Clinical Use of High-sensitivity Cardiac Troponin in Patients With Suspected Myocardial Infarction

Raphael Twerenbold, MD; Jasper Boeddinghaus, MD; Thomas Nestelberger, MD; Karin Wildi, MD; Maria Rubini Gimenez, MD; Patrick Badertscher, MD; Christian Mueller, MD


J Am Coll Cardiol. 2017;70(8):996-1012. 

In This Article

Troponin-based Strategies for Rapid Rule-out or Rule-in of Mi

The most important clinical advantage of the new, more-sensitive cTn assays is their ability to substantially reduce the initial "troponin-blind" interval in the first hours after MI onset and to allow novel, rapid strategies for the early rule-out or rule-in of MI. Several troponin-based strategies rely on serial hs-cTn testing. Two of them, a 0-h to 1-h (0/1h) algorithm and a 0-h to 3-h (0/3h) algorithm, are recommended by the European Society of Cardiology (ESC) with a class I recommendation and deserve in-depth discussion. These novel strategies have been fine tuned to detect MI, but not unstable angina (UA), the acute coronary syndrome (ACS) phenotype at much lower short-term risk of death and/or major arrhythmias, but at substantial long-term risk of MI.[45,46] Therefore, full cardiology workup and intensive lifestyle modification and medical therapy remain crucial in UA.

It is important to highlight 5 aspects when applying troponin-based strategies in clinical practice (Central Illustration). First, they should be used only in conjunction with full clinical assessment, including a pre-test probability assessment to identify those patients at high risk who may not be suitable for early discharge. Second, these strategies should be considered triage strategies rather than definite diagnostic strategies, because additional imaging tests, for example, invasive coronary angiography, stress testing, echocardiography, or computed tomography angiography, may be necessary for a definite diagnosis. Third, the percentage of patients eligible for rule-out or rule-in varies widely from ≈9.8% to 77% depending on the underlying algorithm, the cTn assay used, and the clinical setting, including the prevalence of MI.[24,28] Fourth, these strategies should only be applied after the initial ECG has excluded ST-segment elevation myocardial infarction (STEMI) because these high-risk patients need prompt identification based on the ECG, and immediate reperfusion therapy, without the need for cTn testing.[12] Some rule-out strategies require a completely normal ECG to be applied; others allow also for mild and nonspecific ECG abnormalities. Fifth, all triage strategies should be embedded in the local standard operating procedures of the ED.

Central Illustration.

Central Illustration Patient Assessment With Suspected ACS
The initial assessment is based on the integration of low likelihood and/or high likelihood features derived from clinical setting (i.e., symptoms, vital signs), 12-lead ECG, and cardiac troponin determined at presentation to the ED and serially thereafter. "Other cardiac" includes, among other, myocarditis, Takotsubo cardiomyopathy, or congestive heart failure. "Noncardiac" refers to thoracic diseases such as pneumonia or pneumothorax. Cardiac troponin and its change during serial sampling should be interpreted as a quantitative marker: the higher the 0h-level or the absolute change during serial sampling, the higher the likelihood for the presence of myocardial infarction. In patients presenting with cardiac arrest or hemodynamic instability of presumed cardiovascular origin, echocardiography should be performed/interpreted by trained physicians immediately following a 12-lead ECG. If the initial evaluation suggests aortic dissection or pulmonary embolism, D-dimers and multidetector computed tomography angiography are recommended according to dedicated algorithms. Width of boxes represent the prevalence of the respective disorders among consecutive patients presenting with acute chest pain to the emergency department. ACS = acute coronary syndromes; CPR = cardio-pulmonary resuscitation; ECG = electrocardiography; hs-cTn = high-sensitivity cardiac troponin; MI = myocardial infarction; NSTEMI = non–ST-segment elevation myocardial infarction; STEMI = ST-segment elevation myocardial infarction. Adapted with permission from Roffi et al..[27]

