Objective To identify patients with type 2 myocardial infarction (MI) and patients with non-ischaemic myocardial injury (NIMI) and to compare their prognosis with those of patients with type 1 MI.
Methods A retrospective observational study was performed in 1010 patients admitted to the emergency department of a university hospital with at least one troponin I test between 2012 and 2013. Participants were identified using laboratory records and divided into three groups: type 1 MI (rupture of atheromatous plaque), type 2 MI (imbalance between myocardial oxygen supply and/or demand) and NIMI (patients who did not meet diagnostic criteria for type 1 or type 2 MI). Clinical characteristics and 2-year outcomes were analysed.
Results Patients with type 2 MI and NIMI were older, with higher proportion of women and more comorbidities than patients with type 1 MI. Absolute mortality and the adjusted risk for all-cause mortality in both groups were significantly higher than that of patients with type 1 MI (39.7%, HR: 1.41 95% CI 1.02 to 1.94, p=0.038 and 40.0%, HR: 1.54 95% CI 1.16 to 2.04, p=0.002, respectively). Patients with type 2 MI and NIMI tended to present more readmissions due to heart failure (16.5%, HR: 1.55 95% CI 0.87 to 2.76, p=0.133 and 12.3%, HR: 1.15 95% CI 0.70 to 1.90, p=0.580) and less readmission rates due to acute coronary syndrome (ACS) than patients with type 1 MI (2.1%, HR: 0.11 95% CI 0.04 to 0.31, p<0.001 and 4.3%, HR: 0.22 95% CI 0.12 to 0.41, p<0.001),
Conclusions Patients diagnosed with type 2 MI and NIMI have higher rates of mortality and lower readmission rates for ACS compared with patients with type 1 MI.
Statistics from Altmetric.com
The universal definition of myocardial infarction (MI) introduced the term type 2 MI to describe several clinical conditions that are associated with ischaemic myocardial injury in the absence of complicated atheromatous plaques.1 The prognosis of these patients is directly related to the underlying disease.2 ,3 Due to the huge variety of clinical conditions involved in type 2 MI, the clinical reports referring to this entity are heterogeneous and do not allow to precisely delineate the mechanisms and prognosis of this disease.
However, there is a group of patients with elevated levels of troponin that do not fulfil the diagnostic criteria of type 1 or type 2 MI. These patients are often discharged with a variety of clinical diagnoses such as necrosis without MI,4 myocardial injury5 or myocardial damage,6 and their prognosis is not well established. Indeed, several clinical studies have shown that the presence of high troponin levels is associated with a worse prognosis.7–9 It is therefore conceivable that the outcome of patients with high troponin levels but without typical symptoms of myocardial ischaemia would not be better than the outcome of patients with type 2 MI, however, this has not yet been reported. If this hypothesis were confirmed, then the universal definition of MI should be reviewed accordingly.
Therefore, the aim of our study was to compare the 2-year all-cause mortality and readmissions rates for acute coronary syndrome (ACS) and heart failure between patients with type 2 MI and patients with high troponin levels but without typical myocardial ischaemia related symptoms. Since the prognosis of patients with type 1 MI is well established, a sample of this population is included in this study as reference group.
This is a retrospective cohort study involving all patients who were admitted at the emergency department of a university hospital between 1 January 2012 and 31 December 2013 and who underwent at least one troponin I (TnI) test. The patients were identified using the laboratory records. The troponin tests were performed according to the chest pain protocol of the centre, although these biochemical analyses were also performed in patients with atypical symptoms or in those with suspected ACS. For patients with more than one TnI test we selected the highest TnI value and for those who were admitted to the emergency department several times, we included the highest value recorded during the first admission episode. The exclusion criteria were: (A) age below 18 years, (B) patients recovered from cardiac arrest, (C) clinical conditions known to be associated with high troponin levels such as myocarditis or pulmonary embolism10 ,11 and (D) patients living away from our referral area.
