Background The measurement of cardiac troponin is central for the diagnosis of myocardial infarction (MI). It is recommended that a coefficient of variation of ≤10% is achieved at the diagnostic threshold and significant change between serial measurements reported. Many modern laboratories use multiple analysers linked by automation where samples are randomly assigned to an analyser. It is therefore important to consider the combined effect of all analysers on the analytical performance of troponin measurement. Method The performance of a contemporary troponin-I (cTn-I) assay run on three analysers, linked by an automated track, was undertaken across a range of cTn-I concentrations. The data for the three analysers were aggregated to obtain the combined analytical coefficient of variation (CVA) and reference change values (RCVs). Results The CVA improved with increasing concentration and calculated RCVs ranged from 67.2% (±13 ng/L) to 32% (±160 ng/L) between cTn-I values 20 ng/L and 500 ng/L. Although there were significant differences in cTn-I measurement between analysers around the diagnostic threshold ( P < 0.05), the CVA was 13.6%. Conclusions We demonstrate that there are significant differences between the performances of analysers which can impact the biochemical criteria for the diagnosis of MI. We also show that the RCV varies according to baseline cTn-I values and that reporting a single RCV across the analytical range of cTn-I may not be appropriate.
Keywords: Troponin; analytes; laboratory methods; statistics.