Can changes in troponin results be useful in diagnosing myocardial infarction?

In 1979, the WHO defined 3 specific criteria for use in the diagnosis of myocardial infarction (MI): (a) clinical symptoms typical for ischemic heart disease, (b) specific electrocardiogram changes, and (c) a typical change in serial measurements of cardiac enzymes (1 ). Later definitions have continued in the mold of listing specific criteria, and in the current definition of MI, a characteristic increase and/or decrease in a cardiac biomarker with at least 1 result 99 percentile of a healthy population and evidence of myocardial ischemia (either specific symptoms or electrocardiographic or ultrasound findings) are mandatory (2 ). During the last decade, troponin T or I molecules, which are found only in myocardial cells and released after cell necrosis, have been the preferred cardiac biomarkers because of their high sensitivity and specificity (3 ), although myocardial cell necrosis can be seen in conditions other than myocardial ischemia (renal failure, sepsis, treatment with cardiotoxic drugs, myocarditis, and so on). A common problem in clinical practice is how to diagnose an acute MI in a patient who already has increased troponin concentrations. The National Academy of Clinical Biochemistry has suggested that a 20% increase in troponin results can be used to diagnose an ischemic event. This recommendation is based on analytical variation for troponin analysis alone, without taking biological variation into account (4 ). Recently, troponin assays with better analytical sensitivity have become commercially available. These assays not only have opened up new diagnostic opportunities but also have challenged our understanding of ischemic heart disease, because the assays can detect low troponin concentrations, even in the healthy population (5, 6 ). The natural fluctuations of troponin concentrations around a homeostatic set point (i.e., within-person biological variation) as well as the between-person biological variation can be established, as is reported by Vasile et al. in this issue of Clinical Chemistry for a highly sensitive troponin T assay (7 ) and was reported in 2009 by Wu et al. for a highly sensitive troponin I assay (8 ). A low ratio of the within-person (plus analytical) variation to the between-person variation (as seen in these 2 studies) would suggest that the circulating concentration of troponin is a highly individualistic measure. For this reason, this ratio has been termed the “index of individuality,” and the results of tests with low ratios are quite person specific. Given a low index of individuality for troponin, it is easy to understand how an MI may occur in persons whose natural troponin concentration is quite low and in whom a troponin concentration indicative of MI might still fall below the 99th percentile. Likewise, because of this individuality, an increased troponin result in some persons might not imply an MI, despite being 99th percentile (5, 9 ). Thus, after the implementation of the new, more-sensitive assays, one can anticipate that the problem of differentiating between troponin increases due to nonischemic causes and those due to true ischemic events will occur more frequently (10 ). With the highly sensitive troponin assays, it is conceivable that measurement of changes in troponin results might provide a better— or an additional—way to diagnose MI. To judge whether changes in serial test results reflect true differences, one must know the biological variation of the measured constituents, especially the within-person biological variation (CVi). Although data on CVi exist for many constituents, these data need to be interpreted with caution because not all have been obtained with a strict methodology. For example, in a recent study of urine albumin, the CVi information for the albumin/creatinine ratio obtained from a review of 25 reports ranged from 4% to 103% (11 ). It is therefore of the utmost importance that reports on biological variation strictly define the methods used to obtain the data (e.g., exclusion of outliers, tests for homogeneous variances, sampling time, posture, preanalytical conditions, time of measurement, and storage of the samples before measurement). Different critical-appraisal forms exist for evaluating reports of diagnostic accuracy (12 ), and similar critical1 Laboratory of Clinical Biochemistry, Haukeland University Hospital, Bergen, Norway; 2 The Norwegian Quality Improvement of Laboratory Services in Primary Care (NOKLUS), Department of Public Health and Primary Health Care, University of Bergen, Bergen, Norway. * Address correspondence to this author at: Laboratory of Clinical Biochemistry, Haukeland University Hospital, Helse Bergen HF, Postbox 1, 5021 Bergen, Norway. Fax 47-55975976; e-mail [email protected]. Received April 23, 2010; accepted April 23, 2010. Previously published online at DOI: 10.1373/clinchem.2010.146639 3 Nonstandard abbreviations: MI, myocardial infarction; CVi, within-person biological variation; RCV, reference change value; CVa, underlying analytical imprecision; LR, likelihood ratio. Clinical Chemistry 56:7 1047–1049 (2010) Editorials