Author + information
- Received March 12, 2019
- Revision received May 6, 2019
- Accepted June 3, 2019
- Published online August 12, 2019.
- Thomas Nestelberger, MDa,b,∗@thomas_nest,
- Jasper Boeddinghaus, MDa,b,c,∗@J_Boeddinghaus,
- Desiree Wussler, MDa,b,c,
- Raphael Twerenbold, MDa,b,
- Patrick Badertscher, MDa,b,d,
- Karin Wildi, MDa,b,e,
- Òscar Miró, MDb,f,
- Beatriz López, MDf,
- F. Javier Martin-Sanchez, MDb,g,
- Piotr Muzyk, MDb,h,
- Luca Koechlin, MDa,b,i,
- Benjamin Baumgartner, MDa,b,
- Mario Meier, MDa,b,
- Valentina Troester, MDa,b,
- Maria Rubini Giménez, MDa,b,
- Christian Puelacher, MDa,b,c,
- Jeanne du Fay de Lavallaz, MDa,b,
- Joan Walter, MDa,b,
- Nikola Kozhuharov, MDa,b,
- Tobias Zimmermann, MDa,b,
- Danielle M. Gualandro, MDa,b,
- Eleni Michou, MDa,b,
- Eliska Potlukova, MDa,c,
- Nicolas Geigy, MDj,
- Dagmar I. Keller, MDk,
- Tobias Reichlin, MDb,l,
- Christian Mueller, MDa,b,∗ (, )@CRIBasel,
- for the APACE Investigators
- aCardiovascular Research Institute Basel (CRIB) and Department of Cardiology, University Hospital Basel, University of Basel, Basel, Switzerland
- bGREAT Network
- cDepartment of Internal Medicine, University Hospital Basel, University of Basel, Basel, Switzerland
- dDivision of Cardiology, University of Illinois at Chicago, Chicago, Illinois
- eCritical Care Research Institute, The Prince Charles Hospital, Brisbane and University of Queensland, Brisbane, Queensland, Australia
- fEmergency Department, Hospital Clinic, Barcelona, Spain
- gServicio de Urgencias, Hospital Clínico San Carlos, Madrid, Spain
- h2nd Department of Cardiology, School of Medicine with the Division of Dentistry, Zabrze, Medical University of Katowice, Katowice, Poland
- iDepartment of Cardiac Surgery, University Hospital Basel, University of Basel, Basel, Switzerland
- jEmergency Department, Kantonsspital Baselland, Liestal, Switzerland
- kEmergency Department, University Hospital Zürich, Zürich, Switzerland
- lDepartment of Cardiology, University Hospital Bern, University of Bern, Bern, Switzerland
- ↵∗Address for correspondence:
Dr. Christian Mueller, CRIB and Department of Cardiology, University Hospital Basel, Petersgraben 4, CH-4031, Basel, Switzerland.
Background Early and accurate detection of short-term major adverse cardiac events (MACE) in patients with suspected acute myocardial infarction (AMI) is an unmet clinical need.
Objectives The goal of this study was to test the hypothesis that adding clinical judgment and electrocardiogram findings to the European Society of Cardiology (ESC) high-sensitivity cardiac troponin (hs-cTn) measurement at presentation and after 1 h (ESC hs-cTn 0/1 h algorithm) would further improve its performance to predict MACE.
Methods Patients presenting to an emergency department with suspected AMI were enrolled in a prospective, multicenter diagnostic study. The primary endpoint was MACE, including all-cause death, cardiac arrest, AMI, cardiogenic shock, sustained ventricular arrhythmia, and high-grade atrioventricular block within 30 days including index events. The secondary endpoint was MACE + unstable angina (UA) receiving early (≤24 h) revascularization.
Results Among 3,123 patients, the ESC hs-cTnT 0/1 h algorithm triaged significantly more patients toward rule-out compared with the extended algorithm (60%; 95% CI: 59% to 62% vs. 45%; 95% CI: 43% to 46%; p < 0.001), while maintaining similar 30-day MACE rates (0.6%; 95% CI: 0.3% to 1.1% vs. 0.4%; 95% CI: 0.1% to 0.9%; p = 0.429), resulting in a similar negative predictive value (99.4%; 95% CI: 98.9% to 99.6% vs. 99.6%; 95% CI: 99.2% to 99.8%; p = 0.097). The ESC hs-cTnT 0/1 h algorithm ruled-in fewer patients (16%; 95% CI: 14.9% to 17.5% vs. 26%; 95% CI: 24.2% to 27.2%; p < 0.001) compared with the extended algorithm, albeit with a higher positive predictive value (76.6%; 95% CI: 72.8% to 80.1% vs. 59%; 95% CI: 55.5% to 62.3%; p < 0.001). For 30-day MACE + UA, the ESC hs-cTnT 0/1 h algorithm had a higher positive predictive value for rule-in, whereas the extended algorithm had a higher negative predictive value for the rule-out. Similar findings emerged when using hs-cTnI.
