PATIENTS AND METHODS
Two hundred forty-six patients (mean ± SD age 46 ± 6 years, 187 women and 59 men) with CTDs referred by rheumatologists, internists, or lung physicians for CMR over 8 years were retrospectively selected on the basis of excellent CMR image quality and a normal echocardiography study. A normal echocardiography study was defined by normal ejection fraction (EF) of both ventricles and a lack of wall motion or systolic thickening abnormalities. The presence of mild, trivial mitral regurgitation was not considered as an abnormality. Forty percent of patients with normal echocardiograms were judged to have technically limited studies because of a bad acoustic window, obesity, and/or lung disease.
Twenty-seven patients with inadequate CMR quality because of motion or respiratory artifacts and 32 patients with heart failure documented by echocardiography were excluded from the study. One hundred forty-six of 246 patients presented with recent TCS, including shortness of breath, chest pain, palpitations, recent echocardiography changes, and/or an increase in troponin 1 (group A). This group included 9 patients with inflammatory myopathy (IM), 35 with sarcoidosis, 30 with systemic sclerosis (SSc), 14 with systemic lupus erythematosus (SLE), 10 with rheumatoid arthritis (RA), and 48 with small to medium vessel vasculitis (15 with microscopic polyangiitis, 20 with granulomatosis with polyangiitis [Wegener's] [GPA], and 13 with eosinophilic granulomatosis with polyangiitis [Churg-Strauss] [EGPA]). The remaining 100 patients, who presented with ATCS, included 25 patients with RA, 20 with SLE, 20 with sarcoidosis, 15 with SSc, 10 with IM, and 10 with GPA (group B). ATCS included easy fatigue, diffuse thoracic pain, and feeling unwell. The CMR examination was performed within 10–20 days after the onset of symptoms and echocardiography assessment. Patients' clinical characteristics are shown in Tables 1 and 2.
Table 1. Clinical characteristics of patients with connective tissue diseases at the time of cardiovascular magnetic resonance*
|Age, mean ± SD years||45 ± 3||41 ± 3||50 ± 2||32 ± 4||55 ± 3||43 ± 2|
|Sex, no. F/M||46/12||42/13||38/7||32/2||20/15||9/10|
|Diabetes mellitus, no.||5||4||3||5||6||0|
|CRP, mean ± SD mg/liter (normal range 0–5)||10 ± 3||11 ± 3||9 ± 5||10 ± 4||9 ± 2||6 ± 3|
|ESR, mean ± SD mm/hour (normal range 0–10)||17 ± 3.4||19 ± 5||12 ± 8||17 ± 7||20 ± 5||10 ± 7|
|Family history of IHD, no.||5||4||3||3||6||1|
|Disease duration, mean ± SD years||3 ± 2||4 ± 2||5 ± 3||2 ± 1||4 ± 3||3 ± 1|
|ACE/angiotensin II receptor||50||30||10||30||35||0|
|Current nonbiologic DMARDs||80||0||90||90||50||50|
|Any current biologic DMARD||0||0||0||0||60||0|
Table 2. Data of connective tissue disease patients with positive cardiovascular magnetic resonance*
|SRC (n = 13)||41 ± 4||–||11 TCS, 2 ATCS||157.7 ± 44.9||47.5 ± 16.4||58.5 ± 12||2 ± 0.8||6 ± 1||6 diff subendo, 7 subepi|
|RA (n = 10)||53 ± 6.5||1+, 9−||7 TCS, 3 ATCS||166 ± 26||69.4 ± 17.7||57.6 ± 7||2 ± 0.8||7.7 ± 2.8||3 diff subendo, 4 subepi, 1 intra, 2 trans|
|SLE (n = 12)||37 ± 4||–||9 TCS, 3 ATCS||166.2 ± 44||71.8 ± 30||57.8 ± 7.2||1.9 ± 0.5||8.3 ± 2.2||2 diff subendo, 6 subepi, 4 trans|
|EGPA (n = 1)a||38||+||TCS||174||100||43||3.5||8||Diff subendo|
|GPA (n = 1)a||45||–||ATCS||210||122||42||2.2||7||Subepi|
|SSc (n = 18)||46.5 ± 5.4||–||14 TCS, 4 ATCS||150.2 ± 30||53.4 ± 13.6||62.6 ± 4.6||1.85 ± 0.5||6.5 ± 1||5 diff subendo, 13 subepi|
|IM (n = 7)||39 ± 2||–||5 TCS, 2 ATCS||144.4 ± 21||56.8 ± 12||61 ± 4||1.68 ± 0.16||6.5 ± 2.3||7 subepi|
