Troponin and Heart Failure: An Early Warning System Worth Listening to?

Do we need more risk factors for heart failure?

Heart failure is common, costly and deadly [1], and the development and clinical use of prediction tools for early initiation of preventive treatment is essential. The major risk factors for heart failure are long since known, and are easily assessed in primary care. These include diabetes mellitus, left ventricular hypertrophy, valvular disease, ischemic heart disease and hypertension. Of these, hypertension confers the highest population-attributable risk, and it also underlies many of the other risk factors [2]. Using combinations of such easily accessible risk factors, three out of four people that develop heart failure can be identified [3]. These risk factors are identified and treated to an unsatisfactory extent today, and also in those at high risk for heart failure [4]. Hence, the unmet need for new risk factors for prediction of heart failure in the general population could be considered to be rather low, and one could argue that more efforts should be spent implementing existing knowledge into clinical practice rather than searching for new risk factors.

Increasing demands on assessment of clinical usefulness of new risk factors

In order to take a new risk factor from bench to bedside, more is needed than observations of an association of levels of biomarker to an adverse outcome. For clinical usefulness, the biomarker should also be demonstrated to add power to discriminate those who will suffer adverse outcomes from those who will not, beyond known risk factors [5].

Discriminative capacity has, until now, usually been investigated using the area under the receiver operating characteristics curve of sensitivity and specificity (the C-statistic). Used for this purpose, the C-statistic is the probability that, in a given pair of patients, a statistical model will assign a higher risk to the patient who in fact will suffer an adverse outcome than to the patient who will not. The use of this statistic for clinical discrimination purposes is associated with a few problems, mainly because the extent of the changes in predicted risk and the absolute risk of the patients are not captured by the C-statistic. The C-statistic is also a conservative method: many risk factors that we know are clinically important do not affect the C-statistic very much [6,7], which may lead to an underestimation of the importance of the risk factors [8].

As a solution to these problems, new measures of added discriminatory capacity have been proposed [5], namely, the net reclassification improvement (NRI) and the integrated discrimination improvement (IDI). NRI measures the reclassification of patients from one risk category to another by the addition of a new risk factor to a model with established risk factors. The NRI is attractive from a clinical point of view, because the investigator must decide which are the most clinically relevant risk thresholds. If a patient‘s risk is more correctly classified according to these thresholds by measuring a new risk factor, the risk factor has the potential to be clinically useful. The IDI is similar, but does not consider risk thresholds.

Troponin-I has passed such tests of usefulness for discrimination of risk for cardiovascular death [8] and heart failure [9] in the general population.

Circulating troponin levels indicate risk in heart failure & risk of heart failure

The troponins are located on the thin filaments in the sacomere and are involved in the regulation of the contractile function of the sarcomere. Peripheral blood levels of cardiac troponins have a major clinical use as a tool for diagnosing myocardial damage, mainly that from myocardial infarctions [10].

Higher circulating troponin levels indicate elevated risk of mortality in patients with acute coronary syndromes [11], and may also indicate high mortality risk in other clinical settings, such as in end-stage renal disease [12], as well as in the general population [13].

High circulating troponin levels have been associated with high mortality risk both in the setting of acute decompensated heart failure [14–16] and chronic stable heart failure [17–20]. Persistently increased or increasing troponin levels on serial measurements predict worse outcome than decreasing levels [15,18,19].

Only a few studies have investigated the associations of troponin levels with subsequent heart failure incidence. In the setting of an acute coronary syndrome, high troponin levels, a few days after symptom onset, predict subsequent left ventricular systolic dysfunction [21]. In one community-based sample of 70-year-old men free from previous heart failure, higher troponin-I levels were associated with a higher risk of subsequent heart failure [9]. This relation was independent of established and more recently described risk factors for heart failure, such as N-terminal prohormone brain natriuretic peptide (NT-proBNP). In this study, troponin-I also predicted subsequent nonischemic heart failure.

