Clin Res Cardiol DOI 10.1007/s00392-014-0770-7
ORIGINAL PAPER
Diagnostic value of CMR in young patients with clinically suspected acute myocarditis is determined by cardiac enzymes Anca Florian • Tim Scha¨ufele • Anna Ludwig • Sabine Ro¨sch • Ina Wenzelburger • Handan Yildiz Udo Sechtem • Ali Yilmaz
•
Received: 4 February 2014 / Accepted: 1 October 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Aims Cardiovascular magnetic resonance (CMR) has become a valuable diagnostic tool for non-invasive diagnosis of acute myocarditis. However, since CMR studies are time- and cost-intensive and its diagnostic accuracy still not perfect, additional parameters are warranted to preselect and identify those individuals in whom a CMR study is likely to add crucial information regarding correct and timely diagnosis of acute myocarditis. The diagnostic value of CMR was evaluated in a population of young patients with clinically suspected acute myocarditis in relation to ECG and serum cardiac enzyme findings. Methods and results Only young patients aged B40 years in whom acute myocarditis was highly suspected based on their clinical symptoms, resting ECG findings and/or levels of cardiac enzymes (at presentation) were included to this study. After ruling out obstructive coronary artery disease, a multi-parametric CMR study was performed as part of the diagnostic work-up. The CMR protocol comprised cine sequences, T2-weighted edema imaging and late gadolinium enhancement (LGE) imaging on a 1.5-T MR scanner. 89 patients (28 ± 7 years, 89 % male) were included to this study presenting with symptoms of chest pain (85 %), dyspnea (26 %), fatigue (23 %) and/or palpitations (18 %). Pathological ECG changes were present in 72 patients
A. Florian A. Yilmaz (&) Department of Cardiology and Angiology, University Hospital Mu¨nster, Albert-Schweitzer-Campus 1, Building A1, 48149 Mu¨nster, Germany e-mail: [email protected] T. Scha¨ufele A. Ludwig S. Ro¨sch I. Wenzelburger H. Yildiz U. Sechtem A. Yilmaz Division of Cardiology, Robert-Bosch-Krankenhaus, Stuttgart, Germany
(81 %). An elevated serum troponin level was measured in 45 patients (51 %). Pathological CMR findings (presence of edema and/or LGE) were detected in 35 patients (39 %). In detail, pathological CMR findings were detected in 36 % of patients with resting ECG changes and in 73 % of patients with troponin rise. In contrast, normal CMR results were obtained in 95 % of patients with negative troponin at presentation, but only in 41 % of patients with normal ECG. On multivariable analysis, a positive serum troponin was the only independent predictor for a pathological CMR finding (OR = 33.26, 95 % CI = 3.04–363.35, p = 0.004). Conclusions The clinical use of non-invasive CMR in the work-up of clinically suspected ‘‘acute’’ myocarditis in young patients is only helpful and appropriately indicated in those ones with elevated cardiac enzymes. A preselection of such patients for CMR based on serum cardiac enzymes—but not on ECG recordings—may prevent a meaningless overuse of CMR. Keywords
Myocarditis Troponin CMR LGE ECG
Introduction Current recommendations in acute myocarditis suggest that the diagnosis should be based on an integrated approach including history, clinical assessment and a range of noninvasive tests comprising ECG, serum biomarkers and noninvasive cardiac imaging [1]. Nevertheless, a definite diagnosis remains a difficult challenge in a large proportion of patients (in particular in those with mild forms), since clinical presentation is multifaceted and ECG-changes as well as biomarkers tend to be non-specific [1–4]. Cardiovascular magnetic resonance (CMR) imaging with its arsenal of different sequences allowing non-invasive
123
Clin Res Cardiol
myocardial tissue characterization has evolved as a valuable tool for the diagnosis of myocarditis. T2-weighted imaging can detect tissue edema while T1-weighted imaging before and early (within the first minutes) after gadolinium contrast administration may detect myocardial hyperemia and capillary leakage, both hallmarks of acute myocardial inflammation [1, 5, 6]. In addition, late gadolinium enhancement (LGE) imaging by T1-weighted inversion recovery sequences accurately depicts myocardial damage and allows the differentiation of ischemic from non-ischemic injury with either subepicardial or midwall patterns of distribution commonly described in myocarditis [7–9]. At present, the recommended CMR diagnostic criteria for active myocarditis (‘‘Lake Louise Criteria’’) support the combined use of all the three above-mentioned sequences—but are based on a limited number of trials without large-scale multicenter data. Moreover, it is acknowledged that despite the good accuracy of the ‘‘any-two’’ out of three criteria approach, certain clinical settings might require even higher test sensitivities [1]. In the past decade, the number of non-invasive cardiac imaging studies has been intensively growing with a nonneglectable increase in associated costs. Regarding both increasing costs and limited health-care budgets, the use of any imaging modality including CMR requires justification to both its clinical diagnostic value and its costeffectiveness [10]. Hence, additional parameters are warranted to easily and successfully pre-select and identify those individuals in whom a CMR study is likely to add crucial information regarding appropriate diagnosis of, e.g., acute myocarditis. Therefore, the aim of this study was to evaluate the diagnostic performance of CMR in a population of young patients (B40 years) with a high clinical suspicion of acute myocarditis in relation to the presence or absence of (a) ECG changes and (b) elevated levels of cardiac enzymes as additional markers of acute myocardial injury.
Methods Study population Between 2009 and 2012, only young patients aged B40 years in whom acute myocarditis was highly suspected based on their clinical symptoms, resting ECG findings and/or levels of cardiac enzymes (at presentation) were included to this study. Patients aged [40 years were excluded to minimize confounding CMR findings possibly caused by other cardiovascular diseases apart from myocarditis. To be included into this study, patients had to fulfil all of the following criteria: (1) new onset (\14 days) of at least
123
one of the following symptoms: chest pain, dyspnea, unexplained fatigue and/or palpitations; (2) history of recent (\8 weeks) flu-like symptoms; (3) evidence for recent or ongoing myocardial injury by any of the two: ECG abnormalities (as detailed below) or elevated cardiac troponin; (4) obstructive coronary artery disease (CAD) ruled out by invasive coronary angiography or by clinical assessment. Patients with a prior diagnosis of pericarditis/ myocarditis, previously known significant valvular, myocardial or CAD, as well as patients with cardiovascular instability at presentation were excluded. Moreover, patients with rather unspecific non-cardiac chest pain and/ or only minor and quickly disappearing symptoms were excluded. Only patients with a high clinical likelihood of acute myocarditis were included to this study. Patients with the combination of normal ECG and negative enzymes were not included in this study due to the low likelihood of myocarditis. Moreover, in case of logistic limitations regarding the availability of CMR studies, patients with a troponin rise and normal ECG had a higher priority in getting a CMR study (and consequently being included to this study) than patients with normal troponin but abnormal ECG. From the 195 patients initially screened, 91 patients underwent a CMR study as part of the diagnostic work-up. In the remaining 104 patients, CMR was not possible due to logistic reasons, contraindicated or not consented. In two patients, CMR data were not assessable due to poor image quality. The final analysis included 89 patients. ECG analysis Standard 12-lead ECG tracings recorded at hospital presentation were interpreted by two cardiologists. The following new or persistent abnormalities were considered positive signs of myocardial injury: [1] conduction abnormalities; [2] ST-segment elevation (C0.1 mV) or depression (C0.05 mV) in two or more contiguous leads; [3] isolated inverted T-waves (C0.1 mV, negative in any leads except III, aVR and V1). In addition, the presence of arrhythmia and pathologic Q-waves, as well as heart rate and PR-, QRS-, QT-interval duration were noted. ECG analysis was detailed elsewhere [11]. Patients showing both ST–segment elevation and depression were classified as having ST-segment elevation. The presence of an early repolarization pattern in precordial leads was not considered as positive ECG sign after agreement between readers. Cardiac magnetic resonance: data acquisition and analysis CMR imaging was performed in breath-hold on a 1.5-T unit (Aera, Siemens Medical Solutions, Erlangen,
Clin Res Cardiol
Germany) using commercially available cardiac software, electrocardiographic triggering, and cardiac-dedicated surface coils. CMR included steady-state-free-precession cine imaging, T2-weighted STIR ‘‘edema’’ imaging and LGE imaging 5–10 min after intravenous contrast administration (0.15 mmol/kg MagnevistÒ) as previously described [12]. Image analysis was performed off-line by two experienced readers blinded to patient data. Ventricular volumes, ejection fraction and left ventricular mass were derived by contouring endo- and epicardial borders on the short-axis cine images. The presence of regional wall motion abnormalities (WMA) was visually evaluated both on the shortand long-axis cines using the standard AHA 17- segment model. In the T2-weighted STIR images, regional myocardial edema was visually identified while global edema was quantitatively assessed as signal intensity of the myocardium normalized to that of the skeletal muscle greater than 1.9 [6]. Poor quality T2-weighted STIR images were excluded from the analysis after consensus between readers. The diagnosis of positive LGE required visual identification of bright, hyper-enhanced signal within the myocardium in two orthogonal views. Using the same 17-segment model, all segments were dichotomously classified with respect to LGE presence. Positive LGE segments were further characterized according to myocardial damage localization as subendocardial, midwall, subepicardial or transmural. In addition, pericardial enhancement was noted, if present. Pericardial effusion was evaluated on cine images, as present at end diastole. Serum biomarker assessment Serum biomarkers of myocardial injury [creatine kinase MB (CK-MB) and cardiac troponin] were measured and considered elevated if at least one of the determinations on the day of admission exceeded the upper laboratory reference limit. Our hospital laboratory measured troponin T until 2011 and switched to troponin I measurements in 2012. C-reactive protein (CRP) was considered elevated if serum levels at admission exceeded the upper laboratory reference limit. Statistical analysis Continuous variables were expressed as mean ± SD. Skewed variables were expressed as median and interquartile range. Categorical variables were expressed as frequency with percentage t Student test was used for the between group comparison according to ECG and troponin status. Levene’s test was used for testing equality of variances. Mann-Withney U test was used for not normally distributed variables. The Chi-square test and Fischer’s
exact test were used to compare categorical variables expressed as proportions. To find independent predictors for a diagnostic CMR, first, univariable analysis was performed. Second, the parameters with significant p-values were introduced in the multivariable analysis. Statistical analysis was performed using SPSS software for Windows (version 18, SPSS, Chicago Illinois, US). All tests were 2-sided and a p value below 0.05 was considered statistically significant.