CMR was performed by a 1.5T scanner (Signa CV/i, GE Medical Systems) using echocardiography-triggered steady-state, free precession breath-hold cine (echo time [TE] 1.6 msec, repetition time [TR] 3.2 msec, flip angle 60°) in long-axis planes and sequential 8-mm short-axis slices (3-mm gap) from the atrioventricular ring to the apex. STIR T2-weighted images (triple inversion recovery; TE 60 msec, TR 2 × R-R interval, inversion time 170 msec, slice thickness 20 mm, flip angle 180°, pixel size 2.3 × 1.3 mm) were acquired in short-axis planes for edema imaging. Finally, LGE images were acquired 10 minutes after intravenous gadolinium diethylenetriaminepentaacetic acid (Schering; 0.2 mmoles/kg) in identical short-axis planes using an inversion-recovery gradient-echo sequence for fibrosis detection. Inversion times were adjusted to the null normal myocardium (typically 320–440 msec; pixel size 1.7 × 1.4 mm). Ventricular volumes and function were analyzed using specialized software (Medis) ([10-13]).
CMR was evaluated independently by 2 experienced interpreters blinded to clinical data (SM, KB). Scans were reviewed for ventricular volumes/function using images from the steady-state free precession sequence. The T2 ratio was calculated by the ratio of myocardial to skeletal muscle signal intensity from STIR T2-weighted images (). Finally, LGE images were assessed for subendocardial/transmural enhancement in the distribution of a coronary artery compatible with myocardial infarction (), for intramural/subepicardial enhancement compatible with myocarditis (), and for diffuse subendocardial fibrosis compatible with vasculitis (). Patients were further subclassified into an acute (T2 ratio >2) or nonacute (T2 ratio ≤2) stage ([14, 15]). A combination of a T2 ratio >2 with positive LGE was considered positive for an acute myocardial lesion, whereas a T2 ratio ≤2 with positive LGE was considered positive for a chronic myocardial lesion ([7, 8]). According to the location and morphology of LGE, CTD patients were categorized as having 1) diffuse subendocardial LGE, indicative of subendocardial vasculitis; 2) intramural/subepicardial LGE not following the distribution of coronary arteries, indicative of myocarditis; and 3) subendocardial/transmural LGE following the distribution of coronary arteries, indicative of myocardial infarction. The possibility of amyloidosis in patients with diffuse subendocardial fibrosis was excluded because the classic presentation of diffuse subendocardial amyloidosis is best visualized on early (3–5 minutes) delayed imaging, but would not appear with the methods used here; additionally, it is usually accompanied by dilated atria, small hypertrophied ventricles, and thickening of the interatrial septum. None of these findings were seen in any of our patients. Furthermore, if cardiac amyloidosis is suspected and LGE imaging is difficult due to diffuse infiltration, absolute quantification of myocardial distribution volume of an extracellular agent as an indication of extracellular matrix size may have additional value. However, myocardial biopsy still remains the gold standard for amyloidosis documentation (). Scans with completely normal range volumes and function with no LGE/T2 abnormalities were considered as normal.
In T2-weighted imaging, the signal ratio was measured from the region of interest covering the left ventricular (LV) myocardium as well as a skeletal muscle in the same slice. To assess the contrast-enhanced images (LGE), all short-axis slices from the base to the apex were inspected visually to identify areas of normal (completely nulled) myocardium. Mean ± SD signal intensity was derived and a threshold of 0.4 SD exceeding the mean was used to define areas of LGE. Summing the planimetered areas of LGE in all short-axis slices yielded the total volume, which was also expressed as a proportion of the total LV myocardium. Cine images were used for the evaluation of LVEF. LV endocardial borders were outlined on the end-systolic and end-diastolic short-axis view images covering the entire LV. LVEF was calculated as follows: LVEF = [(volume at end diastole − volume at end systole)/volume at end diastole].
