Practical implications of coexistent different technologies in clinical chemical laboratories. Solid phase chemistry and conventional analysis.

If different analytical methods are alternatively used for the determination of the same analyte, basic differences in test methodology can give rise to an increased number of deviating results. Such coexistence of methods might be necessary, for example, during a transition phase while upgrading to new technologies. We have exemplarily investigated this topic for the comparison of solid phase chemistry ("dry chemistry") versus conventional methods ("wet chemistry"). The Kodak Ektachem 700XR clinical chemistry analyser was compared with the Hitachi 737 analyser from Boehringer Mannheim using 18 clinical chemical analytes and specimens submitted for routine analysis. Before the start of the evaluation, the Ektachem 700XR was adjusted ("calibrated") by the manufacturer for optimal agreement with the Hitachi 737. Satisfactory agreement was obtained for most investigated analytes as judged by correlation coefficients and three commonly applied regression methods (linear regression, principal components, and Passing/Bablok method). For some analytes, however, strongly deviating results were often obtained. Quality control-derived limits (maximum acceptable inaccuracy) and data from biological variation (critical differences) were used for the assessment of the inter-instrument bias for diagnosis and patient monitoring, respectively. For enzymes, 0% (amylase) to 22% (creatine kinase) of all pathologic results differed by more than the maximum acceptable analytical inaccuracy (21%-27%) of these analytes. If more stringent limits derived from biological variation were used, 24% (creatine kinase)--62% (aspartate aminotransferase) of all differences between paired measurements exceeded the critical difference for enzymes. Deviations greater than the critical differences were also marked for serum concentrations of sodium, calcium, and creatinine.(ABSTRACT TRUNCATED AT 250 WORDS)