocol. Method comparison and diagnostic accuracy were assessed using samples obtained from 85 patients who were referred for diagnostic imaging and D-dimer testing due to clinically suspected VTE. Within-run coefficients of variation (CVs) were 3.0%, 0.5% and 0.5% at D-dimer concentrations of 0.54, 1.42 and 2.68 mg/L FEU, while respective between-run CVs were 2.0%, 3.4% and 2.7%, hence fulfilling the desirable biological variation criteria for imprecision (<12.6%). Passing-Bablok regression analysis yielded a small proportional difference between the two compared assays (y = 1.09 (95% confidence interval (CI): 1.01-1.18) x + 0.09 (95%CI: -0.09 to 0.16)), while Bland-Altman analysis showed significant negative absolute (-0.6 mg/L FEU, 95%CI: -0.9 to -0.3) and relative mean bias (-14.1%, 95%CI: -20.3 to -7.9). Spearman's ρ was 0.979 (95%CI: 0.967-0.986). Inter-assay agreement relative to the cut-off was 92% (kappa coefficient = 0.547 (95%CI: 0.255-0.839)). Diagnostic sensitivity, specificity, positive and negative predictive values of the Abbott assay were 100%, 9.2%, 25.3% and 100%, respectively, compared to the following data for the INNOVANCE assay: 95.0%, 15.4%, 25.7% and 90.9%. Abbott D-dimer assay has shown excellent analytical precision, high comparability with the INNOVANCE D-dimer and high NPV at manufacturer's cut-off.
Asunto(s)
Tromboembolia Venosa , Humanos , Tromboembolia Venosa/diagnóstico , Productos de Degradación de Fibrina-Fibrinógeno/análisis , Valor Predictivo de las Pruebas , Química ClínicaRESUMEN
The aim of this study was to perform the analytical validation of Alinity c and i analyzers (Abbott Laboratories, Chicago, IL, USA) for 39 clinical chemistry tests and 17 immunoassays. Precision was evaluated at least at two concentration levels for 5 days in quintuplicate, following CLSI EP15-A3. Method comparison included parallel analysis of leftover routine samples on Alinity analyzers and the previously used Cobas c501 and e601 (Roche Diagnostics, Mannheim, Germany). Linearity was tested by preparing sequential sample dilutions with high analyte concentration, following the CLSI EP6 document. For clinical chemistry tests, within-run coefficients of variation (CV) were up to 6.0% (beta-2-microglobulin), while between-run CVs up to 5.4% (immunoglobulin M). Among immunoassays, the highest within-run CV was obtained for vitamin B12 (6.9%), while between-run for CA 19-9 (4.3%). Complete agreement with Roche analyzers was observed for 16 (41%) clinical chemistry assays and 6 (35%) immunoassays. Half of all evaluated assays did not meet the desirable biological variation criteria for bias, being especially exceeded for alpha1-antitrypsin, apolipoprotein A1, ceruloplasmin, complement C3 and C4, hemoglobin A1c, lipoprotein (a) and myoglobin, as well as some tumor markers (CA 125, CEA, fPSA, AFP, and ferritin), hormones (cortisol, DHEA-S, insulin) and vitamins (25-OHD). Linearity in the tested ranges was confirmed. Overall, this study revealed that precision criteria derived from manufacturer's claims were not satisfied for all assays while comparison study for some assays yielded differences that imply the need for additional assay evaluation prior to introduction into routine practice.