Myocardial infarction with normal coronary arteries is common and associated with normal findings on cardiovascular magnetic resonance imaging: results from the Stockholm Myocardial Infarction with Normal Coronaries study
Södersjukhuset, Stockholm, Sweden
Correspondence: Olov Collste, Department of Cardiology, Södersjukhuset, 118 83 Stockholm, Sweden. (fax: +46 8 616 30 31; e-mail: email@example.com).
Myocardial infarction with angiographically normal coronary arteries (MINCA) is an important subtype of myocardial infarction; however, the prevalence, underlying pathophysiology, prognosis and optimal management of this condition are still largely unknown. Cardiovascular magnetic resonance (CMR) imaging has the potential to clarify the underlying pathology in patients with MINCA. The objective of this study was to investigate the diagnostic value of CMR imaging in this group of patients.
The prospective, multicentre, observational Stockholm Myocardial Infarction with Normal Coronaries (SMINC) study.
Coronary care units in the Stockholm metropolitan area.
Patients between 35 and 70 years of age with MINCA were consecutively included in the screening phase of the SMINC study. All patients had a typical clinical presentation, fulfilling the universal definition of myocardial infarction and had normal coronary angiography finding. Patients with known structural or coronary heart disease or other known causes of elevated troponin levels were excluded.
In total, 176 patients with MINCA were screened from 2007 to 2011. Of these, 152 underwent CMR imaging. The investigation was performed a median of 12 (interquartile range 6–28) days after hospital admission; 67% of the findings were normal, whereas 19% of patients had signs of myocardial necrosis and 7% had signs of myocarditis. The remaining patients (7%) had either unrecognized hypertrophic cardiomyopathy or could not be classified.
In this consecutive series of patients with MINCA, CMR imaging may help to differentiate between those with myocarditis, myocardial necrosis and normal myocardium. The incidence of MINCA was higher than previously reported. After excluding cases of myocarditis, MINCA consists of a large group of patients with normal CMR imaging results and a smaller group with myocardial necrosis. The aetiologies of these different imaging findings need to be explored.
Myocardial infarction with normal coronary arteries (MINCA) is an important subtype of myocardial infarction. Interest and awareness of MINCA has increased recently because of the frequent use of coronary angiography, the description of Takotsubo stress cardiomyopathy and the introduction of new sensitive troponin assays. However, despite an increasing interest, there have been few studies of MINCA, hence the prevalence remains uncertain. It was recently reported that approximately 4% of all patients with suspected myocardial infarction may have angiographically normal coronary arteries , whereas we have reported an incidence of 7% . Because myocarditis may mimic myocardial infarction, it is essential to exclude this condition in patients with MINCA. An accurate diagnosis is important to reduce unnecessary medications and investigations as well as stigmatization for the patient.
It has been shown that cardiovascular magnetic resonance (CMR) imaging is able to differentiate between myocarditis and myocardial infarction [3-6]. Previous studies of MINCA have not used CMR imaging to exclude myocarditis [7, 8]. The aim of this study, which is part of the screening phase of the Stockholm Myocardial Infarction Normal Coronaries (SMINC) study, was to prospectively investigate the value of CMR imaging in patients with MINCA. A further aim was to define, as accurately as possible, subgroups of MINCA and exclude myocarditis, which has been included and may have affected the results of previous studies of MINCA.
From June 2007 to May 2011, a total of 176 patients were screened at five different coronary care units in the Stockholm metropolitan area. Patients between 35 and 70 years of age with typical chest pain suggestive of myocardial infarction, an ECG showing sinus rhythm on admission, an elevated troponin level and a coronary angiography with no or minimal signs of atheromatosis according to the angiographer were eligible for the study. Minimal atheromatosis was defined as small irregularities in the coronary vessel wall, not giving rise to any stenosis >30%.