Results Patient characteristics and clinical presentation The study population of 89 patients comprised predominantly males (89 %) with a mean age of 28 ± 7 years. The distribution of different cardiovascular risk factors is shown in Table 1. 20 patients (22 %) had fever ([38.5 °C) at hospital admission. All patients were stable and 14 (16 %) showed tachycardia ([100/min) at presentation. ECG findings As shown in Table 2, 72 patients (81 %) had at least one pathological ECG sign as previously defined. The majority of patients (97 %) were in sinus rhythm with only three (3 %) showing atrial fibrillation. Conduction abnormalities were present as follows: two patients had sinuatrial block, one patient AV block II°, two patients AV block III° and Table 1 Patient characteristics All patients (n = 89) Male, n (%) Age (years)
79 (89) 28 ± 7
Cardiovascular risk factors, n (%) Family history of CAD
12 (13)
Arterial hypertension Diabetes mellitus
4 (5) 0 (0)
High cholesterol
6 (7)
Current smoker
36 (40)
Obesity
6 (7)
Symptoms, n (%) Chest pain
76 (85)
Dyspnea
23 (26)
Fatigue
20 (23)
Palpitations
18 (20)
Clinical parameters on admission HR (bpm) BP systolic/diastolic (mmHg)
79 ± 18 129 ± 19/78 ± 13
CAD coronary artery disease, HR heart rate, BP blood pressure
123
Clin Res Cardiol
one patient newly diagnosed left bundle branch block. Additional ST/T changes were present in two patients with sinuatrial block and AV block II°, respectively. Regarding ST/T abnormalities, 49 patients (55 %) showed ST–segment elevation, 13 (15 %) had ST-segment depression and eight (9 %) had isolated negative T-waves. Pathological Q-waves could be depicted in three patients (3 %). In six patients (7 %) runs of non-sustained ventricular tachycardia were recorded upon admission. Laboratory findings Overall, 45 patients (51 %) had elevated troponin levels and 39 (44 %) elevated CK–MB levels. As shown in Fig. 1, patients with chest pain were more likely to have Table 2 ECG and laboratory findings
elevated troponin compared to patients without such a complaint (55 vs. 23 %, p = 0.039). In contrast, the presence of fatigue or palpitations was associated more often with a negative troponin (32 vs. 13 % for fatigue, p = 0.045; 36 vs. 4 % for palpitations, p \ 0.0001). 28 patients (31 %) had both elevated troponin and ECG abnormalities, of whom 22 (25 %) had ST-segment elevation, two (2 %) had ST-segment depression and 5 (6 %) had isolated negative T-waves. None of the patients presenting with conduction abnormalities only had elevated troponin. No significant differences were found regarding the presence of abnormal ECG changes between troponinpositive and -negative patients. 61 patients (69 %) had elevated CRP levels on admission. Elevated CRP was more frequently associated with a positive troponin (p \ 0.0001). CMR findings
All patients (n = 89) ECG on admission, n (%) Pathological ECG Conduction abnormalities
72 (81) 6 (7)
ST-segment elevation
49 (55)
ST-segment depression
13 (15)
Isolated T-wave inversion
8 (9)
Pathological Q-waves
3 (3)
Atrial fibrillation
3 (3)
Laboratory work up Elevated troponin, n (%)
45 (51)
Elevated CK-MB, n (%) Elevated CRP, n (%)
39 (44) 61 (69)
Coronary angiography performed, n (%)
CMR was performed at a median of 2 (IQR 1–3) days after admission. The mean values for left ventricular (LV) volumes, ejection fraction and mass are illustrated in Tables 3 and 4. 16 (18 %) patients had an LV ejection fraction \60 %. Three patients with reduced LV ejection fraction demonstrated associated regional WMA. Regional WMA were found in 9 % of the cases (n = 8) and were all associated with positive LGE in the respective segments. 35 (39 %) patients showed positive LGE with a median extent of 4 (IQR 3–6) segments with the following Table 3 CMR results in relation to ECG
30 (34)
ECG electrocardiogram, CK creatine kinase, CRP C reactive protein Admission to CMR, days
Fig. 1 Bar graphs showing the prevalence of symptoms in patients with vs. without troponin elevation
123
Total (n = 89)
ECG positive (n = 72)
ECG negative (n = 17)
p value
2 (1–3)
2 (1–3)
2 (1–3)
0.61
LV-EDV (mL)
139 ± 26
137 ± 26
146 ± 27
0.21
LV-ESV (mL)
52 ± 14
52 ± 14
53 ± 11
0.66
LV-EF (%)
63 ± 7
63 ± 7
64 ± 5
0.58
LV-EF \60 %, n (%)
16 (18)
14 (19)
2 (12)
0.73
Regional WMA, n (%) Edema, n (%)
8 (9)
6 (8)
2 (12)
0.65
(A) Assessable
52 (58)
42 (58)
10 (59)
1.00
(B) Positive
0.015
13 (15)
7 (10)
6 (35)
LGE positive, n (%)
35 (39)
26 (36)
10 (59)
0.10
Edema and/or LGE positive, n (%)
35 (39)
26 (36)
10 (59)
0.10
Pericardial effusion, n (%)
13 (15)
11 (15)
2 (12)
1.00
CMR cardiac magnetic resonance, LV left ventricular, EDV left ventricular end diastolic volume, ESV end systolic volume, EF ejection fraction, WMA wall motion abnormalities, LGE late gadolinium enhancement
Clin Res Cardiol Table 4 CMR results in relation to troponin
Table 5 CMR results in relation to ECG and troponin
Total (n = 89)
Trop. positive (n = 45)
Trop. negative (n = 44)
Admission to CMR, days
2 (1–3)
1 (1–2.5)
2.5 (1.25–4)
LV-EDV (mL)
139 ± 26
142 ± 25
135 ± 27
0.21
LV-ESV (mL)
52 ± 14
56 ± 14
48 ± 12
0.005
65 ± 6
0.004
LV-EF (%)
63 ± 7
61 ± 7
p value
0.002
LV-EF \60 %, n (%)
9 (10)
8 (18)
1 (2)
0.030
Regional WMA, n (%) Edema, n (%)
8 (9)
8 (18)
0 (0)
0.006
(A) Assessable
52 (58)
25 (56)
27 (61)
(B) Positive
13 (15)
13 (29)
0 (0)
\0.001
0.67
LGE positive, n (%)
35 (39)
33 (73)
2 (5)
\0.001
Edema and/or LGE positive, n (%)
35 (39)
33 (73)
2 (5)
\0.001
Pericardial effusion, n (%)
13 (15)
9 (20)
4 (9)
0.23
CMR cardiac magnetic resonance, LV left ventricular, EDV left ventricular end diastolic volume, ESV end systolic volume, EF ejection fraction, WMA wall motion abnormalities, LGE late gadolinium enhancement
patterns: 18 (51 %) subepicardial, 6 (17 %) midwall, 1 (3 %) transmural and 10 (29 %) mixed (subepicardial and midwall). An ischemic pattern of LGE was depicted in none of the patients. LGE was more frequently present in patients with impaired LV function (69 %) than in patients with normal ejection fraction (34 %; p = 0.022). T2weighted edema images were assessable in 52 patients and presence of edema was documented in 13 (25 %) of them. All patients with proof of edema also showed related presence of LGE. However, ten patients (26 %) without edema had nevertheless evidence for LGE. Taken together, 35 patients (39 %) demonstrated presence of edema and/or LGE. A mild to moderate pericardial effusion was observed in 13 (15 %) patients. A moderate to strong pericardial enhancement could be depicted in two patients. CMR findings according to ECG abnormalities Out of 72 patients with ECG abnormalities at presentation, only 26 (36 %) exhibited at least one CMR positive finding suggestive of myocarditis. Notably, all of these 26 patients had positive LGE. Global and regional systolic dysfunction was depicted in 14 (19 %) and six (8 %) patients with pathological ECG, respectively. ECG changes in LGEpositive patients (n = 26) were as follows: 20 (77 %) had
Number (%)
ECG negative trop. positive (n = 17)
ECG positive trop. positive (n = 28)
ECG positive trop. negative (n = 44)
CMR neg.
CMR pos.
CMR neg.
CMR pos.
CMR neg.
CMR pos.
