Abstract
Aims:
First, describe how acute myocardial infarction criteria are used to diagnose type 1 (T1MI) and 2 (T2MI) myocardial infarction. Second, determine whether subjective or objective criteria are used for T2MI. Third, examine outcomes for T2MI based on the presence or absence of objective evidence of myocardial ischemia compared with myocardial injury.
Methods and results:
Post-hoc analysis of UTROPIA (NCT02060760), a prospective, observational, cohort study involving 1640 consecutive emergency department patients with serial high-sensitivity cardiac troponin I among whom 74 (4.5%) had T1MI, 103 (6.3%) T2MI, and 245 (15%) myocardial injury. Compared with T1MI, patients with T2MI were less likely to have ischemic symptoms (97% vs. 83%), Q waves (24% vs. 1%), new ST-T wave changes (74% vs. 51%), new regional wall motion abnormality (64% vs. 11%), and a culprit lesion on coronary angiography (59% vs. 0%) (all p <0.05). T2MIs were more likely to be diagnosed using subjective criteria (symptoms alone) than T1MI (42% vs. 12%, p <0.0001). Patients with objective T2MI, but not subjective T2MI, had a two-fold increase in early mortality compared with myocardial injury, with 30- and 60-day hazard ratios (95% confidence interval) of 2.3 (0.9, 6.2) and 2.0 (0.9, 4.7) respectively.
Conclusions:
Among patients with T2MI, many cases are diagnosed using subjective criteria. The presence of objective evidence of myocardial ischemia may identify a higher-risk group of T2MI patients in whom early outcomes are worse than myocardial injury. Emphasis on using objective evidence of myocardial ischemia to diagnose T2MI may result in a more precise and specific disease definition.
Introduction
Acute myocardial infarction (MI) is a clinical syndrome affecting over 800,000 individuals in the United States (US) each year. 1 Numerous efforts have been made over time to define and re-define the diagnosis of acute MI based on our improved understanding of the distinct mechanisms leading to MI, as well as the development and availability of new diagnostic modalities that impact how we recognize this condition.2–4 Early definitions of acute MI were based on history, electrocardiographic (ECG) changes, and serum enzymes. 2 In addition to symptoms, 12-lead ECG, and cardiac troponin (cTn) as the preferred cardiac biomarker to support the diagnosis of acute MI, the diagnostic criteria have since evolved to include also non-invasive imaging and coronary angiography.3,4
Recognizing the distinct pathophysiological etiologies leading to acute MI, which is not always due to acute coronary atherosclerotic plaque disruption, the Global Task Force for the Universal Definition of MI (UDMI) introduced a novel subtype classification, 4 amongst which type 2 MI (T2MI) has received much attention because of its frequency and impact on prognosis.5–12
T2MI occurs in instances of acute myocardial injury with clinical evidence of acute myocardial ischemia where a condition other than an acute atherothrombotic event leads to an imbalance between myocardial oxygen supply and/or demand.3–5 Much debate exists surrounding how to diagnose T2MI, with a wide range in incident rates observed across studies, partly due to ambiguity and variation in how the diagnosis is established.5–16 Challenges persist regarding its diagnosis in routine clinical practice and its adjudication in research studies. 6 Moreover, various studies have suggested that long-term outcomes between T2MI and myocardial injury (INJ) are similar.8,17,18
The goals of our study were three-fold. First, we described and compared how the MI diagnostic criteria were utilized to support the diagnosis of type 1 (T1MI) and T2MI. Second, we examined whether subjective (symptoms alone) or objective criteria were used to diagnose T2MI. Third, we examined outcomes for T2MI according to the presence or absence of objective features of myocardial ischemia compared with INJ.
Methods
Post-hoc analysis of the Use of TROPonin In Acute coronary syndromes (UTROPIA) (NCT02060760) study, a prospective, observational cohort study involving consecutive, unselected patients in whom pre-set serial cTnI measurements were ordered on clinical indication in the emergency department (ED) at Hennepin County Medical Center (Minneapolis, MN, USA) between 4 February 2014 and 9 May 2014 to rule-in and rule-out acute MI. The design and primary results have been reported previously.8,19 For inclusion, patients needed a baseline cTnI measurement at presentation and at least one additional cTnI measured within 24 h of presentation before discharge, and at least one 12-lead ECG performed. Patients <18 years, pregnant, with trauma, those who declined to participate on research as documented on information disclosure, and those who did not present through the ED, or were transferred from an outside hospital, were excluded. For patients with more than one presentation during the study period, we included only the first.