Myocarditis was excluded by CMR imaging; some patients with classical clinical features of myocarditis or perimyocarditis did not undergo coronary angiography and were therefore not eligible for the study. The first 100 patients underwent chest computed tomography (CT) to exclude those with pulmonary embolism. Because all CT examinations were negative for pulmonary embolism, the protocol was changed to measurement of D-dimer generally and chest CT examination only in the case of suspicion of pulmonary embolism. Patients with a history of structural or coronary heart disease, a pacemaker, chronic obstructive lung disease or renal disease were not screened. Acute myocardial infarction was defined according to the ESC/ACC/AHA universal definition of myocardial infarction  and the diagnosis of Takotsubo stress cardiomyopathy was based on the Mayo Clinic criteria . Patients fulfilling the criteria for diagnosis of either acute myocardial infarction or Takotsubo stress cardiomyopathy were included in the study. Troponin levels were calculated relative to the upper limit of normal. The following troponin assays were used in this study: AccuTnI (Beckman Coulter, Brea, CA, USA), Elecsys Troponin T and Elecsys Troponin T high sensitive assays (both Roche Diagnostics, Basel, Switzerland) and TnI Ultra and Acute Care Troponin I (both Siemens Healthcare Diagnostics, Erlangen, Germany)., The SWEDEHEART registry, which includes all patients admitted to a coronary care unit because of acute coronary syndromes , was used to determine the total number of patients with myocardial infarction and the rate of normal coronary angiograms in the Stockholm metropolitan area during the study period.
CMR imaging protocol
The CMR imaging protocol included standard steady-state free precession (SSFP) cine imaging, T2-weighted oedema imaging and late gadolinium enhancement (LGE) for fibrosis detection; the protocol differed slightly between the different sites. The investigation was performed in the supine position with a cardiac coil using one of three 1.5 T systems (General Electric Healthcare; Signa Excite TwinSpeed, Waukesha, WI, USA; Siemens Sonata, Erlangen, Germany; and Philips Intera CV, Best, the Netherlands) during vector ECG monitoring. The imaging protocol included scout imaging, localization of the short axis and then acquiring retrospectively gated cine SSFP images covering the whole left ventricle (LV) 1.6–3.3 ms, repetition time (TR) 2.8–3.6 ms, flip angle 60°, 25 phases, 8 mm slice, no gap, matrix 160–226 × 141–226 and T2-weighted [triple inversion recovery, TE 60–80 ms, TR two R–R intervals, TI 150–170 ms, slice thickness 8 or 14 mm (14 mm with GE TwinSpeed), gap 8 mm, flip angle 90–180°, matrix 226–256 × 226–256] images were acquired in the same long- and short-axis planes. LGE images were acquired 15–20 min after contrast injection of intravenous gadolinium-DTPA (0.2 mmol kg−1) using a 2D or 3D (3D with Philips Intera CV) inversion recovery gradient echo sequence (TE 1.1–3.3 ms, TR 3.8–7.0 ms, inversion time 180–300 ms to null the signal of myocardium, 8 mm slice, no gap, matrix 240–256 × 180–192) in the same slice orientation as cine SSFP images. Each slice was obtained during end-expiratory breath holding. Two-, three- and four-chamber views were also obtained to confirm the findings.
CMR imaging analysis
All CMR images were analysed using offline, freely available segmentation software (segment v.1.8 r1405; http://segment.heiberg.se/). The CMR images were examined and interpreted by two independent experienced experts blinded to all clinical data and the imaging report from the investigating hospital, thereby minimizing interhospital variation. In case of disagreement, a third CMR imaging specialist was consulted for consensus.
End-diastolic and end-systolic volumes were measured in the phase with the largest and smallest LV volumes, respectively. LV ejection fraction, stroke volume and LV mass were calculated on cine SSFP sequences using manual delineation of the endocardial and epicardial borders including papillary muscles and trabeculations when contiguous with the LV. To calculate LV mass, the myocardial volume was multiplied by the density of myocardial tissue (1.05 g mL−1). All volumes were determined relative to body surface area.
