Article Text

Download PDFPDF

Original research article
Cardiac magnetic resonance in patients with elevated troponin and normal coronary angiography
  1. Subir Bhatia1,
  2. Christopher Anstine1,
  3. Allan S Jaffe2,
  4. Bernard J Gersh2,
  5. Krishnaswamy Chandrasekaran2,
  6. Thomas A Foley2,3,
  7. David Hodge4,
  8. Nandan S Anavekar2,3
  1. 1 Department of Internal Medicine, Mayo Clinic, Rochester, Minnesota, USA
  2. 2 Department of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota, USA
  3. 3 Department of Radiology, Mayo Clinic, Rochester, Minnesota, USA
  4. 4 Department of Health Sciences Research, Mayo Clinic, Jacksonville, Florida, USA
  1. Correspondence to Dr Nandan S Anavekar; anavekar.nandan{at}


Background Invasive angiography in the setting of cardiac troponin elevation may reveal non-obstructive coronary arteries leading to uncertainty in diagnosis. Cardiac MR (CMR) may aid in diagnosis, however, the spectrum of diagnostic findings in the patient presenting with symptoms of cardiac ischaemia, elevated cardiac biomarkers and a negative invasive coronary angiogram is yet to be completely described.

Methods We queried the Mayo Clinic, Rochester inpatient record from 1 January 2000 to 31 December 2016 to identify patients who: (1) had an elevated troponin T during admission, (2) underwent coronary angiography within 30 days of troponin T elevation which was considered negative for obstructive coronary arterial disease and (3) underwent CMR within 30 days of troponin T elevation. CMR diagnoses were classified as either (1) myocarditis, (2) small area myocardial infarction, (3) stress cardiomyopathy, (4) non-ischaemic cardiomyopathy or (5) normal.

Results Of 215 patients, the spectrum of disease seen on CMR was myocarditis (32%), small area infarction (22%), non-ischaemic cardiomyopathy (20%) and stress cardiomyopathy (9.3%).

Conclusion In the largest single-centre study assessing the role of CMR in patients admitted with elevated troponin T with a non-obstructive coronary disease on an angiogram, small area infarction was seen in 22% of patients.

  • cardiac magnetic resonance (cmr) imaging
  • acute myocardial infarction

Statistics from


Troponin measurement is used in the diagnostic evaluation and risk stratification of patients presenting with symptoms compatible with acute coronary syndrome (ACS) infarction.1 Despite the use of higher sensitivity cardiac troponin (cTn) assays to evaluate for ACS, non-obstructive coronary artery disease (CAD) and/or angiographic coronary arteries in the absence of severe coronary stenoses are reported in nearly 10% of patients undergoing immediate cardiac catheterisation.2–4 This clinical situation can present a diagnostic dilemma as there are multiple aetiologies for acute myocardial infarction (AMI) in the absence of severe coronary atherosclerosis and other causes for troponin elevation with this type of acute presentation, such as myocarditis and Takotsubo’s cardiomyopathy. Regardless of the cause, these patients remain at higher risk for re-infarction or death, emphasising the need for accurate diagnosis and if possible disease-specific treatment.3 Cardiovascular magnetic resonance (CMR) imaging has emerged as a diagnostic adjunct in this setting. Prior studies have assessed the spectrum of disease using CMR of patients presenting with elevated troponin values found to have non-obstructive coronary disease. However, these studies have been limited by small numbers and prolonged time from troponin elevation to the performance of CMR.

The objective of this study was to determine the diagnostic yield of CMR in patients with troponin elevation who are subsequently found to have non-obstructive CAD or coronary arteries in the absence of severe coronary stenosis by invasive angiography.


Patient population

All adult (age ≥18) patients with cardiac troponin elevation, admitted for an inpatient evaluation, over the time period 1 January 2000 through 31 December 2016 were considered eligible (n=60 421). Further criteria for inclusion included (1) coronary angiography within 30 days of troponin T elevation which was considered negative for obstructive coronary arterial disease (angiogram negative) and (2) cardiac MRI within 30 days of troponin T elevation. Patients were identified by screening the medical record for an increased troponin T greater than the upper limit of normal. This list was then filtered to identify patients who underwent coronary angiography and cardiac MRI within 30 days of troponin T elevation. The resulting 215 charts were reviewed manually to identify the cardiac MRI diagnoses. Equivocal cases of cardiac MRI diagnoses were adjudicated by a cardiac radiologist who was blinded to the laboratory and clinical data. Patients who underwent any form of revascularisation therapy including lytic therapy, percutaneous coronary intervention or coronary artery bypass graft surgery during the hospitalisation were excluded from analyses.