Conclusions The ESC hs-cTn 0/1 h algorithm better balanced efficacy and safety in the prediction of MACE, whereas the extended algorithm is the preferred option for the rule-out of 30-day MACE + UA. (Advantageous Predictors of Acute Coronary Syndromes Evaluation [APACE]; NCT00470587).
- acute myocardial infarction
- clinical assessment
- high-sensitivity cardiac troponin
- major adverse cardiac events
Acute chest discomfort accounts for ∼10% of all presentations to the emergency department (ED) (1–4). Among patients with acute chest discomfort, the early detection of acute myocardial infarction (AMI) has high priority because AMI is common, associated with high mortality within the first hours, and amendable with effective treatment (1–3). Efficient risk stratification is mainly based on 3 diagnostic columns: detailed clinical assessment (including chest pain characteristics), the 12-lead electrocardiogram (ECG), and cardiac troponin (cTn) as a blood biomarker of cardiomyocyte injury (2–4).
The clinical introduction of cTn assays with higher analytical sensitivity (hs-cTn) has allowed the sensitivity deficit for AMI at ED presentation associated with conventional cTn assays to be largely overcome and, thereby, substantially increased the early diagnostic accuracy for AMI (3,5–7). Very high diagnostic accuracy for AMI within the first hours after ED presentation has permitted the shortening of the time interval to the second hs-cTn measurement and reduced the time to rule-out and/or rule-in AMI.
Among the rapid hs-cTn–based strategies, the European Society of Cardiology (ESC) hs-cTnT/I measurements at presentation and after 1 h (0/1 h algorithm) are particularly appealing, as they seem to combine very high safety with high efficacy in the early rule-out or rule-in of AMI (8–10).
Although it is mandatory to use the ESC hs-cTnT/I 0/1 h algorithms, as with all other hs-cTn–based strategies, in conjunction with full clinical assessment and the 12-lead ECG (3), it is unknown how to best combine these diagnostic variables and how possible combinations would affect its performance characteristics. As the ESC hs-cTnT/I 0/1 h algorithms have been optimized for the rule-out and rule-in of AMI, the incremental value of the additional diagnostic variables seems to be small for AMI (8–10). However, it might be substantial for the accurate prediction of major adverse cardiac events (MACE) within 30 days, the second key task for physicians in the ED necessary for the decision of whether early discharge and outpatient management or in-hospital diagnostic and therapeutic assessment is indicated. A pilot study provided a pragmatic suggestion on how to combine the hs-cTnT 0/1 h algorithm with clinical assessment and the ECG findings (i.e., “the extended algorithm”) (11).
The goal of the current study was to externally validate in a large prospective multicenter study the performance characteristics of the extended algorithm in the prediction of short-term MACE and compare it versus that of the ESC hs-cTnT/I 0/1 h algorithms.
Study design and patient population
APACE (Advantageous Predictors of Acute Coronary Syndrome Evaluation) is a prospective, international, multicenter diagnostic study with 12 centers in 5 European countries aiming to advance the early diagnosis of AMI (NCT00470587) (12–14). Adult patients presenting to the ED with symptoms suggestive of AMI (e.g., acute chest discomfort, angina pectoris) with an onset or peak within the last 12 h were recruited. Enrollment was independent of renal function, although patients with terminal kidney failure on chronic dialysis were excluded. For this analysis, patients with ST-segment elevation MI, patients in whom the final diagnosis remained unclear even after central adjudication and at least 1 elevated hs-cTnT concentration possibly indicating AMI, and patients with no available hs-cTnT/I concentrations determined upon presentation to the ED and after 1 h were also ineligible and therefore excluded. The most common reasons for missing samples after 1 h were early transfer to the catheter laboratory or coronary care unit and diagnostic procedures around the 1-h window that precluded blood draw at 1 h.
The study was conducted according to the principles of the Declaration of Helsinki and approved by the local ethics committees. Written informed consent was obtained from all patients. The Standards for the Reporting of Diagnostic Accuracy Studies Checklist is given in Online Table 1 (15).
Routine clinical assessment
Patients underwent clinical assessment that included medical history, physical examination, standard blood test including serial measurements of local (hs)-cTn, 12-lead ECG, chest radiography, continuous ECG rhythm monitoring, and pulse oximetry. Management of patients was independent of this diagnostic study and left to the discretion of the attending physician.
Reference standard: Adjudicated 30-day MACE
Two independent cardiologists adjudicated the presence (or absence) of 30-day MACE, including all-cause death, cardiac arrest, AMI, cardiogenic shock, sustained ventricular arrhythmia, and high-grade atrioventricular block (the primary endpoint), including the index event. The secondary endpoint was 30-day MACE plus unstable angina (UA) receiving early (≤24 h) revascularization. Any early (≤24 h) revascularization was considered, including revascularization performed for mainly medical reasons such as recurrent episodes of UA as well as revascularizations performed mainly for convenience reasons such as an available catheter laboratory slot. Because UA has a much better prognosis than AMI, it was not included in the primary endpoint (3).