T2-weighted images were visually examined to detect areas of high signal compatible with oedema. LGE images were assessed for subendocardial enhancement in the distribution of a coronary artery suggesting myocardial infarction, or midwall/subepicardial enhancement suggesting myocarditis. Patients with patchy involvement on LGE (intramyocardial, including both subepicardial and subendocardial) were considered to have myocarditis. Images showing normal volumes and function and with no LGE or T2-weighted abnormalities were considered to be ‘normal CMR images’.
spss version 20 (IBM, Armonk, New York, NY, USA) was used for statistical analysis. A P-value of <0.05 was considered significant. Kruskal–Wallis test was used for continuous nonparametric variables. If the P-value was <0.05 with the Kruskal–Wallis test, the Mann–Whitney U-test was used for pairwise comparison between groups. Fisher′s exact two-sided test was used for categorical variables.
According to the SWEDEHEART registry, there were 4801 patients with myocardial infarction as their primary diagnosis in the Stockholm metropolitan area during the study period from June 2007 to May 2011. During the same period, 4412 (92%) of these patients underwent in-hospital coronary angiography, 277 (6%) of whom were found to have normal arteries or minimal atheromatosis on coronary angiography. Of the 277 eligible patients, 176 (64%) were screened for the study; the remaining eligible patients were not screened because of missed opportunities by the treating physicians. Patient selection and study flow are shown in Fig. 1. Of the 176 patients screened for the study, 152 (86%) underwent CMR imaging; reasons for not imaging were claustrophobia (n =9), CMR unavailable (n =10) or others (n =5). The CMR imaging was performed a median of 12 days [interquartile range (IQR 6–28 days)] after initial presentation to hospital.
The baseline characteristics of screened patients and those who underwent CMR imaging are shown in Table 1. There were no significant differences in clinical characteristics between patients in these two groups. The median age of the patients was 58 years and the majority were women. Approximately, 50% of the patients were previous or current smokers. Almost one-third had a family history of coronary heart disease, but only a small proportion had diabetes mellitus. One-third of the screened patients had hypertension.
Values are given as number of patients (percentage), unless stated otherwise. Maximum troponin level (times greater than the upper limit of normal) and age are median values. All groups were first compared to determine any significant difference. *P =0.02 vs. group V; **P <0.01 vs. group V. Kruskal–Wallis test was used for continuous nonparametric variables. If the P-value was <0.05 for the Kruskal–Wallis test, Mann–Whitney U-test was used for pairwise comparison between the groups. Fisher′s exact two-sided test was used for categorical variables.
Of 152 patients who underwent CMR imaging, 11 (7%) had signs of myocarditis, 29 (19%) showed a typical pattern of myocardial necrosis and 102 (67%) were considered normal (i.e. no signs of myocarditis, myocardial necrosis or cardiomyopathy). The remaining 10 had other or unclear findings (as shown in Fig. 2). Amongst patients considered to have a normal CMR imaging result, 33 (32%) were found to have typical signs and symptoms of Takotsubo stress cardiomyopathy including reversible LV dysfunction. Of all patients screened with CMR imaging, 33 (22%) were found to have Takotsubo stress cardiomyopathy. In addition, there were a number of patients fulfilling some but not all of the criteria of this condition.
Baseline characteristics differed between patients with myocarditis, those with myocardial necrosis and those with normal CMR images (Table 1). Patients with normal images, more than two-thirds of whom were women, had the lowest maximum troponin levels. There was a difference in maximum troponin levels between the normal and the myocardial necrosis groups. Patients with myocarditis were younger and had a lower prevalence of treated hypertension.
