Methods Comparison and Bias Estimation of Three Distinct Biochemistry Analytical Systems in One Clinical Laboratory Using Patient Samples.

Abstract

BACKGROUND In China, one laboratory usually owns more than one diagnostic device or reagent kit measuring the same analyte and this situation causes great troubles for quality control and traceability. To determine if the different devices yield equivalent results, the correlation coefficients and predicted bias between three distinct bio-chemistry analyzers in our lab were evaluated. METHODS 40 analytes were detected and used to evaluate method performance and comparability of results between different analyzers. The Vitros5600 and Hitachi7170 analyzers were separately compared with the Cobas 8000 analyzer according to Clinical and Laboratory Standards Institute (CLSI) guideline (EP9-A2). Within-day and between-day imprecision of the three analytical systems were calculated in accordance with CLSI guidelines EP15-A2. RESULTS Comparing the Hitachi7170 with Cobas8000 analyzer, except for calcium, magnesium, chloride ion (CL-), and carbon dioxide, the other 36 analytes were closely correlated (R2 > 0.95), while 4 of the 36 analytes' predicted bias exceeded the acceptable criteria. As for the Vitros5600 and Cobas8000, except for albumin, sodium ion (NA+), magnesium, and chloride (CL), the other 13 analytes were closely correlated (R2 > 0.95), while 5 of the 13 analytes' predicted bias exceeded the acceptable criteria. CONCLUSIONS Significant differences for several analytes between distinct analyzers were found; for some analytes the predicted bias between dry chemical analyzer and conventional wet chemical analyzer can reach 30%, which is worthy of our concern. When one analyte was detected on more than one device, a strict method comparison study should be performed regularly. Reference intervals should be validated and transferred from the Cobas 8000 to Vitros5600 when the bias cannot be adjusted.