7 (41)
10 (59)
4 (14)
24 (86)
42 (95)
2 (5)
CMR cardiac magnetic resonance
ST-segment elevation, three (12 %) ST-segment depression and two (8 %) conduction abnormalities, of which one demonstrated both sinuatrial block and ST-segment elevation. The second patient with AV block III° exhibited only limited septal midwall LGE involving one segment. Table 3 shows CMR findings according to ECG status. No significant differences for LV dimensions and function as well as for global or regional impairment were noted. CMR findings according to troponin status Out of 45 patients with elevated troponin at admission, 33 (73 %) demonstrated at least one pathological CMR finding, with all 33 patients showing positive LGE. Tables 4 and 5 show CMR findings according to the troponin status. Troponin-positive patients had a significantly shorter admission-to-CMR time compared to troponin-negative ones. Mean LV end systolic volume was significantly higher and mean LV ejection fraction significantly lower in patients with troponin elevation. Only one patient without troponin rise had an impaired LV ejection fraction which was mild (50 %). None of the troponin-negative patients showed WMA or (when assessable) edema. 31 of 33 patients with positive LGE had increased troponin levels. The two patients without increased troponin had limited LGE involving one and two segments, respectively, and did not show any edema (that was assessable and negative in both). There was no significant difference in pericardial effusion presence between the patients with and without troponin rise. CMR was able to verify myocarditis (defined as presence of edema and/or LGE) in 74 % (n = 31) of patients with chest pain and elevated troponin and in 81 % (n = 17) of patients with chest pain, ECG abnormalities and elevated troponin. Predictors for a diagnostic CMR Table 6 shows the results of the regression analysis regarding independent predictors for a diagnostic CMR. Although in univariable analysis, several variables were associated with a pathological CMR result, at multivariable
123
Clin Res Cardiol Table 6 Predictors of a positive edema and/or LGE CMR study
CMR cardiac magnetic resonance, LV left ventricular, EDV left ventricular end diastolic volume, ESV end systolic volume, EF ejection fraction, WMA wall motion abnormalities, LGE late gadolinium enhancement, CK creatine kinase, CRP C reactive protein
(n = 89)
Univariable analysis OR (95 % CI)
p value
Age
0.989 (0.928–1.054)
Male gender
6.800 (0.822–56.281)
0.08
Admission to CMR, days
0.688 (0.506–0.935)
0.017
Chest pain
1.550 (0.438–5.484)
0.50
Dyspnea
0.333 (0.111–1.004)
0.05
Pericardial effusion
2.904 (0.864–9.758)
0.09
LV-EDV (mL)
1.004 (0.987–1.021)
0.67
1.036 (1.001–1.072)
0.044
LV-EF (%)
0.914 (0.849–0.983)
0.016
Positive ECG
0.396 (0.134–1.164)
0.09
ST-segment elevation Elevated troponin
1.034 (0.442–2.423) 57.750 (12.077–276.155)
0.94 \0.001
Elevated CK-MB
12.732 (4.036–40.165)
\0.001
6.786 (2.016–22.844)
0.002
Elevated CRP
Discussion The main finding of the present study that included young, previously healthy individuals with a high clinical likelihood of acute myocarditis (in whom CAD was excluded or unlikely) is that troponin elevation at presentation was the only independent predictor for a pathological CMR finding—regardless of symptoms and ECG status. In particular, a CMR study added almost no diagnostic information in troponin-negative patients—despite a high clinical likelihood of myocarditis. Based on the results of this study, we suggest that at least in young patients (aged B400 years) a pre-selection for CMR imaging based on the cardiac troponin level may prevent a meaningless overuse of CMR. The selection of patients for this study was somewhat biased due to the inclusion criteria emphasizing the presence of clinical symptoms such as chest pain, ECG changes and (above all) elevated cardiac enzymes. However, this approach reflects our daily clinical practice and allows a successful identification of those patients who are most likely suffering from severe myocarditis—from a clinical point of view. Obviously, the present study population and the chosen approach in patient selection are not ideally suited to test the diagnostic value of ECG or cardiac enzymes in the work-up of myocarditis, since (a) we did not include patients with the combination of normal ECG and negative enzymes and (b) patients with a troponin rise and normal ECG were more likely to get a CMR study and to be included to this study than patients with normal
OR (95 % CI)
p value
33.256 (3.044–363.354)
0.004
0.73
LV-ESV (mL)
analysis, the only independent predictor was an elevated cardiac troponin value.
123
Multivariable analysis
troponin but abnormal ECG. However, testing the diagnostic value of ECG or cardiac enzymes in the work-up of myocarditis was not the primary aim of this study; instead, the ‘‘additional’’ diagnostic value of CMR was tested in those young patients who were most likely suffering from myocarditis. Clinical, ECG and laboratory findings Choosing the optimal diagnostic approach in patients with suspected acute myocarditis is still challenging [1, 3]. Patient history and clinical presentation, the first ‘‘bricks’’ in the process of building a correct diagnosis, unfortunately lack specificity. In accordance with other publications, the most frequent symptom encountered in our young patient population was chest pain [13, 14]. In such a setting, an acute coronary syndrome related to obstructive CAD had to be ruled out before study inclusion, leaving a predominantly male population with low prevalence of cardiovascular risk factors and high clinical suspicion for acute myocarditis based on a recent history of viral illness and presence of at least one sign of on-going myocardial injury. For comparison, Pellaton et al. [15] showed that myocarditis is diagnosed in up to 88 % of patients less than 40 years old presenting to the emergency department with chest pain and elevated troponin after acute MI was excluded. Pathological ECG findings were relatively frequent in our study population due to our patient selection and inclusion criteria, however, in accordance with existing data not very specific in the work-up of acute myocarditis [1]. 81 % of the included patients had at least one pathological ECG sign (that was previously not known) suggestive of myocardial injury, however, with only 39 % of
Clin Res Cardiol
All study patients underwent CMR imaging as part of the diagnostic work-up of suspected myocarditis, in accordance with the current available recommendations [1]. First, a low prevalence for global and regional LV dysfunction was encountered. This is probably explained by the inclusion of clinically stable patients with limited
extent of LGE and predominant subepicardial distribution (Figs. 2, 3). Second, while all patients with evidence of myocardial edema demonstrated corresponding LGE, a substantial proportion of patients with positive LGE had assessable, however, negative edema images (29 %). As the probability for presence of acute myocardial inflammation in these patients was high based on their clinical picture and troponin values, our positive LGE findings in a substantial number of edema-negative patients may reflect the more robust and therefore superior diagnostic performance of LGE compared to T2-weighted edema imaging. Published data on this topic are conflicting, with some studies reporting superior accuracies and others showing inferior accuracies for edema imaging compared to LGE in acute myocarditis [6, 13, 17, 18]. For example, in a similar population of 81 patients with suspected acute myocarditis and positive LGE, Di Bella et al. [11] report 25 % of patients as not having detectable edema.
Fig. 2 31-year-old male patient complaining of chest pain and dyspnea and flu-like symptoms with troponin T elevation (1.197 pg/ mL; ULN 14 pg/mL). ECG on admission showed ST-segment elevation in I and V2–V6. Late gadolinium enhancement (LGE)
CMR (same day) in short axis (upper images) and long axis (lower images) showed extensive hyperenhancement with subepicardial distribution in the inferior and lateral walls and mid-myocardial distribution in the anterior wall and septum
them demonstrating corresponding troponin elevation. STsegment elevation and T-wave inversion, somewhat sensitive ECG abnormalities (shown to be present in approximately 50 % of patients with histological active myocarditis), were observed in 62 % of our study population [1, 16]. Noteworthy, conduction abnormalities were rare (7 % of patients) and when isolated, were not associated with troponin elevations. CMR findings and their relationship to ECG and troponin status
123
Clin Res Cardiol
Fig. 3 23-year old male patient complaining of chest pain after an upper respiratory tract infection episode with troponin T elevation (200 pg/mL; ULN 14 pg/mL). ECG on admission showed no STsegment/T-wave abnormalities. Late gadolinium enhancement (LGE)
CMR (next day) in short axis (upper images) showed limited subepicardial hyperenhancement in the basal half of the lateral wall. T2-weighted edema (lower images) showed edema presence in the same segments
In the present study, pathological CMR results were detected in only one-third of the patients with ECG abnormalities. And vice versa, ECG-negative patients demonstrated pathological CMR findings in up to 59 %. At first glance, this finding is surprising. Obviously, this result is (partly) due to our approach in (a) selecting patients and (b) enabling CMR studies as mentioned previously. However, our observation that ECGs were normal despite presence of LGE (suggestive of myocardial damage) in a substantial proportion of patients is neither new nor surprising: recently, Di Bella et al. [1.] demonstrated ‘‘normal’’ ECG findings in up to 32 % of 77 patients with clinically suspected acute myocarditis and corresponding presence of LGE. Moreover, there was no relationship between the location of myocardial damage deduced from the ECG leads showing ST/T changes and the myocardial segments showing presence of LGE in that study. Hence, both the sensitivity as well as specificity of ECG changes
has to be seriously questioned in patients with clinical suspicion of acute myocarditis.
