Assessment of the Analytical Performance of 14 Analytes using the Epoc® Blood Analysis System

Objective: To evaluate the analytical performance of Epoc® Blood Analysis System for 14 analytes (pH, pCO2, pO2, HCO3-, BE, sO2, Na+, K+, iCa2+, Cl-, Glu, Lac, Crea and BUN) Material and Methods: The coefficient of variation (CV%) was calculated based on a between-day replication study using internal quality control material at two concentrations. The relative mean difference (BIAS%) was calculated based on method comparisons of 53 to 55 arterial patient samples using ABL 835 Flex Blood Gas Analyzer (Radiometer) and Dimension Vista 1500 System (Siemens Healthineers). The total analytical error (TAE%) was estimated by calculation of the 95% confidence interval, which incorporates the observed CV% from the replication study and BIAS% from the method comparison study. Each analyte’s precision, trueness and accuracy were assessed by comparing the observed CV%, BIAS% and TAE% to the analytical performance specifications (APS) from Westgard for imprecision (I%), bias (B%) and total allowable error (TE%), respectively. The analytical performance using the Epoc were considered acceptable in clinical settings if at least the minimum specifications for accuracy were achieved. Results: pH, BE, K+, Glu, Lac and BUN fulfilled the minimum specifications for precision, while pCO2, HCO3-, Na+, iCa2+, Cl- and Crea did not. pH, pCO2, Na+, K+, Glu, Lac and BUN fulfilled the minimum specifications for trueness, while HCO3-, iCa2+, Cl- and Crea did not. pH, pCO2, BE, K+, Glu, Lac and BUN fulfilled minimum specifications for accuracy, while iCa2+ did not. No specifications were specified for pO2 and sO2. Conclusions: pH, pCO2, BE, K+, Glu, Lac and BUN showed analytical performances considered acceptable for use in clinical settings, since at least the minimum specifications regarding accuracy were achieved. iCa2+ showed unacceptable analytical performance for use in clinical settings, whereas the results for HCO3-, Na+, Cl- and Crea were inconclusive.