Categorisation of the study population
Patients were classified by consensus of two cardiologists according to the third universal definition of MI12 into the following three categories: (1) type 1 MI was diagnosed when the physiopathological mechanism was thought to indicate the rupture of an atheromatous plaque and the resulting intracoronary thrombus; (2) type 2 MI was diagnosed when a condition other than suspected coronary artery disease led to an imbalance between myocardial oxygen supply and/or demand according to the criteria proposed previously by Saaby et al.4 Conditions with decreased oxygen supply were: anaemia defined as a haemoglobin concentration <90 g/L for men and <80 g/L for women; shock defined as systolic blood pressure <90 mm Hg together with signs of organ dysfunction (ie, metabolic acidosis, arterial oxygen saturation <90%, oliguria (diuresis <30 mL/h for at least 3 hours) or encephalopathy); bradycardia requiring medical treatment or cardiac pacing; coronary embolus in the presence of an increased risk of embolism (left heart endocarditis, intracardiac mural thrombus, documented venous thrombus and a patent foramen ovale or atrial septum defect); respiratory failure with an arterial oxygen saturation <90% and clinical signs of acute respiratory failure lasting 20 min. Conditions with increased oxygen demand were: ventricular tachyarrhythmia lasting 20 min; supraventricular tachyarrhythmia lasting 20 min with a ventricular rate >150 bpm; hypertensive pulmonary oedema defined as the presence of a systolic blood pressure >160 mm Hg, signs of pulmonary oedema and a need for treatment with nitrates or diuretics; arterial hypertension with a systolic blood pressure >160 mm Hg and concomitant left ventricular hypertrophy identified by echocardiography or ECG. Finally, the third group, non-ischaemic myocardial injury (NIMI) included patients with high troponin levels who did not meet the diagnostic criteria for type 1 MI or type 2 MI.
Study clinical variables
The electronic clinical records of all patients were reviewed and we included their demographic variables, Charlson Index, physical examination at the initial evaluation in the emergency department, ECG findings, laboratory test and main cardiac explorations (echocardiogram, exercise stress test and coronary angiography). The glomerular filtration rate was calculated using the Modification of Diet in Renal Disease 4 formula. The principal diagnoses on discharge were recorded for patients with type 2 MI and NIMI.
Plasma troponin I
All TnI measurements were carried out in the same laboratory using the immunoassay technique (TnI-Ultra from Siemens, Advia Centaur). In accordance with the manufacturer, the lower and upper detection limits were 0.006 μg/mL and >50 μg/mL, respectively. Levels under the detection limit were given a 0 value and those above 50 μg/mL were given a value of 50. The reference range for a positive TnI test was >0.039 μg/mL, corresponding to the 99th centile of a reference control group, with a coefficient of variation <10%.
The primary outcome of the study was all-cause mortality at 2-year follow-up. The secondary outcomes were readmission rates for heart failure and ACS at 2-year follow-up. The follow-up events were obtained from patients' electronic clinical records and from death registers.
The baseline characteristics of patients in the three categories were compared using the Kruskal-Wallis test for continuous variables and Pearson's c2 test for categorical variables. The Wilcoxon rank-sum test was used for continuous variables when patients with type 2 MI were compared directly with patients with NIMI. Data are presented as medians and IQRs for continuous variables and as counts with percentages for categorical variables. Cox proportional hazards regression analysis was used in univariate and multivariate mortality analyses. Proportional hazards assumption was evaluated assessing the constancy of the parallel plotted lines in the log-log graph and by Schoenfeld residuals. Death from any cause can be considered a competing event of readmission for heart failure and ACS; for that reason, a multivariate competing risk model was used to obtain the HRs for readmission for heart failure and ACS in patients with type 2 MI and NIMI compared with type 1 MI. Cumulative incidence curves using the competing risk model were plotted. Differences were considered statistically significant at p<0.05. STATA V.13.0 (College Station, Texas, USA) was used for all analyses.
From a total of 3790 patients assessed for TnI in the emergency department, 1010 fulfilled the inclusion criteria and were included in the study (figure 1). Patients with type 2 MI and NIMI were found to be older and to have higher comorbidities as expressed by the Charlson Index than patients with type 1 MI. However, there were no differences between the three groups in the documented history of diabetes or hypertension. Patients with type 1 MI were more likely to have a history of smoking. The principal symptoms presented by patients with type 2 MI and NIMI were dyspnoea and atypical symptoms, in contrast to patients with type 1 MI, who were more likely to present with chest pain. Patients with type 2 MI and NIMI also presented with greater deterioration in renal function, lower levels of haemoglobin and higher rates of atrial fibrillation than patients with type 1 MI. The highest TnI levels and hospitalisation rates were observed in patients with type 1 MI (table 1).
Clinical profile of patients with type 2 MI and NIMI
No differences were found in terms of age, proportion of men or medical history between patients with type 2 MI and NIMI, being the risk profile similar for both groups (table 1). Patients with type 2 MI were more likely to present with dyspnoea, in contrast to patients with NIMI, who were more likely to present with atypical symptoms. Patients with type 2 MI were more likely to have tachycardia, lower oxygen saturation and slightly lower levels of haemoglobin. Both groups presented with deteriorated baseline renal function, this being slightly worse in cases of type 2 MI. Almost half of the patients needed to be admitted to hospital and cardiology assessment were similar for both groups. Echocardiogram was performed more frequently in patients with type 2 MI.
Table 2 shows the main diagnoses for patients with type 2 MI and NIMI at discharge. Heart failure, heart rhythm disorders (tachyarrhythmia, bradycardia) and anaemia were most prevalent in patients with type 2 MI, whereas syncope, cerebrovascular disease, and digestive and infectious pathologies were more common in patients with NIMI.
