This longitudinal observational study included consecutive patients aged 18 years or older with a working diagnosis of MINOCA, as per the ESC definition2 who underwent CMR at the University Hospital Germans Trias i Pujol (Badalona, Spain). The inclusion period was between January 2009 and March 2022 and the patients were identified retrospectively from the institutional CMR database. Patients presenting with heart failure as the main clinical symptom or an electrocardiogram without sinus rhythm at admission were excluded. Other CMR-related exclusion criteria were patients with nonmagnetic resonance-conditional pacemakers or implantable cardioverter-defibrillators, claustrophobia, or severe renal impairment (estimated glomerular filtration rate <30mL/min/1.73 m2).

Myocardial injury was defined as levels of troponin I ≥ 0.5 ng/mL and high-sensitivity troponin I (hsTnI) ≥ 30 pg/mL. A typical rise and fall in levels in serial assessments of these biomarkers were required to define MI.10 During the inclusion period, the assay used to analyze the levels of troponin changed. Therefore, the maximal ratio of troponin calculated as the maximal troponin value at admission divided by the upper limit of the normal value for the troponin assay used is reported. Nonobstructed coronary arteries were defined as <50% stenosis in any epicardial coronary artery,1 either on invasive coronary angiography or on coronary computed tomography angiography.

The study was performed in accordance with the Declaration of Helsinki and was reviewed and approved by the Institutional Review Board (ethics code: PI-22-048 and PI-19-257). Due to the retrospective analysis of clinically acquired data, the local ethics committee waived the need for patient written informed consent.

Patient-related data, including baseline demographics, cardiovascular risk factors and comorbidities as well as electrocardiogram and biomarkers throughout admission were obtained from medical reports. Obesity was defined as a body mass index ≥ 30kg/m2.11 Dyslipidemia was defined as total cholesterol >250mg/dL and/or low-density lipoprotein-cholesterol >130mg/dL12 and diabetes as an HbA1c >6.5% and/or fasting plasma glucose >126mg/dL.13 Previous coronary artery disease included prior history of acute MI and/or percutaneous coronary intervention or coronary artery bypass graft surgery. Renal disease was defined as a reduced glomerular filtration rate (< 90mL/min/1.73 m2) using the CKD-EPI formula.14 Psychiatric disorders included depressive syndrome, anxiety, and personality disorders. Previous use of aspirin and other antiplatelet therapies, anticoagulation treatment, beta-blockers, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers was reviewed.

CMR imaging was performed on a 1.5 Tesla magnetic resonance imaging scanner (Achieva, Philips, The Netherlands). Time to CMR was defined as the number of days from admission with MINOCA to CMR study. The CMR data acquisition protocol included functional cine imaging using breath-hold balanced steady-state free precession sequences in the 3 long-axis views and in short-axis orientation covering both ventricles from the base to the apex of the left ventricle (8-mm slice thickness and 2-mm slice gap). T2-weighted short-tau inversion-recovery (T2-STIR) images were used to detect myocardial edema and were acquired in the same long- and short-axis views as cine sequences. After 2 to 3minutes of intravenous administration of gadolinium-based contrast agent gadobutrol [0.15 mmol/kg; Gadovist (Bayer-Schering Pharma, Germany)], early gadolinium enhancement images were acquired in the 3 long-axis views to evaluate the presence of hyperemia, microvascular obstruction, or intraventricular thrombus. To assess myocardial necrosis/fibrosis, late gadolinium enhancement (LGE) images acquired 8 to 10minutes after contrast administration with inversion-recovery segmented gradient echo sequences in the same long- and short-axis views as cine sequences were used. Inversion time was adjusted to null healthy myocardium over time. The same standardized protocol was maintained throughout the inclusion period with no significant changes in the sequences performed (supplementary data).

Quantification of ventricular volumes and ejection fraction of both ventricles and left ventricular (LV) mass was performed using the short-axis cine stack as described previously, using the Portal IntelliSpace software (Philips Healthcare, The Netherlands).15 Ventricular volumes and LV mass were indexed using the body surface area calculated with the Dubois formula.16

The presence and distribution of regional and global LV wall motion abnormalities using the 17-segment model of the American Heart Association were evaluated.17 In addition, the distribution of myocardial edema identified as high-intensity signal areas in T2-STIR images was analyzed. Finally, the pattern and extent of myocardial LGE (subendocardial, mid-wall, or subepicardial) were evaluated, allowing the classification of the patients into 5 CMR-based diagnostic groups: acute myocarditis, acute MI, tako-tsubo cardiomyopathy, cardiomyopathy group, or normal CMR. Acute myocarditis was defined as the presence of subepicardial and/or transmural edema with patchy mid-wall or subepicardial LGE with or without associated wall motion abnormalities, based on the Lake Louise criteria.18 Acute MI was defined by the presence of regional LV wall motion abnormalities with edema and subendocardial to transmural LGE pattern in the territory supplied by a coronary artery. Tako-tsubo cardiomyopathy was diagnosed by the presence of apical and/or mid-ventricular wall motion abnormality in cine sequences beyond an epicardial vascular distribution with transmural myocardial edema in the same segments and in the absence of significant LGE.19 The cardiomyopathy group included diagnoses of dilated cardiomyopathy, hypertrophic cardiomyopathy, sarcoidosis, or arrhythmogenic right ventricular cardiomyopathy based on CMR criteria.20–22 CMR was considered normal if there were no wall motion abnormalities, myocardial edema or specifical LGE patterns and no other structural or functional abnormalities (figure 1).

