Physicians use tests to inform decision-making. Whether this is a bedside test using a stethoscope, the seemingly ancient technology of recording an electrocardiogram (ECG), or the most advanced imaging modalities and biochemical panels available – all pursue diagnostic clarity. But, more frequently than we might like to admit, the results do not illuminate a clear path of treatment.
Measuring diagnostic accuracy
The value of any test depends on its diagnostic accuracy, which we express in terms of sensitivity and specificity values – metrics that are heavily dependent on the pre-test likelihood and a clear gold-standard for the presence of the condition in question. The evolution of cardiac Troponin (cTn) as a biomarker to rule-in and rule-out myocardial infarction (MI) is testament to this: embraced as far superior to creatine kinase-MB (CK-MB), it had all the properties to change our approach to chest pain triage – and so it did. Evidenced by the endorsement from the Universal Definition of Myocardial Infarction,1 it has become seemingly impossible to diagnose MI unless there is cTn elevation. Justifiably, in the era of sensitive cTn assays we enjoyed its use as a dichotomous test – as an elevated cTn did invariably represent significant myocardial injury (with matching high specificity for MI).
As the assay developers overcame the sensitivity-gap with high-sensitivity assays, we were suddenly able to not only detect acute myocardial injury earlier, but also ‘monitor’ chronic, sub-clinical disease fluctuation (see heart failure2). While appealing from an analytic standpoint, physicians suddenly face the impact of the sensitivity/specificity interplay in clinical reality: whereas the emergency physician demands ‘perfect’ sensitivity (to avoid inappropriate discharge), the cardiologist requires specificity to implement an appropriate treatment strategy.
As guidelines attempt to please both camps, this leaves a large proportion of patients in a diagnostic grey-zone: decision-limits, which are widely-spaced around the 99th centile, are calibrated such that an ‘undetectable troponin’ exceeds a sensitivity of 99%, and a rule-in threshold five-fold the upper reference limit of a healthy population (for high-sensitivity [hs] cTnT3) yields a specificity for MI >75%. In clinical reality, the 25–40% of patients4-6 seen in this observational zone of diagnostic uncertainty are a heterogeneous group. This is further complicated by mounting evidence that a detectable cTn level (whether below or above the 99th centile) has important prognostic implications: long-term survival seems to track that of the group with the highest cTn values,7 which probably reflects the impact of important comorbidities that yield an elevation of exquisitely sensitive biochemical signals8 – rather than an underlying, primary cardiac cause for ill health. Now this naturally leads to further questions down the road, some of which more recent publications are trying to answer. How predictive is cTn in otherwise ‘stable’ patients, and – more importantly – can we change outcomes by aiming for a lower cTn level, much like we target other risk factors such as low-density lipoprotein (LDL)-cholesterol and blood pressure?
Blurring the grey-zone
In the UK-based WOSCOPS (West of Scotland Coronary Prevention) study,9 investigators described the predictive power of elevated cTn levels in an otherwise asymptomatic (male) population with raised LDL-levels: elevated hs-cTnI at a threshold of 5.2 ng/L (99th centile: 26 ng/L)10 identified a subgroup of patients at two- to three-fold long-term risk (for MI or cardiovascular death). A cynic might speculate that the resulting black and white threshold at 5.2 ng/L results from statistically significant hazard ratios more than true biology – ignorant as to whether the assay is capable of reliably detecting decimal points (it cannot11). A more recent – Swedish – study adds further fuel:12 the authors followed >19,000 patients with ‘stable’ hs-cTnT levels for a mean of 3.3 years – notably 62% of participants had hs-cTnT values below the limit of detection. Irrespective of whether a graded presentation of cumulative mortality adequately reflects the predictive strength of a continuous variable, the take-home message appears clear-cut: the higher the troponin level, the worse the outcome. Importantly, the 10 most common causes of death contained a substantial proportion of cancer- and heart failure-related deaths. And it probably comes as no surprise that median troponin levels were directly correlated with age – which further complicates chest pain triage in clinical reality.
The patients attending accident and emergency (A&E) departments with medical complaints are of older age, and represent a very different cohort with a higher prevalence of comorbidities than healthy volunteers who were enrolled to guide assay manufacturers in derivation of the clinically illusive 99th centile. This poses an everyday challenge: many patients have detectable cTn levels in the diagnostic grey-zone, and now we have established that increasing cTn levels are indicative of poor prognosis even if there is no acute cause identified! But not all is bad – as long as the medical community embraces the latest evidence as a meaningful addition to the diagnostic arsenal.