All measurements were expressed as the mean ± SD. Statistical significance of the differences was investigated using the unpaired Student's t-test. Statistical significance was considered for P values less than 0.05.
Here we described our experience in CTD patients with normal echocardiography referred for evaluation of TCS or ATCS. By combining edema and fibrosis imaging, we identified heart lesions that could influence risk stratification, undetected by the routine noninvasive cardiac evaluation (Table 2). Classifying patients according to the amount of LGE, we found that 49 of 62 CMR-positive CTD patients had LGE >5% of LV. According to recent data, these patients were at a higher risk for future cardiac events compared to those without LGE or with LGE ≤5% of LV ().
The main findings of the present study show that 1) CMR identified myocardial edema, indicative of acute myocardial lesions, in 3.25% of CTD patients; 2) CMR detected myocardial fibrosis in 25.2% of CTD patients with either TCS or ATCS and normal echocardiography; 3) CMR assessed impaired LVEF in 9 of 47 patients with TCS and 3 of 15 patients with ATCS undiagnosed by echocardiography; and 4) using the combination of edema and fibrosis imaging, CMR identified heart pathophysiology, undetected by the routine noninvasive evaluation.
In the past, edema could not be used as a diagnostic tool because even histology failed to provide reliable information on its presence. Extensive studies have confirmed a close correlation between T2 and edema (). Adding T2 to a standard CMR protocol (function, perfusion, and scar) increased the specificity, positive predictive value, and overall accuracy for detection of an acute coronary syndrome from 84% to 96%, 55% to 85%, and 84% to 93%, respectively (). Furthermore, using LGE, CMR not only detects myocardial infarction in as little as 1 cm3 of tissue, substantially less than other in vivo methods, but also has excellent agreement with histology in animal and human studies ([20, 21]). Finally, CMR was also proven useful in detecting small myocardial scars and diffuse subendocardial fibrosis that were missed by other imaging techniques (). Even a small area of LGE (<2% of LV mass) was associated with a >7-fold increase in risk for a major adverse cardiac event ().
According to previous studies, CMR identified cardiac involvement in vasculitis (), sarcoidosis (), SLE ([24, 25]), SSc ([26, 27]), RA (), and EGPA () with or without abnormal echocardiography. However, to our knowledge, this is the first report presenting CMR findings in CTD patients with recent onset of cardiac symptoms/signs and normal echocardiography.
There are many studies emphasizing the role of LGE in the diagnosis of cardiac sarcoidosis (). In all of our sarcoidosis patients with acute heart involvement, prompt immunosuppressive treatment was started; however, 1 patient developed episodes of ventricular tachycardia during the next 8 months and a defibrillator was implanted. Our findings were in agreement with previous studies supporting the finding that CMR reveals early lesions undetected by standard assessment. Additionally, CMR may potentially motivate early cardiac treatment in the future to avoid evolution to overt heart failure, according to the American College of Cardiology (ACC)/American Heart Association (AHA) guidelines (). However, at the moment, although there are abundant observational data that LGE may identify situations of increased risk, there are as yet no randomized trials using LGE to institute or change treatment.
In SLE, premature damage occurs in both macro- and microvasculature () and relates to disease activity and duration (). Furthermore, pathology studies have demonstrated the high incidence of vasculitis (). A comparison between echocardiography and CMR in SLE showed CMR superiority over echocardiography (). Additionally, echocardiography in SLE can be misleading regarding the etiology of myocardial lesions (). There is evidence that an imaging approach combining T2 and LGE is useful to assess myocardial involvement in SLE (). Furthermore, due to subclinical presentation of SLE myocarditis and the serious limitations of EMB, CMR may be the best alternative for diagnosis (). In our SLE patients with TCS, 1 had acute subendocardial vasculitis and 8 had myocardial fibrosis due to diffuse subendocardial vasculitis, myocarditis, and past transmural myocardial infarction. The patient with acute vasculitis had a recent history of disease flare. In the SLE patients with ATCS, CMR identified 1 patient with clinically silent, acute myocarditis verified by EMB, and 2 with myocardial fibrosis due to past myocarditis and past myocardial infarction, unnoticed by both the patients and physicians.