123
Troponin measurement as gate-keeper for CMR? Regarding the relation between CMR findings and troponin status (only 73 % of troponin-positive patients demonstrated presence of edema and/or LGE), we think that our findings reflect the superior sensitivity of troponin measurements compared to LGE in depicting even small amounts of acute myocardial damage. Troponin assays reflect global myocardial injury and allow detection of even minimal myonecrosis [19]. In comparison, even though LGE-CMR can depict focal ischemic myocardial damage of as low as 2 % of the myocardium, it might miss more subtle, patchy non-ischemic pathology that is often present in myocarditic processes [12, 20]. The only two patients with limited positive LGE and without troponin
Clin Res Cardiol
elevation had no chest pain at presentation and showed only non-specific symptoms of fatigue and dyspnea with some ST-segment depression. In this constellation, a past/ sub-acute episode of myocarditis—rather than an acute one—is a plausible explanation. The ECG is expected to have a longer ‘‘memory’’ capacity of preceding myocardial changes in response to myocarditis (at least in some individuals) compared to cardiac enzymes that are only elevated in the acute setting. Based on our results, the diagnostic as well as prognostic value of CMR studies in young patients with clinical suspicion of myocarditis, but ‘‘negative’’ cardiac enzymes has to be questioned: in 95 % of troponin-negative patients, the CMR study was completely normal. Hence, in these patients the additionally performed CMR study only caused additional costs—without adding any further diagnostic information. In contrast, in patients with normal resting ECGs (but elevated troponin), diagnostic/conclusive CMR findings were reported in up to 59 %. Hence, to dispense with a CMR study can be suggested in troponinnegative patients, but not in ECG-negative/troponin-positive ones since this would result in not verifying the diagnosis of myocarditis in a substantial number of patients. The importance of establishing the presence or absence of LGE was recently shown by Gruen et al. [21] in a follow-up study performed in patients with biopsy-proven myocarditis: the parameter LGE was the best predictor of all-cause as well as cardiac mortality. Patients without LGE demonstrated an excellent outcome in that study. Thus, as ‘‘normal’’ cardiac enzymes in a patient with suspected acute myocarditis are highly suggestive of LGE ‘‘absence’’, this finding would confer a reassuring excellent long-term prognosis in such patients even without performing CMR. The good outcome in these patients with NYHA class 0 or 1, narrow QRS complexes and normal range LV end-diastolic diameter and EF has also been shown in previous studies [22–24]. Finally, one may also question the ‘‘diagnostic’’ value of CMR even in patients with elevated troponin and argue that (a) the diagnosis of acute myocarditis may already be made based on the clinical picture and measurement of raised enzymes and that (b) CMR seems to be less sensitive compared to troponin measurements. Apart from verifying the clinically suspected diagnosis of myocarditis, CMR allows to rule out (amongst others) an ischemic origin of myocardial damage, non-ischemic cardiomyopathies such as hypertrophic cardiomyopathy (HCM), dilative cardiomyopathy (DCM) or takotsubo cardiomyopathy that may be associated with a troponin rise, an acute aortic disease (such as aortic dissection or aortitis) and of course congenital heart diseases. Moreover, CMR enables an accurate delineation of the pattern and extent of myocardial damage caused by myocarditis. This pattern of myocardial damage
may not only give some clues to the underlying pathogen causing myocarditis, but also enable prognostication of the future course of myocarditis [8, 21]. These information in turn may help to decide (a) whether more invasive diagnostic approaches such as endomyocardial biopsy are required [25], (b) which biopsy approach (LV vs. RV biopsy) will be the best regarding optimal conclusive diagnostic results (in case of indicated biopsy) [26] and (c) whether follow-up studies are required/helpful in order to assess the disease course and the effect of implemented therapy [27]. Normalization of elevated troponin levels in myocarditis patients usually occurs within a few days— while myocardial inflammation/damage may still be there or even be on going. Therefore, disappearance of myocardial inflammation and the extent of myocardial recovery can only be accurately assessed based on follow-up CMR studies. Study limitations In contrast to the current recommended criteria for work-up of myocarditis (‘‘Lake Louise Criteria’’), our CMR protocol was based on T2-weighted edema and LGE imaging only. Several studies have addressed the diagnostic performance of different combinations from the three available techniques, with the ‘‘any-two’’ approach yielding the highest accuracy [1, 6]. Nevertheless, it is a well-known fact that early gadolinium enhancement (EGE) and T2weighted edema sequences are often skipped in institutional standard protocols due to inconsistent image quality. Moreover, a recent study by Chu et al. [28] showed that in patients with a relevant pre-test likelihood for active myocarditis, skipping EGE imaging does not change overall diagnostic accuracy. Second, only 58 % of our patients had assessable T2weighted STIR images, the rest discarded due to poor quality. This again is in agreement with previous reports regarding the inconsistent image quality of T2-weighted spin-echo images with susceptibility to motion artefacts and other confounders [29]. Finally, no endomyocardial biopsy data for confirmation of the acute myocarditic process were available in our patient population. Considering the clinical presentation without hemodynamic instability or overt heart failure, there was no compelling clinical indication for endomyocardial biopsy [30].
Conclusion The clinical use of non-invasive CMR in the work-up of clinically suspected acute myocarditis in young patients is only helpful and appropriately indicated in those ones with
123
Clin Res Cardiol
elevated cardiac enzymes. A pre-selection of such patients for CMR based on serum cardiac enzymes—but not on ECG recordings—may prevent a meaningless overuse of CMR. Acknowledgments This work was financially supported by a Grant from the Robert-Bosch-Stiftung (Grant to A.Y. and US). Conflict of interest
None.
References 1. Friedrich MG, Sechtem U, Schulz-Menger J et al (2009) Cardiovascular magnetic resonance in myocarditis: a JACC white paper. J Am Coll Cardiol 53:1475–1487 2. Cooper LT Jr (2009) Myocarditis. N Engl J Med 360:1526–1538 3. Kindermann I, Barth C, Mahfoud F et al (2012) Update on myocarditis. J Am Coll Cardiol 59:779–792 4. Lauer B, Niederau C, Kuhl U et al (1997) Cardiac troponin T in patients with clinically suspected myocarditis. J Am Coll Cardiol 30:1354–1359 5. Schumm J, Greulich S, Sechtem U, Mahrholdt H (2014) T1 mapping as new diagnostic technique in a case of acute onset of biopsy-proven viral myocarditis. Clin Res Cardiol 103:405–408 6. Abdel-Aty H, Boye P, Zagrosek A et al (2005) Diagnostic performance of cardiovascular magnetic resonance in patients with suspected acute myocarditis: comparison of different approaches. J Am Coll Cardiol 45:1815–1822 7. Mahrholdt H, Wagner A, Judd RM, Sechtem U, Kim RJ (2005) Delayed enhancement cardiovascular magnetic resonance assessment of non-ischaemic cardiomyopathies. Eur Heart J 26:1461–1474 8. Mahrholdt H, Wagner A, Deluigi CC et al (2006) Presentation, patterns of myocardial damage, and clinical course of viral myocarditis. Circulation 114:1581–1590 9. Yilmaz A, Ferreira V, Klingel K, Kandolf R, Neubauer S, Sechtem U (2012) Role of cardiovascular magnetic resonance imaging (CMR) in the diagnosis of acute and chronic myocarditis. Heart Fail Rev 10. Kwon DH, Hachamovitch R, Popovic ZB et al (2012) Survival in patients with severe ischemic cardiomyopathy undergoing revascularization versus medical therapy: association with endsystolic volume and viability. Circulation 126:S3–S8 11. Di BG, Florian A, Oreto L et al (2012) Electrocardiographic findings and myocardial damage in acute myocarditis detected by cardiac magnetic resonance. Clin Res Cardiol 101:617–624 12. Baccouche H, Mahrholdt H, Meinhardt G et al (2009) Diagnostic synergy of non-invasive cardiovascular magnetic resonance and invasive endomyocardial biopsy in troponin-positive patients without coronary artery disease. Eur Heart J 30:2869–2879 13. Lurz P, Eitel I, Adam J et al (2012) Diagnostic performance of CMR imaging compared with EMB in patients with suspected myocarditis. JACC Cardiovasc Imaging 5:513–524 14. Yilmaz A, Mahrholdt H, Athanasiadis A et al (2008) Coronary vasospasm as the underlying cause for chest pain in patients with PVB19 myocarditis. Heart 94:1456–1463
123
15. Pellaton C, Monney P, Ludman AJ, et al. (2012) Clinical features of myocardial infarction and myocarditis in young adults: a retrospective study. BMJ Open 2 16. Morgera T, Di LA, Dreas L et al (1992) Electrocardiography of myocarditis revisited: clinical and prognostic significance of electrocardiographic changes. Am Heart J 124:455–467 17. Leurent G, Langella B, Fougerou C et al (2011) Diagnostic contributions of cardiac magnetic resonance imaging in patients presenting with elevated troponin, acute chest pain syndrome and unobstructed coronary arteries. Arch Cardiovasc Dis 104: 161–170 18. Rottgen R, Christiani R, Freyhardt P et al (2011) Magnetic resonance imaging findings in acute myocarditis and correlation with immunohistological parameters. Eur Radiol 21:1259–1266 19. Agewall S, Giannitsis E, Jernberg T, Katus H (2011) Troponin elevation in coronary vs. non-coronary disease. Eur Heart J 32:404–411 20. Ricciardi MJ, Wu E, Davidson CJ et al (2001) Visualization of discrete microinfarction after percutaneous coronary intervention associated with mild creatine kinase-MB elevation. Circulation 103:2780–2783 21. Grun S, Schumm J, Greulich S et al (2012) Long-term follow-up of biopsy-proven viral myocarditis: predictors of mortality and incomplete recovery. J Am Coll Cardiol 59:1604–1615 22. Kindermann I, Kindermann M, Kandolf R et al (2008) Predictors of outcome in patients with suspected myocarditis. Circulation 118:639–648 23. McNamara DM, Starling RC, Cooper LT et al (2011) Clinical and demographic predictors of outcomes in recent onset dilated cardiomyopathy: results of the IMAC (intervention in myocarditis and acute cardiomyopathy)-2 study. J Am Coll Cardiol 58:1112–1118 24. Ukena C, Mahfoud F, Kindermann I, Kandolf R, Kindermann M, Bohm M (2011) Prognostic electrocardiographic parameters in patients with suspected myocarditis. Eur J Heart Fail 13:398–405 25. Baccouche H, Mahrholdt H, Meinhardt G et al (2009) Diagnostic synergy of non-invasive cardiovascular magnetic resonance and invasive endomyocardial biopsy in troponin-positive patients without coronary artery disease. Eur Heart J 30:2869–2879 26. Yilmaz A, Kindermann I, Kindermann M et al (2010) Comparative evaluation of left and right ventricular endomyocardial biopsy: differences in complication rate and diagnostic performance. Circulation 122:900–909 27. Elezkurtaj S, Lassner D, Schultheiss HP, Escher F (2014) Vascular involvement in cardiac giant cell myocarditis: a new pathophysiological aspect. Clin Res Cardiol 103:161–163 28. Chu GC, Flewitt JA, Mikami Y, Vermes E, Friedrich MG (2013) Assessment of acute myocarditis by cardiovascular MR: diagnostic performance of shortened protocols. Int J Cardiovasc Imaging 29:1077–1083 29. Friedrich MG (2008) Tissue characterization of acute myocardial infarction and myocarditis by cardiac magnetic resonance. JACC Cardiovasc Imaging 1:652–662 30. Cooper LT, Baughman KL, Feldman AM et al (2007) The role of endomyocardial biopsy in the management of cardiovascular disease: a scientific statement from the American Heart Association, the American College of Cardiology, and the European Society of Cardiology. Circulation 116:2216–2233
