Troponin assays: More questions than answers?

See also pp. 465–71 The introduction of cardiac troponin assays has revolutionized the diagnosis, definition and treatment of acute coronary syndromes (ACS). They are both highly sensitive for the detection of myocardial damage and, in the setting of chest pain, powerful prognostic indicators of future adverse cardiac events.1 Over the decade since their introduction, understanding of troponin has evolved. Troponin elevation was initially thought to be highly specific for cardiac damage. It is now increasingly recognized that levels can be elevated in conditions other than ACS (Table 1).2,3 In some, but not all of these, an association with increased mortality has been shown. Tests have also become more accurate, resulting in lowering of definitions for the upper limit of normal, for example, for troponin I, from the order of 2 to 0.04 µg/L: a 50-fold change! This has resulted in increased sensitivity for ACS, but at the expense of specificity.4,5 Concerns about diagnostic specificity and test performance have led some ACS experts to suggest that a threshold level of two to five times the upper limit of normal should be used to guide therapeutic intervention, rather than simply a level about the so-called normal range.6 This is supported by a small retrospective study that found that 49% of patients with an initial troponin I in this ‘grey zone’ had a troponin rise pattern consistent with ACS on serial testing, whereas 36% had a pattern consistent with chronic troponin elevation without acute myocardial damage, and 15% had a rapid rise and fall pattern not indicative of myocardial injury.7 The findings reported by Flindell and colleagues add to both our understanding and potentially our confusion about what an abnormal troponin assay in the ED means.8 Interestingly, they have looked at the significance of an abnormal troponin assay from a fresh angle – that of outcome of any patient who had the test, rather than confining their study to a chest pain or cardiac symptom cohort. This approach has both strengths and weaknesses. Their results clearly demonstrate a strong relationship between an abnormal troponin level and increased short-term mortality and hospital re-admission. This in itself is not surprising. More challenging is the finding that, for the small number of patients discharged from the ED with an abnormal troponin level, short-term mortality was so high (mortality 11.3% at 2 days and 19.6% at 28 days; odds ratio compared with patients discharged with normal troponin 60.4 at 2 days and 19.4 at 28 days). The reason for this finding remains unexplained. The authors report that 79% of the discharged troponin-positive group had discharge diagnoses other than ischaemic heart disease, but do not report what they were. They also do not report what proportion had serial biomarker testing or the proportions with acute and chronic troponin elevation patterns. Importantly, the research design used in the study does not allow for individual case review, so assessment of clinical decision-making and its relationship with cause of death cannot be made. What proportion of these cases truly represents missed preventable mortality remains unanswered. We also need to put these alarming numbers in perspective. The discharged troponin-positive group represents 0.92% of patients discharged from the ED and 0.22% of the total sample of 73 965 presentations. The authors make a case for hospital admission for all patients with an abnormal troponin, irrespective of cause, on the basis of potentially preventable mortality. Hospital admission of itself does not ensure adequate assessment or management of patients. Indeed, a narrow focus on an abnormal troponin as potentially indicating ACS might result in inappropriate investigations or treatment, at risk to the patient and cost to health services. Perhaps of more concern, it could also result in a delay to appropriate treatment for the patient's condition. What is required is appropriate clinical assessment, serial biomarker assays (if indicated) and a management plan tailored to the patient's condition. The study's results cannot tell us one way or the other whether appropriate assessments were made in the cases in question, or had the patients been admitted whether a different outcome might have occurred. A significant proportion of troponin-positive patients, particularly those with chronic troponin elevations, might be quite appropriately managed in the community. This does not underplay their mortality risk. Rather, it accepts that these patients have chronic illnesses requiring long-term supervision. The study also does not tell us what proportion of patients chose to leave hospital against medical advice. Studies such as this are powerful tools to alert clinicians to potential problems in systems of care. Their weakness is the inability to provide data about the contributing factors to the outcome found. They are a starting point. The hypotheses they raise, in this case that there is preventable mortality associated with a positive troponin level in patients discharged from the ED, require detailed analysis at the case level to confirm or refute the hypothesis, and to identify improvements to systems of care. We need to understand ‘why’ if we are going to fix it. The troponin story, particularly from the emergency medicine perspective, is far from over. Debate continues about the significance and management of troponin levels in the one to five times upper limit of normal zone. The use of troponin for risk stratification of conditions other than ACS is also under active exploration. Timing of serial troponin assays is in flux as more sensitive tests become more widely available. And research continues to identify even better biomarkers for identification of ACS. Watch this space!