The adjudicators reviewed all available medical records (i.e., patient history, physical examination, results of laboratory testing, radiologic testing, ECG, echocardiography, cardiac exercise stress test, lesion severity and morphology in coronary angiography) pertaining to the patient from the time of ED presentation to 90-days’ follow-up. Follow-up information was obtained directly from patients by telephone calls or in a written form, as well as from the patient’s hospital notes, the family physician’s or involved cardiologists’ records, and the national registry on mortality. In situations of disagreement about the diagnosis, cases were reviewed and adjudicated in conjunction with a third cardiologist. Adjudication was performed centrally in a core laboratory. The diagnosis of AMI included 2 sets of serial cTn measurements: serial cTn measurements obtained as part of routine clinical care locally (several cTn assays), and serial measurements of hs-cTnT from study blood draws performed centrally in a core laboratory to take advantage of the higher sensitivity and higher overall diagnostic accuracy offered by the hs-cTnT assays (16).
AMI was defined and hs-cTn levels interpreted as recommended in current guidelines (4). In brief, AMI was diagnosed when there was evidence of myocardial necrosis in association with a clinical setting consistent with myocardial ischemia. Myocardial necrosis was diagnosed by at least 1 hs-cTnT value above the 99th percentile together with a significant rise and/or fall. Absolute changes in hs-cTnT were used to determine significant changes based on the diagnostic superiority of absolute over relative changes (17,18). Based on studies of the biological variation of cTnT (19,20), as well as on data from previous chest pain cohort studies (21–23), a significant absolute change was defined as a rise or fall of at least 10 ng/l within 6 h or 6 ng/l within 3 h.
UA was diagnosed in patients with ischemic symptoms at rest or minor exercise with normal (hs)-cTn levels or mild elevations without dynamic changes (criteria for AMI not met) but requiring urgent (≤24 h) revascularization. The following criteria were interpreted as increasing the likelihood for UA versus noncardiac chest pain: typical angina pectoris at rest, worsening/deterioration of a previously stable angina, results of a cardiac stress test showing myocardial ischemia, coronary angiography revealing a diameter stenosis of at least 70%, fractional flow reserve documenting functional significance of a coronary lesion, and sudden cardiac death or AMI occurring during 60 days’ follow-up.
The Online Methods describe the blood sampling, laboratory methods, and the ESC hs-cTnT/I 0/1 h algorithms, which were optimized for the early rule-out and rule-in of AMI and incorporated assay-specific hs-cTnT/I concentrations at ED presentation and their absolute changes within 1 h (3).
The extended algorithm classified patients as rule-out, if in addition to the hs-cTnT/I 0/1 h algorithm triage recommendation, clinical assessment also revealed only a low or moderate likelihood for acute coronary syndrome (ACS) (<70%) and the ECG did not show possible ischemic changes, including ST-segment depression (≥1 mV in at least 2 contiguous leads), pathological Q waves, or T-inversion. Patients with at least one of these criteria were reclassified toward observe or rule-in. To qualify for rule-in, patients had to fulfill the criteria of the ESC hs-cTnT/I 0/1 h algorithm triage recommendation for rule-in or an hs-cTnT/I concentration above the 99th percentile combined with either a high likelihood for ACS (≥70%) and/or an ischemic ECG. The rationale for modifying the rule-in criteria was that in these patients with a high pre-test probability, an hs-cTnT level >14 ng/l (hs-cTnI >26 ng/l) should have a positive predictive value (PPV) sufficient for rule-in. The remaining patients were classified to the observe zone (Online Figures 2A and 2B).
Pre-test probability for an ACS as the cause of the presenting symptoms was quantified 90 min after presentation by the treating ED physician using a visual analog scale (VAS). At this time point, the ED physician had completed his or her clinical assessment, including patient history, chest pain characteristics, detailed physical examination including vital signs, and reviewed the ECG and the first local cTn measurement. A low or moderate likelihood for ACS was defined as <70% and a high likelihood was defined as ≥70%.
Safety for rule-out of MACE or MACE + UA was quantified according to the resulting negative predictive value (NPV) and the likelihood ratio; accuracy for rule-in of MACE or MACE + UA was quantified by the resulting PPV and the likelihood ratio; and efficacy was defined as the percentage of patients triaged to either rule-out or rule-in of MACE or MACE + UA. We used the cross-tables derived by the application of assay and algorithm-specific cutoff criteria for rule-out or rule-in to calculate diagnostic performance parameters and their 95% confidence intervals (CIs). Likelihood ratios with 95% CIs were calculated to assess the specific value of each specific rule-out or rule-in algorithm. McNemar’s statistics or generalized score statistics were used as appropriate to compare sensitivity, specificity, NPV, and PPV between both algorithms and predefined subgroups. Subgroup analyses have been performed in patients presenting within 2 h from chest pain onset, sex, age, known coronary artery disease (CAD), and estimated glomerular filtration rate.
