The universal definition of myocardial infarction: some issues and concerns: reply

We are pleased to increase the educational impact of the new universal definition of myocardial infarction (MI)1–3 by responding to Lanza, who wonders why consistency was not the only criteria we used. The obvious response is that some important scientific information is unknown; there are data in some areas that suggest different approaches and strongly held convictions in others that needed to be balanced against blind obedience to consistency. The diagnosis of spontaneous MI remains unchanged from our previous document in the year 2000,4,5 mandating a rise and/or fall of troponin values with at least one value above the 99th percentile of the upper reference limit together with evidence of myocardial ischaemia.1–5 However, the diagnosis of PCI-related MI has changed somewhat from the original document.1–5 The reasons for mandating a normal baseline value and how to proceed when that value is elevated but stable are described in the document.1–3 The criterion employed when the baseline sample is at or below the 99th percentile is a compromise solution developed to accommodate the interventional community and the clinical trials groups who pointed to a vast literature on the use of three-fold increases to optimize prognostic accuracy. At first glance, such criteria may seem different from those proposed for spontaneous MI, but they are far closer if one includes the need to develop criteria to define the rise and/or fall. At the lower range of measurement such as the 99th percentile, there is greater imprecision for troponin assays than at higher ranges.6 Furthermore, at the lower level of measurement of troponin, a large percentage change (three standard deviations of the variability) is required for to be sure that the values are analytically different. If one also appreciates that values usually are reported to 100ths and thus the need for rounding, it becomes clear that it is necessary to show a very substantial increment of change for the analytical difference. For some assays as for the troponin T assay, that is a value fully three-fold higher than the 99th percentile range. For other assays, it may not be quite that high.