In SSc, vasculitis can present with concurrent pleuropericarditis and may predispose to subsequent renal failure ([27, 28]). Two of our SSc patients with acute vasculitis also had pleuropericarditis and responded well to steroid treatment. In SSc, fibrosis is the result of microvascular disease and can appear at any stage of the disease (). Different CMR fibrotic patterns have been described previously ([28, 37]). These patterns were also found in our SSc patients, although their routine, noninvasive evaluation was normal.
Myocarditis is a rare but well recognized form of RA () and can be associated with secondary cardiomyopathy (). Corticosteroids and antimalarials, commonly used in the treatment of RA, rarely have been associated with cardiac injury (). Secondary vasculitis also has been described in RA. In a pathology study of 161 RA patients, systemic vasculitis was observed in 22.4% and the most frequently involved organ was the heart (66.7%) (). However, clinically detected vasculitis in RA is rather low. Finally, RA patients are twice as likely to develop myocardial infraction, irrespective of age, history of prior cardiovascular disease events, and traditional risk factors (). In our RA patients with TCS, we identified 1 patient with acute myocarditis and 6 with fibrosis due to past myocarditis, diffuse subendocardial fibrosis, and past myocardial infarction. Our findings were in agreement with previous studies identifying these types of pathology in RA ([38-41]). The evaluation of those with RA and ATCS revealed 1 patient with acute myocarditis and 2 with fibrosis due to past myocarditis and diffuse subendocardial vasculitis.
In a study of 32 EGPA patients in remission who were previously unaware of cardiac involvement, CMR revealed a 62% prevalence of cardiac involvement compared with 3% in controls (). In our vasculitis patients with TCS, 1 patient with EGPA had a clinical presentation mimicking acute myocardial infarction. Therefore, she underwent a coronary angiography with normal results. Although guideline recommendations about CMR in acute coronary syndromes are still missing, if the CMR evaluation was included as the first noninvasive technique in the diagnostic algorithm, an unnecessary interventional coronary angiography could be avoided. In vasculitis patients with ATCS, CMR detected a silent, acute myocardial inflammation in 1 GPA patient, unnoticed by clinical and echocardiographic evaluation. It is rather strange that we found only 2 vasculitis patients with abnormal CMR findings. However, there are some possible explanations for this. The usually dramatic clinical presentation of these patients orders the prompt coronary angiography, bypassing any sophisticated assessment, such as CMR. Additionally, after heart involvement, most of the patients have impaired cardiac function and therefore were not included in this study.
Finally, in IM, cardiac involvement is an important cause of mortality. In previously published studies by our group, we assessed that CMR can unveil both acute and chronic silent myocardial lesions in IM ([43, 44]). In the current study, CMR identified fibrotic lesions indicative of past myocarditis in 7 IM patients (5 with TCS and 2 with ATCS).
Collectively, comparing patients with positive and negative CMR, we found that those with positive CMR had significantly higher inflammatory indices, although their LV function was not significantly different. This supports that LV dysfunction is a late event; therefore, inflammatory indices and CMR findings are more important for decision making of heart involvement in CTDs. Additionally, the lack of significant differences in risk factors between CTD patients with positive and negative CMR supports the multifactorial nature of heart involvement in CTDs.
Notably, in patients with normal CMR, the application of stress techniques revealed that 20% of the patients had evidence of coronary artery disease. This finding suggests that a noninvasive stress evaluation also should be included in the diagnostic evaluation of CTD patients with persistent TCS or ATCS and normal resting studies, either by echocardiography or by CMR, to exclude the possibility of coronary artery disease.
Based on the findings shown here, we propose that the additive value of CMR in the evaluation of CTD patients is focused on the detection of cardiac lesion acuity even in cases with subclinical presentation, as well as on the detection of myocardial fibrosis in CTD patients who are considered normal by echocardiography study. The LGE location and distribution are of value to assess the pathophysiology of cardiac lesions and schedule further diagnostic evaluation. More importantly, the extent of LGE is of great value for cardiac risk assessment. A study of 137 patients evaluated for ICD implantation proved that in patients with an LVEF >30%, the presence of LGE >5% of LV identifies a high-risk cohort similar in risk to those with an LVEF ≤30%. Conversely, in patients with an LVEF ≤30%, minimal or no scarring identifies a low-risk cohort, similar to those with an LVEF >30% (). Along these lines, positive CMR results may potentially motivate cardiac treatment in the future to avoid evolution to overt heart failure, according to ACC/AHA guidelines (), but randomized controlled trial data to support such practice are currently lacking.