Among the multitude of available triage algorithms for patients with suspected MI, 6 novel strategies based on cTn have been studied and validated in large, methodologically robust, multicenter diagnostic studies including several thousand patients and rigorous MI adjudication using serial hs-cTn testing. These warrant consideration for clinical use in the appropriate clinical setting of patients presenting to the ED with acute chest discomfort and/or suspected MI. The main performance metrics of the studies include safety (quantified by the negative predictive value [NPV] and sensitivity for MI) and efficacy (percentage of patients triaged early) for rule-out, as well as the PPV and specificity for MI (Table 2). Four algorithms use the absolute change between 2 measurements in addition to the hs-cTn concentrations determined at presentation to the ED to take advantage of the full diagnostic information provided. Rising and/or falling cTn levels differentiate acute from chronic myocardial injury. Absolute rather than relative changes seem to be the best metric to differentiate MI from other causes of chest pain.[16,23–25] The larger the absolute (unsigned) cTn change within 1h, 2h, or 3h, the higher the likelihood for the presence of MI.[16,23–25]

Two strategies require the use of a pre-defined risk score (0/3h-ESC algorithm and 2h advanced diagnostic pathways [ADP]), whereas the remaining 4 strategies do not.

0/3h-ESC Algorithm

MI is ruled out if concentrations of hs-cTn remain in the normal range (below the respective 99th percentiles) in the blood sample drawn at presentation and 3 h after presentation, and if the patient fulfils 2 additional requirements: to be pain-free and to be at low risk of in-hospital mortality as quantified by a Global Registry of Acute Coronary Events (GRACE) score below 140.[27] In patients presenting more than 6 h after chest pain onset, in whom chest pain onset can be reliably quantified, 1 single blood draw at presentation is considered to be sufficient. Patients are ruled in if they have a clearly elevated hs-cTn blood concentration at presentation, or if the 3-h sample shows a relevant change. This approach has been recommended by the ESC guidelines since 2011 and is the standard of care in many institutions worldwide (Figure 1A).[10,27] Its use regarding rule-out of MI seems to be safe for all hs-cTn assays and possibly also some s-cTn assays.[47] The exact performance for rule-in cannot be quantified, as no precise definitions of its rule-in cutoff levels are given. Given the average turnaround time for hs-cTn of about 1 h, the hs-cTn measurement performed at 3 h after ED presentation would become available at about 4 h after ED presentation and would allow clinical decision making regarding hospitalization versus outpatient management about 4 h after ED presentation in the majority of patients. In a recent study, this strategy enabled outpatient management in 56% of patients, with a median time in the ED of about 5 h in the overall population, and 4.5 h in those patients managed as outpatients.[48]

Figure 1.

ESC Rapid Triage Algorithms
(A) The European Society of Cardiology (ESC) 0/3h rule-out and rule-in algorithm of non–ST-segment elevation acute coronary syndromes using high-sensitivity cardiac troponin (hs-cTn) assays. aΔ change, dependent on assay. Highly abnormal hsTn defines values beyond 5-fold the upper limit of normal. (B) The ESC 0/1h rule-out and rule-in algorithm using hs-cTn assays in patients presenting with suspected non–ST-segment elevation myocardial infarction (NSTEMI) to the emergency department. (C) The ESC 0/1h rule-out and rule-in algorithm with assay-specific cutoff values for the Elecsys hs-cTnT assay (Roche Diagnostics, Rotkreuz, Switzerland). (D) The ESC 0/1h rule-out and rule-in algorithm with assay-specific cutoff values for the Architect hs-cTnI assay (Abbott Laboratories, Chicago, Illinois). Adapted with permission from Roffi et al..[27] CPO = chest pain onset; GRACE = Global Registry of Acute Coronary events score; ULN = upper limit of normal.

