The Promise of High-sensitivity Cardiac Troponins in the Rapid Exclusion Algorithm for Acute Coronary Syndrome

By Tan Hong Jie Gabriel, MBBS, Singapore, and Jack Tan Wei Chieh, MBBS, Singapore, MMed (Int Med), MRCP, UK

Chest pain is one of the leading presentations in the Emergency Department (ED), and the historically conservative approach to avoid missing a potential acute myocardial infarction (AMI) has led clinicians to admit or prolong the ED dwell time of many more patients than are subsequently found to have an AMI. This consequently leads to crowding in the ED, which is associated with adverse outcomes for all patients, with or without AMI.

Approximately 30 per cent of chest pain related ED visits will have a final diagnosis of a myocardial infarction (MI). Clinicians are obligated to exclude myocardial ischemia with a high degree of certainty. To manage costs and the adverse effects of overcrowding in the ED, it is of high priority to be able to rapidly and safely discharge patients with a sufficiently low probability for acute coronary syndrome (<0.5 per cent – 1 per cent).

The 2012 third universal definition of a MI is the detection of a rise and/or fall of cardiac troponins with at least one value above the 99th percentile upper reference limit (URL) in addition to at least one of the following: 1) symptoms of ischaemia; 2) new electrocardiogram (ECG) changes of ST-T segments, new left bundle branch block or development of pathological Q waves; 3) imaging evidence of new regional wall motion abnormality or loss of viable myocardium; and/or 4) identification of an intracoronary thrombus by angiography or autopsy. 

To diagnose AMI, myocardial injury must be accompanied by clinical indicators of an ischemic mechanism. On the contrary, an AMI cannot be present in the absence of myocardial injury. Hence, a MI is easier to rule out than to rule in. In addition, sufficient time must be allowed for the release of myocardial structural proteins (i.e. cardiac troponins, cTn) into the circulation in measurable quantities to be detectable post-myocardial injury. Therefore, the absence of high-sensitivity troponin (hsTn)>99th percentile URL and ≥6 hours after the onset of ischemia has conventionally been used to rule out MI. 

In the 1960s, aspartate transaminase (AST) was the first biomarker widely used in diagnosing AMI, followed by creatine kinase (CK) and lactate dehydrogenase (LDH) by the 1970s, all of which were not specific to cardiac muscle and hence detection of these were not specific to myocardial injury. Myoglobin, found in the heart and striated skeletal muscles, was later developed in 1978 as a cardiac biomarker as its serum level rises following acute myocardial injury. This was replaced in the era of electrophoresis advancement during which cardiac isoenzymes CK-MB, LDH 1 and 2 could be detected. 

Troponin is a complex within the contractile apparatus of cardiac and skeletal myocytes discovered in 1965. It was detectable with a reliably sensitive radioimmunoassay developed in the late 1980s. Several generations of troponin assays have since been developed, each reportedly with increasing diagnostic utility as well as the ability to rule out MI. Cardiac troponin (cTn) is a collective term referring to serum troponins T and I, which are isoforms highly sensitive and specific to cardiac myocytes, the sensitivity for detection of which approaches 100 per cent when sampled 6 – 12 h after acute onset of chest pain. Therefore, to reliably rule out AMI, patients with acute chest pain have had repeat troponin samples 6 – 12 h after the initial assessment. Consequently, patients were increasingly admitted for observation.

The latest generation of hsTn assays is defined as assays that have a coefficient variant (CV) of 10 per cent or less, at the 99th percentile URL, with the ability to detect cTn levels in at least 50 per cent of the reference normal population. Compared with earlier-generation assays, hsTn can establish biochemical evidence of myocardial injury at a much lower concentrations (10- to 100-fold lower) and thus earlier after the onset of ischemia. It is also able to discriminate small changes in concentration starting within the normal reference range, above the limit of detection (LoD) of the assay but below the URL. Small dynamic increases (delta values) are associated with a higher probability of subsequent rises above the URL and future major adverse cardiac events (MACE); whereas stable concentrations of hsTn in the detectable range below the URL are associated with structural heart disease, atherosclerosis risk factors, and a higher risk of future MACE. Consequently, non-detectable or very low hsTn concentrations identify patients with lower cardiovascular risk. 

