Management and outcomes with 5‐year mortality of patients with mildly elevated high‐sensitivity troponin T levels not meeting criteria for myocardial infarction

To examine management and outcomes of patients presenting to EDs with symptoms suggestive of acute coronary syndrome, who have mild non‐dynamically elevated high‐sensitivity troponin T (HsTnT) levels, not meeting the fourth universal definition of myocardial infarction (MI) criteria (observation group).

• ED patients with MI neither ruled-in or ruled-out (observation group), compared to those with MI had similar 1-year, but higher 5-year mortality.• Observation group patients were predominantly adjudicated as having chronic myocardial injury and had high non-cardiac death rates at 1 year.• Of 547 observation group patients, 62% were admitted, 42% had non-invasive tests (36% echocardiography) and 16% had coronary angiography.

Introduction
The majority of patients presenting to EDs potentially with symptoms of myocardial ischaemia, including chest pain or dyspnoea, do not have an acute coronary syndrome (ACS) as the final diagnosis.2][3][4][5] Early 'rule out' of MI is feasible in a large proportion of patients undergoing suspected ACS assessment protocols, which contributes to effective resource utilisation, improved ED patient flow and hospital efficiency. 3,6owever, 20-30% of patients require further observation and investigations 7 as they often have minor elevations in HsTn levels, which are often non-dynamic, and have MI neither ruled-in nor ruled-out.These patients represent a significant burden on busy EDs and hospital services and have similar 1-year mortality to those with MI. 2 The management of these patients, sometimes called in the 'observation group', 2 is heterogeneous in the absence of evidencebased guidelines, both with respect to non-invasive and invasive investigations and pharmaco-therapies.Also, whether further assessment and management should be undertaken in inpatient or outpatient settings needs clarification.
In the present study, we sought to characterise disposition, management including investigations and outcomes of patients with minor but non-dynamic troponin T elevations following initial assessment for a suspected ACS.

Methods
The study cohort consisted of consecutive patients presenting to the ED at Liverpool Hospital, Sydney, Australia over a 4-month period in 2014 (1 March 2014-30 June 2014) who were assessed for a suspected ACS using a previously reported algorithm, 8,9 and were deemed to need 'further observation'.South Western Sydney Local Health District Human Research Ethics Committee approval number was 12/050.Most patients had >1 HsTnT levels ≥15 ng/ L, but did not meet MI criteria based on the fourth universal definition of MI (4th UDMI). 10Laboratory HsTnT assays (Roche, Basle, Switzerland) were performed with an upper reference limit (URL) of 14 ng/L.
After initial assessment, patient classification was MI 'ruled-in', MI 'ruled-out', or those requiring further observation and evaluation (Fig. 1), called the 'observation group' when MI was neither ruled-in nor ruled-out.These patients mostly had baseline HsTnT levels between 15-49 ng/L and <3 ng/L/h change (particularly when the HsTnT testing interval was >3 h), 6,11 or an absolute HsTnT change of 5 ng/L < X < 10 ng/ L, or relative delta (10% < X < 50%).The causes of presentation were adjudicated according to the 4th UDMI. 10,12Diagnoses, cardiac and non-cardiac, were adjudicated by at least two cardiologists (four in total), according to the 4th UDMI, 12 as previously described. 9An MI at presentation was defined as occurring within the first 18 h.The main reasons for not repeating an HsTnT level included presentation at >6 h from the symptom onset, unknown symptom onset time, very low risk of MI on medical review, or ruled-in MI with high suggestive symptoms and/or ECG changes consistent with acute ischemia.In the MI ruled-in group, 12 patients had first sample haemolysis, and then two valid measurements and 38 patients no repeated second sample.Among ruled-out MI patients, a patient with previous coronary artery bypass grafting, had two initial HsTnT levels <14 ng/L, although he reported stuttering symptoms with dynamic ST segment changes, and was subsequently diagnosed as gastrointestinal sepsis associated with type 2 MI and acute myocardial injury.*Type 2 MI includes acute myocardial injury. 9HsTnT, highsensitivity troponin T; MI, myocardial infarction.
Patient disposition was categorised as discharged when patients were solely under the care of the ED team or admitted, either to cardiology or other in-patient services.Cardiac investigations within the study period (and also where indicated 6 months prior to, or post, this period) performed in Liverpool Hospital, were recorded.Late outcomes, including death, MI, rehospitalisation, were determined by review of medical records, clinician contact and patient contact (if necessary).The cause of death was defined as cardiac (e.g.fatal MI, heart failure or sudden death), where there was a documented cardiac cause, or if it was unexplained after review of all available documents, as previously described, 6,11 although deaths due to peripheral vascular causes not included.However, if a patient had a terminal malignancy, for example, as the underlying cause of death but they died of an arrhythmia, the death was classified as non-cardiac.MI was recorded within 30 days and 1 year, and rehospitalisation and cardiacassociated rehospitalisation within a month.Significant angiographic coronary artery disease was defined as ≥50% stenosis in an artery large enough to qualify for Synergy between percutaneous coronary intervention with Taxus and Cardiac Surgery scoring (SS); this definition includes significant side branch arteries. 13linical presentation, demographic descriptors and use of investigations were obtained from the cardiology department or hospital records.Non-invasive cardiac tests, including exercise stress testing (EST), echocardiography (including post-stress), myocardial perfusion scans, cardiac magnetic resonance imaging, computerised tomography coronary angiography (CTCA) and invasive procedures, as well as outcomes were determined by examining hospital records and reviewed by cardiologists.
Statistical analyses were performed using SPSS version 24.0 (SPSS Inc., Chicago, IL, USA).Categorical variables are expressed as numbers and percentages, and continuous variables as mean AE standard deviations or medians and interquartile range (IQR) for skewed variables.For group comparisons, Pearson's χ 2 test or Fisher's exact test were used as appropriate for unpaired categorical variables.The Student's t-tests or the Mann-Whitney U tests (for skewed distribution) were used for related categorical variables.Survival analysis used the Kaplan-Meier method, and P-values <0.05 (two-sided) were considered statistically significant.