ORIGINAL PAPER
Diagnostic value of CMR in young patients with clinically suspected acute myocarditis is determined by cardiac enzymes Anca Florian • Tim Scha¨ufele • Anna Ludwig • Sabine Ro¨sch • Ina Wenzelburger • Handan Yildiz Udo Sechtem • Ali Yilmaz
•
Received: 4 February 2014 / Accepted: 1 October 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Aims Cardiovascular magnetic resonance (CMR) has become a valuable diagnostic tool for non-invasive diagnosis of acute myocarditis. However, since CMR studies are time- and cost-intensive and its diagnostic accuracy still not perfect, additional parameters are warranted to preselect and identify those individuals in whom a CMR study is likely to add crucial information regarding correct and timely diagnosis of acute myocarditis. The diagnostic value of CMR was evaluated in a population of young patients with clinically suspected acute myocarditis in relation to ECG and serum cardiac enzyme findings. Methods and results Only young patients aged B40 years in whom acute myocarditis was highly suspected based on their clinical symptoms, resting ECG findings and/or levels of cardiac enzymes (at presentation) were included to this study. After ruling out obstructive coronary artery disease, a multi-parametric CMR study was performed as part of the diagnostic work-up. The CMR protocol comprised cine sequences, T2-weighted edema imaging and late gadolinium enhancement (LGE) imaging on a 1.5-T MR scanner. 89 patients (28 ± 7 years, 89 % male) were included to this study presenting with symptoms of chest pain (85 %), dyspnea (26 %), fatigue (23 %) and/or palpitations (18 %). Pathological ECG changes were present in 72 patients
A. Florian A. Yilmaz (&) Department of Cardiology and Angiology, University Hospital Mu¨nster, Albert-Schweitzer-Campus 1, Building A1, 48149 Mu¨nster, Germany e-mail: [email protected] T. Scha¨ufele A. Ludwig S. Ro¨sch I. Wenzelburger H. Yildiz U. Sechtem A. Yilmaz Division of Cardiology, Robert-Bosch-Krankenhaus, Stuttgart, Germany
(81 %). An elevated serum troponin level was measured in 45 patients (51 %). Pathological CMR findings (presence of edema and/or LGE) were detected in 35 patients (39 %). In detail, pathological CMR findings were detected in 36 % of patients with resting ECG changes and in 73 % of patients with troponin rise. In contrast, normal CMR results were obtained in 95 % of patients with negative troponin at presentation, but only in 41 % of patients with normal ECG. On multivariable analysis, a positive serum troponin was the only independent predictor for a pathological CMR finding (OR = 33.26, 95 % CI = 3.04–363.35, p = 0.004). Conclusions The clinical use of non-invasive CMR in the work-up of clinically suspected ‘‘acute’’ myocarditis in young patients is only helpful and appropriately indicated in those ones with elevated cardiac enzymes. A preselection of such patients for CMR based on serum cardiac enzymes—but not on ECG recordings—may prevent a meaningless overuse of CMR. Keywords
Myocarditis Troponin CMR LGE ECG
Introduction Current recommendations in acute myocarditis suggest that the diagnosis should be based on an integrated approach including history, clinical assessment and a range of noninvasive tests comprising ECG, serum biomarkers and noninvasive cardiac imaging [1]. Nevertheless, a definite diagnosis remains a difficult challenge in a large proportion of patients (in particular in those with mild forms), since clinical presentation is multifaceted and ECG-changes as well as biomarkers tend to be non-specific [1–4]. Cardiovascular magnetic resonance (CMR) imaging with its arsenal of different sequences allowing non-invasive
123
Clin Res Cardiol
myocardial tissue characterization has evolved as a valuable tool for the diagnosis of myocarditis. T2-weighted imaging can detect tissue edema while T1-weighted imaging before and early (within the first minutes) after gadolinium contrast administration may detect myocardial hyperemia and capillary leakage, both hallmarks of acute myocardial inflammation [1, 5, 6]. In addition, late gadolinium enhancement (LGE) imaging by T1-weighted inversion recovery sequences accurately depicts myocardial damage and allows the differentiation of ischemic from non-ischemic injury with either subepicardial or midwall patterns of distribution commonly described in myocarditis [7–9]. At present, the recommended CMR diagnostic criteria for active myocarditis (‘‘Lake Louise Criteria’’) support the combined use of all the three above-mentioned sequences—but are based on a limited number of trials without large-scale multicenter data. Moreover, it is acknowledged that despite the good accuracy of the ‘‘any-two’’ out of three criteria approach, certain clinical settings might require even higher test sensitivities [1]. In the past decade, the number of non-invasive cardiac imaging studies has been intensively growing with a nonneglectable increase in associated costs. Regarding both increasing costs and limited health-care budgets, the use of any imaging modality including CMR requires justification to both its clinical diagnostic value and its costeffectiveness [10]. Hence, additional parameters are warranted to easily and successfully pre-select and identify those individuals in whom a CMR study is likely to add crucial information regarding appropriate diagnosis of, e.g., acute myocarditis. Therefore, the aim of this study was to evaluate the diagnostic performance of CMR in a population of young patients (B40 years) with a high clinical suspicion of acute myocarditis in relation to the presence or absence of (a) ECG changes and (b) elevated levels of cardiac enzymes as additional markers of acute myocardial injury.
Methods Study population Between 2009 and 2012, only young patients aged B40 years in whom acute myocarditis was highly suspected based on their clinical symptoms, resting ECG findings and/or levels of cardiac enzymes (at presentation) were included to this study. Patients aged [40 years were excluded to minimize confounding CMR findings possibly caused by other cardiovascular diseases apart from myocarditis. To be included into this study, patients had to fulfil all of the following criteria: (1) new onset (\14 days) of at least
123
one of the following symptoms: chest pain, dyspnea, unexplained fatigue and/or palpitations; (2) history of recent (\8 weeks) flu-like symptoms; (3) evidence for recent or ongoing myocardial injury by any of the two: ECG abnormalities (as detailed below) or elevated cardiac troponin; (4) obstructive coronary artery disease (CAD) ruled out by invasive coronary angiography or by clinical assessment. Patients with a prior diagnosis of pericarditis/ myocarditis, previously known significant valvular, myocardial or CAD, as well as patients with cardiovascular instability at presentation were excluded. Moreover, patients with rather unspecific non-cardiac chest pain and/ or only minor and quickly disappearing symptoms were excluded. Only patients with a high clinical likelihood of acute myocarditis were included to this study. Patients with the combination of normal ECG and negative enzymes were not included in this study due to the low likelihood of myocarditis. Moreover, in case of logistic limitations regarding the availability of CMR studies, patients with a troponin rise and normal ECG had a higher priority in getting a CMR study (and consequently being included to this study) than patients with normal troponin but abnormal ECG. From the 195 patients initially screened, 91 patients underwent a CMR study as part of the diagnostic work-up. In the remaining 104 patients, CMR was not possible due to logistic reasons, contraindicated or not consented. In two patients, CMR data were not assessable due to poor image quality. The final analysis included 89 patients. ECG analysis Standard 12-lead ECG tracings recorded at hospital presentation were interpreted by two cardiologists. The following new or persistent abnormalities were considered positive signs of myocardial injury: [1] conduction abnormalities; [2] ST-segment elevation (C0.1 mV) or depression (C0.05 mV) in two or more contiguous leads; [3] isolated inverted T-waves (C0.1 mV, negative in any leads except III, aVR and V1). In addition, the presence of arrhythmia and pathologic Q-waves, as well as heart rate and PR-, QRS-, QT-interval duration were noted. ECG analysis was detailed elsewhere [11]. Patients showing both ST–segment elevation and depression were classified as having ST-segment elevation. The presence of an early repolarization pattern in precordial leads was not considered as positive ECG sign after agreement between readers. Cardiac magnetic resonance: data acquisition and analysis CMR imaging was performed in breath-hold on a 1.5-T unit (Aera, Siemens Medical Solutions, Erlangen,
Clin Res Cardiol
Germany) using commercially available cardiac software, electrocardiographic triggering, and cardiac-dedicated surface coils. CMR included steady-state-free-precession cine imaging, T2-weighted STIR ‘‘edema’’ imaging and LGE imaging 5–10 min after intravenous contrast administration (0.15 mmol/kg MagnevistÒ) as previously described [12]. Image analysis was performed off-line by two experienced readers blinded to patient data. Ventricular volumes, ejection fraction and left ventricular mass were derived by contouring endo- and epicardial borders on the short-axis cine images. The presence of regional wall motion abnormalities (WMA) was visually evaluated both on the shortand long-axis cines using the standard AHA 17- segment model. In the T2-weighted STIR images, regional myocardial edema was visually identified while global edema was quantitatively assessed as signal intensity of the myocardium normalized to that of the skeletal muscle greater than 1.9 [6]. Poor quality T2-weighted STIR images were excluded from the analysis after consensus between readers. The diagnosis of positive LGE required visual identification of bright, hyper-enhanced signal within the myocardium in two orthogonal views. Using the same 17-segment model, all segments were dichotomously classified with respect to LGE presence. Positive LGE segments were further characterized according to myocardial damage localization as subendocardial, midwall, subepicardial or transmural. In addition, pericardial enhancement was noted, if present. Pericardial effusion was evaluated on cine images, as present at end diastole. Serum biomarker assessment Serum biomarkers of myocardial injury [creatine kinase MB (CK-MB) and cardiac troponin] were measured and considered elevated if at least one of the determinations on the day of admission exceeded the upper laboratory reference limit. Our hospital laboratory measured troponin T until 2011 and switched to troponin I measurements in 2012. C-reactive protein (CRP) was considered elevated if serum levels at admission exceeded the upper laboratory reference limit. Statistical analysis Continuous variables were expressed as mean ± SD. Skewed variables were expressed as median and interquartile range. Categorical variables were expressed as frequency with percentage t Student test was used for the between group comparison according to ECG and troponin status. Levene’s test was used for testing equality of variances. Mann-Withney U test was used for not normally distributed variables. The Chi-square test and Fischer’s
exact test were used to compare categorical variables expressed as proportions. To find independent predictors for a diagnostic CMR, first, univariable analysis was performed. Second, the parameters with significant p-values were introduced in the multivariable analysis. Statistical analysis was performed using SPSS software for Windows (version 18, SPSS, Chicago Illinois, US). All tests were 2-sided and a p value below 0.05 was considered statistically significant.
