The performance of high sensitivity troponin for the diagnosis of acute myocardial infarction is underestimated

BACKGROUND: Many papers evaluating high sensitivity troponin assays make the diagnosis of myocardial infarction based on conventional troponin assays in clinical use at the time of recruitment. Such analyses often do not show superiority of high sensitivity assays compared with contemporary assays meeting precision guidelines. METHODS: Three hundred and twenty-two patients presenting to the emergency department between November 2006 and April 2007 for evaluation for acute coronary syndrome had serial (0 h and >6 h) bloods taken to compare troponin assays (Roche hsTnT, Abbott TnI, Roche TnT and Vitros TnI). The diagnosis of myocardial infarction was made using each troponin assay separately with which that same assay was analysed for diagnostic performance. RESULTS: The rate of myocardial infarction would be 38.9% using serial hsTnT, 31.3% using serial Abbott TnI, 27.1% using serial TnT and 26.4% using serial Vitros TnI. The baseline sensitivities (0 h) are 89.9% (85.2-93.3) for hsTnT, 77.9% (71.0-87.5) for Abbott TnI, 73.0% (65.6-78.7) for TnT and 86.8% (74.6-94.4%) for Vitros TnI. The specificities (peak 0 h and >6 h samples) are 93.1% (91.2-93.1) for hsTnT, 88.3% (86.5-88.3) for Abbott TnI, 92.2% (90.5-92.2) for TnT and 90.6% (70.1-90.6) for Vitros TnI. CONCLUSIONS: hsTnT has superior sensitivity for myocardial infarction than even assays at or near guideline precision requirements (Abbott and Vitros TnI). The specificity of hsTnT assay is not as poor as previous analyses suggest.