Results
Among 2738 patients with suspected ACS who had ≥1 reported HsTnT level, 1355 had two valid HsTnT results (reasons for only one sample see Fig. 1); the median time between the HsTnT levels was 3.8 h (IQR 3.13-5.37).Of these patients, 623 (46%) had MI ruled-out, 208 (8%) had MI ruled-in, and 547 patients had minor non-dynamic elevations in troponin T levels, representing 40% of patients with two samples, or 20% of all these with suspected ACS.These patients with minor non-dynamic HsTnT elevations, in comparison to those in whom MI was ruled-out, were older (P < 0.0001), and were more likely to have hypertension, hyperlipidaemia, chronic kidney disease with estimated glomerular filtration rate <60 mL/min/1.73m 2 , congestive heart failure or prior coronary heart disease (all P < 0.05) (Table 1).
Of patients with minor nondynamic HsTnT elevations, 352 (60%) had ≥2 HsTnT levels between 15 and 49 ng/L, whereas the remainder had at least one level outside this range.Of these patients, 8/547 (1.5%) were adjudicated as an MI at presentation, four due to late presentation with stable elevated HsTnT levels and four stuttering symptoms; a further 16 patients had unstable angina but the elevated HsTnT levels were adjudicated as chronic myocardial injury.Final diagnoses of patients with chronic myocardial injury are shown in Table S1.Among these patients, 204 (37%) were discharged at a median of 8.2 h (IQR 6.6-14.3),and 343 patients were admitted, of whom 179 (52%) were admitted to cardiac services and 164 (48%) to non-cardiac services.
Of the 623 patients in whom an MI was ruled-out, 131 (21%) were admitted, 93 (15%) being admitted to cardiac services, of whom 19 (3%) had a diagnosis of unstable angina.No cardiac deaths or further MI events among ruled-out MI patients occurred in 1 year.Furthermore, 1-year mortality was 1% overall (1% among hospitalised and 1% among discharged; P = 0.826), and many were returned to their general practitioner's care, although 78 (12%) had no community follow-up recorded.Only 31 (4.9%) ruled-out MI patients were hospitalised at 30 days, including 4 (0.6%) due to cardiac causes (Table 2).