Results Patient characteristics and clinical presentation The study population of 89 patients comprised predominantly males (89 %) with a mean age of 28 ± 7 years. The distribution of different cardiovascular risk factors is shown in Table 1. 20 patients (22 %) had fever ([38.5 °C) at hospital admission. All patients were stable and 14 (16 %) showed tachycardia ([100/min) at presentation. ECG findings As shown in Table 2, 72 patients (81 %) had at least one pathological ECG sign as previously defined. The majority of patients (97 %) were in sinus rhythm with only three (3 %) showing atrial fibrillation. Conduction abnormalities were present as follows: two patients had sinuatrial block, one patient AV block II°, two patients AV block III° and Table 1 Patient characteristics All patients (n = 89) Male, n (%) Age (years)
79 (89) 28 ± 7
Cardiovascular risk factors, n (%) Family history of CAD
12 (13)
Arterial hypertension Diabetes mellitus
4 (5) 0 (0)
High cholesterol
6 (7)
Current smoker
36 (40)
Obesity
6 (7)
Symptoms, n (%) Chest pain
76 (85)
Dyspnea
23 (26)
Fatigue
20 (23)
Palpitations
18 (20)
Clinical parameters on admission HR (bpm) BP systolic/diastolic (mmHg)
79 ± 18 129 ± 19/78 ± 13
CAD coronary artery disease, HR heart rate, BP blood pressure
123
Clin Res Cardiol
one patient newly diagnosed left bundle branch block. Additional ST/T changes were present in two patients with sinuatrial block and AV block II°, respectively. Regarding ST/T abnormalities, 49 patients (55 %) showed ST–segment elevation, 13 (15 %) had ST-segment depression and eight (9 %) had isolated negative T-waves. Pathological Q-waves could be depicted in three patients (3 %). In six patients (7 %) runs of non-sustained ventricular tachycardia were recorded upon admission. Laboratory findings Overall, 45 patients (51 %) had elevated troponin levels and 39 (44 %) elevated CK–MB levels. As shown in Fig. 1, patients with chest pain were more likely to have Table 2 ECG and laboratory findings
elevated troponin compared to patients without such a complaint (55 vs. 23 %, p = 0.039). In contrast, the presence of fatigue or palpitations was associated more often with a negative troponin (32 vs. 13 % for fatigue, p = 0.045; 36 vs. 4 % for palpitations, p \ 0.0001). 28 patients (31 %) had both elevated troponin and ECG abnormalities, of whom 22 (25 %) had ST-segment elevation, two (2 %) had ST-segment depression and 5 (6 %) had isolated negative T-waves. None of the patients presenting with conduction abnormalities only had elevated troponin. No significant differences were found regarding the presence of abnormal ECG changes between troponinpositive and -negative patients. 61 patients (69 %) had elevated CRP levels on admission. Elevated CRP was more frequently associated with a positive troponin (p \ 0.0001). CMR findings
All patients (n = 89) ECG on admission, n (%) Pathological ECG Conduction abnormalities
72 (81) 6 (7)
ST-segment elevation
49 (55)
ST-segment depression
13 (15)
Isolated T-wave inversion
8 (9)
Pathological Q-waves
3 (3)
Atrial fibrillation
3 (3)
Laboratory work up Elevated troponin, n (%)
45 (51)
Elevated CK-MB, n (%) Elevated CRP, n (%)
39 (44) 61 (69)
Coronary angiography performed, n (%)
CMR was performed at a median of 2 (IQR 1–3) days after admission. The mean values for left ventricular (LV) volumes, ejection fraction and mass are illustrated in Tables 3 and 4. 16 (18 %) patients had an LV ejection fraction \60 %. Three patients with reduced LV ejection fraction demonstrated associated regional WMA. Regional WMA were found in 9 % of the cases (n = 8) and were all associated with positive LGE in the respective segments. 35 (39 %) patients showed positive LGE with a median extent of 4 (IQR 3–6) segments with the following Table 3 CMR results in relation to ECG
30 (34)
ECG electrocardiogram, CK creatine kinase, CRP C reactive protein Admission to CMR, days
Fig. 1 Bar graphs showing the prevalence of symptoms in patients with vs. without troponin elevation
123
Total (n = 89)
ECG positive (n = 72)
ECG negative (n = 17)
p value
2 (1–3)
2 (1–3)
2 (1–3)
0.61
LV-EDV (mL)
139 ± 26
137 ± 26
146 ± 27
0.21
LV-ESV (mL)
52 ± 14
52 ± 14
53 ± 11
0.66
LV-EF (%)
63 ± 7
63 ± 7
64 ± 5
0.58
LV-EF \60 %, n (%)
16 (18)
14 (19)
2 (12)
0.73
Regional WMA, n (%) Edema, n (%)
8 (9)
6 (8)
2 (12)
0.65
(A) Assessable
52 (58)
42 (58)
10 (59)
1.00
(B) Positive
0.015
13 (15)
7 (10)
6 (35)
LGE positive, n (%)
35 (39)
26 (36)
10 (59)
0.10
Edema and/or LGE positive, n (%)
35 (39)
26 (36)
10 (59)
0.10
Pericardial effusion, n (%)
13 (15)
11 (15)
2 (12)
1.00
CMR cardiac magnetic resonance, LV left ventricular, EDV left ventricular end diastolic volume, ESV end systolic volume, EF ejection fraction, WMA wall motion abnormalities, LGE late gadolinium enhancement
Clin Res Cardiol Table 4 CMR results in relation to troponin
Table 5 CMR results in relation to ECG and troponin
Total (n = 89)
Trop. positive (n = 45)
Trop. negative (n = 44)
Admission to CMR, days
2 (1–3)
1 (1–2.5)
2.5 (1.25–4)
LV-EDV (mL)
139 ± 26
142 ± 25
135 ± 27
0.21
LV-ESV (mL)
52 ± 14
56 ± 14
48 ± 12
0.005
65 ± 6
0.004
LV-EF (%)
63 ± 7
61 ± 7
p value
0.002
LV-EF \60 %, n (%)
9 (10)
8 (18)
1 (2)
0.030
Regional WMA, n (%) Edema, n (%)
8 (9)
8 (18)
0 (0)
0.006
(A) Assessable
52 (58)
25 (56)
27 (61)
(B) Positive
13 (15)
13 (29)
0 (0)
\0.001
0.67
LGE positive, n (%)
35 (39)
33 (73)
2 (5)
\0.001
Edema and/or LGE positive, n (%)
35 (39)
33 (73)
2 (5)
\0.001
Pericardial effusion, n (%)
13 (15)
9 (20)
4 (9)
0.23
CMR cardiac magnetic resonance, LV left ventricular, EDV left ventricular end diastolic volume, ESV end systolic volume, EF ejection fraction, WMA wall motion abnormalities, LGE late gadolinium enhancement
patterns: 18 (51 %) subepicardial, 6 (17 %) midwall, 1 (3 %) transmural and 10 (29 %) mixed (subepicardial and midwall). An ischemic pattern of LGE was depicted in none of the patients. LGE was more frequently present in patients with impaired LV function (69 %) than in patients with normal ejection fraction (34 %; p = 0.022). T2weighted edema images were assessable in 52 patients and presence of edema was documented in 13 (25 %) of them. All patients with proof of edema also showed related presence of LGE. However, ten patients (26 %) without edema had nevertheless evidence for LGE. Taken together, 35 patients (39 %) demonstrated presence of edema and/or LGE. A mild to moderate pericardial effusion was observed in 13 (15 %) patients. A moderate to strong pericardial enhancement could be depicted in two patients. CMR findings according to ECG abnormalities Out of 72 patients with ECG abnormalities at presentation, only 26 (36 %) exhibited at least one CMR positive finding suggestive of myocarditis. Notably, all of these 26 patients had positive LGE. Global and regional systolic dysfunction was depicted in 14 (19 %) and six (8 %) patients with pathological ECG, respectively. ECG changes in LGEpositive patients (n = 26) were as follows: 20 (77 %) had
Number (%)
ECG negative trop. positive (n = 17)
ECG positive trop. positive (n = 28)
ECG positive trop. negative (n = 44)
CMR neg.
CMR pos.
CMR neg.
CMR pos.
CMR neg.
CMR pos.