Of the emerging applications for hsTn, the rapid rule-out of MI in the ED is the application most likely to be incorporated by clinicians. Accelerated diagnostic protocols (ADP) utilising hsTn can facilitate earlier triage while maintaining an acceptable negative predictive value (NPV). For example, the ADP developed by Meller et al (2015) utilising high-sensitivity troponin T (hsTnT) identified 35-40 per cent of patients to be at extremely low risk of MACE and hence ideal candidates for early outpatient management. 

ST-elevation myocardial infarction (STEMI) is an ECG diagnosis in the context of a patient with typical cardiac chest pain, and clinicians do not wait for biochemical evidence of myocardial damage before emergency revascularisation is instituted. Non-ST-elevation myocardial infarction (NSTEMI), on the other hand, can be ruled out as early as four hours after symptom onset with existing hsTn assays, allowing shorter ED or inpatient stay for patients without raised levels of troponin and earlier intervention for those with a confirmed AMI. 

ADPs incorporating the clinical history, electrocardiogram, and cTn concentrations provide a framework to rapidly evaluate and triage chest pain patients suspicious for ischemia. The National Institute for Health and Care Excellence (NICE) updated guidelines in 2016 on the evaluation of patients with suspected AMI recommended clinicians to consider ruling out MI if a patient has very low concentrations of cTn at presentation. They further recommended clinicians to apply the LoD as a threshold below which MI can be safely ruled out at presentation, provided patients are deemed to be at low risk of MI by an appropriate risk stratification system. Evidence from studies including both the Thrombolysis in Myocardial Infarction (TIMI) score and the Global Registry of Acute Coronary Events (GRACE) score were considered. Having cTn concentrations embedded in both scores were initially derived and validated in patients with confirmed MI for prognostication, but over time, these scores have been implemented to risk stratify patients with suspected MI as well. Although NICE ultimately recommended the TIMI risk score, it was yet to be validated with hsTn, using LoD at presentation alone. 

In 2018, Carlton et al found when a hsTnT of <5ng/L (LoD) was applied alongside a TIMI score of 0 and a non-ischaemic ECG, the sensitivity and NPV were extremely high, at 99.5 per cent and 99.6 per cent respectively. For the high-sensitivity Troponin I, (hsTnI), using the LoD (<2ng/L) and a TIMI score of 0 alongside a non-ischaemic ECG, the sensitivity was 98.9 per cent and NPV was 99.5 per cent. These strategies would identify 17.9 per cent and 21 per cent of low-risk patients respectively. 

The European Society of Cardiology (ESC) has previously advocated the use of a 0/3-hour algorithm in conjunction with the ECG and GRACE (Global Registry of Acute Coronary Event) risk score in clinical practice. However, a recent meta-analysis by Pickering et al (2017) have concluded that a single hsTnT concentration below the LoD in combination with a non-ischemic ECG provided excellent NPV (99.3 per cent) and sensitivity (98.7 per cent) for MI, hence may successfully rule out AMI in patients presenting to EDs with possible ACS without the need for additional risk score. Indeed, the ESC now advocates use of the LoD at presentation in conjunction with ECG, but do not recommend the addition of clinical risk scores. Nevertheless, application of clinical risk score is still widespread in most settings, probably due to the perceived assurance it provides diagnostically. 

One of the most widely used risk scores, the history, ECG, age, risk factors, troponins (heart) score, was developed and validated in a population with suspected ACS. A recent meta-analysis of 11,217 patients demonstrated that this score had a sensitivity of just 96.7 per cent, below the threshold of 99 per cent, which most ED clinicians deem acceptable. It is unclear whether this score risk stratifies better than with hsTn alone, hence comparative studies between risk stratification thresholds of hsTn alone or in combination with risk scores are required to determine if patient safety can be improved. 

The evolution of hsTn-based rule-out strategies in the ED is aimed at reliably excluding myocardial injury as early as possible through staged measurement of hsTn in conjunction with other clinical assessments for the probability of MI. Emerging components include movement of serial hsTn samples to earlier time points, addition of criteria for an absolute delta hsTn changes between measurements, and integration of very low decision limits well below the 99th percentile URL at the early time points. For example, the ESC 2015 practice guidelines also included an alternative (Class I) strategy, reducing the sampling interval from three to one hour when a validated hsTn assay with 0/1- hour algorithm is used. Such an algorithm incorporates all three components listed above, including a very low cut-off applied at the initial hsTn value aimed at excluding MI after the first sample in patients who arrive >three hours after symptom onset, and a delta criteria for patients with dynamic 

hsTn concentrations subjecting them to additional testing. These reports have shown that using very low hsTn on the first sample can reasonably exclude MI in 40-50 per cent of low-risk patients having presented >two hours from symptom onset. Similarly, a comparative study by Chapman et al (2017) between the High-STEACS (High-Sensitivity Troponin in the Evaluation of Patients With Acute Coronary Syndrome) 0/3/6-hr algorithm and the standard ESC 0/3-hr algorithm to rule out MI, also value added on the use of a very low 0-hour cut-off to facilitate earlier rule-out, and a delta criterion to exclude increasing values with absolute concentrations below the URL among patients requiring three-hour sampling. 