All hypothesis testing was 2-tailed, and p values <0.05 were considered to indicate statistical significance. No adjustments for multiple comparisons were made. All statistical analyses were performed by using IBM SPSS Statistics for Windows and MAC version 25.0 (IBM SPSS Statistics, IBM Corporation, Armonk, New York) and R 3.3.1 (R Foundation for Statistical Computing, Vienna, Austria).
From April 2006 to August 2015, a total of 3,123 patients eligible for this analysis were enrolled (Online Figure 1A). Patients triaged toward rule-out according to the ESC hs-cTnT 0/1 h algorithm and the extended algorithm were younger and less often had cardiovascular risk factors, pre-existing CAD, ECG abnormalities, and cardiovascular medication compared with patients triaged toward observe and rule-in (Online Tables 2 and 3).
The ESC hs-cTnT 0/1 h algorithm triaged 1,880 patients (60%; 95% CI: 58.5% to 61.9%) toward rule-out. In 209 (11%) of these patients, integrated clinical judgment for ACS was ≥70%, 72 (4%) patients had ST-segment depression, 131 (7%) patients had significant Q waves, and 103 (6%) patients had T-wave inversion. At least 1 of these criteria was present in 487 (26%) patients in the rule-out group. The extended algorithm reclassified these patients to observe (relative increase of 66%) and, additionally, 297 previously classified observe patients were reclassified to rule-in (relative increase of 61%) (Central Illustration, Figure 1).
Prognostic performance for MACE
Within 30 days, MACE occurred in 524 (17%) patients (Tables 1 and 2). The ESC hs-cTnT 0/1 h algorithm triaged significantly more patients toward rule-out compared with the extended algorithm (60%; 95% CI: 58.5% to 61.9% vs. 45%; 95% CI: 42.9% to 46.4%; p < 0.001) while maintaining similar 30-day MACE rates (0.6%; 95% CI: 0.3% to 1.1% vs. 0.4%; 95% CI: 0.1% to 0.9%; p = 0.429), resulting in similar NPV (99.4%; 95% CI: 98.9% to 99.6% vs. 99.6%; 95% CI: 99.2% to 99.8%; p = 0.097) and likelihood ratios (0.03; 95% CI: 0.02 to 0.06 vs. 0.02; 95% CI: 0.01 to 0.04) (Figures 2A to 2C).
Among the 487 patients reclassified toward observe according to the extended algorithm, the 30-day MACE rate was 1.1% (6 patients, 2 patients with type 1 MI) and 16% (86 patients, 55 patients with type 1 MI, 15 with type 2 MI) among patients reclassified toward rule-in (Central Illustration). These 30-day MACE rates were significantly lower compared with the other observe and rule-in patients triaged by using the hs-cTnT component (both, p < 0.001). Detailed information on patients triaged toward rule-out of AMI with MACE is given in Online Table 4.
The ESC hs-cTnT 0/1 h algorithm ruled-in fewer patients (16%; 95% CI: 14.9% to 17.5% vs. 26%; 95% CI: 24.2% to 27.2%; p < 0.001) compared with the extended algorithm, albeit with a higher PPV (76.6%; 95% CI: 72.8% to 80.1% vs. 59%; 95% CI: 55.5% to 62.3%; p < 0.001) and a higher likelihood ratio for 30-day MACE (16.3; 95% CI: 13.5 to 19.5 vs. 7.1; 95% CI: 6.4 to 7.9). With the ESC hs-cTnT 0/1 h algorithm, fewer patients remained in the observe zone compared with the extended algorithm (24%; 95% CI: 22.2% to 25.2% vs. 30%; 95% CI: 28.1% to 31.3%; p < 0.001).
After excluding type 2 MI from MACE, a similar NPV was reported for rule-out (99.4%; 95% CI: 98.9% to 99.6% vs. 99.4%; 95% CI: 98.6% to 99.9%; p = 0.997) but a lower PPV for rule-in (76.6%; 95% CI: 72.8% to 80.1% vs. 69.5%; 95% CI: 65.4% to 73.4%; p < 0.001) for MACE and MACE without type 2 MI, respectively. Similar findings were reported when the extended algorithm was used (NPV, 99.6%; 95% CI: 99.2% to 99.8% vs. 99.6%; 95% CI: 99.2% to 99.8%; p = 0.989; and PPV, 68.5%; 95% CI: 65.2% to 71.6% vs. 53.2%; 95% CI: 49.8% to 56.7%; p < 0.001) for MACE and MACE without type 2 MI, respectively (Online Table 10).