An alternative 0/3h algorithm using lower cutoff criteria than the 99th percentile for rule-out has recently been developed specifically for hs-cTnI.[49]

2h-ADP With Risk Scores

This is the best validated strategy, combining serial cTn testing with a pre-defined clinical risk score, the Thrombolysis In Myocardial Infarction (TIMI) score.[28–32] The TIMI score was developed about 2 decades ago as a prognostic score for patients with ACS to identify those who may benefit most from anticlotting therapy and was subsequently validated for the use in ED patients.[50,51] The original 2h-ADP combines a TIMI score of 0 with a nonischemic ECG and negative conventional cTn testing at 0 and 2 h, classifying 9.8% of patients as low risk with a sensitivity and NPV for major adverse cardiac events (MACE) within 30 days of ≥99%.[28] A modified protocol using hs-cTnI and a TIMI score of ≤1 could safely discharge 40% of patients with comparable MACE incidence.[30]

0/2h Algorithm Without Risk Scores

The alternative 0/2h strategy exclusively uses hs-cTn data to triage patients without the use of a specific clinical risk score and thereby achieves a comparable NPV and sensitivity for rule-out by also taking into account absolute concentration changes within 2 h.[22,52,53] The lack of a relevant absolute change from presentation to 2 h combined with the fact that both concentrations need to be normal obviates the need of a pre-defined score and allows one to safely rule out MI even in patients with pre-existing CAD. Accordingly, this strategy is more effective and allows the rapid rule-out of MI in up to 60% of patients.[22,52,53] Moreover, this strategy includes a rule-in algorithm that provides a PPV above 75% for MI and allows the early rule-in in about 10% to 15% of acute chest pain patients within 2 h to 3 h of presentation.

0/1h-ESC Algorithm

The concept of the 0/1h-ESC algorithm is identical to that of the 0/2h algorithm and is also based exclusively on information provided by hs-cTn blood concentrations (Figure 1B). The decision points derived and validated for each assay are assay-specific (Figures 1C and 1D).[24,27,34,35,54] The 0/1h-ESC algorithm obviates the need for formal use of clinical scores and allows safe rule-out of MI even in patients with pre-existing CAD or mild, nonspecific, and often pre-existing ECG abnormalities. Accordingly, this strategy is very effective and allows an accurate early triage in about 75% of patients: 60% towards rule-out and in 15% towards rule-in of MI. Again, given the average turnaround time for hs-cTn of about 1 h, the hs-cTn measurement performed at 1 h after ED presentation will become available at about 2 h after ED presentation and will facilitate clinical decision making regarding hospitalization versus outpatient management about 2 to 3 h after ED presentation in the majority of patients. The application of the 0/1h-ESC algorithm is also possible in institutions with a median turnaround time of more than 1 h because the 1 h only refers to the time interval between the serial blood sampling. In this situation, the second blood draw would need to be performed while still awaiting the results from the first blood draw. In order to further demonstrate the simplicity of the 0/1h-ESC algorithm, 5 common clinical scenarios are described as case reports in the Online Appendix.

Alternative 1h Algorithm

A modification of the 0/1h-ESC algorithm has recently been developed for hs-cTnI.[55] Safety of rule-out and overall efficacy are very high and comparable to the 0/1h-ESC algorithm. In contrast to the 0/1h-ESC algorithm, this algorithm does not allow for direct rule-out or rule-in on the basis of the 0 h sample only.

Undetectable/Very low Baseline hs-cTn Concentrations

Undetectable or very low blood concentrations of hs-cTn at presentation to the ED have a very high (98.6% to 100%) NPV for MI. This approach has unique simplicity, as it requires only a single blood draw of an inexpensive and widely available biomarker. Because the lower limit of detection is assay-dependent and varies among the clinically available hs-cTn assays, "very low concentrations" (e.g., below the 30% percentile of healthy individuals) may be the preferred metrics to identify biological-equivalent values. Four large studies and a recent meta-analysis have provided consistent results for hs-cTnT, whereas 3 studies showed comparable findings for 3 hs-cTnI assays.[36,37,56–60] Because the release of cTn is a time-dependent phenomenon, this approach should only be used in patients with a chest pain onset of at least 2 to 3 h before ED presentation, because safety was reduced in these very early presenters in a recent observation.[60] In the 2015 ESC guidelines, this approach is recommended in combination with the 0/1h algorithm as the preferred rule-out strategies due to their excellent balance between speed and accuracy.[27]