Detailed methods describing the high-sensitivity cTnI (hs-cTnI) assay used and event adjudication processes are included in the Supplementary Material online. In brief, the present study analysis is based on adjudications performed using a hs-cTnI assay (Abbott) (sex-specific 99th percentiles: 16 ng/L for women and 34 ng/L for men). All cases with at least one hs-cTnI >99th percentile were adjudicated according to the Third UDMI consensus recommendations 3 by two clinicians following review of all available medical records including 12-lead ECG, echocardiography, angiography, hs-cTnI concentrations, and clinical presentation. Cases with an adjudication discrepancy were reviewed and adjudicated by a third senior clinician. Following the UDMI, in addition to a rise and/or fall of hs-cTnI with at least one value >99th percentile, adjudicators were required to identify the presence of ⩾1 criteria to support making the diagnosis of acute MI. For the purpose of this study, subjective MI refers to those participants that were diagnosed based on subjective criteria (ischemic symptoms alone) in addition to the rise and/or fall of hs-cTnI levels; whereas objective MI refers to participants that have at least one objective criterion in addition to the rise and/or fall in hs-cTnI levels as the basis for the diagnosis, as shown in Supplementary Material Table 1.
T1MI was defined as MI related to atherosclerotic plaque rupture, ulceration, fissuring, erosion, or dissection with resulting intraluminal thrombus. 3 Following examination of the catheterization (if performed) report, adjudicators were required to determine whether there was angiographic evidence of a culprit lesion (i.e. atherosclerotic plaque disruption as described above). T2MI was defined as MI secondary to an ischemic imbalance between myocardial oxygen supply and/or demand not due to atherothrombosis and was adjudicated following the use of broad criteria,3,5 an approach in which adjudicators evaluate all contributing supply/demand variables and give a diagnosis of T2MI without applying strict parameters. For T2MI to be adjudicated cases were required to have definite evidence or documentation of a supply/demand imbalance, with adjudicators required to identify the source(s) of increased demand and/or decreased supply. The term INJ applies to any cTnI increase >99th percentile that was not adjudicated as MI. For those with hs-cTnI increases >99th percentile, mortality was examined until two-year follow-up.
Statistical analysis
Categorical variables were compared using Pearson’s chi-square and Fisher’s exact test where 50% of cells were <5. Continuous variables were compared using the F-test. Diagnostic agreement rates (%) were evaluated using kappa (κ) statistics with respective 95% confidence intervals (CIs). Diagnostic criteria were calculated based on available adjudication forms for each participant. If any adjudicator said that the criterion was present then the participant was considered to have that criterion; only when all adjudicators agreed that the criterion was not present was the participant deemed negative for the criterion. Cumulative survival curves were plotted using the Kaplan–Meier method and compared using the log-rank test. Cox proportional hazard ratios with 95% CIs are presented. Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).
Results
Among 1640 patients, 177 (10.8%) patients had acute MI, of which 74 (4.5%) had T1MI and 103 (6.3%) T2MI. For the entire cohort of 1640 patients, excellent diagnostic agreement occurred in 92% of cases (κ: 0.81, 95% confidence interval (CI): 0.78–0.84) (Supplementary Table 2). For those with increased hs-cTnI (n=422) values, moderate diagnostic agreement occurred in 69% of cases (κ: 0.44, 95% CI: 0.31–0.48), with agreement rates of 95% (κ: 0.78, 95% CI: 0.68–0.87) for T1MI, 72% (κ: 0.21, 95% CI: 0.11, 0.32) for T2MI, and 71% (κ: 0.37, 95% CI: 0.27–0.46) for INJ.
All patients had serial hs-cTnI measurements and at least one 12-lead ECG following study design, with transthoracic echocardiography and coronary angiography performed in 84% and 65% of patients with T1MI and 58% and 19% of T2MI, respectively. Patients with T1MI were more likely than T2MI to have ischemic symptoms (97% vs. 83%, p=0.002), had a higher rate of developing pathological Q waves (24% vs. 1%, p<0.0001), electrocardiographic changes indicative of new ischemia (74% vs. 51%, p=0.002), imaging evidence of new loss of viable myocardium or new regional wall motion abnormality (64% vs. 11%, p<0.0001), and a culprit lesion on coronary angiography (59% vs. 0%, p<0.0001) (Figure 1).
Diagnostic criteria for type 1 and 2 myocardial infarction.