Discussion
Of all patients with suspected ACS attending EDs ≥20% had MI neither confirmed nor ruled-out after initial assessment protocols.These patients who minor non-dynamic HsTnT elevations, have lack an evidence-base to guide their management.We found $60% were admitted presumably due to clinician concerns about their 1-year event rates. 14,15Chronic myocardial injury was the common adjudicated diagnosis.As an elevated troponin is generally associated with a worse prognosis, 3,15 some form of cardiac testing occurred, mainly noninvasive echocardiography, in 42% of these 'observation group' patients.
The mortality rates for 'observation group' patients were 2% at 30 days and 13% at 1 year, which was similar to MI type 1 (11%) and similar to other studies. 14,16We also found that among those patients with chronic myocardial injury, a non-cardiac diagnosis had a higher mortality rate than those with a cardiac diagnosis, also reported in other studies. 17,18ollow-up at 5 years found patients with chronic myocardial injury or type 2 MI had higher late mortality rates than those with type 1 MI, also reported in other studies. 9,19,20Bardaji et al. 21studied a similar cohort of patients and found all-cause 4-year mortality rates among patients with chronic myocardial injury of 52% similar to 48% at 5-year mortality rates among 'observation group' patients who were adjudicated as chronic myocardial in our study.
We adjudicated that most patients with non-dynamic HsTnT levels as having chronic myocardial injury. 10A small proportion of observation group patients 8/547 (1.5%) had in retrospect type 1 MI at presentation, four due to late presentation and four stuttering symptoms.In the following year 19/547 ($4%) of all 'observation group' patients had MI.In contrast, among the 'observational cohort' of 285 patients in the TRAPID-AMI study, 14 which was approximately half the size of our observation group cohort in the present study, 22.5% of patients had an MI, although due to consent requirements of patient selection in TRAPID-AMI was likely to have led to recruitment of patients with somehow different characteristics.While both studies used change in HsTnT levels, TRAPID-AMI sampling used a 0-1 h whereas ours was scheduled for 0-3 h.We used velocity change in HsTnT, 22 whereas others have reported absolute changes in HsTn levels were diagnostically superior to relative changes. 23,24While HsTnT testing was liberally applied in our study including in patients with medical conditions such as sepsis, atrial fibrillation, heart failure and renal failure (50% of patients in our study with chronic myocardial injury had a reduction of estimated glomerular filtration rate), the positive predictive value of the particular presentation representing an acute ischemic event was likely to be reduced. 25However, the risk of late events may be higher. 3,9hether in patients with stable minor HsTnT elevations routine functional testing or coronary angiography (including CTCA) should be undertaken, and whether this should occur as an inpatient or out-patient, needs clarification.Cardiac imaging has often been performed in ED on actively symptomatic patients, including echocardiography and radionuclide myocardial perfusion imaging. 26In our study in these patients, 99/189 ($52%) of echocardiograms were performed on non-admitted patients, whereas 5/19 (26%) of MPS were conducted as inpatients.Also, our study revealed that half of stress tests among observation group patients (18 EST, five MPS and one stress echo) were conducted as inpatient.On the other hand, some studies suggested that outpatient stress Of 'observation group' patients, 198/231 (85%) of NITs were conducted during the study period; whereas 96% of NITs in ruled-out MI were performed during the same period.Echocardiography was the most conducted test which contributed for 415/659 (60%) of all tests in all groups.The second requested test was stress testing which accounted for 237/695 (34%) of NITs, and mainly among ruled-out MI patients 171/237 (72%).Exercise stress tests were performed on 27/547(5%) of observation group patients, of whom 12 were normal, seven were non-diagnostic, two were borderline and six were positive, of whom four had significant angiographic coronary artery disease requiring revascularisation.Myocardial perfusion scans (MPS) were performed in 19 (4%) patients (four also had echocardiography; 13 during the study period), and of those 18 were either normal or had non-significant abnormalities; the MPS positive patient had significant angiographic coronary artery disease.CTCA, computerised tomography coronary angiography; Echo, echocardiography; EST, exercise stress test; MBS, stress myocardial perfusion scan; TOE, transoesophageal echocardiography; TTE, transthoracic echocardiography.
testing is a reasonable alternative for appropriately screened patients with low probability of ACS, who are compliant and reliable and have proper instructions with close follow-up, with an outpatient stress test scheduled to be performed within 2 weeks of ED discharge. 27mong patients in whom MI was ruled-out, 2% had an ACS diagnosis (unstable angina).][31] Of patients in our study, 50% (1302/2738) had only first HsTnT level recorded, for a variety of reasons, often late presentation.As HsTnT testing at 1 h has been subsequently increasingly adopted, caution is required in early presenting patients, and an additional HsTnT level >4 h after symptom onset may be necessary. 6Approximately, three quarters of stress tests in ruled-out MI patients (113 EST) were conducted in ED.
Our study had very low rates of CTCA in those with stable mild troponin T elevations (2/547), and patients with MI (2/208) or MI ruled-out (7/623).This alternate approach to functional testing imaging safely facilitates higher rates of discharge from the ED, 32 and can also assist to rule out aortic dissection and pulmonary embolism.][35] However, the 500-patient BEACON trial, which is the only trial involving both CTCA and HsTnT testing, and included those with mild HsTnT elevations, did not show shortened lengths of ED stay. 36ur study had several limitations, including while being a prospective cohort study, necessarily adjudication of diagnoses was retrospective, so unappreciated biases could have occurred.The onset time of the symptoms was sometimes not Although exercise testing without inducible ischemia suggests that obstructive coronary artery disease is unlikely, a prior negative stress test is not useful to rule out MI in patients with active chest pain in the ED. 37,38A larger, ideally multicentre study would be required to determine the reproducibility of our disposition decision, outcomes and rate of non-invasive testing.

