Troponin measurement is not straightforward

EDITOR—McKenna and Forfar describe how the proposed redefinition of myocardial infarction has made the measurement of cardiac specific troponin central to its diagnosis.1 In a patient presenting with chest pain a troponin concentration above the 99th centile of normal is now sufficient to diagnose myocardial infarction, irrespective of any electrocardiographic changes. Previously, as the authors state, a patient had to show the development of Q waves on electrocardiography or an increase in creatine kinase activity to more than twice the upper reference limit before this judgment was made. The widespread introduction of troponin T and troponin I measurements is an undoubted improvement, but limitations exist with the assays, which most clinicians are unaware of and may give rise to diagnostic difficulties.

One issue concerns the use of the 99th centile of normal as the cut off point for myocardial infarction. Most assays are not sensitive enough to measure values as low as this. For example, the 99th centile for troponin T (Roche Diagnostics, Lewes, UK) is about 0.01 μg/l, but the laboratory assay in routine use is unable to measure reliably below 0.03 μg/l. This means that patients with troponin T concentrations between 0.01 μg/l and 0.03 μg/l, who would be defined by the new criteria as having had a myocardial infarction, are currently being missed and included in the low risk category. McKenna and Forfar mention the use of bedside as opposed to laboratory troponin measurement, but near patient tests are less sensitive again, with the troponin T example being suitable for measurement only down to 0.1 μg/l.

A second issue is that troponin measurement involves immunoassay analytical techniques as opposed to the more robust enzymatic methods used with traditional cardiac markers such as creatine kinase. As Ismail and Barth pointed out, immunoassay methods can—unpredictably—lead to wrong results in some people and in the context of troponins these are likely to be difficult to identify.2

These and other limitations of troponin assays should not detract from the value they provide in most patients tested. They do raise concerns, however, if these tests are being used as the sole means of diagnosing myocardial infarction in a patient presenting with chest pain.

Variations in access to and interpretation of troponin assays are wide

EDITOR—Cardiospecific markers, such as troponin T and troponin I and creatine kinase MB are more sensitive than creatine kinase in detecting ischaemic myocardial necrosis and predicting prognosis in acute coronary syndromes and coronary interventions.1 2 3 In 2000 a joint committee of the European Society of Cardiology and American College of Cardiology recommended changing the diagnostic criteria for acute myocardial infarction to take account of these findings.4

According to the new criteria, acute myocardial infarction should be diagnosed by a raised concentration of troponin T, troponin I, or creatine kinase in addition to typical symptoms, changes on electrocardiography, or coronary intervention. Even minor amounts of myocardial necrosis would thus be classified as myocardial infarction. As a result, some patients previously diagnosed as having unstable angina will instead be classified as having acute myocardial infarction.

Changes to the diagnostic criteria may lead to misinterpretation of epidemiological studies and clinical audits in which trends are analysed over time.5 Furthermore, piecemeal implementation of the new criteria may lead to erroneous comparisons between hospitals, particularly in respect of clinical indicators by which standards of care are often judged.

We undertook a postal survey of all 71 cardiologists and physicians with an interest in cardiology in Scotland to determine whether they currently have access to troponin T, troponin I, or creatine kinase MB assays, and whether the new criteria for acute myocardial infarction are being implemented in their units. Sixty four (90%) clinicians responded, providing at least one response from each of the 29 relevant hospitals. Five respondents had no access to troponin or creatine kinase MB assays. Twenty two had access to troponin T, 23 to troponin I, and 27 to creatine kinase MB. Access was often by special request only. Of the 59 respondents with access to at least one assay, only 36 (61%) routinely used it for all patients presenting with possible acute myocardial infarction.

Across Scotland percutaneous coronary interventions are undertaken in six NHS cardiac catheterisation laboratories. In two, troponin concentrations were checked routinely after all procedures. In the remaining four, they were checked only in selected patients. There was no consensus on the diagnostic label that should be given to a patient presenting with typical chest pain who had a significant rise in troponin in the absence of typical electrocardiogram changes or a significant rise in cardiac enzymes. Twenty respondents diagnosed myocardial infarction, 25 unstable angina, and 17 used the non-specific label of acute coronary syndrome.

Clinicians and hospitals are currently adopting the new criteria inconsistently. All clinicians need to have access to assays for cardiospecific markers, and consistent interpretation of the results is essential to prevent epidemiological chaos.