Angiogram negative was defined by having non-obstructive epicardial CAD on invasive coronary angiography. Furthermore, the non-obstructive disease was defined by atheromatous lesions causing stenosis of diameter <50% as assessed by visual assessment reported by the interventional cardiologist. An increased troponin T was defined as having an elevated troponin T greater than the upper limit of the normal range with the fourth-generation assay (>0.01 ng/mL). Our institution utilised a sampling protocol for cTnT that included a zero-hour, a 3-hour and a 6-hour sample. This was completed in 100% of the patients at zero hour, 91% of patients at the 3-hour sample, and 91% of patients at the 6-hour sample. Fifteen patients had only one troponin value.

Cardiac MRI protocol

Patients were studied using a 1.5 T clinical CMRI General Electric scanner (GE scanner, with steady-state free-precession cine images acquired in three long-axis planes and a stack of short-axis images that encompassed the entire left ventricle [LV]). Gadolinium dimeglumine was then administered intravenously at a dose of 0.2 mmol/kg of body weight. Long-axis and short-axis LGE-CMRI images were acquired after a 10 min delay using an inversion recovery segmented gradient echo sequence. Typical voxel size was 23 mm3. A consultant cardiac radiologist reported all the images on dedicated workstations. LV end-diastolic and end-systolic volumes were generated using dedicated post-processing software and these were used to calculate the LV ejection fraction (EF).

CMR diagnoses were classified as either (1) myocarditis, (2) small area myocardial infarction, (3) stress cardiomyopathy, (4) non-ischaemic cardiomyopathy or (5) normal.


The baseline demographics of patients included in the study were assessed by manual review of individual patient charts on admission to the Mayo Clinic on the date of troponin elevation. Overall comparisons between groups were completed with a Fisher’s exact test for categorical variables. Continuous variables were compared between the five groups of interest using the analysis of variance.

Additionally, the EF of patients as determined by two-dimension transthoracic echocardiogram (TTE) was assessed at three time points: (1) initial hospital admission at the time of troponin elevation, (2) 3 months post-troponin elevation and (3) 1 year post-troponin elevation. 71%, 27% and 25% of the cohort had undergone TTE at initial hospitalisation, 3 months post-troponin elevation and 1 year post-troponin elevation, respectively.


Of 60 421 patients who were admitted to Mayo Clinic, Rochester with an elevated troponin T during inpatient admission, 18 480 patients underwent coronary angiogram within 30 days from the date of admission. Of the 7286 patients who were angiogram negative, 215 (2.9%) underwent CMR (figure 1). The baseline characteristics of patients stratified by cardiac MRI diagnoses are listed in table 1. Patients with a CMR diagnosis of ischaemic cardiomyopathy were more likely to be older, have hypertension, hyperlipidaemia and have undergone the previous percutaneous coronary intervention (PCI). Patients with a small size infarction seen on CMR were significantly older than those with myocarditis (62 vs 46 years, p<0.001). Additionally, those with a small infarction on CMR were more likely to have known cardiovascular comorbidities including hypertension, hyperlipidaemia and diabetes mellitus.

Figure 1

CABG, coronary artery bypass grafting; NICM, non-ischaemic cardiomyopathy; PCI, percutaneous coronary intervention; RWA, regional wall abnormalities; Stress CM, stress cardiomyopathy.

Table 1 also demonstrates the overall mean number of days from troponin elevation to cardiac catheterisation and CMR, 1.13 and 3.68 days, respectively. While patients with small areas of infarction on CMR were found to have underwent cardiac catheterisation and CMR faster compared with the overall study group (0.78 and 3.56 days, respectively), there was no statistical significance among the five CMR subgroups in time from troponin elevation to catheterisation or CMR (p=0.51 and 0.98, respectively). The mean EF at baseline, 3 months and 1 year from initial troponin T rise was 47%, 53% and 52% for the entire cohort, respectively. There was a significant difference in EF at baseline, 3 months and at 1 year among the patient subgroups; patients found to have a small area infarct had a similar EF at baseline but lower EF at 3 months and 1 year when compared with those with normal CMR. Table 1 also demonstrates that among patients with small areas of infarction on CMR, 89%, 60%, 51% and 87% of patients were discharged on beta-blocker, dual antiplatelet therapy, angiotensin-converting enzyme-inhibitor/angiotensin-receptor blocker or statin on discharge, respectively.

