Assessment of myocardial abnormalities in rheumatoid arthritis using a comprehensive cardiac magnetic resonance approach: a pilot study

Rheumatoid arthritis (RA) is a multi-organ inflammatory disorder affecting approximately 1% of the adult general population. A reduction in life expectancy in RA patients is primarily due to an increase of cardiovascular events [1‐4] associated with both ischemic heart disease and congestive heart failure [5, 6]. Importantly, myocardial disease is typically clinically silent [7], only manifesting as myocardial dysfunction after an extended preclinical phase. Myocardial dysfunction may arise from a number of distinct processes, including micro- and macrovascular coronary ischemia, myocardial inflammation (myocarditis), and/or myocardial fibrosis [8], any of which may be active in RA. Histopathologic studies have confirmed an increased prevalence of each of these findings in RA [9‐11], yet are limited by their source (autopsy) and age (most from the 1950s and 1960s).

Considering these issues, an updated appraisal of myocardial disease is warranted in RA, preferably using a technique that is non-invasive and can be serially followed over time for progression, potentially identifying patients at the greatest risk of cardiac-related morbidity and mortality. Although some non-invasive methods, such as conventional and tissue-doppler echocardiography and single photon emission CT (SPECT), may reveal functional abnormalities [12‐16], spatial resolution is not sufficiently accurate to identify the range of potential myocardial abnormalities in RA. Contrast-enhanced cardiac magnetic resonance imaging (MRI) is a non-invasive tool for the diagnosis of ischemic and non-ischemic heart diseases. Contrast agents diffused to the interstitial space will be resorbed into the capillary bed and undergo renal excretion. When the myocardial tissue is damaged, the resorption rate of contrast will be diminished. Specifically, delayed myocardial enhancement (delayed enhancement: DE), particularly when there is a delay in contrast washout within the tissue can indicate myocardial inflammation, fibrosis or myocardial infarction (MI) [17‐20]. Adding pharmacological stress can identify myocardial perfusion impairment in various forms of cardiomyopathy as well as ischemic heart disease. Areas of reduced perfusion within the myocardium (perfusion defects: PD) when pharmacologic stress is applied indicate vascular impairment. Segmental PDs (that is, corresponding to the distribution territory of an epicardial coronary artery) suggest macrovascular involvement, while diffuse or non-segmental PDs are highly suggestive of microvascular impairment [21, 22]. Combining contrast enhanced MRI with pharmacologic stress can aid in identifying underlying pathophysiological features more accurately than each approach alone.

Recently, several investigators have reported on cardiac MRI for the assessment of myocardial abnormality in patients with systemic sclerosis [23‐25] and systemic lupus erythematosus [26]; however, only one report of cardiac MRI has been published in RA [27]. In this case, cardiac MRI was used successfully to differentiate myocarditis from myocardial infarction [27]. We sought to identify myocardial abnormalities in RA patients using contrast enhanced cardiac MRI combined with pharmacologic stress perfusion. Further, we sought to explore the associations of RA disease-related characteristics with cardiac MRI abnormalities.

Characteristics of the 18 participants are summarized in Table 2. Participants were mostly female (78%) with a mean age of 57 ± 10 years. RA duration ranged from 5 months to 16 years (median 2.7 years) with 10 participants (55.6%) categorized as having early RA (disease duration less than three years). While most (88.9%) participants were seropositive for rheumatoid factor (RF), only nine (50%) had anti-cyclic citrullinated peptide (CCP) antibodies. The mean DAS28 for the cohort was 3.96, with six participants (33.3%) falling into the low disease activity category (DAS28 < 3.2) and six participants (33.3%) falling into the high disease activity category (DAS28 > 5.1). TNF inhibitors were prescribed in seven (38.9%) of the participants. The median Framingham 10-year hard CVD event risk score was 2%. CV risk factors were generally low, consistent with the study design in which participants were deliberately selected based on having a low risk CV risk profile.

Myocardial DE was observed in seven patients (38.9%). Among these, enhancement was observed in the middle layer in four patients (57.1%) (Figures 2 and 3), the subendocardial layer in two (28.6%), and the subepicardial layer in one patient (14.3%) (Figure 4). Patterns of enhancement included: patchy in four patients (57.1%) (Figure 2), nodular in two patients (28.6%) (Figure 3), and linear in one patient (14.3%) (Figure 4). Anatomically, enhancement was observed throughout the left-ventricle, occurring in the inferolateral wall (n = 2), lateral wall (n = 1), inferior wall (n = 1), anterior wall (n = 1), and in the septum (n = 1).