Elevated troponin levels as a warning of myocardial damage

Several studies have now demonstrated that circulating troponin levels in a range previously assumed to be ‘normal‘, in fact may signify higher risk than that associated with even lower troponin levels. This is quite reasonable. Very few biological risk factors have distinct thresholds for increased risk; most have more or less continuous relations with risk, albeit more or less linear. Circulating troponin levels are believed to reflect cardiomyocyte injury or death [22], and cardiomyocyte death is continuously ongoing in the myocardium. Men without heart disease appear to lose 1 g of myocardium per year, corresponding to 64 million cardiomyocytes [23]. It is not known how much of this death is through necrosis or through apoptosis, and it is not known how these processes are reflected in circulating troponin levels. Nevertheless, a certain level of troponins in the circulation should probably be considered normal, and future highly sensitive troponin assays will help us characterize this normal level in various settings.

The observed relations of circulating troponin levels with risk of subsequent heart failure remain to be characterized on the myocardial level [9]. In those with heart failure, both cardiomyocyte necrosis and apoptosis have been documented, with sevenfold higher necrosis than apoptosis rates reported in severe heart failure [24]. Besides cardiomyocyte death, increased circulating troponin levels may also signify leakage of unbound sarcoplasmic troponin through damaged cardiomyocyte membranes [25] or may be the result of troponin assays detecting higher levels of cleaved troponin peptides [26].

A wide range of clinical substrates for the observed associations of elevated troponin levels to heart failure risk [9] are conceivable. These include left ventricular hypertrophy [27,28] and myocarditis [29], conditions that may be asymptomatic and are known precursors of heart failure [30–32]. Numerous other sources are possible [33,34], mainly involving an oxygen supply/demand mismatch, and many of these will be defined as type 2 (secondary) myocardial infarctions according to the new universal definition of myocardial infarction [10]. In that context, it should be noted that elevated troponin levels have been demonstrated to predict adverse outcome [19] and to be related to BNP levels [35] also in heart failure patients angiographically free from coronary disease, as well as to predict nonischemic heart failure in the general population [9].

Future perspective

Troponin testing in the setting of acute chest pain is here to stay, but still has the potential for refinement. Troponin levels previously assumed to be elevated can be detected in a large fraction of seemingly healthy elderly individuals [9,36], and a clinical problem arises when these individuals present with chest pain. The compulsory dynamic troponin pattern for the diagnosis of myocardial infarction is attractive, although the timing and magnitude of this rise and/or fall may be discussed [10]. The other diagnostic component, a troponin level above a certain percentile, is more problematic. Hopefully, the future will see percentile-based decision limits abandoned and replaced by limits based on discriminatory value for subsequent adverse prognosis [9]. Studies of such decision limits need to be replicated in several samples of diverse ethic and age compositions.

In order to investigate if troponin testing could be useful in settings other than acute chest pain, a series of studies could be considered. The most constructive would be randomized intervention studies investigating whether treatment decisions based on knowledge of troponin values are superior to decisions made without this knowledge, in terms of long-term prognosis. The ‘Breathing Not Properly‘ study can give inspiration for such a study design [37], which could be extended to also include treatment decisions and long-term follow-up. Very few new risk factors have been investigated in such a way without having a clear treatment correlate. Therefore, intervention studies aiming to decrease troponin levels likely need to be undertaken first. Several treatment options are conceivable. C-reactive protein is an example of such a risk factor without a specific treatment correlate, which is currently being promoted for clinical use; the C-reactive protein-lowering effects of statins are mainly lipid-dependent [38], and statin studies [39] have not been designed to study whether knowledge of C-reactive protein levels aid in treatment decision making. It should be noted that a risk factor could be useful for treatment decision making even if the risk factor itself cannot be treated, as exemplified by the powerful risk factors age and sex. In such cases, one must hope that the necessary studies will be made despite probable lack of funding from the pharmaceutical industry.

Financial & competing interests disclosure

The author has received grants from the Swedish Research Council (grant 2007–5942) and the Swedish Heart Lung Foundation (grant 20041151). The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.

Additional information

Funding

The author has received grants from the Swedish Research Council (grant 2007–5942) and the Swedish Heart Lung Foundation (grant 20041151). The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed