Precision, accuracy, cross reactivity and comparability of serum indices measurement on Abbott Architect c8000, Beckman Coulter AU5800 and Roche Cobas 6000 c501 clinical chemistry analyzers

Background:The aim of our study was to perform verification of serum indices on three clinical chemistry platforms.Methods:This study was done on three analyzers: Abbott Architect c8000, Beckman Coulter AU5800 (BC) and Roche Cobas 6000 c501. The following analytical specifications were verified: precision (two patient samples), accuracy (sample with the highest concentration of interferent was serially diluted and measured values compared to theoretical values), comparability (120 patients samples) and cross reactivity (samples with increasing concentrations of interferent were divided in two aliquots and remaining interferents were added in each aliquot. Measurements were done before and after adding interferents).Results:Best results for precision were obtained for the H index (0.72%–2.08%). Accuracy for the H index was acceptable for Cobas and BC, while on Architect, deviations in the high concentration range were observed (y=0.02 [0.01–0.07]+1.07 [1.06–1.08]x). All three analyzers showed acceptable results in evaluating accuracy of L index and unacceptable results for I index. The H index was comparable between BC and both, Architect (Cohen’s κ [95% CI]=0.795 [0.692–0.898]) and Roche (Cohen’s κ [95% CI]=0.825 [0.729–0.922]), while Roche and Architect were not comparable. The I index was not comparable between all analyzer combinations, while the L index was only comparable between Abbott and BC. Cross reactivity analysis mostly showed that serum indices measurement is affected when a combination of interferences is present.Conclusions:There is heterogeneity between analyzers in the hemolysis, icteria, lipemia (HIL) quality performance. Verification of serum indices in routine work is necessary to establish analytical specifications.