Cardiovascular magnetic resonance imaging data are shown in Table 2. The median LV ejection fraction amongst all patients who underwent CMR imaging was 59%. Patients with a normal CMR images had a higher ejection fraction than patients with myocardial necrosis or myocarditis. Wall motion abnormalities and oedema were observed in the vast majority of patients with myocardial necrosis and in about half of those with myocarditis. All patients with myocardial necrosis and myocarditis showed LGE. The median time from admission to imaging was 13, 10 and 14 days for the normal, myocardial necrosis and myocarditis groups, respectively.
Table 2. Cardiovascular magnetic resonance (CMR) imaging data
In the present study, we were able to screen the whole Stockholm metropolitan area for a period of 4 years to obtain a reliable estimate of the incidence of MINCA. To date, this study included the largest cohort of patients with MINCA undergoing CMR imaging. We found that in patients with MINCA, CMR imaging can differentiate between myocarditis, myocardial necrosis and normal myocardium. Approximately, two-thirds of CMR images appeared completely normal, <10% showed signs of myocarditis and about 20% had signs of myocardial necrosis. Moreover, at least one in five patients had typical signs of Takotsubo stress cardiomyopathy. Finally, overall in this population of patients with acute myodardial infarction, MINCA is more common than previously reported.
The combination of CMR imaging and clinical evaluation using the Mayo criteria for Takotsubo stress cardiomyopathy made it possible to make a definite diagnosis in approximately 50% of patients. Furthermore, an initial diagnosis of acute coronary syndrome was changed in two of three patients. This is in accordance with two previous studies [3, 12] in which the initial diagnosis or therapy was modified after CMR imaging in 65% and 32% of cases, respectively. Thus, it seems appropriate to consider CMR imaging as a standard investigational tool in patients with MINCA.
Cardiovascular magnetic resonance imaging can differentiate between normal myocardium and minor areas of myocardial necrosis . However, the lower limit for detecting myocardial necrosis is unknown, as is the sensitivity to detect myocarditis. It is likely that when troponin levels are raised, just above the upper limit of normal, the diagnostic abilities of CMR imaging are challenged. This will be especially true for the new high-sensitivity troponin assays. Thus, in these patients, we might see false-negative results for LGE or myocardial oedema affecting the diagnosis of myocardial necrosis and myocarditis. The median time from admission to CMR imaging was 12 days. It is possible that CMR imaging patterns may normalize after the acute event, and therefore, the results may depend on the timing of imaging. However, in the present study, the mean number of days between admission and CMR imaging was the same for patients with normal results and for those with either myocardial necrosis or myocarditis, thereby excluding this as a possible source of bias. On the other hand, although a correct diagnosis on CMR is of importance, it is known that a normal CMR imaging result after a presumed myocardial infarction is a good prognostic indicator [14, 15].
We found a lower prevalence of myocarditis in comparison with previous studies. There are two possible reasons for this. First, our study population was older (mean age 57 years) in comparison with previous studies and it has been shown that myocarditis is more common in younger cohorts [3, 5, 12]. The age range in this study, 35–70 years, was chosen to minimize the number of patients with myocarditis and to ensure a high proportion of patients undergoing coronary angiography. In the previous studies by Assomull et al.  and Laraudogoitia Zaldumbide et al. , younger patients were included resulting in a lower mean age (44–48 years) and a higher prevalence of myocarditis. Gerbaud et al.  on the other hand included 130 patients with a mean age of 54 years, 26% of whom had myocarditis, supporting the validity of our results. Second, the majority of patients with MINCA (64%) from the Stockholm metropolitan area over a 4-year period were included in the present study, which was therefore more representative of all patients with this condition than previous studies. In addition, the present study represents the largest cohort of patients with MINCA investigated with CMR imaging, thus increasing the accuracy of the results.
