Effect of recalibration of the hs-TnT assay on diagnostic performance

Abstract

To the Editor,Apple and Jaffe [1] recently reported on an adjustment to the high-sensitivity cardiac troponin T (hs-TnT) assay and questioned the consequent clinical implications. We report further on the effect of restandardisation of the assay on diagnostic performance. In our study between November 2008 and February 2011 (Trial registration ACTRN12610000053022) the diagnostic accuracy of the hs-TnT assay for acute myocardial infarction (AMI) was determined in a well-characterised but unselected cohort of patients presenting to the emergency department with chest pain suggestive of acute coronary syndrome (ACS) in whom blood was drawn for assay at presentation and after 2 h [2]. Serum was stored at −80C for later cTnT °analysis in 2012 using the Roche cobas e601 hs-TnT assay (Roche Diagnostics, Sydney, Australia; reagent lot no. 00163704; calibrator lot no. 00165094). Clinical endpoint adjudication was performed by one of two cardiologists using all clinical information available up to at least 30 days following presentation including results of the hospital’s usual troponin assay performed at presenta-tion and after 6–8 h (Beckman Coulter AccuTnI, deci-sion cut-off 40 ng/L cTnI). The outcome for the analysis of diagnostic accuracy was AMI at presentation and was diagnosed according to the Universal Definition of Myo-cardial Infarction published at the time of adjudication [3]. Briefly, this was defined as evidence of myocardial necrosis together with clinical evidence of myocar-dial ischaemia (ischaemic symptoms, ECG changes, or imaging evidence). Necrosis was diagnosed on the basis of a rising or falling pattern of the laboratory cTnI concentrations, with at least one value above the 99th percentile.Following completion of the hs-TnT analysis in 2012 a technical bulletin (No: 12–023) was published by Roche Diagnostics (Penzberg, Germany) calling for adjustment to the calibration of the hs-TnT assay when performed with certain reagent lots [1]. All study samples were found to require adjustment. This was performed according to the manufacturer’s method in which a new value assignment of the calibrator set used in this study was required. The result was that following recalibration the number of values that were below the limit of detection (5 ng/L) decreased from 1178 (71.0%) to 560 (33.8%). Eighty-six of 1659 values were in the cTnT concentration range 8.71–13.94 ng/L, and when recalcu-lated were on average 5 ng/L above the decision cut-off of 14 ng/L in the range 14.04–18.93 ng/L. The difference between recalculated and original values was +38% at the decision limit of 14 ng/L compared with +15.3% at 30 ng/L cTnT (Figure 1).Seven hundred and thirty-seven patients were included in the clinical analysis and an adjudicated diagnosis of AMI at presentation was made in 50 (6.8%)