During follow-up, 39.7% of patients with type 2 MI and 40.0% of patients with NIMI died, in contrast to 19.7% of patients with type 1 MI (figure 2; table 3). After adjusting for age, sex, history of MI, heart failure, hypertension, diabetes, Charlson Index and glomerular filtration, the risk of mortality in patients with type 2 MI was higher than in patients with type 1 MI (HR: 1.41 95% CI 1.02 to 1.94, p=0.038). Likewise, the risk for mortality was higher in patients with NIMI than in patients with type 1 MI (HR: 1.54 95% CI 1.16 to 2.04, p=0.002). Readmission rates for heart failure were higher in patients with type 2 MI and NIMI, although not statistically significant (HR: 1.55 95% CI 0.87 to 2.76, p=0.133 and HR: 1.15 95% CI 0.70 to 1.90, p=0.580); readmission rates for ACS were significantly lower in patients with type 2 MI and NIMI compared with patients with type 1 MI (HR: 0.11 95% CI 0.04 to 0.31, p<0.001 and HR: 0.22 95% CI 0.12 to 0.41, p<0.001) (figure 3; table 3).
This study reveals that patients admitted to the emergency department with high plasma troponin levels but free of ischaemia related symptoms have a 2-year mortality and readmission risk comparable to patients with raised troponin levels secondary to myocardial ischaemic injury who fulfil the type 2 MI criteria of the universal definition of MI. Thus a population of high-risk patients with abnormal troponin levels fails to be included in the universal definition of MI types.
Diagnostic criteria for type 2 MI
The continuous technical refinement of laboratory troponin measurement brought about this test as an ideal biomarker for early detection of MI and this has favoured a progressive increase in the number of patients with high troponin levels presenting with acute or chronic clinical conditions other than ACS.13 This heterogeneous clinical scenario led to a review of the classical diagnostic criteria of MI. A consensus document including the definition of five different types of MI was published in 20071 and later revisited.2 ,4 ,14–17 We followed the criteria used by Saaby et al4 to classify patients into the type 2 MI category since it afforded the most reliable pathophysiological description of this clinical entity. In the present study we deal with the problem of a key group of patients with high troponin levels who are not currently assigned to either type 1 or type 2 MI and therefore are excluded from most publications. Tentatively, we have assigned these patients to the NIMI category assuming that the myocardial injury did not result from a specific ischaemic phenomenon but rather from systemic processes leading to the death of cardiomyocytes. Possible mechanisms for elevated troponins include myocardial stretch, cytokine-mediated myocardial injury and catecholamine-mediated myocardial toxicity.18 The majority of cardiac troponin is found in the contractile apparatus and is released via proteolytic degradation. Although 6–8% of cardiac troponin has been reported to occur as a free cytosolic component,19 the precise amount is not certain, because studies have used tissues from variable sources (organ donor tissue, postmortem tissue or atrial tissue). Release from a cytosolic pool has been suggested to occur with ischaemia; however, the current accepted evidence supports the concept that the release of cardiac troponin is due to cell death and is irreversible. Other proposed mechanisms to explain cardiac troponin release include apoptosis, normal myocyte turnover, increased cell wall permeability, and formation and release of membranous blebs.20 In cases of sepsis or infectious diseases, local and circulating inflammatory markers may lead to direct myocardial injury by cytotoxic effects.21
Our study describes the clinical profile and carries out a follow-up of patients diagnosed with NIMI and, furthermore, directly compares such patients with patients diagnosed with type 2 MI. The prevalence of the NIMI in our study was about 43% and this is in contrast with the higher incidence in the study by Saaby et al.4 The prevalence of the NIMI in our study was about 43% and this is in contrast with the higher incidence in the study by Saaby et al. Although these authors did not separately analyse the “non-ischaemic” group, their incidence was about 71%. One possible explanation for this discrepancy is that our study included only patients who had been admitted to the hospital through the emergency department and not the total number of patients who had been admitted to the hospital through different ways (ie, scheduled hospital admissions or direct admission to hospital ward). The prevalence of type 2 MI in our study was about 19% and this was comparable with that found in previous studies, where the range was from 1.6% to 29%.22 ,23 However, comparisons with other publications should be viewed with caution given the heterogeneity of the study populations.