Prognostic performance for MACE + UA
Within 30 days, MACE + UA occurred in 806 (26%) patients. Among patients triaged toward rule-out, the ESC hs-cTnT 0/1 h algorithm had a lower NPV and a higher negative likelihood ratio compared with the extended algorithm: 91.7% (95% CI: 90.4% to 92.9%) versus 95.3% (95% CI: 94.0% to 96.3%), and 0.26 (95% CI: 0.23 to 0.3) versus 0.14 (95% CI: 0.11 to 0.18; all, p < 0.001). Among patients triaged toward rule-in, the ESC hs-cTnT 0/1 h algorithm had a higher PPV and a positive likelihood ratio compared with the extended algorithm: 77% (95% CI: 73.2% to 80.5%) versus 68.5% (95% CI: 65.2% to 71.6%) and a positive likelihood ratio 9.5 (95% CI: 8.0 to 11.7) versus 6.2 (95% CI: 5.5 to 7.1; all, p < 0.001) (Figures 2A, 2B, 3A, and 3B).
A total of 830 patients (27%) presented within 2 h of chest pain onset. In these patients, the ESC hs-cTnT 0/1 h algorithm classified 532 patients (77%) as rule-out. For MACE, NPV was 99.3% (95% CI: 98.1% to 99.7%). Using the extended algorithm, 263 patients (38%) were classified as rule-out. For MACE, NPV was 100% (95% CI: 98.6% to 100%; p = 0.955 for comparison, respectively) (Online Figure 4, Online Table 11). Both algorithms showed consistent and overall favorable performance characteristics in additional subgroup analyses according to sex, age, known CAD, and estimated glomerular filtration rate.
Overall, the results for the ESC hs-cTnI 0/1 h algorithm and its extended algorithm were comparable and are shown in the Online Results.
This prospective, multicenter diagnostic study enrolling unselected patients presenting with acute chest discomfort to the ED used central adjudication to externally validate the performance of an extended algorithm, a pragmatic approach combining the ESC hs-cTnT/I 0/1 h algorithm with quantified clinical assessment and the ECG findings for the prediction of 30-day MACE. We report 5 major findings.
First, adding the 2 additional criteria (VAS for ACS <70% and ischemic ECG findings) as a requirement for triage toward rule-out significantly reduced the percentage of patients triaged toward rule-out compared with the ESC hs-cTnT 0/1 h algorithm only (45% vs. 60%). Among all reclassified patients, noncardiac causes of chest pain were the most common final adjudicated diagnoses, and 30-day MACE occurred in only 1.1%. These estimates will help clinicians to appropriately manage patients triaged toward rule-out according to the ESC hs-cTnT 0/1 h algorithm, in whom either the VAS for ACS or the ECG still suggests the presence of an ACS. As an alternative to subjective clinical judgment by the treating physician, a formal score could be used to help in the prediction of 30-day MACE. This strategy of using well-validated scores such as the HEART (history, ECG, age, risk factors, troponin) score or the Thrombolysis In Myocardial Infarction risk score in addition to early hs-cTnT/I–based algorithms has been evaluated in 2 recent studies and found not to provide an incremental value but substantially reduced rule-out efficacy (24,25). Current guidelines recommend a conservative approach and advocate prolonged monitoring, including an additional measurement of hs-cTnT/I 3 to 6 h after presentation in these patients (3).
Second, among patients triaged toward rule-out, the extended algorithm achieved a similar NPV for 30-day MACE (both >99%) compared with the ESC hs-cTnT 0/1 h algorithm alone, whereas sensitivity was higher when using the extended algorithm (99.0% vs. 97.7%; p = 0.014). Therefore, the ESC hs-cTnT 0/1 h algorithm seemed to better balance safety and efficacy overall. It is important to highlight that the MACE rate (including all-cause death, cardiac arrest, AMI, cardiogenic shock, sustained ventricular arrhythmia, and high-grade atrioventricular block) in the rule-out group of the ESC hs-cTnT 0/1 h algorithm only was 0.6%, and therefore within the range requested by ED physicians in a survey for patients considered for discharge from the ED (26).
Third, the extended algorithm had a higher rule-in rate, albeit a lower PPV, compared with the ESC hs-cTnT 0/1 h algorithm (59% vs. 77%). Because AMI at presentation was the dominant event contributing to 30-day MACE, it is a matter of debate if PPV is sufficient to justify a homogeneous management in the rule-in group, including early coronary angiography.
Fourth, including patients with UA receiving early revascularization to MACE led to an increase in MACE from 524 to 804 patients (relative increase of 53%) and a MACE event rate of 26% in the overall cohort. This reflects the enrollment of unselected (both high-risk and low-risk) patients with suspected AMI and the adjudication of AMI, including type 1 and type 2. Furthermore, excluding type 2 MI from MACE resulted in a similar NPV for rule-out but a lower PPV for rule-in. For 30-day MACE + UA, the extended algorithm significantly increased NPV compared with the ESC hs-cTnT 0/1 h algorithm, whereas the latter again achieved a higher PPV for rule-in.
