Random uncertainty of photometric determination of hemolysis index on the Abbott Architect c16000 platform

Abstract

BackgroundAutomatic photometric determination of the hemolysis index (HI) on serum and plasma samples is central to detect potential interferences of in vitro hemolysis on laboratory tests. When HI is above an established cut-off for interference, results may suffer from a significant bias and undermine clinical reliability of the test. Despite its undeniable importance for patient safety, the analytical performance of HI estimation is not usually checked in laboratories. Here we evaluated for the first time the random source of measurement uncertainty of HI determination on the two Abbott Architect c16000 platforms in use in our laboratory. MethodsFrom January 2016 to September 2017, we collected data from daily photometric determination of HI on a fresh-frozen serum pool with a predetermined HI value of ~100 (corresponding to ~1g/L of free hemoglobin). Monthly and cumulative CVs were calculated. ResultsDuring 21months, 442 and 451 measurements were performed on the two platforms, respectively. Monthly CVs ranged from 0.7% to 2.7% on c16000–1 and from 0.8% to 2.5% on c16000-2, with a between-platform cumulative CV of 1.82% (corresponding to an expanded uncertainty of 3.64%). Mean HI values on the two platforms were just slightly biased (101.3 vs. 103.1, 1.76%), but, due to the high precision of measurements, this difference assumed statistical significance (p<0.0001). ConclusionsEven though no quality specifications are available to date, our study shows that the HI measurement on Architect c16000 platform has nice reproducibility that could be considered in establishing the state of the art of the measurement.