None of the participants demonstrated perfusion defects at rest. Perfusion defects under pharmacologic stress were observed in two patients (11.1%). In one (Figure 5), a circumferential patchy perfusion defect in the subendocardial layer was observed and was not associated with delayed enhancement, a pattern consistent with microvascular coronary artery disease without accompanying myocarditis or fibrosis. In the second participant with a perfusion defect under stress, the defect was non-segmental patchy in the subendocardial layer, a pattern consistent with microvascular impairment. This participant also demonstrated nodular delayed enhancement (Figure 3). Of note, microvascular impairment was not observed in the remaining six participants with delayed enhancement.

In this pilot study, we used cardiac MRI to assess myocardial abnormalities in RA patients with no cardiac symptoms. To our knowledge, this is the first investigation to combine both perfusion and delayed enhancement cardiac MRI for assessing cardiac involvement in asymptomatic patients with RA. Our preliminary data lend support for a low prevalence of microvascular impairment in the myocardium with RA patients, and a high prevalence of delayed enhancement, suggesting myocarditis or focal fibrosis. These findings have potential implications for research and clinical care. For research, they suggest that distinct pathologic processes affecting the myocardium may be identified and tracked in RA patients using a safe, non-invasive assessment tool, allowing for a more complete understanding of the RA disease process. For clinical care, early identification of myocardial disease may permit earlier intervention, potentially reducing the impact of myocardial dysfunction on cardiovascular morbidity and mortality in RA patients.

The clinical diagnosis of myocardial disease in RA remains problematic. Endomyocardial biopsy (EMB) may increase the frequency of detection; however, the sensitivity of EMB may be as low as 35%, primarily because of patchy involvement or focality of myocarditis [36]. Additionally, EMB carries a 1 in 250 risk of perforation and a 1 in 1,000 risk of death, even in experienced hands [37]. Therefore, less-invasive tools are needed to diagnose and track cardiac abnormalities. Recent advances in cardiac MRI enable dynamic first-pass perfusion MRI of the entire left-ventricular myocardium with improved image quality and higher spatial resolution compared to SPECT. Pharmacological stress perfusion MRI has been shown to be an accurate established method for detecting coronary artery disease (CAD) [38]. Fenchel et al. reported that stress examination increased the perfusion differences of MRI between normal and ischemic myocardial area compared with rest imaging [39]. Klem et al. [40] reported that the addition of delayed enhanced MRI to the stress-rest perfusion examination would help distinguish true perfusion defects from artifacts and improve test reliability to the point that rapid visual interpretation could be performed with high accuracy. Moreover combining stress perfusion and delayed enhanced MRI is an accurate method for characterizing pathologic features [33, 35].

Other non-invasive modalities for myocardial imaging are available, but not as sensitive or versatile as MRI. Conventional echocardiography allows for estimation of global and regional myocardial function, but cannot distinguish specific etiologies of dysfunction with accuracy. In addition, echocardiography requires careful standardization, both for acquisition and interpretation, a limitation that is considerably less for MRI. For these reasons, in our study, echocardiography was performed solely to exclude patients with low ejection fraction and/or diastolic dysfunction and not compared to the cardiac MRI studies. SPECT has been used in two published papers in patients with RA [16, 41]. Banks et al. reported, using adenosine stressed SPECT, that ischemic heart disease was twice as common and more often silent in 67 RA patients compared with 37 matched controls with osteoarthritis [41]. Noninvasive assessment of myocardial blood flow by adenosine stressed SPECT showed reversible ischemia and diffusely poor myocardial perfusion, suggesting myocardial microvascular disease [16]. Although myocardial perfusion SPECT is widely used in clinical practice and the high sensitivity of SPECT has been reported, the spatial resolution of this modality is limited and the radiation dose is problematic. Furthermore, Ishida et al. [42] reported that pharmacological stress perfusion MRI was superior to SPECT for the assessment of myocardial ischemia in the general population.

Non-segmental PD during pharmacological stress was seen in 2 out of 18 patients (11%). Non-segmental PD, not corresponding to any epicardial coronary artery distribution, is highly suggestive of microvascular impairment. Coronary arteritis is seen in about 20% of RA patients at autopsy [11]. Inflammation with edema of the intima of the artery may lead to severe narrowing or occlusion of its lumen, to necrosis, and angina or infarction [43]. Nevertheless, myocardial infarction secondary to coronary arteritis is rare in patients with RA [9]. In contrast to atherosclerosis, rheumatoid vasculitis frequently involves intramyocardial arteries, indicating microvascular impairment [9, 11, 44]. Raza et al. reported reversible ischemia and diffusely poor myocardial perfusion in one patient with RA using noninvasive assessment of myocardial blood flow by adenosine stressed SPECT [16]. Repeat assessment after intensive immunosuppression therapy revealed increased myocardial perfusion. Coronary angiography revealed no significant atheroma, suggesting that myocardial microvascular disease was responsible for the ischemia. Pharmacological stress perfusion MRI is a non-invasive tool to assess the microvascular circulation of myocardium [21, 22]. And cardiac MRI can be repeatedly performed for the assessment of therapeutic effect or follow-up examination, because there is no radiation exposure. Our findings would support the prospective study of the effect of RA treatment, and associated change in joint disease activity, on myocardial perfusion defects.