Fifth, using hs-cTnI, the performance of the extended algorithm relative to the ESC hs-cTnI 0/1 h algorithm overall was similar to that observed for hs-cTnT. Similar findings have been found in several predefined subgroups, especially in early presenters within 2 h after chest pain onset (NPV 99.3% and 100%) for the ESC hs-cTnT 0/1 h algorithm and the extended algorithm, respectively.
These findings extend and corroborate insights gained in the single-center pilot study of the extended-algorithm (11). In those study patients triaged toward rule-out for 30-day MACE, the NPV was 99.9% (95% CI: 99.2% to 100%) for the ESC hs-cTnT 0/1 h algorithm and 100% (95% CI: 99.4% to 100%) for the extended algorithm. In patients triaged toward rule-in, the PPV was 65.4% (95% CI: 55.2% to 74.5%) and 52.7% (95% CI: 44.3% to 61.1%), respectively. Again, for 30-day MACE + UA, the extended algorithm had higher NPV, but lower PPV, compared with the ESC hs-cTnT 0/1 h algorithm.
The management of patients in whom the 3 diagnostic pillars for suspected AMI (including the ESC 0/1 h algorithm, the ECG findings, and clinical assessment summarized in the integrated clinical judgment) consistently suggest rule-out of AMI is usually straightforward and consists of rapid discharge from the ED followed by appropriate outpatient management (3). In contrast, in the subgroup of patients in whom the ESC hs-cTnT 0/1 h algorithm suggests rule-out of AMI but in whom either the VAS for ACS or the ECG still suggests the presence of an ACS, management remains controversial. Although prolonged initial monitoring and additional hs-cTnT/I measurements at ∼3 h are reasonable next steps for the vast majority of these patients, the most critical decision is whether to admit or to refer for outpatient management in case the 3-h measurements still do not provide evidence of AMI.
The low but not very low risk of MACE observed in these patients in this large study should provide for joined informed decision-making by the physician and the patient. Patient preferences, availability of same-day or next-day noninvasive imaging and/or coronary angiography, and absence or presence of comorbidities possibly decreasing the benefit/risk ratio of an early invasive approach (including dementia, very high age, severe renal dysfunction, and known atherosclerosis in the aortic arch) will have to be considered. These arguments are similar to those guiding management regarding possible ACS in patients triaged toward the observe zone (12).
Our findings corroborate and extend previous research regarding the development and validation of algorithms for the safe and effective rule-out and rule-in of MACE in patients with symptoms suggestive of AMI (11,24,25). Pioneering work by research groups in Asia-Pacific 15 years ago, at a time when only cTn assays with poor sensitivity were available, resulted in the first accelerated diagnostic protocols (11,27,28). These required the use of a formal risk score in addition to the ECG and cTn to achieve a very high NPV for MACE, and allowed the rapid rule-out in 9% of patients. The development of hs-cTn assays then enabled 2 advances of major clinical relevance. The first advance was to obviate the need of formal risk stratification for safe rule-out, while maintaining high NPV for MACE, which increased the likelihood for adaptation and adherence, and also substantially increased the percentage of patients eligible for early rule-out (11,24,25). For example, the ESC hs-cTnT/I 0/1 h algorithm, the ESC hs-cTnT/I 0/2 h algorithm, and the High-STEACS (High-Sensitivity Troponin in the Evaluation of Patients With Acute Coronary Syndrome) algorithm triage >50% of patients toward rule-out and provide comparable and very high safety (3,11,29). The second major advance was the addition of a rule-in component that also provides well-structured guidance to physicians for the early identification of those patients with acute chest discomfort requiring immediate hospitalization in a monitored unit, early coronary revascularization, early initiation of high-intensity statin, and potent dual antiplatelet therapy (11). Thereby, from being only helpful for the early identification of a small group of low-risk patients, these algorithms have matured toward highly effective clinical decision support tools that provide detailed guidance for patient management in ∼75% of unselected patients presenting with acute chest discomfort to the ED.
When diagnosed by using hs-cTnT/I assays, UA seems to more closely reflect the pathophysiology and outcomes of stable CAD than AMI (3,30). Patients with UA have substantially lower 30-day mortality, including fatal MI, compared with patients with non–ST-segment elevation MI (NSTEMI) (31). Likely due to these differences, early invasive strategies and potent dual antiplatelet therapy seem to provide less incremental benefit compared with patients with NSTEMI (3,32,33). Still, the appropriate detection of UA is of considerable importance, and the extended algorithm seems to be a valuable tool in this regard. Because the ESC hs-cTnT/I 0/1 h algorithms were designed to detect AMI, and because UA, by definition, is not associated with acute cardiomyocyte injury, the ESC hs-cTnT/I 0/1 h algorithms cannot detect UA. As an alternative to formal quantification of integrated clinical judgment, performed as part of the extended algorithm, UA can of course also be assessed and diagnosed using individualized clinical judgment based on history, chest pain characteristics, physical examination, and the ECG findings by the treating physician. It is worth mentioning that in our cohort, patients with UA underwent early percutaneous coronary intervention, which was likely combined with dual antiplatelet therapy and sometimes together with anticoagulation if indicated; this approach probably contributed to the clinical stabilization of these patients and their low risk of experiencing other MACE at 30 days. This finding should not misinterpret that patients with UA can be safely discharged from the ED without further evaluation. Further clinical trials are necessary to compare different management strategies based on these algorithms to clarify the best treatment in these patients.