Conclusion
Most unselected patients undergoing HsTnT testing in ED who had minor non-dynamic elevations had chronic myocardial injury.Of these patients, 60% were admitted and 40% had some form of cardiac testing.Mortality rates at 1 and 5 years of these 'observation group' patients were higher than those with confirmed type 1 MI but lower than those with type 2 MI.Patients with non-cardiac causes of TnT elevations had high readmission rates, which were similar to those with cardiac causes.Furthermore, the timing, preferred location options of non-invasive testing including CTCA in the assessment of these patients should be prospectively examined.

Figure 1 .
Figure 1.Study flow diagram.The diagram shows the patient disposition and diagnoses according to HsTnT testing, of whom 81 had first sample haemolysis.Of the 1302 patients who had one HsTnT level, this was ≤14 ng/L in 968, ≥15-49 ng/L in 253 and ≥50 ng/L in 81.The main reasons for not repeating an HsTnT level included presentation at >6 h from the symptom onset, unknown symptom onset time, very low risk of MI on medical review, or ruled-in MI with high suggestive symptoms and/or ECG changes consistent with acute ischemia.In the MI ruled-in group, 12 patients had first sample haemolysis, and then two valid measurements and 38 patients no repeated second sample.Among ruled-out MI patients, a patient with previous coronary artery bypass grafting, had two initial HsTnT levels <14 ng/L, although he reported stuttering symptoms with dynamic ST segment changes, and was subsequently diagnosed as gastrointestinal sepsis associated with type 2 MI and acute myocardial injury.*Type 2 MI includes acute myocardial injury. 9HsTnT, highsensitivity troponin T; MI, myocardial infarction.

Figure 2 .
Figure 2. Rate of non-invasive tests among observation group, myocardial infarction (MI) ruled-in and MI ruled-out.The flow diagram shows non-invasive tests (NITs) for the whole 1378 patients were 695 (50%).Of 'observation group' patients, 198/231(85%) of NITs were conducted during the study period; whereas 96% of NITs in ruled-out MI were performed during the same period.Echocardiography was the most conducted test which contributed for 415/659 (60%) of all tests in all groups.The second requested test was stress testing which accounted for 237/695 (34%) of NITs, and mainly among ruled-out MI patients 171/237 (72%).Exercise stress tests were performed on 27/547(5%) of observation group patients, of whom 12 were normal, seven were non-diagnostic, two were borderline and six were positive, of whom four had significant angiographic coronary artery disease requiring revascularisation.Myocardial perfusion scans (MPS) were performed in 19 (4%) patients (four also had echocardiography; 13 during the study period), and of those 18 were either normal or had non-significant abnormalities; the MPS positive patient had significant angiographic coronary artery disease.CTCA, computerised tomography coronary angiography; Echo, echocardiography; EST, exercise stress test; MBS, stress myocardial perfusion scan; TOE, transoesophageal echocardiography; TTE, transthoracic echocardiography.

Figure 3 .
Figure 3. Kaplan-Meier curves for mortality among observation, myocardial infarction (MI) ruled-in, and MI ruled-out group.Late survival to 60 months is shown among patients with ruled-in MI, ruled-out MI and in the observation group.Numbers at risk in 10-monthly intervals are shown.
*The only significant P values comparing patients having further observation and MI type 2 were for reduced eGFR (both P

TABLE 2 .
Patient outcomes in the observation group, myocardial infarction and myocardial infarction ruled-out © 2023 The Authors.Emergency Medicine Australasia published by John Wiley & Sons Australia, Ltd on behalf of Australasian College for Emergency Medicine.LATE MORTALITY AFTER NON-DYNAMIC TROPONIN T RISES specified, potentially confounding our interpretation of HsTnT serial measurements.Also, due to our ED's liberal HsTnT testing strategy, and organisational characteristics of our institution, which has an emergency short-stay unit, but only limited general medical services, extrapolation of our findings to institutions 1-hour algorithm to rule-out and rule-in acute myocardial infarction using a high-sensitivity cardiac troponin T assay.CMAJ 2015; 187: E243-52.6. Chew DP, Lambrakis K, Blyth A et al.A randomized trial of a 1-hour troponin T protocol in suspected acute coronary syndromes: the rapid assessment of possible acute coronary syndrome in the emergency department with high-sensitivity troponin T study (RAPID-TnT).Circulation 2019; 140: 1543-56.7. Chiang C, Chiang C, Lee GH et al.Safety and efficacy of the European Society of Cardiology 0/1-hour algorithm for diagnosis of myocardial infarction: systematic review and meta-analysis.Heart 2020; 106: 985-91.8. Saad YM, McEwan J, Shugman IM et al.Use of a high-sensitivity troponin T assay in the assessment and disposition of patients attending a tertiary Australian emergency department: a cross-sectional pilot study.Emerg.Med.Australas.2015; 27: 405-11.9. Etaher A, Gibbs OJ, Saad YM et al.