Five patients (28%) had areas of DE in the middle or subepicardial layer of the left ventricle, suggesting myocarditis. The detection of DE, suggesting myocarditis or fibrosis, in 28% of our study participants is striking, particularly considering that all of the patients were asymptomatic, had no risk factors other than RA for these findings, and had normal conventional echocardiographic assessments. Myocarditis is a rare but recognized form of rheumatoid cardiac disease and is a known cause of congestive heart failure [45]. Rheumatoid myocarditis can assume a granulomatous form that is considered specific for RA, or a nonspecific form that may also be observed in other disorders. When present, the granulomas show a predilection for the left ventricle and are morphologically identical to the subcutaneous nodules of RA [9]. In contrast, the nonspecific inflammatory pattern involves the collagenous interstitium of the heart, and is composed predominantly of lymphocytes, plasma cells, and histiocytes [9, 46]. In cardiac MRI, relatively greater gadolinium contrast enhancement in the myocardium has been reported in patients with active myocardial inflammation by histopathology [47]. There may be either focal or patchy diffuse enhancement of the left ventricular (LV) myocardium, typically with a subepicardial or midwall distribution [48‐50]. However, isolated subendocardial distribution was not demonstrated. The LV lateral free wall is more often involved, followed by the ventricular septum [51]. Consistent with this, in our study, two out of five cases with myocarditis pattern had DE in the lateral free wall, and two had DE in the septum. In a study of individuals without RA, the extent of myocardial enhancement was roughly correlated with LV dysfunction, and it may decrease during healing and essentially disappear after recovery [47].

Two patients in our study (11%) had a subendocardial nodular DE, suggesting focal myocardial infarction. RA patients are less likely to report symptoms of angina and are more likely to experience unrecognized myocardial infarction and sudden death [52]. Interestingly, RA patients were more likely to be hospitalized for acute myocardial infarction or to have experienced unrecognized myocardial infarctions prior to their diagnosis of RA [52]. In myocardial infarction, cardiac MRI typically shows subendocardial DE [48, 53]. In contrast, myocarditis shows a characteristic MRI pattern of contrast enhancement, which originates from the epicardium, sparing the subendocardial layer [53]. Our patients had no cardiac symptoms, no history of cardiac diseases, and normal ECG studies; therefore these DE findings may correspond to myocardial scarring from unrecognized myocardial infarctions. Barbier et al. [54] reported that 49 of 259 (19.8%) randomly chosen 70-year-old subjects had MI scars in delayed enhanced MRI. On the other hand, only one unrecognized myocardial infarction was found with delayed enhanced MRI in 298 subjects with a mean age of 50 years [55]. The higher frequency of findings suggesting myocardial scarring in our study compared to the above published report is compelling, but deserves study in a larger cohort of RA patients.

Interestingly, we detected a significant association between RA disease activity and inflammation and myocardial DE in cardiac MRI. In several autopsy studies of RA patients, those with myocarditis were noted to have active arthritis, endocarditis, or pericarditis and vascular changes, which strongly suggests that myocarditis is associated with active forms of RA [9‐11]. Longitudinal follow-up is required to explore whether changes in disease activity are associated with corresponding changes in DE and the contribution of RA pharmacotherapies to dynamic changes in DE.

Our study had limitations. This was a pilot study and the size is too small for definitive conclusions. The lack of non-RA controls limits our ability to strongly conclude that our findings are unique to RA patients; however, the much higher frequency of myocardial abnormalities observed in the present study compared to studies conducted in non-RA patients with similar demographic characteristics and CVD risk factors is suggestive of a true disease effect. We did not perform T2-weighted MR images, which are more reliable in differentiating edema induced from acute myocarditis from fibrosis [56]. Nonetheless our preliminary data lend support for a high prevalence of myocarditis with asymptomatic RA patients.

To the best of our knowledge, our study is the largest patient series to date to report cardiac abnormalities in asymptomatic patients with RA using pharmacological stress and rest perfusion, and delayed-enhanced MRI. The detection and histological confirmation of cardiac involvement in RA is often difficult because of a lack of symptoms and invasive diagnostic method such as EMB. Comprehensive cardiac MRI might be considered to be a useful and less-invasive diagnostic tool for assessing cardiac involvements in asymptomatic patients with RA. Further studies of this diagnostic tool should be evaluated to discuss its utility.