Patients with T1MI were more likely than T2MI to have multiple criteria supporting the diagnosis of acute MI, as follows: a) two criteria: 82% vs. 42%, p<0.0001, b) three criteria: 68% vs. 3.9%, p<0.0001, c) four criteria: 53% vs. 0%, p<0.0001, d) five criteria: 15% vs. 0%, p<0.0001, respectively. Only 12% of patients with T1MI were diagnosed based on the presence of subjective criteria (symptoms alone), whereas 42% of T2MI were diagnosed based on the presence of subjective criteria without any objective evidence of myocardial ischemia (p<0.0001). Among those with objective T2MI (n=60), the components used to support the diagnosis (according to frequency) were: electrocardiographic changes indicative of new ischemia (88%), ischemic symptoms (70%), imaging evidence of new loss of viable myocardium or new regional wall motion abnormality (18%), and development of pathological Q waves (1.7%).
Baseline characteristics are shown in Table 1. No differences were noted in the rates of congestive heart failure and renal insufficiency between subjective T2MI and INJ. Conversely, patients with INJ were more likely than objective T2MI to have a history of congestive heart failure (36% vs. 17%, p=0.004) and renal insufficiency (38% vs. 17%, p=0.002) (driven by a higher rate of end-stage renal disease on hemodialysis among those with INJ compared with objective T2MI, 18% vs. 3.3%, p=0.004). No differences were observed in the rate of chest discomfort among subgroups; however, patients with subjective T2MI were more likely to have dyspnea as the presenting symptom compared with patients with objective T2MI (58% vs. 33%, p=0.01) (Table 2).
Baseline characteristics.
T2MI: type 2 myocardial infarction; INJ: myocardial injury; PCI: percutaneous coronary intervention; CABG: coronary artery bypass grafting
Diagnostic evaluation including symptoms, 12-lead electrocardiogram and high-sensitivity cardiac troponin I results.
T2MI: type 2 myocardial infarction; INJ: myocardial injury; hs-cTnI: high-sensitivity cardiac troponin I
No differences were noted among subgroups across baseline hs-cTnI concentrations. Patients with objective T2MI (mean hs-cTnI 786 ng/L), but not subjective T2MI (mean hs-cTnI 264 ng/L), had significantly higher maximum hs-cTnI concentrations than those with INJ (193 ng/L, p=0.003) (Table 2). Echocardiographic and angiographic findings are shown in Supplementary Table 3.
There was no difference in mortality rates between subjective T2MI and INJ at any time point during the two-year follow-up. Conversely, patients with objective T2MI had a two-fold increase in early all-cause mortality as compared with patients with INJ (30- and 60-day hazard ratios (95% CI) of 2.3 (0.9, 6.2) and 2.0 (0.9, 4.7) respectively) (Table 3, Figure 2). No differences in subsequent outcomes existed among subgroups during two-year follow-up (Table 3, Figure 3).
Outcomes.
T2MI: type 2 myocardial infarction; INJ: myocardial injury; CI: confidence interval
Thirty-day mortality, objective type 2 myocardial infarction (T2MI) versus myocardial injury.
Two-year mortality, objective type 2 myocardial infarction (T2MI) versus myocardial injury.
Discussion
The clinical syndrome of acute MI is increasingly recognized in patients with an imbalance in myocardial oxygen supply and/or demand not due to atherosclerotic plaque disruption.5,6 T2MI can be due to either a primarily coronary etiology such as coronary vasospasm, coronary embolism, spontaneous coronary artery dissection, or coronary endothelial dysfunction, or it can, alternatively, be secondary to another acute condition or illness that contributes to a myocardial oxygen supply–demand imbalance and consequent myocardial ischemia such as severe anemia or significant tachy-brady-arrhythmias. T2MI, however, has not been fully embraced by clinicians and researchers, partly due to ambiguity in how the diagnosis is established. Such ambiguity has led to various efforts and proposals to refine its diagnosis. Some have proposed to use strict pre-defined criteria to define what constitutes a reduced supply and/or increased demand. 9 Others have suggested that more emphasis be placed on the presence of coronary artery disease (CAD) and to make CAD a requirement for the diagnosis.14,15,20 Others have simply suggested that such nomenclature should not be used. 13 Due to the existing challenges in diagnosing T2MI, it is both timely and important to address how the diagnosis is established using existing criteria following the UDMI consensus recommendations.
Our study, UTROPIA, a prospective, observational, cohort study, required adjudicators to identify each criterion that supported making the diagnosis of acute MI. Such design allowed for this important analysis focused on exploring the diagnostic and prognostic classification of patients with T2MI.