Clearly, we are moving away from a much-revered black and white biomarker test to the new reality of graded risk stratification. Much like we are happy to interpret and investigate a low haemoglobin or elevated leucocyte count, we should evaluate an elevated cardiac troponin in the clinical context and not scream ‘myocardial infarction’ where a biomarker elevation is the only finding. An undetectable troponin level is clearly a reassuring result (and prognostically relevant too), but we better get comfortable with the new reality which increasingly sensitive assays pose – as ultra-sensitive troponin tests13 (and more novel markers14) are just round the corner. Ultimately, we need to know the answer to the question, chronic elevation of cTn is bad, but what is the treatment path it illuminates?
Conflict of interest
1. Alpert JS, Chaitman BR, Jaffe AS et al. Third universal definition of myocardial infarction. J Am Coll Cardiol 2012;60:1581–98. https://doi.org/10.1016/j.jacc.2012.08.001
2. Basuray A, French B, Ky B et al. Heart failure with recovered ejection fraction: clinical description, biomarkers, and outcomes. Circulation 2014;129:2380–7. https://doi.org/10.1161/CIRCULATIONAHA.113.006855
3. Roffi M, Patrono C, Collet J-P et al. 2015 ESC guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation: Task Force for the Management of Acute Coronary Syndromes in Patients Presenting without Persistent ST-Segment Elevation of the European Society of Cardiology (ESC). Eur Heart J 2016;37:267–315. https://doi.org/10.1093/eurheartj/ehv320
4. Mueller C, Giannitsis E, Christ M et al. Multicenter evaluation of a 0-hour/1-hour algorithm in the diagnosis of myocardial infarction with high-sensitivity cardiac troponin T. Ann Emerg Med 2016;68:76–87.e4. https://doi.org/10.1016/j.annemergmed.2015.11.013
5. Mokhtari A, Borna C, Gilje P et al. A 1-h combination algorithm allows fast rule-out and rule-in of major adverse cardiac events. J Am Coll Cardiol 2016;67:1531–40. https://doi.org/10.1016/j.jacc.2016.01.059
6. Reichlin T, Schindler C, Drexler B et al. One-hour rule-out and rule-in of acute myocardial infarction using high-sensitivity cardiac troponin T. Arch Intern Med 2012;172:1211–18. https://doi.org/10.1001/archinternmed.2012.3698
7. Nestelberger T, Wildi K, Boeddinghaus J et al. Characterization of the observe zone of the ESC 2015 high-sensitivity cardiac troponin 0h/1h-algorithm for the early diagnosis of acute myocardial infarction. Int J Cardiol 2016;207:238–45. https://doi.org/10.1016/j.ijcard.2016.01.112
8. Marjot J, Kaier TE, Martin ED et al. Quantifying the release of biomarkers of myocardial necrosis from cardiac myocytes and intact myocardium. Clin Chem 2017;63:990–6. https://doi.org/10.1373/clinchem.2016.264648
9. Ford I, Shah ASV, Zhang R et al. High-sensitivity cardiac troponin, statin therapy, and risk of coronary heart disease. J Am Coll Cardiol 2016;68:2719–28. https://doi.org/10.1016/j.jacc.2016.10.020
10. Aw T-C, Phua S-K, Tan S-P. Measurement of cardiac troponin I in serum with a new high-sensitivity assay in a large multi-ethnic Asian cohort and the impact of gender. Clin Chim Acta 2013;422:26–8. https://doi.org/10.1016/j.cca.2013.03.034
11. Apple FS, Sandoval Y, Jaffe AS, Ordonez-Llanos J, IFCC Task Force on Clinical Applications of Cardiac Bio-Markers. Cardiac troponin assays: guide to understanding analytical characteristics and their impact on clinical care. Clin Chem 2017;63:73–81. https://doi.org/10.1373/clinchem.2016.255109
12. Roos A, Bandstein N, Lundbäck M, Hammarsten O, Ljung R, Holzmann MJ. Stable high-sensitivity cardiac troponin T levels and outcomes in patients with chest pain. J Am Coll Cardiol 2017;70:2226–36. https://doi.org/10.1016/j.jacc.2017.08.064
13. Kaess BM, Las Heras Gala de T, Zierer A et al. Ultra-sensitive troponin I is an independent predictor of incident coronary heart disease in the general population. Eur J Epidemiol 2017;32:583–91. https://doi.org/10.1007/s10654-017-0266-7
14. Kaier TE, Twerenbold R, Puelacher C et al. Direct comparison of cardiac myosin-binding protein C with cardiac troponins for the early diagnosis of acute myocardial infarction. Circulation 2017;136:1495–508. https://doi.org/10.1161/CIRCULATIONAHA.117.028084