Table 1

Baseline characteristics of study population

Median troponin T values ±1SD by CMR diagnosis at baseline, 3 and 6 hours from baseline were 0.14, 0.27 and 0.28, respectively, for the entire cohort (figure 2). Patients with small areas of infarction on CMR had higher median troponin T values at each time point compared with the overall study group (0.16, 0.32 and 0.35, respectively) while patients found to have a normal CMR had the lowest troponin T values of any subgroup at baseline, 3 hours and 6 hours (0.13, 0.22 and 0.21, respectively). There was no statistical difference in the proportion of patients with a rise or fall of troponin measurements between CMR diagnoses. Figure 3 demonstrates the breakdown of CMR diagnoses based on the presence or absence of plaque on coronary angiography.

Figure 2

Median troponin T values ± 1 SD by CMR diagnosis at baseline, 3 hours, and 6 hours from baseline.

Figure 3

Breakdown of CMR diagnoses in two subgroups, (1) those with non-obstructive coronary artery disease (but still with plaque present) and (2) those with normal coronary arteries (smooth arteries with no plaque). CMR, cardiovascular magnetic resonance.

The most common clinical diagnoses were out of hospital arrest or presentation with ventricular tachycardia or ventricular fibrillation (n=9; 24%) and myocarditis (n=8; 22%). Four patients had paroxysmal supraventricular tachycardia (11%), specifically, atrial fibrillation (two patients), Wolff-Parkinson-White (one patient) and atrioventricular nodal re-entry (one patient). No aetiology could be found for the troponin T elevation in 10 patients. Three patients had respiratory failure, two patients had coronary vasospasm and one had acute congestive heart failure.


In the largest single-centre study assessing the role of cardiac MRI in patients admitted with increased values for cardiac troponin T, signs or symptoms warranting coronary catheterisation and non-obstructive cardiac disease, we found:

  1. 22% of patients had a small area of infarction as determined by CMR. This is significantly higher than seen in previous studies.

  2. In those patients identified with a small area of infarction, troponin elevations were modestly higher than in those with other aetiologies, but the overlap was substantial and inadequate to provide for the distinction of this group compared with others.

  3. As in other studies, myocarditis was the most common diagnosis although the incidence in our study is lower than in many others.

Cardiac troponin is the recommended biomarker test to assess for myocardial injury, but it is not synonymous with myocardial infarction or ischaemia. Troponin elevations are observed in patients with renal failure, those with myocarditis, sepsis, atrial fibrillation, structural heart disease and heart failure.5 However, it is often difficult to discern the aetiology for an elevated cTn value in patients presenting with symptoms potentially indicative of unstable CAD. As these patients benefit from an aggressive strategy including antiplatelet agents and early coronary angiography per guidelines, diagnostic coronary angiography is often warranted.6 7 However, the angiogram may reveal no obstructive coronary atherosclerosis, a distinct clinical syndrome termed myocardial infarction with no obstructive coronary atherosclerosis (MINOCA). Causes of MINOCA include both epicardial causes, such as CAD and coronary dissection, and microvascular causes including coronary spasm, Takotsubo syndrome, myocarditis, coronary embolism and thrombophilia. Numerous diagnostic tests are available to aid in elucidating the specific cause of MINOCA such as intracoronary ergonovine testing for suspicion of coronary artery spasm.8 Previous studies have identified cardiac MRI as a useful tool to identify both those situations due to acute ischaemia and alternative aetiologies that are non-ischaemic nature.9

Our study population is the largest in the literature. Previous study cohorts have ranged from 60 patients to 125 patients. In addition, our CMRs were completed closer to the time of presentation which may explain why we were able to identify a diagnosis in 83% of cases, which is in line with other studies where MRI was done soon after presentation.10 The most common diagnosis of CMR in our study was myocarditis (32%) in keeping with other reports. However, the second most common diagnosis were small areas of infarction based on the presence of focal late enhancement in a subendocardial distribution of a coronary artery. Figures 4 and 5 demonstrate the CMR and angiographic findings of two patients from our cohort. Our rate of infarction was significantly higher than that found in previous studies, which found the frequency of myocardial infarction to be between 5% and 16%.11–13 There are several potential reasons for this observation. First, the patient population in this investigation had a higher rate of comorbidities compared with patient populations studied in prior investigations. Specifically, the rate of known cardiac risk factors such as hypertension, hyperlipidaemia and diabetes in this study was 48%, 40% and 10%, which is comparable with that found in previous studies which ranged from 17% to 48% for hypertension, 15% to 37% for hyperlipidaemia and 3% to 12% for diabetes. Furthermore, the current patient population was older as the mean age of our patient cohort was 55 years compared with previous studies’ whose patients had an average age of 44–55 years.

Figure 4

(Panels A and B) Arrowhead demonstrating area of near transmural infarction. (Panels C and D) Still frames of coronary angiogram demonstrating widely patent epicardial coronary arteries.

Figure 5

(Panels A and B) Arrowhead demonstrating multifocal subepicardial areas of post-contrast delayed enhancement consistent with the clinical diagnosis of myocarditis. (Panels C and D) Still frames of coronary angiogram demonstrating widely patent epicardial coronary arteries.

Second, it is possible our study found a higher frequency of small infarctions based on CMR because the mean time from troponin elevation to cardiac MRI in our study was 3.6 days compared with the previous studies with mean times to CMR ranging from 7 days to 2 months.7 11–13 While our study used the time from initial troponin elevation to time of cardiac MRI, previous studies instead used index time defined as when a diagnosis of ‘ACS’ was made which may have led to more subjectivity between studies. Previous research has shown that LGE is more extensive in the first days after an AMI compared with the chronic phase where it is well known that the area of infarction often shrinks.14 In our study, the time from troponin elevation to CMR for the entire cohort was not only shorter than other studies, but those found to have a small area of infarction had a shorter time to CMR compared with the entire cohort, raising the possibility that the detection rate of small areas of infarction was possibly due to faster time to imaging.

In our study, rupture of non-obstructive plaque is a potential cause of small area infarct. Other proposed mechanisms of myocardial infarction in the absence of culprit CAD include coronary embolism, coronary artery spasm and distal embolisation.15 Patients with atrial fibrillation often have microvascular disease in addition to those with structural heart disease and heart failure. Thus, in addition to CMR, diagnostic interventions in the catheterisation lab such as a measure of coronary flow reserve and acetylcholine reactivity are helpful to aid in diagnosis in this subgroup.16 Functional testing may be helpful in patients with out of hospital arrest or presentation with ventricular tachycardia or ventricular fibrillation.

There are several limitations to our study. Although we found the diagnostic yield of CMR to be >80%, the study was performed in <3% of patients who had elevated troponins and were angiogram negative. This is a source of bias as the compelling factors to order versus withhold ordering a CMR are unknown. Furthermore, there may have been added suspicion of stress cardiomyopathy or myocarditis leading to the CMR being performed, therefore skewing the prevalence of diagnoses on CMR. Given the paucity of data among patients found to have ACS with non-obstructive arteries, recent guidelines propose aspirin, statins and, with evidence of vasospasm, calcium channel blockers as routine medical treatment among patients with no clear aetiology of troponin elevation on CMR as these medications would also benefit possibly underlying thromboembolism, coronary plaque disruption and coronary artery vasospasm.15 In our study, the use of secondary prevention among patients who had a small area infarction on CMR was not standardised and was dependent on management preference of the discharging physician leading to bias which cannot be accounted for. Finally, we were unable to provide an objective distinction between patients who had a type 1 or type 2 MI as this distinction was not consistently made by the treating provider during the admission.


Among admitted patients with elevated cardiac biomarkers who have non-obstructive coronary arterial disease, small areas of infarction were seen in 22% of cases.

Key messages

What is already known on this subject?

  • Non-obstructive coronary artery disease and/or angiographic normal coronary arteries are reported in nearly 10% of patients undergoing immediate cardiac catheterisation. Cardiovascular magnetic resonance (CMR) imaging has emerged as a potentially useful diagnostic tool in this setting. Few studies to date have assessed the CMR imaging characteristics and outcomes of patients presenting with cardiac ischaemia, elevated cardiac biomarkers and a ‘negative’ invasive coronary angiogram.

What might this study add?

  • Among patients with the acute coronary syndrome who have non-obstructive disease on coronary catheterisation, myocarditis was the most common cardiac MRI diagnosis (32%) and small area infarction was seen in 22% of patients.

How might this impact on clinical practice?

  • Cardiac MRI may be a beneficial diagnostic tool in patients with myocardial infarction with non-obstructive coronary arteries.



  • Contributors SB and CA: involved in data collection, study design and manuscript preparation. AJ, BG, KC and TF: involved in manuscript review. DH: involved in statistical support. NA: involved in study design and manuscript review.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Ethics approval The study received approval from the Mayo Clinic Institutional Review Board (IRB).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

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.

Linked Articles

  • Editorial
    Udo Sechtem Andreas Seitz Peter Ong