To our knowledge, there are no studies comparing CMR data of CTD patients with ATCS and TCS. However, according to published CMR experience, it seems that myocardial inflammation is more common than was initially thought and in most cases, it is accompanied by a normal LVEF ([36, 45]). This is an important finding suggesting that myocardial involvement also can be found in CTD patients with ATCS and therefore should not be underestimated by the clinicians, and a normal LVEF and a lack of typical echocardiography findings and/or abnormal troponin 1 cannot exclude heart lesions. Since it is a technique able to detect early tissue changes before any LV dysfunction takes place, CMR could be considered as an important tool for early detection of heart involvement in CTD patients.
Furthermore, by classifying CTD patients according to LGE >5% or ≤5% of LV, we assessed that the majority of our CTD patients were at a higher risk for future cardiac events than patients with the same LVEF without LGE and/or with LGE ≤5% of LV (). This unique information, given only by CMR, changes the heart assessment algorithm from functional evaluation, which is a late finding, to tissue characterization, which is a very early finding.
Finally, we should also mention that CMR detected impaired LVEF in 9 CTD patients with TCS and 3 with ATCS, unnoticed by echocardiography. This was not unusual because echocardiography is an operator-dependent technique, influenced by both the operator's experience and limitations of the acoustic window ().
At the moment, strict criteria about the selection of CTD patients for CMR have not been established. The current experience from cardiology supports that CMR is the best noninvasive technique to detect myocardial edema in acute myocardial inflammation () and myocardial fibrosis in myocarditis, myocardial infarction, and/or cardiomyopathies ([15, 47, 48]). This experience was used by the referring physicians, who were aware of the CMR application and believed that it could clarify heart pathophysiology in CTD patients. Our findings support the necessity of CMR teaching not only in cardiologists, but also in other subspecialties of internal medicine.
Our results followed the current experience, already documented in cardiology, by proving the CMR diagnostic value in distinguishing disease acuity and patterns of myocardial fibrosis. Therefore, CMR may play an important role in the diagnostic evaluation of CTD patients. However, we emphasize that this is a descriptive case series, which is potentially useful for presenting the diagnostic and therapeutic challenges posed by the utilization of CMR in patients with CTDs, but it is not at this stage sufficiently generating compelling data to lead to practice changes. Further prospective, multicenter studies are needed to establish the full clinical implications of CMR in CTD patients.
Some limitations of the study include 1) a lack of internationally accepted criteria for the selection of CTD patients for CMR; only physicians aware of CMR advantages were referred to us and therefore, there is a significant bias that does not allow us to identify the true incidence of heart involvement in CTD patients; 2) a high likelihood for selection bias, since patients undergoing echocardiography and CMR at a referral center usually differ from the general CTD population; 3) patients' detailed echocardiography data were not available for a face-to-face comparison with CMR and new echocardiography techniques have not been applied; and 4) a lack of EMB and short- to long-term followup in the majority of our CTD patients.
In conclusion, tissue characterization by CMR in CTD patients with either TCS or ATCS and a normal LV can identify different patterns of myocardial lesions, including myocarditis, diffuse subendocardial vasculitis, and myocardial infarction. In parallel, it can assess acuity of heart involvement and classify patients as low or high risk according to fibrosis amount. This unique noninvasive information may have important clinical implications in early and accurate assessment of cardiac lesions in CTD patients that require further assessment in research protocols developed specifically for this purpose.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Mavrogeni had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Mavrogeni, Sfikakis, Gialafos, Bratis, Karabela, Stavropoulos, Spiliotis, Sfendouraki, Panopoulos, Bournia, Kolovou, Kitas.
Acquisition of data. Mavrogeni, Sfikakis, Gialafos, Bratis, Sfendouraki, Panopoulos, Bournia, Kolovou.
Analysis and interpretation of data. Mavrogeni, Sfikakis, Gialafos, Bratis, Karabela, Stavropoulos, Spiliotis, Sfendouraki, Panopoulos, Bournia, Kolovou, Kitas.