7 (41)
10 (59)
4 (14)
24 (86)
42 (95)
2 (5)
CMR cardiac magnetic resonance
ST-segment elevation, three (12 %) ST-segment depression and two (8 %) conduction abnormalities, of which one demonstrated both sinuatrial block and ST-segment elevation. The second patient with AV block III° exhibited only limited septal midwall LGE involving one segment. Table 3 shows CMR findings according to ECG status. No significant differences for LV dimensions and function as well as for global or regional impairment were noted. CMR findings according to troponin status Out of 45 patients with elevated troponin at admission, 33 (73 %) demonstrated at least one pathological CMR finding, with all 33 patients showing positive LGE. Tables 4 and 5 show CMR findings according to the troponin status. Troponin-positive patients had a significantly shorter admission-to-CMR time compared to troponin-negative ones. Mean LV end systolic volume was significantly higher and mean LV ejection fraction significantly lower in patients with troponin elevation. Only one patient without troponin rise had an impaired LV ejection fraction which was mild (50 %). None of the troponin-negative patients showed WMA or (when assessable) edema. 31 of 33 patients with positive LGE had increased troponin levels. The two patients without increased troponin had limited LGE involving one and two segments, respectively, and did not show any edema (that was assessable and negative in both). There was no significant difference in pericardial effusion presence between the patients with and without troponin rise. CMR was able to verify myocarditis (defined as presence of edema and/or LGE) in 74 % (n = 31) of patients with chest pain and elevated troponin and in 81 % (n = 17) of patients with chest pain, ECG abnormalities and elevated troponin. Predictors for a diagnostic CMR Table 6 shows the results of the regression analysis regarding independent predictors for a diagnostic CMR. Although in univariable analysis, several variables were associated with a pathological CMR result, at multivariable
123
Clin Res Cardiol Table 6 Predictors of a positive edema and/or LGE CMR study
CMR cardiac magnetic resonance, LV left ventricular, EDV left ventricular end diastolic volume, ESV end systolic volume, EF ejection fraction, WMA wall motion abnormalities, LGE late gadolinium enhancement, CK creatine kinase, CRP C reactive protein
(n = 89)
Univariable analysis OR (95 % CI)
p value
Age
0.989 (0.928–1.054)
Male gender
6.800 (0.822–56.281)
0.08
Admission to CMR, days
0.688 (0.506–0.935)
0.017
Chest pain
1.550 (0.438–5.484)
0.50
Dyspnea
0.333 (0.111–1.004)
0.05
Pericardial effusion
2.904 (0.864–9.758)
0.09
LV-EDV (mL)
1.004 (0.987–1.021)
0.67
1.036 (1.001–1.072)
0.044
LV-EF (%)
0.914 (0.849–0.983)
0.016
Positive ECG
0.396 (0.134–1.164)
0.09
ST-segment elevation Elevated troponin
1.034 (0.442–2.423) 57.750 (12.077–276.155)
0.94 \0.001
Elevated CK-MB
12.732 (4.036–40.165)
\0.001
6.786 (2.016–22.844)
0.002
Elevated CRP
Discussion The main finding of the present study that included young, previously healthy individuals with a high clinical likelihood of acute myocarditis (in whom CAD was excluded or unlikely) is that troponin elevation at presentation was the only independent predictor for a pathological CMR finding—regardless of symptoms and ECG status. In particular, a CMR study added almost no diagnostic information in troponin-negative patients—despite a high clinical likelihood of myocarditis. Based on the results of this study, we suggest that at least in young patients (aged B400 years) a pre-selection for CMR imaging based on the cardiac troponin level may prevent a meaningless overuse of CMR. The selection of patients for this study was somewhat biased due to the inclusion criteria emphasizing the presence of clinical symptoms such as chest pain, ECG changes and (above all) elevated cardiac enzymes. However, this approach reflects our daily clinical practice and allows a successful identification of those patients who are most likely suffering from severe myocarditis—from a clinical point of view. Obviously, the present study population and the chosen approach in patient selection are not ideally suited to test the diagnostic value of ECG or cardiac enzymes in the work-up of myocarditis, since (a) we did not include patients with the combination of normal ECG and negative enzymes and (b) patients with a troponin rise and normal ECG were more likely to get a CMR study and to be included to this study than patients with normal
OR (95 % CI)
p value
33.256 (3.044–363.354)
0.004
0.73
LV-ESV (mL)
analysis, the only independent predictor was an elevated cardiac troponin value.
123
Multivariable analysis
troponin but abnormal ECG. However, testing the diagnostic value of ECG or cardiac enzymes in the work-up of myocarditis was not the primary aim of this study; instead, the ‘‘additional’’ diagnostic value of CMR was tested in those young patients who were most likely suffering from myocarditis. Clinical, ECG and laboratory findings Choosing the optimal diagnostic approach in patients with suspected acute myocarditis is still challenging [1, 3]. Patient history and clinical presentation, the first ‘‘bricks’’ in the process of building a correct diagnosis, unfortunately lack specificity. In accordance with other publications, the most frequent symptom encountered in our young patient population was chest pain [13, 14]. In such a setting, an acute coronary syndrome related to obstructive CAD had to be ruled out before study inclusion, leaving a predominantly male population with low prevalence of cardiovascular risk factors and high clinical suspicion for acute myocarditis based on a recent history of viral illness and presence of at least one sign of on-going myocardial injury. For comparison, Pellaton et al. [15] showed that myocarditis is diagnosed in up to 88 % of patients less than 40 years old presenting to the emergency department with chest pain and elevated troponin after acute MI was excluded. Pathological ECG findings were relatively frequent in our study population due to our patient selection and inclusion criteria, however, in accordance with existing data not very specific in the work-up of acute myocarditis [1]. 81 % of the included patients had at least one pathological ECG sign (that was previously not known) suggestive of myocardial injury, however, with only 39 % of
Clin Res Cardiol
All study patients underwent CMR imaging as part of the diagnostic work-up of suspected myocarditis, in accordance with the current available recommendations [1]. First, a low prevalence for global and regional LV dysfunction was encountered. This is probably explained by the inclusion of clinically stable patients with limited
extent of LGE and predominant subepicardial distribution (Figs. 2, 3). Second, while all patients with evidence of myocardial edema demonstrated corresponding LGE, a substantial proportion of patients with positive LGE had assessable, however, negative edema images (29 %). As the probability for presence of acute myocardial inflammation in these patients was high based on their clinical picture and troponin values, our positive LGE findings in a substantial number of edema-negative patients may reflect the more robust and therefore superior diagnostic performance of LGE compared to T2-weighted edema imaging. Published data on this topic are conflicting, with some studies reporting superior accuracies and others showing inferior accuracies for edema imaging compared to LGE in acute myocarditis [6, 13, 17, 18]. For example, in a similar population of 81 patients with suspected acute myocarditis and positive LGE, Di Bella et al. [11] report 25 % of patients as not having detectable edema.
Fig. 2 31-year-old male patient complaining of chest pain and dyspnea and flu-like symptoms with troponin T elevation (1.197 pg/ mL; ULN 14 pg/mL). ECG on admission showed ST-segment elevation in I and V2–V6. Late gadolinium enhancement (LGE)
CMR (same day) in short axis (upper images) and long axis (lower images) showed extensive hyperenhancement with subepicardial distribution in the inferior and lateral walls and mid-myocardial distribution in the anterior wall and septum
them demonstrating corresponding troponin elevation. STsegment elevation and T-wave inversion, somewhat sensitive ECG abnormalities (shown to be present in approximately 50 % of patients with histological active myocarditis), were observed in 62 % of our study population [1, 16]. Noteworthy, conduction abnormalities were rare (7 % of patients) and when isolated, were not associated with troponin elevations. CMR findings and their relationship to ECG and troponin status
123
Clin Res Cardiol
Fig. 3 23-year old male patient complaining of chest pain after an upper respiratory tract infection episode with troponin T elevation (200 pg/mL; ULN 14 pg/mL). ECG on admission showed no STsegment/T-wave abnormalities. Late gadolinium enhancement (LGE)
CMR (next day) in short axis (upper images) showed limited subepicardial hyperenhancement in the basal half of the lateral wall. T2-weighted edema (lower images) showed edema presence in the same segments
In the present study, pathological CMR results were detected in only one-third of the patients with ECG abnormalities. And vice versa, ECG-negative patients demonstrated pathological CMR findings in up to 59 %. At first glance, this finding is surprising. Obviously, this result is (partly) due to our approach in (a) selecting patients and (b) enabling CMR studies as mentioned previously. However, our observation that ECGs were normal despite presence of LGE (suggestive of myocardial damage) in a substantial proportion of patients is neither new nor surprising: recently, Di Bella et al. [1.] demonstrated ‘‘normal’’ ECG findings in up to 32 % of 77 patients with clinically suspected acute myocarditis and corresponding presence of LGE. Moreover, there was no relationship between the location of myocardial damage deduced from the ECG leads showing ST/T changes and the myocardial segments showing presence of LGE in that study. Hence, both the sensitivity as well as specificity of ECG changes
has to be seriously questioned in patients with clinical suspicion of acute myocarditis.
