Evaluation of troponin T on AQT90 Flex and COBAS 8000 as a rule in/out tool in an emergency ward

BACKGROUND: Troponin measurement is the gold standard for diagnosis of Acute Myocardial Infarction (AMI). Troponin (highly sensitive (hs), T or I) is measured by immunochemistry instrument or by Point of Care (POCT). POCT can be useful in emergency lab or ward for a faster diagnosis of patients with chest pain. Our study compared analytical performance of a POCT AQT90 Flex (Radiometer Medical) (AQT) and TnThs Cobas 8000 (Roche Diagnostics) (Cobas). We also compared the clinical performance of both methods at recommended cut-off (14 ng/L for Cobas and 30 ng/
L for AQT).
METHODS: We selected 104 patients (296 samples) (range: 6-13822 ng/L) admitted in the Emergency ward for which at least 1 troponin determination (Cobas 8000) had been re-quested in the past 24 hours according to rule in/out procedure applied by this ward. Samples were then measured with the AQT. Inter-assay CV was maximum 8.6% and 9.6% for Cobas and AQT respectively. The cut-off defined as the 99th percentile for Roche was 14 ng/L and the recommended decision threshold value was 30 ng/L for Radiometer. Retrospective analysis of final diagnostic was obtained for all
participants: we considered as “true positive” patients for whom a final diagnostic was ST segment-Elevation Myocardial Infarction (STEMI) or non STEMI (NSTEMI).
RESULTS: On the whole range of measure, the 2 methods showed a good correlation (r2=0.98). Regression equation was Cobas = 0.98 AQT + 31 ng/L (95%CI of the intercept: (26.7;37.7) and 95% CI of the slope (0.96;1)). When we stratified, for the values <54 ng/L, the equation became Cobas = 0.52 AQT +1.1 ng/L (95%CI of the intercept: (-4.8;5.5) and 95% CI of the slope (0.39;0.69)). Bland and Altman plot did not show any bias. At admission [2-7 hours], 78 (81%) of admitted
patients were finally considered as AMI, sensitivity was 92 % [96%] for Cobas and 78% [91%] for AQT. Specificity was 15% for Cobas (cut-off 14ng/L) or 73% (cut-off 54 ng/L) and 76% for AQT.
CONCLUSIONS: Overall, there was a good correlation between the 2 methods. However, using a cut-off of 14 ng/L for Cobas is questionable for a rule in/out procedure in an emergency ward. Using 54 ng/L for Roche and 30 ng/L for AQT would have led to the best discrimination between patients presenting AMI or not.