Several findings are unique to our study. First, a significant proportion of T2MI diagnoses are established using subjective criteria (symptoms alone) following the UDMI consensus recommendations. These recommendations allow the diagnosis of acute MI in the appropriate clinical setting when there is only one additional diagnostic criterion and this may be limited to ischemic symptoms alone. 3 The additional criterion may also be new or presumed new significant ST-T wave changes or new left bundle branch block, development of pathological Q waves in the ECG, imaging evidence of new loss of viable myocardium or new regional wall motion abnormality, and identification of an intracoronary thrombus by angiography. 3 There is paucity of data, however, informing how the diagnosis of MI is established in clinical practice or in research studies using these criteria, mainly among patients with T2MI.
In the context of typical chest pain and high pre-test probability, symptoms alone are sufficient to diagnose acute MI if combined with a rise and/or fall of cTn. Many patients, however, undergo evaluation for suspected MI, but suffer more often from subtle, atypical symptoms. The challenge for clinicians evaluating such patients relies on the fact that patients with acute MI can indeed present with more atypical presentations or even without symptoms. 21 This is especially true for women, the elderly, those with diabetes mellitus, patients in the post-operative setting, or the critically ill.3,21 Consequently, in practice, clinicians often pursue cTn measurements in such scenarios to rule-in and rule-out acute MI. This practice, however, creates challenges for both clinicians and adjudicators in research studies, as there are patients with rising and/or falling cTn concentrations suggestive of an acute event where either there may not be objective evidence of myocardial ischemia or patients may not undergo testing assessing for such. Consequently, clinicians and adjudicators rely solely on the clinical presentation and the subjective assessment of symptoms to either make or not make the diagnosis of acute MI.
Our study shows that nearly 90% of adjudicated T1MI have objective evidence of myocardial ischemia supporting the diagnosis, and only a minor subset was diagnosed solely on the presence of symptoms. On the other hand, a significant proportion of T2MI diagnoses are established using subjective criteria (symptoms alone). These observations highlight how the distinction between INJ and T2MI is challenging and subtle, with such distinction largely based on whether the symptoms and clinical presentation are judged to be ischemic or not. In this regard, our study demonstrates that the main source of discordant diagnoses stems from disagreements between T2MI and INJ, highlighting the need for more precise diagnostic criteria.
Previous studies examining the incidence of T2MI have shown wide incidence rates, partly explained by differences in diagnosis definitions, differences in cTn assays and cutoff values used, and study populations. 6 Our study, consistent with the findings from Gard et al., 16 shows that agreement rates for T2MI following UDMI diagnostic criteria are at the most fair-to-moderate. Our current study elaborates on these challenges by showing how the existing UDMI diagnostic criteria and the interpretation of symptoms, a subjective criterion, may also influence the incidence of T2MI in both clinical practice and research studies. Further, we also demonstrate that patients with subjective T2MI and INJ share certain clinical features, as shown by the similar rates of congestive heart failure and renal insufficiency, as well as the similar biomarker profile shown by maximum hs-cTnI concentrations; results that differ from those categorized as objective T2MI.
We suggest a definition that places more emphasis on the use of objective features of myocardial ischemia to diagnose T2MI; an approach that may potentially represent a more practical disease definition that can be replicated across studies, facilitate disease coding, and overcome the existing diagnostic and prognostic overlap with INJ, while also identifying a subset of patients at higher short-term risk. It may be debated that an approach that increases diagnostic specificity for T2MI would cause a reciprocal reduction in diagnostic sensitivity; the proposed diagnostic approach, however, would still identify these patients as having INJ, a population known to have an increased risk for adverse events, and therefore such reclassification should not lead to under-appreciation of risk.
Our reported rate of T2MI is consistent with several other recent US-based studies. For example, in a study by Meigher et al. examining 96,612 US ED patients, 792 patients had acute MI, amongst which 57% had T2MI. 22 Similarly, Gaggin et al. examined consecutive patients undergoing coronary and peripheral angiographic procedures and showed that among 206 patients with MI on follow-up, 74% had at least one T2MI. 12 Conversely, studies using more selective inclusion criteria (e.g. chest pain cohorts or those admitted to cardiac units) tend to show lower rates of T2MI and higher rates of T1MI. The increased rate of T2MI observed in several US studies may be partly explained by cTn testing being used more broadly in the US across scenarios in which INJ or T2MI is more likely, as compared with outside the US, where cTn is used more selectively and T1MI is more frequent; observations supported by the findings of Shah et al. showing how patient selection for cTn testing varies across healthcare settings. 23 The implementation of the ICD-10 code for T2MI will hopefully facilitate the epidemiological assessment of T2MI across distinct centers and patient populations, if used appropriately.
