A-161 Analytical Performance of Reagents for Determination of Activated Partial Thromboplastin Time: A Sigma Metric Perspective

Abstract

Abstract Background Sigma Metric (SM) assesses the degree of deviation of lab. tests from established specifications, combining Total Acceptable Error (ETa), imprecision and analytical bias. Quality criteria, defined by models of desirable specification, are used to determine the analytical performance of lab. tests. The concepts of biological variation are used as the preferred approach for establishing specifications. Reagents (RE) for the determination of activated partial thromboplastin time (APTT) use thromboplastin obtained from different tissue extracts and activation factors, being standardized to reduce the variability of the tests, and ensure equivalence of clinical diagnosis. This work evaluated three RE for APTT and the fulfillment of the desirable specifications of analytical quality, using the SM as an analysis tool. Material & methods APTT determinations were performed on the Sysmex®CS-2500 (Siemens Healthineers) with Actin®FS (AFS), Actin®FSL (FSL), and Pathromtin®SL (PSL) RE and bi-level Control Plasma N (CN) and Ci-Trol®2 Control (CI). Two different lots were used for each RE and control (CT), and the results were analyzed in RE-CT combination. We determined (a)analytical bias (Bias), from intra-assay imprecision determined by 30 repetitions in a single analytical run and by comparison with manufacturer's data, and (b)analytical coefficient of variation (CVa), from inter-assay imprecision, by triplicate in 3 periods each 3 h, for the 2 CT levels. Analytical quality specifications for APTT were selected in the literature, adopting desirable ETa values of 4.5% and minimum of 6.7%, obtained from a database for biological variation and acceptable limit of 15%, based on CLIA-CAP. SM was calculated with formula: Sigma = ETa–Bias/CVa, with graphic representation adapted from the Westgard’s model. Results For CVa, AFS and FSL results with CN and CI were <1.0%. PSL presented results above 1.0% for CN and CI. All RE presented results lower than those recommended by the manufacturer, respectively 4.0% for AFS; 3.0% for FSL; and 2.8% for PSL. For bias, a greater variation was observed for the combinations of RE and CT, ranging from 1.16% to 4.07%, being related to the differences between the averages obtained and the values defined in the leaflets. For the desirable specification of ETa (4.5%), 3 combinations of RE and CT level AFS-CN, FSL-CN and PSL-CI presented Sigma values >3 (respectively, 3.61, 3.87 and 3.33). When using the minimum specification (6.7%), the PSL-CN combination presented a Sigma = 2.06. With the CLIA-CAP specification, all RE and CT combinations had Sigma >6. Discussion All RE showed similar analytical performance when compared by SM, demonstrating that different commercial formulations for APTT determination do not show differences in analytical Bias and CVa determinations. The graphical representation by SM allows you to visually compare analytical imprecision and inaccuracy values. Conclusion All RE showed no differences in analytical performance, with SM being a visual tool for analyzing and comparing RE performance.