First, our study was conducted in ED patients with symptoms suggestive of AMI. Further studies are required to quantify the utility of both algorithms in patients with either higher (e.g., in a coronary care unit setting) or lower (e.g., in a general practitioner setting) pre-test probability for AMI. Second, some patients did not have a 1-h sample and therefore were excluded from this analysis. It is very unlikely that the performance of both algorithms would be worse in these patients, particularly as a common reason for a missing blood sample at 1 h were logistic issues related to, for example, early transfer to the catheter laboratory. Third, although we used a very stringent methodology to adjudicate the presence or absence of AMI, including central adjudication by experienced cardiologists, imaging, and serial measurements of hs-cTn, we still may have misclassified a small number of patients (4). However, it is unlikely that these rare misclassifications affected the relative performance of the extended algorithm versus the ESC hs-cTnT/I 0/1 h algorithm. Fourth, in this prospective study, no specific sample size calculation was performed. Although this secondary analysis from a multicenter study is one of the largest ever performed, it may still have been underpowered for some subgroup analyses. Fifth, we cannot generalize our findings to patients with terminal kidney failure undergoing chronic dialysis because they were excluded from this study. Sixth, these algorithms were derived and validated in patients capable of providing informed consent. The use of these algorithms is discouraged in extreme conditions, such as in patients after cardiopulmonary resuscitation or in shock or respiratory failure, as they have not been tested in these subjects.
The ESC hs-cTn 0/1 h algorithm better balanced efficacy and safety in the prediction, both rule-out and rule-in, of 30-day MACE, whereas the extended algorithm is the preferred option for the rule-out of 30-day MACE + UA.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: When patients are diagnosed based on hs-cTnT/I assays, the pathophysiology and outcomes of UA more closely parallel stable coronary disease than AMI. Patients with UA have lower 30-day mortality than those with NSTEMI, and early invasive strategies and potent dual antiplatelet therapy provide less benefit than for patients with NSTEMI. The hs-cTnT/I 0/1 h MI rule-out algorithms do not detect UA, the diagnosis of which depends on the clinical characteristics of chest pain and ECG interpretation.
TRANSLATIONAL OUTLOOK: Randomized trials are needed to compare the outcomes associated with alternative management strategies for patients with chest pain based on hs-cTnT/I algorithms.
The authors are indebted to the patients who participated in the study and to the ED staff as well as the laboratory technicians of all participating sites for their most valuable efforts. In addition, they thank Kirsten Hochholzer, MSc, Melanie Wieland, RN, Irina Klimmeck, RN, and Fausta Chiaverio, RN (all University Hospital Basel, Basel, Switzerland); Esther Garrido, RN, and Helena Mañé Cruz (Hospital Clinic, Barcelona, Spain); and Miguel Angel García Briñón and María Suárez Cadenas (Hospital Clínico San Carlos, Madrid, Spain). Other APACE Investigators and contributors to this paper can be found in the Online Appendix.
↵∗ Drs. Nestelberger and Boeddinghaus have contributed equally and should be considered first authors.