123
Troponin measurement as gate-keeper for CMR? Regarding the relation between CMR findings and troponin status (only 73 % of troponin-positive patients demonstrated presence of edema and/or LGE), we think that our findings reflect the superior sensitivity of troponin measurements compared to LGE in depicting even small amounts of acute myocardial damage. Troponin assays reflect global myocardial injury and allow detection of even minimal myonecrosis [19]. In comparison, even though LGE-CMR can depict focal ischemic myocardial damage of as low as 2 % of the myocardium, it might miss more subtle, patchy non-ischemic pathology that is often present in myocarditic processes [12, 20]. The only two patients with limited positive LGE and without troponin
Clin Res Cardiol
elevation had no chest pain at presentation and showed only non-specific symptoms of fatigue and dyspnea with some ST-segment depression. In this constellation, a past/ sub-acute episode of myocarditis—rather than an acute one—is a plausible explanation. The ECG is expected to have a longer ‘‘memory’’ capacity of preceding myocardial changes in response to myocarditis (at least in some individuals) compared to cardiac enzymes that are only elevated in the acute setting. Based on our results, the diagnostic as well as prognostic value of CMR studies in young patients with clinical suspicion of myocarditis, but ‘‘negative’’ cardiac enzymes has to be questioned: in 95 % of troponin-negative patients, the CMR study was completely normal. Hence, in these patients the additionally performed CMR study only caused additional costs—without adding any further diagnostic information. In contrast, in patients with normal resting ECGs (but elevated troponin), diagnostic/conclusive CMR findings were reported in up to 59 %. Hence, to dispense with a CMR study can be suggested in troponinnegative patients, but not in ECG-negative/troponin-positive ones since this would result in not verifying the diagnosis of myocarditis in a substantial number of patients. The importance of establishing the presence or absence of LGE was recently shown by Gruen et al. [21] in a follow-up study performed in patients with biopsy-proven myocarditis: the parameter LGE was the best predictor of all-cause as well as cardiac mortality. Patients without LGE demonstrated an excellent outcome in that study. Thus, as ‘‘normal’’ cardiac enzymes in a patient with suspected acute myocarditis are highly suggestive of LGE ‘‘absence’’, this finding would confer a reassuring excellent long-term prognosis in such patients even without performing CMR. The good outcome in these patients with NYHA class 0 or 1, narrow QRS complexes and normal range LV end-diastolic diameter and EF has also been shown in previous studies [22–24]. Finally, one may also question the ‘‘diagnostic’’ value of CMR even in patients with elevated troponin and argue that (a) the diagnosis of acute myocarditis may already be made based on the clinical picture and measurement of raised enzymes and that (b) CMR seems to be less sensitive compared to troponin measurements. Apart from verifying the clinically suspected diagnosis of myocarditis, CMR allows to rule out (amongst others) an ischemic origin of myocardial damage, non-ischemic cardiomyopathies such as hypertrophic cardiomyopathy (HCM), dilative cardiomyopathy (DCM) or takotsubo cardiomyopathy that may be associated with a troponin rise, an acute aortic disease (such as aortic dissection or aortitis) and of course congenital heart diseases. Moreover, CMR enables an accurate delineation of the pattern and extent of myocardial damage caused by myocarditis. This pattern of myocardial damage
may not only give some clues to the underlying pathogen causing myocarditis, but also enable prognostication of the future course of myocarditis [8, 21]. These information in turn may help to decide (a) whether more invasive diagnostic approaches such as endomyocardial biopsy are required [25], (b) which biopsy approach (LV vs. RV biopsy) will be the best regarding optimal conclusive diagnostic results (in case of indicated biopsy) [26] and (c) whether follow-up studies are required/helpful in order to assess the disease course and the effect of implemented therapy [27]. Normalization of elevated troponin levels in myocarditis patients usually occurs within a few days— while myocardial inflammation/damage may still be there or even be on going. Therefore, disappearance of myocardial inflammation and the extent of myocardial recovery can only be accurately assessed based on follow-up CMR studies. Study limitations In contrast to the current recommended criteria for work-up of myocarditis (‘‘Lake Louise Criteria’’), our CMR protocol was based on T2-weighted edema and LGE imaging only. Several studies have addressed the diagnostic performance of different combinations from the three available techniques, with the ‘‘any-two’’ approach yielding the highest accuracy [1, 6]. Nevertheless, it is a well-known fact that early gadolinium enhancement (EGE) and T2weighted edema sequences are often skipped in institutional standard protocols due to inconsistent image quality. Moreover, a recent study by Chu et al. [28] showed that in patients with a relevant pre-test likelihood for active myocarditis, skipping EGE imaging does not change overall diagnostic accuracy. Second, only 58 % of our patients had assessable T2weighted STIR images, the rest discarded due to poor quality. This again is in agreement with previous reports regarding the inconsistent image quality of T2-weighted spin-echo images with susceptibility to motion artefacts and other confounders [29]. Finally, no endomyocardial biopsy data for confirmation of the acute myocarditic process were available in our patient population. Considering the clinical presentation without hemodynamic instability or overt heart failure, there was no compelling clinical indication for endomyocardial biopsy [30].
Conclusion The clinical use of non-invasive CMR in the work-up of clinically suspected acute myocarditis in young patients is only helpful and appropriately indicated in those ones with
123
Clin Res Cardiol
elevated cardiac enzymes. A pre-selection of such patients for CMR based on serum cardiac enzymes—but not on ECG recordings—may prevent a meaningless overuse of CMR. Acknowledgments This work was financially supported by a Grant from the Robert-Bosch-Stiftung (Grant to A.Y. and US). Conflict of interest
None.
References 1. Friedrich MG, Sechtem U, Schulz-Menger J et al (2009) Cardiovascular magnetic resonance in myocarditis: a JACC white paper. J Am Coll Cardiol 53:1475–1487 2. Cooper LT Jr (2009) Myocarditis. N Engl J Med 360:1526–1538 3. Kindermann I, Barth C, Mahfoud F et al (2012) Update on myocarditis. J Am Coll Cardiol 59:779–792 4. Lauer B, Niederau C, Kuhl U et al (1997) Cardiac troponin T in patients with clinically suspected myocarditis. J Am Coll Cardiol 30:1354–1359 5. Schumm J, Greulich S, Sechtem U, Mahrholdt H (2014) T1 mapping as new diagnostic technique in a case of acute onset of biopsy-proven viral myocarditis. Clin Res Cardiol 103:405–408 6. Abdel-Aty H, Boye P, Zagrosek A et al (2005) Diagnostic performance of cardiovascular magnetic resonance in patients with suspected acute myocarditis: comparison of different approaches. J Am Coll Cardiol 45:1815–1822 7. Mahrholdt H, Wagner A, Judd RM, Sechtem U, Kim RJ (2005) Delayed enhancement cardiovascular magnetic resonance assessment of non-ischaemic cardiomyopathies. Eur Heart J 26:1461–1474 8. Mahrholdt H, Wagner A, Deluigi CC et al (2006) Presentation, patterns of myocardial damage, and clinical course of viral myocarditis. Circulation 114:1581–1590 9. Yilmaz A, Ferreira V, Klingel K, Kandolf R, Neubauer S, Sechtem U (2012) Role of cardiovascular magnetic resonance imaging (CMR) in the diagnosis of acute and chronic myocarditis. Heart Fail Rev 10. Kwon DH, Hachamovitch R, Popovic ZB et al (2012) Survival in patients with severe ischemic cardiomyopathy undergoing revascularization versus medical therapy: association with endsystolic volume and viability. Circulation 126:S3–S8 11. Di BG, Florian A, Oreto L et al (2012) Electrocardiographic findings and myocardial damage in acute myocarditis detected by cardiac magnetic resonance. Clin Res Cardiol 101:617–624 12. Baccouche H, Mahrholdt H, Meinhardt G et al (2009) Diagnostic synergy of non-invasive cardiovascular magnetic resonance and invasive endomyocardial biopsy in troponin-positive patients without coronary artery disease. Eur Heart J 30:2869–2879 13. Lurz P, Eitel I, Adam J et al (2012) Diagnostic performance of CMR imaging compared with EMB in patients with suspected myocarditis. JACC Cardiovasc Imaging 5:513–524 14. Yilmaz A, Mahrholdt H, Athanasiadis A et al (2008) Coronary vasospasm as the underlying cause for chest pain in patients with PVB19 myocarditis. Heart 94:1456–1463
123
15. Pellaton C, Monney P, Ludman AJ, et al. (2012) Clinical features of myocardial infarction and myocarditis in young adults: a retrospective study. BMJ Open 2 16. Morgera T, Di LA, Dreas L et al (1992) Electrocardiography of myocarditis revisited: clinical and prognostic significance of electrocardiographic changes. Am Heart J 124:455–467 17. Leurent G, Langella B, Fougerou C et al (2011) Diagnostic contributions of cardiac magnetic resonance imaging in patients presenting with elevated troponin, acute chest pain syndrome and unobstructed coronary arteries. Arch Cardiovasc Dis 104: 161–170 18. Rottgen R, Christiani R, Freyhardt P et al (2011) Magnetic resonance imaging findings in acute myocarditis and correlation with immunohistological parameters. Eur Radiol 21:1259–1266 19. Agewall S, Giannitsis E, Jernberg T, Katus H (2011) Troponin elevation in coronary vs. non-coronary disease. Eur Heart J 32:404–411 20. Ricciardi MJ, Wu E, Davidson CJ et al (2001) Visualization of discrete microinfarction after percutaneous coronary intervention associated with mild creatine kinase-MB elevation. Circulation 103:2780–2783 21. Grun S, Schumm J, Greulich S et al (2012) Long-term follow-up of biopsy-proven viral myocarditis: predictors of mortality and incomplete recovery. J Am Coll Cardiol 59:1604–1615 22. Kindermann I, Kindermann M, Kandolf R et al (2008) Predictors of outcome in patients with suspected myocarditis. Circulation 118:639–648 23. McNamara DM, Starling RC, Cooper LT et al (2011) Clinical and demographic predictors of outcomes in recent onset dilated cardiomyopathy: results of the IMAC (intervention in myocarditis and acute cardiomyopathy)-2 study. J Am Coll Cardiol 58:1112–1118 24. Ukena C, Mahfoud F, Kindermann I, Kandolf R, Kindermann M, Bohm M (2011) Prognostic electrocardiographic parameters in patients with suspected myocarditis. Eur J Heart Fail 13:398–405 25. Baccouche H, Mahrholdt H, Meinhardt G et al (2009) Diagnostic synergy of non-invasive cardiovascular magnetic resonance and invasive endomyocardial biopsy in troponin-positive patients without coronary artery disease. Eur Heart J 30:2869–2879 26. Yilmaz A, Kindermann I, Kindermann M et al (2010) Comparative evaluation of left and right ventricular endomyocardial biopsy: differences in complication rate and diagnostic performance. Circulation 122:900–909 27. Elezkurtaj S, Lassner D, Schultheiss HP, Escher F (2014) Vascular involvement in cardiac giant cell myocarditis: a new pathophysiological aspect. Clin Res Cardiol 103:161–163 28. Chu GC, Flewitt JA, Mikami Y, Vermes E, Friedrich MG (2013) Assessment of acute myocarditis by cardiovascular MR: diagnostic performance of shortened protocols. Int J Cardiovasc Imaging 29:1077–1083 29. Friedrich MG (2008) Tissue characterization of acute myocardial infarction and myocarditis by cardiac magnetic resonance. JACC Cardiovasc Imaging 1:652–662 30. Cooper LT, Baughman KL, Feldman AM et al (2007) The role of endomyocardial biopsy in the management of cardiovascular disease: a scientific statement from the American Heart Association, the American College of Cardiology, and the European Society of Cardiology. Circulation 116:2216–2233