Boeddinghaus et al (2017) moved the serial sample in the 0/3- hour algorithm forward to one hour and comparing the ESC alternative 0/1-hour strategy with other approaches using either a single cut-off at 0 hours or the one-hour strategy. They found that each of these approaches performed similarly in delivering an NPV >99 per cent, comparing favourably with the ESC 0/3- hour algorithm (NPV, 98.4 per cent) among patients presenting >two hours after symptom onset. However, among early presenters, the NPV (98.5 per cent) and sensitivity (94.2 per cent) were lowered with the use of the single 0-hour cut-off (5 ng/L). Hence, it is of utmost importance that very early presenters should have serial testing to support an acceptable NPV, as should patients with other high-risk indicators. 

Although others have also advocated the use of the LoD as necessary to support an acceptable sensitivity with a single sample, a study by Carlton at el (2016) showed that use of LoD (1.2ng/L) ruled out fewer patients (18.8 per cent discharge rate), albeit with higher sensitivity of 99.0 per cent (95 per cent CI, 96.8 per cent – 99.7 per cent) and NPV of 99.5 per cent (95 per cent CI, 98.4 per cent – 99.9 per cent). 

Undoubtedly, there remain several challenges in translating this diagnostic innovation into cost-effective healthcare with improved patient outcomes. Firstly, increased sensitivity of hsTn assays are usually accompanied with reduced specificities and low positive predictive values (PPV) for MI. Although there would be a lower likelihood of a missed MI, these assays would not be suitable for early rule-in decision-making. In other words, clinicians would need to be more prudent and cautious with positive tests given the advent of such assays. Secondly, the reduced specificity and low PPV for MI accompanying the improved analytical precision of the assays leads to greater rate of invasive investigation, i.e. coronary angiography without a proportionate increase in coronary lesion-specific therapy, i.e. revascularisation, although the improved analytical precision enables earlier clinical decision to admit or discharge a patient with suspected ACS. Thirdly, hsTn assays are impractical in the primary care setting due to the long turnaround times for results despite its usefulness in the context of low or intermediate clinical suspicion of MI, provided sufficient time has passed since symptom onset. Also, it is problematic for general practitioners to receive results in a timely manner for execution of an appropriate clinical response. Lastly, improved test precision without a greater discretion in clinical decision making and test interpretation may result in increased costs and inefficiencies, i.e. increased investigative burden of patients tested positive due to the many non-coronary causes of elevated troponin. 

With emerging studies published worldwide optimising the potential application of hsTn, knowledge gaps have still remained to date. For example, there is lack of data among early presenters to provide confidence around and/or lead to revision of current algorithms. Future algorithms may possibly integrate the actual timing of sampling instead of assigning specific time points as per all current algorithms. The time-related rate of rise may become more crucial than the absolute or relative changes. Perhaps coupling hsTn with clinical probability tools such as cardiovascular risk scores or imaging such as CT coronary angiogram can improve diagnostic performance for ACS rule-out. Prospective randomised studies implementing these testing strategies are warranted in view that current studies are mostly observational with reported outcomes based on management according to local standards of care. In view of low cut-offs and small delta criteria optimising NPV but lowering PPV, future studies should also be directed at determining higher delta criteria to rule in MI. These issues should be addressed in the near future as we move toward evidence-based best practice. 

In conclusion, hsTn has greater utility in the early rule-out of MI, potentially at 0-hour presentation for low risk cases clinical outcomes are also improved due to less likelihood of missed AMI in patients presenting with chest pain in the ED. Nevertheless, specificity and PPV remains its main shortfall as hsTn does not differentiate AMI from other non-coronary causes of myocardial necrosis. hsTn should be always be used in tandem with clinical judgement to enable appropriate interpretation of these assays. Lastly, consensus and continuous education in every institution employing hsTn are essential for clinical practice changes in testing strategies to harness the full potential of hsTn in further improving resource efficiency and patient safety.