Figure 1.

The first row shows end-systolic SSFP-cine images in 2-chamber view for each diagnostic. The second row shows T2-weighted images for edema detection (arrows) and third row, T1-weighted images after gadolinium administration to study LGE. In the myocarditis case, there is linear subepicardial edema and LGE in the anterior wall and patchy intramyocardial in the mid-basal inferior wall. The acute myocardial infarction case shows transmural edema in the mid-inferior wall with subendocardial to transmural LGE in the same segment. In the tako-tsubo cardiomyopathy case, there is akinesia and transmural edema in mid-apical segments without LGE. No abnormalities are seen in the normal case. CMR, cardiovascular magnetic resonance; LGE, late gadolinium enhancement; MI, myocardial infarction.

(0.61MB).

Normal distribution of continuous variables was assessed with the Shapiro-Wilk test. Continuous variables with a normal distribution are presented as mean±standard deviation while nonnormally distributed variables are presented as the median [interquartile range]. Categorical variables are presented as frequencies and percentages. The Student t test or variance analysis were used for comparison of continuous variables. The Fisher or Pearson chi-square tests were used for comparison of dichotomous and categorical variables between groups. Stata software package (version 15.1, StataCorp, College Station, United States) was used for statistical analysis.

Clinical characteristics from the overall cohort and for each diagnostic group are presented in table 1. The mean age of the population was 50±18 years and there was a predominance of men (60%). Almost 90% of patients had an abnormal electrocardiogram at presentation with ST-segment elevation as the most usual finding (118 patients [57%]). Concomitant heart failure symptoms at presentation were observed in 14 patients (7%), while there were no patients with pulmonary edema or cardiogenic shock. Male sex was more frequent in the myocarditis and the normal CMR groups (76% and 71%, respectively) whereas female sex was more frequently observed in the tako-tsubo cardiomyopathy group (79%). Age was youngest in the myocarditis group followed by the MI group, and was oldest in the group of tako-tsubo cardiomyopathy.

CMR data according to the diagnostic groups are presented in table 2. Mean left ventricular ejection fraction as well as LV volumes were within the normal values for the overall population. However, when comparing the diagnostic groups, LV ejection fraction was significantly lower and the indexed LV end-diastolic volume significantly larger in the cardiomyopathy group compared with the other groups. Overall, 126 patients (62%) had LV regional wall motion abnormalities. The prevalence of myocardial edema in the whole cohort was 75% based on T2-STIR images, with focal myocardial edema absent in patients with a normal CMR (as per definition). Patients with tako-tsubo cardiomyopathy showed the largest number of LV segments with edema. LGE was observed in 149 patients (72%) and the myocarditis group showed the highest frequency of LGE (94%) followed by patients in the cardiomyopathy group (77%) and in the MI group (74%).

Based on a median time from diagnosis to CMR of 7 days, the cohort was divided in 2 groups: early CMR ( ≤ 7 days, 117 patients) and late CMR group ( >7 days, 88 patients) (figure 3).

Figure 3.

Study design. CMR, cardiovascular magnetic resonance.

(0.06MB).

Age was the only baseline characteristic that showed significant differences between the early vs late CMR groups (47±18 years vs 54±19 years; P=.01). When we compared CMR findings between the 2 groups, the presence of edema in T2-STIR images was more frequent in the patients who underwent early CMR than in those who underwent late CMR (81% vs 67%; P=.03). In contrast, there were no differences in LV regional wall motion abnormalities and the presence of LGE between early and late CMR. LV ejection fraction was slightly lower in the group of patients who underwent early CMR than in those undergoing late CMR (57% vs 60%; P=.03) (table 3).

The diagnostic yield of CMR was higher when CMR was performed within the first week (early group) (96% vs 86%; P=.02). Even though myocarditis remained the most common diagnosis in both the early and late CMR groups, this diagnosis was more frequently achieved in the early CMR group (53% vs 35%; P=.01). No differences were observed between the groups of MI or tako-tsubo cardiomyopathy. On the other hand, a normal CMR was more frequently observed in the late CMR group (14% vs 4%; P=.02) (figure 4).

Figure 4.

Central illustration. CMR diagnostic distribution in early vs late CMR groups. CMR, cardiovascular magnetic resonance.

(0.13MB).

Several limitations should be acknowledged in the present study. This is a single center study with a relatively small number of patients in each diagnostic group, which may underpower the conclusions. Prospective inclusion, but retrospective analysis of patients, may introduce referral bias due to physician decision and logistic considerations for CMR referrals. In 57 (27%) patients, the CMR was performed before coronary angiography, leading to a diagnosis of acute myocarditis in 54 patients. Accordingly, a coronary angiogram was not performed. The remaining 3 patients (1 patient with MI, 1 with tako-tsubo cardiomyopathy, and 1 with cardiomyopathy) had nonobstructive lesions on the coronary angiogram performed after the CMR. Due to the long inclusion period, newer CMR techniques that may increase the diagnostic yield, such as T1 and T2 mapping,18 were not available in all patients.