This study was supported by research grants from the Swiss National Science Foundation, the Swiss Heart Foundation, the European Union, the Cardiovascular Research Foundation Basel, the University Hospital Basel, Abbott, Beckman Coulter, Biomerieux, BRAHMS, Roche, Nanosphere, Siemens, Ortho Diagnostics, and Singulex. The investigated high-sensitivity cardiac troponin assays were donated by the manufacturers, who had no role in the design of the study, the analysis of the data, the preparation of the manuscript, or the decision to submit the manuscript for publication. Dr. Nestelberger has received speaker honoraria from Beckman Coulter. Dr. Boeddinghaus has received research grants from the University of Basel and the Division of Internal Medicine, the Swiss Academy of Medical Sciences, the Gottfried and Julia Bangerter-Rhyner-Foundation; and has received speaker honoraria from Siemens and Roche. Dr. Twerenbold has received research support from the Swiss National Science Foundation (P300PB-167803_1), the Swiss Heart Foundation, the Swiss Society of Cardiology, the University Hospital Basel, and the Cardiovascular Research Foundation; and has received speaker honoraria/consulting honoraria from Abbott, Amgen, Brahms, Roche, Siemens, and Singulex. Dr. Badertscher has received research funding from the “Stiftung für Herzschrittmacher und Elektrophysiologie,” Dr. Koechlin has received a research grant from “Freiwillige Akademische Gesellschaft Basel.” Dr. Rubini Giménez has received speaker honoraria from Abbott; and has received research support from the Swiss National Science Foundation (P400PM_180828) and Swiss Heart Foundation. Dr. Puelacher has received a research grant from Roche Diagnostics. Dr. Gualandro has received speaker honoraria from Servier; and has received research grants from FAPESP (Foundation for Research Support of the State of São Paulo, Brazil). Dr. Walter has received a research grant from the Swiss Academy of Medical Sciences and the Bangerter Foundation (YTCR 23/17). Dr. Martin-Sanchez has received speaker, advisory, or consulting fees from Novartis, MSD, Bristol-Myers Squibb, Pfizer, The Medicine Company, Otsuka, Thermo Fisher, Cardiorentis, and Sanofi; and has received research grants from the Spanish Ministry of Health and FEDER, Mapfre, Novartis, Bayer, MSD, Abbott, and Orion-Pharma. Dr. Reichlin has received research grants from the Goldschmidt-Jacobson-Foundation, the Swiss National Science Foundation (PASMP3-136995), the Swiss Heart Foundation, the Professor Max Cloëtta Foundation, the Uniscientia Foundation Vaduz, the University of Basel, and the Department of Internal Medicine, University Hospital Basel; and has received speaker honoraria from Brahms and Roche. Dr. Mueller has received research support from the Swiss National Science Foundation, the Swiss Heart Foundation, the KTI, the European Union, the University of Basel, the University Hospital Basel, the Stiftung für kardiovaskuläre Forschung Basel, Abbott, Beckman Coulter, Biomerieux, Indorsia, Ortho Clinical Diagnostics, Quidel, Roche, Siemens, Singulex, and Sphingotec; and has received speaker honoraria/consulting honoraria from Abbott, Amgen, AstraZeneca, Biomerieux, Boehringer Ingelheim, Bristol-Myers Squibb, Brahms, Cardiorentis, Indorsia, Novartis, Roche, Sanofi, Siemens, and Singulex. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Listen to this manuscript's audio summary by Editor-in-Chief Dr. Valentin Fuster on JACC.org.
- Abbreviations and Acronyms
- acute coronary syndrome
- acute myocardial infarction
- coronary artery disease
- confidence interval
- cardiac troponin
- emergency department
- European Society of Cardiology
- high-sensitivity cardiac troponin
- major adverse cardiac events
- negative predictive value
- positive predictive value
- unstable angina
- visual analog scale
- Received March 12, 2019.
- Revision received May 6, 2019.
- Accepted June 3, 2019.
- 2019 American College of Cardiology Foundation
- Niska R.,
- Bhuiya F.,
- Xu J.
- Anderson J.L.,
- Adams C.D.,
- Antman E.M.,
- et al.
- Authors/Task Force Members,
- Roffi M.,
- Patrono C.,
- Collet J.P.,
- Mueller C.,
- Valgimigli M.,
- et al.
- Thygesen K.,
- Alpert J.S.,
- Jaffe A.S.,
- et al.
- Al-Saleh A.,
- Alazzoni A.,
- Al Shalash S.,
- et al.
- Boeddinghaus J.,
- Nestelberger T.,
- Twerenbold R.,
- et al.
- Twerenbold R.,
- Neumann J.T.,
- Sörensen N.A.,
- et al.
- Badertscher P.,
- Boeddinghaus J.,
- Twerenbold R.,
- et al.
- Mokhtari A.,
- Borna C.,
- Gilje P.,
- et al.
- Twerenbold R.,
- Wildi K.,
- Jaeger C.,
- et al.
- Bossuyt P.M.,
- Reitsma J.B.,
- Bruns D.E.,
- et al.
- Haaf P.,
- Drexler B.,
- Reichlin T.,
- et al.
- Reichlin T.,
- Irfan A.,
- Twerenbold R.,
- et al.
- Vasile V.C.,
- Saenger A.K.,
- Kroning J.M.,
- Jaffe A.S.
- Wu A.H.B.,
- Quynh A.L.,
- Todd J.,
- Moecks J.,
- Wians F.
- Apple F.S.,
- Pearce L.A.,
- Smith S.W.,
- Kaczmarek J.M.,
- Murakami M.M.
- Hammarsten O.,
- Fu M.L.X.,
- Sigurjonsdottir R.,
- et al.
- Chapman A.R.,
- Hesse K.,
- Andrews J.,
- et al.
- Morawiec B.,
- Boeddinghaus J.,
- Wussler D.,
- et al.
- Than M.,
- Cullen L.,
- Aldous S.,
- et al.
- Chapman A.R.,
- Anand A.,
- Boeddinghaus J.,
- et al.
- Eggers K.M.,
- Jernberg T.,
- Lindahl B.