The clinical profile of patients with NIMI is not substantially different from that of patients with type 2 MI since both groups were characterised by advanced age, slight predominance of men, high prevalence of comorbidities such as heart failure, hypertension and chronic kidney disease, and high prevalence of atrial fibrillation. This clinical profile is consistent with those found in previous studies.14 ,22–23 The number of hospital admissions among these patients is quite low if compared with that of patients with type 1 MI. These results are consistent with the findings by Brunner24 and Fundaro,25 demonstrating that many patients with high troponin levels are not hospitalised. The apparently lower rate of hospital admissions may in fact reflect that the attending physician is prone to institute less aggressive treatments due to the lack of evidence of ongoing myocardial ischaemia. Coronary angiography was performed in a high proportion of patients with type 1 MI while in patients with type 2 MI or NIMI it was performed only in a few cases. These results, similar to those found by Saaby et al4 and Stein et al,16 could be justified for the clinical and ECG presentation of patients with type 1 MI and for the clinical risk profile in the other groups, favouring the implementation of invasive management in patients with type 1 MI. The benefit of the invasive approach in patients with type 2 MI or NIMI has not been studied. Previous studies have reported similar clinical diagnoses at hospital discharge as those in our study and these encompassed various clinical entities, in particular heart failure, heart rhythm disorders and respiratory diseases.
Of particular note is our finding that patients with type 2 MI and patients with NIMI have similar mortality and readmission rates for heart failure during the 2-year follow-up and, moreover, these events occurred more frequently than in patients with type 1 MI; for patients with type 2 MI, this worse prognosis is similar to that described by Saaby26 and it could be explained by the conjunction of several factors which include advanced age and increased presence of comorbidities despite having lower levels of troponin; the worse prognosis is also related to the elevation of TnI, as has been evidenced by Bardají et al,13 and it is also possible that the excess of mortality in this population is determined by the underlying disease. In contrast, readmission rates for ACS during the follow-up of patients diagnosed with type 2 MI and NIMI are lower than those for patients with type 1 MI, which suggests that a high proportion of these patients do not present with significant coronary lesions,27 ,28 a finding that should be investigated in future studies.
The need for a new nomenclature
Patients diagnosed with type 1 MI are treated with medications and invasive procedures that have been shown to increase survival in various clinical trials. However, the distinction between type 2 MI and NIMI (the latter having additional implications for the psychosocial environment of the patient) has not led to the development of rigorously tested therapeutic strategies or procedures aimed at improving their mid-term or short-term prognosis. For this reason, in agreement with other authors,29 ,30 we believe that a review of the current nomenclature of MI is essential to ensuring the homogeneity of clinical studies and thus a thoughtful understanding of this clinical entity.
The following potential limitations are important for interpreting our results. Among these are the fact that this is a retrospective study and we observed only a small number of events during the follow-up. Despite this, the study accurately reflects the real life clinical scenario faced by emergency services in our daily practice. Exclusion of patients with myocarditis or pulmonary embolism may lead to selection bias, despite being a small number of patients. Our patients received differing inhospital medical treatments according to the underlying disease and this may lead to treatment bias. We are also unaware of the treatments that the patients had after the admission to emergency department; their potential influence on mortality and readmission rates during follow up cannot be determined.
Our study shows that a high percentage of patients admitted in emergency departments with high levels of TnI who do not fulfil the present criteria for type 2 MI, thus being classified as NIMI, have in fact a comparable clinical profile and a similar 2-year outcome. Moreover, these patients presented with higher rates of mortality and lower readmission rates for ACS compared with patients with type 1 MI. Our data prompt the need for reviewing the present MI nomenclature, in order to help standardise clinical studies, contribute to a comprehensive understanding of these clinical entities and favour the implementation of more effective therapeutic measures. In this scenario, the presence or absence of elevated troponin should play a key role as a major factor in identifying patients at high mortality risk.
What is already known on this subject?
Few studies have investigated the prognosis of patients with type 2 myocardial infarction (MI) with an important variety of diagnostic criteria in all of them. To date, categorisation and prognosis of patients with positive troponin I who do not meet the criteria for type 1 or type 2 MI is still unclear.
What might this study add?
Clinical risk profile of patients with raised troponin levels and without acute coronary syndrome (ACS) is similar regardless of their final diagnosis and is characterised by higher rates of mortality (39.7% for type 2 MI and 40.0% for non-ischaemic myocardial injury) and lower readmission rates for ACS compared with patients with type 1 MI (2.1% for type 2 MI and 4.3% for non-ischaemic myocardial injury).
How might this impact on clinical practice?
Many patients with raised troponin levels fail to be allocated within the universal definition of MI. A review of the present MI nomenclature should be considered in order to contribute to standardise clinical studies that lead us to a better understanding of these clinical conditions.
The authors thank Dr Juan Cinca for his critical revision and intellectual contribution.
Twitter Follow Germán Cediel at @GermanECediel
Contributors All authors contributed significantly to the work. GC and AB conceived the study. Statistical analysis was carried out by GC. AB supervised data analyses. The manuscript was drafted by GC and AB. MG-d-H, AC, RS and CB revised it critically for intellectual content. All authors provided final approval of the manuscript.
Competing interests None declared.
Ethics approval Local ethical committee (Comité Ético de Investigación Clínica – CEIC).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.