Using another coronary angiography registry, it was previously reported that the incidence of MINCA is approximately 3% in patients with acute myocardial infarction . In the present study, the incidence of MINCA was found to be higher using the SWEDEHEART registry which includes all coronary care units in Stockholm and the rest of Sweden. Moreover, our results may in fact be an underestimation of the true incidence. In patients who are primarily diagnosed with Takotsubo stress cardiomyopathy, it is not likely that myocardial infarction will be recorded as the primary diagnosis in the registry. With inclusion of these patients, we believe that the true incidence of MINCA is in the region of 7% or even 8% of all patients with suspected acute myocardial infarction undergoing coronary angiography.
The data from the SWEDEHEART registry were calculated based on acute myocardial infarction as the primary diagnosis. Our finding of a high incidence of MINCA is even more notable as patients with secondary myocardial infarction caused by atrial fibrillation or supraventricular tachycardia where excluded. In previous studies, these patients were included [3, 12] or not adequately described . Atrial fibrillation or supraventricular tachycardia with a high ventricular rate is a known cause of elevated troponin levels and chest pain and might therefore dilute the number of patients in these studies.
The incidence of Takotsubo stress cardiomyopathy has previously been estimated to be approximately 2% in patients with acute coronary syndromes . The estimated incidence of Takotsubo stress cardiomyopathy in patients with MINCA, however, has been more diverse, ranging from only 1% to 22% [3, 5, 12]. In our study, the incidence of Takotsubo stress cardiomyopathy was 22% of all patients with MINCA. In addition, some patients with normal myocardium according to CMR imaging had symptoms of Takotsubo stress cardiomyopathy but did not fulfil the Mayo Clinic criteria for a diagnosis of this condition . For example, a patient may have experienced a stressful event before admission and have normal CMR imaging findings but without reversible ventricular dysfunction. Therefore, it is possible that a more complex pathology of stress-induced cardiomyopathy may exist, which would imply an even higher incidence of Takotsubo stress cardiomyopathy in patients with MINCA.
There was a risk of selection bias in our prospective study. The decision not to perform coronary angiography was made by the cardiologist at each hospital. Patients with a clinical diagnosis of myocarditis who did not undergo CMR imaging were not screened. As a consequence, some of these patients could potentially have been misdiagnosed. Almost 14% of the screened patients did not undergo CMR imaging, which could have led to a selection bias amongst those who did. However, baseline characteristics between these two groups of patients did not differ significantly.
The validity of the Mayo Clinic criteria used in this study have been questioned by some investigators ; however, at present, there is no consensus on how a new set of criteria should be defined.
Some but not all have suggested that measurement of the early gadolinium enhancement ratio (EGEr) could improve sensitivity and specificity for myocarditis . EGEr and relative signal enhancement on T2-weighted images (comparing images of T2-weighted oedema with those of skeletal muscle) were not used in this study, and therefore, the ability to detect myocarditis may be reduced.
Previous studies have indicated that late-enhancement patterns remain over long periods, even throughout the rest of a patient's life . Nevertheless, an early (within 2–3 days) CMR image showing oedema could have resulted in an increase in the number of patients with pathology [12, 21]. In the future, more sensitive oedema sequences with quantitative signal analysis in combination with EGEr and LGE may help to establish an early and more accurate diagnosis, further reducing the number of MINCA patients without a diagnosis.
In this consecutive series of patients with MINCA, CMR imaging could help to differentiate between myocarditis, myocardial necrosis and normal myocardium. On the basis of our findings, we propose that the incidence of MINCA is 7–8%. After excluding cases of myocarditis, MINCA consists of a large group of patients with normal CMR imaging results and a smaller group with myocardial necrosis. A refined CMR imaging protocol including better oedema sequences and evaluation could further increase the number of MINCA patients with a diagnosis. The aetiologies of these different CMR imaging findings need to be explored.
This study was performed in accordance with the Declaration of Helsinki and good clinical practice. The study was approved by the local ethics committee at the Karolinska Institutet.
Conflict of interest statement
CG is currently employed by Roche AB Sweden. The other authors have no conflicts of interest to declare.
This study was supported by the Swedish Heart-Lung Foundation.