High-Sensitivity Cardiac Troponin Assays—Change Is Important

To the Editor: The recent publication by Aldous and colleagues in Clinical Chemistry (1) starts an important discussion concerning the optimal change criteria that should be used with high-sensitivity (hs)1 cardiac troponin assays. This report suggests that an optimal relative change (i.e., percent δ) in hs cardiac troponin T (hs-cTnT) for predicting a major adverse cardiovascular event over a 1-year period is 3%. This relative difference was not significant, however, with respect to risk stratification (hazard ratio, 1.6; P = 0.052) and would not be considered an analytically robust change, because 3% is within the imprecision of the assay at concentrations measured with the low- and high-quality control materials provided by Roche Diagnostics for this assay. From the data presented, it appears that the percent δ values for hs-cTnT concentrations over 2 h were not important for long-term risk stratification; however, relative changes (e.g., 10% as determined by ROC curve analysis) did improve the detection of the index diagnosis of myocardial infarction (MI) in patients presenting with symptoms suggestive of acute coronary syndrome (1). The authors provide data for using the lower relative change values observed in their study, as opposed to other reports in the literature that used ROC curve analyses for the diagnosis of MI (1). These authors also note that another study indicated that long-term prognostication using change criteria might not be beneficial; however; the proper citation for this finding should be the study by Kavsak et al. (2). Furthermore, the δ (as expressed as ratios) in the Kavsak et al. study, which failed to manifest prognostic significance over 4 years, was assessed only in the patients with documented myocardial injury detected with either the fourth-generation cTnT assay (n = 85) or the second-generation cTnI assay (n = 81), and not hs cardiac troponin assays (2). In a larger analysis from the FAST II (Fast Assessment of Thoracic Pain II) and FASTER I (Fast Assessment Of Thoracic Pain By Neural Networks I) studies, however, Eggers and colleagues (3) demonstrated that cTnI concentrations that exceeded the 99th percentile with at least a 20% change in concentrations identified patients at higher risk for death and myocardial infarction at 6 months and death at a median follow-up of 5.8 years. Interestingly, in this analysis the magnitude of change with this guideline-acceptable assay (i.e., Stratus CS) did not translate into higher event rates (3). These studies highlight the fact that δ is important, with the interpretation of δ being reliant on both the cardiac troponin assay characteristic and the indication for its use (i.e., diagnostic or prognostic). With respect to prognosis, measuring with an hs-cTnI assay and assessing change in a larger chest pain population (cohort n = 223) revealed that δ expressed as either an absolute concentration or a percentage difference was useful for predicting death and/or MI at 1 year (4). Interestingly, absolute δ appeared to be superior to percent δ, because the area under the ROC curve for absolute δ was higher than the area under the curve for percent δ, with only absolute δ providing important information for earlier risk stratification (e.g., at 30 days and 6 months) (4). Aldous and colleagues (1) did not assess absolute δ in their study for either MI diagnosis or subsequent prognosis. A recent publication from the Advantageous Predictors of Acute Coronary Syndromes Evaluation (APACE) trial, however, demonstrated the superiority of absolute δ over percent δ for the diagnosis of MI (5). The findings from the studies of Kavsak et al. and Reichlin et al. (4, 5) suggest that change is also important for hs cardiac troponin assays, with absolute δ possibly providing important diagnostic and prognostic information. Future studies that assess either absolute δ or possibly a combination of both absolute δ and percent δ are required to determine the optimal change for predicting a serious cardiac outcome in patients presenting with chest pain to the emergency department. high-sensitivity high-sensitivity cardiac troponin T (assay) myocardial infarction Fast Assessment of Thoracic Pain II (study) Fast Assessment of Thoracic Pain by Neural Networks I (study) Advantageous Predictors of Acute Coronary Syndromes Evaluation (trial).