Second, we examined whether outcomes for T2MI varied according to the presence or absence of objective features of myocardial ischemia. Consistent with previous studies showing that long-term outcomes are similar between T2MI and INJ, our present analysis demonstrates that at two years, regardless of classification, mortality rates are similar among subgroups. Our study, however, does suggest that the presence of objective evidence of myocardial ischemia may identify a higher-risk subset of patients with T2MI in whom early outcomes are worse than INJ. Among the potential reasons for these findings, it is possible that patients with objective T2MI have worse early outcomes due to the larger amount of myocardium damaged; with patients classified as objective T2MI (mean maximum hs-cTnI: 786 ng/L), but not subjective T2MI (mean maximum hs-cTnI: 264 ng/L), having significantly higher hs-cTnI concentrations than patients with INJ (mean maximum hs-cTnI: 193 ng/L).
Potential limitations exist. First, debate exists surrounding how to define T2MI and INJ with further guidance expected in the Fourth UDMI. Our study does provide unique insights into how the existing UDMI criteria are used uniformly across MI subtypes and is the first to examine whether the criteria applied to spontaneous T1MI should be similarly applied to patients with T2MI; a group of patients that often present more atypically, usually without chest pain. Second, due to our sample size and number of events at follow-up, we acknowledge that there is limited power to conclude whether mortality differences exist among subgroups, with larger studies needed to validate our observations. Third, while symptoms are certainly the most subjective criteria open to interpretation, we recognize that nuances also exist in interpreting 12-lead ECGs, as well as other invasive and non-invasive imaging studies. Fourth, the under-diagnosis of MI is possible as not all patients undergo systematic diagnostic evaluation, with not all patients having a transthoracic echocardiography and/or coronary angiography. This limitation, however, reflects routine clinical care and depends on whether clinicians opt to pursue additional testing based on each individual patient’s clinical circumstances.
Conclusions
Among patients with T2MI, many cases are diagnosed using subjective criteria (symptoms alone). The presence of objective evidence of myocardial ischemia may identify a higher-risk subset of T2MI patients in whom early outcomes are worse than INJ. Emphasis on using objective features of myocardial ischemia to diagnose T2MI may result in a more precise and specific disease definition that can be replicated across studies, avoiding the present diagnostic and prognostic overlap with INJ.
Supplemental Material
SUPPLEMENTAL_MATERIAL_Revision_040218 – Supplemental material for Use of objective evidence of myocardial ischemia to facilitate the diagnostic and prognostic distinction between type 2 myocardial infarction and myocardial injury
Supplemental material, SUPPLEMENTAL_MATERIAL_Revision_040218 for Use of objective evidence of myocardial ischemia to facilitate the diagnostic and prognostic distinction between type 2 myocardial infarction and myocardial injury by Yader Sandoval, Stephen W Smith, Anne Sexter, Karen Schulz and Fred S Apple in European Heart Journal: Acute Cardiovascular Care
Footnotes
Acknowledgements
The authors thank all the staff at the Cardiac Biomarker Trials Laboratory, Minneapolis Medical Research Foundation and the Department of Laboratory Medicine, Hennepin County Medical Center for their contributions, as well as other collaborators, including: Drs Sarah E Thordsen, Charles A Bruen, Michelle D Carlson, Kenneth W Dodd, Brian E Driver, Katherine Jacoby, Benjamin K Johnson, Sara A Love, Johanna C Moore, and Nathaniel L Scott.
Conflict of interest
YS: has been in the past a non-salaried advisor for Roche Diagnostics. SWS: Consultant: Siemens. FSA: Consultant: Metanomics Healthcare; Board of Directors, HyTest Ltd.; Honorarium: Instrumentation Laboratory, Abbott Diagnostics; Research PI through Minneapolis Medical Research Foundation (MMRF) not salaried: Abbott Diagnostics, Roche Diagnostics, Siemens Healthcare, Alere/Quidel, Ortho-Clinical Diagnostics, Becton Dickinson, Beckman Coulter, Abbott POC, Davita, Amgen; Other: Associate Editor Clinical Chemistry. All other authors have nothing to disclose.
Funding
This work was supported by a) Abbott Diagnostics and b) the Minneapolis Medical Research Foundation.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
