Cardiac involvement in eosinophilic granulomatosis with polyangitis: a retrospective study in Chinese population

Background: Cardiac involvement in EGPA indicates poor prognosis and high mortality, while few data about cardiac involvement of EGPA in Chinese population are available. We conducted this study to figure out the clinical characteristics and overall outcome of EGPA patients with cardiac involvement in Chinese population. Methods: We retrospectively collected the clinical data of 83 patients diagnosed with EGPA and analyzed the differences between patients with and without cardiac involvement. Results: The prevalence of cardiac involvement of EGPA in this cohort was 27.7%. Compared with those who without cardiac involvement, EGPA patients with cardiac involvement tended to have younger onset age (mean ± SD, 38.4 ± 10.5 vs. 42.1 ± 15.9 years, p=0.039), higher eosinophil count [median (IQR), 5810 (4020-11090) vs. 2880 (1530-6570) n/μL , p=0.004)], higher disease activity [median (IQR), 20 (16-28) vs. 15 (12-18), p=0.001)] and poorer prognosis (FFS≥1, 100% vs. 38.3%, p=0.001). The most common cardiac manifestation was chest pain (56.5%), and 43.5% of patients were asymptomatic but cardiac abnormalities could be detected by cardiac examinations. With appropriate treatment, the overall outcome of EGPA patients with cardiac involvement in our cohort turned out to be good, with 3 ( 13% ) patients died at acute phase and no patient died during follow-up. Conclusions: Cardiac involvement was common in EGPA and was associated with younger onset age, higher eosinophil count, higher disease activity and poorer prognosis. Comprehensive cardiac examinations and appropriate treatment are essential to improve the prognosis of cardiac involvement.

Methods treatment and follow up were recorded. Laboratory findings at diagnosis and after treatment were recorded retrospectively, such as complete blood count, serum immunoglobulin E (IgE) levels, hypersensitive C-reactive protein (hsCRP), erythrocyte sedimentation rate (ESR), ANCA test, myocardial enzyme levels, N-terminal prohormone of brain natriuretic peptide (NT-proBNP), liver function test, serum creatinine and so on. Imaging examinations included electrocardiogram (ECG), 24 hours dynamic electrocardiogram (DCG), ultrasound scan, magnetic resonance imaging (MRI), computed tomography (CT) scan, neuroelectrophysiological examination etc. Plus, some patients underwent lung/skin/renal/heart biopsy upon indication and the pathology results were also recorded.

Statistical analysis
Continuous variables were shown as mean ± standard deviation (SD) or median (interquartile range, IQR), while categorical variables were shown as numbers (percentage). Continuous variables were analyzed using Student's t test or Mann-Whitney U test, and categorical variables used chi-square test or Fisher's exact test as required. Type-1 error was decided to be 5%. Statistics were analyzed via a SPSS version 23.0 software. There were no differences between the two groups considering organ involvement and the level of hsCRP, ESR as well as serum IgE. These data are shown in Table 1 in detail.

Demographic data and clinical manifestations
All patients with cardiac involvement underwent ECG test: 4 (17.4%) patients were found to have sinus tachycardia and 5 (21.7%) have ST-segment or T wave abnormalities. Among them, one (4.3%) patient who had ST-segment elevation and progressed to cardiac arrest survived due to timely cardiopulmonary resuscitation. One (4.3%) patient was found to have third-degree atrioventricular block and needed the support of temporary pacemaker. Ten (43.5%) patients did not present with ECG abnormalities, but had elevated myocardial enzymes or echocardiographic changes. Thirteen patients (56.5%) had elevated cardiac Troponin I (cTn I) levels and elevated NT-proBNP levels were found in 12 patients (52.2%).
Echocardiography was performed in all patients with cardiac involvement, and 21 (95.5%) patients had different kinds of cardiac disorders. Six (26.1%) patients were found to have decreased LVEF [median (IQR), 34.1% (31%-46%)], among which one patient needed extracorporeal membrane oxygenation (ECMO) support because of cardiogenic shock caused by extremely low LVEF (11%). Eleven (47.8%) patients had diastolic function abnormality or segmental dyskinesia, and 6 (26.1%) patients had mitral or tricuspid insufficiency. Mild to moderate pericardial effusion was observed in 7 (30.4%) patients. Only 2 (8.7%) patients underwent cardiac MRI examination, and both were found to have late gadolinium enhancement of the myocardium and one had decreased LVEF (43.8%) which was not identified by the echocardiography. Six patients had coronary artery CTA examination, none of which had major abnormalities that can account for cardiac involvement. These findings can be seen in Table 2.

Disease activity, treatment, outcome and follow-up
In contrast with patients without cardiac involvement, patients with cardiac involvement had higher As for the outcome of EGPA patients with cardiac involvement, one patient died of infectious shock due to gut perforation, one patient died of multi-organ failure and one died of cerebral hernia. Twenty (87%) patients achieved remission after treatment, with LVEF improved, myocardial enzymes decreased or ECG become normal.
The median follow-up period of patients with cardiac involvement was 38 months (IQR, 25-69.5), with one patient lost to follow up. During the follow up, 3 (15.8%) patients relapsed and were admitted into the hospital again. The relapsed organs were heart, ear and lung respectively, and all patients achieved remission again after treatment. No relationship was observed between the the presence of relapse and disease activity as described by BVAS or FFS at baseline. One (5.3%) patient progressed to heart failure, 15 (78.9%) patients achieved long term remission and no patients died during follow up. These findings are shown in Table 3.    [1,31], and eosinophilia played a more significant role than ANCA in cardiac involvement of EGPA [15]. In our study, we also noted a higher level of eosinophil count in the group of cardiac involvement, which was consistent with the findings of previous series [15,25]. Though not significant, the percentage of ANCA positivity was lower in the group of cardiac involvement in this study. A number of patients had taken glucocorticoids treatment before diagnosis, which may account for this phenomenon.
The diagnosis of cardiac involvement was made upon the combination of clinical manifestations, ECG, myocardial enzyme levels, coronary artery CTA, echocardiography and cardiac MRI. Chest pain was a common syndrome caused by cardiac involvement in our series, which is in accordance with the series of Vinit, J., et al. [15]. However, almost half of the patients were asymptomatic or had normal ECG, while there were cardiac abnormalities detected by echocardiography or cardiac MRI. As a result, an overall evaluation of heart condition should be taken to exclude cardiac involvement despite normal ECG and absence of cardiac manifestations, as recommended by previous studies [14,17], since cardiac involvement in EGPA indicates poor prognosis and high mortality.
EGPA patients with cardiac involvement tended to have higher FFS as previously reported[9, 25], which was also noted in our cohort, suggesting a poor prognosis in patients with cardiac involvement.
It was also identified in our study that patients with cardiac involvement had higher BVAS, which predicted higher disease activity and severe organ damage. However, the application of glucocorticoids and the combination of immunosuppressive agents such as CTX in severe cases significantly improved the outcome of EGPA [5,19]. Thus, all patients with cardiac involvement in our study received the induction treatment of high dose glucocorticoids combined with CTX and achieved remission in acute phase. Consequently, the overall outcome of patients with cardiac involvement in our cohort turned out to be good. Therefore, early diagnosis and appropriate treatment are indispensable to prevent the acceleration of cardiac involvement in EGPA [17,32,33].
This is the first study with a relatively large sample concerning the cardiac involvement of EGPA patients in Chinese population according to our knowledge. We identify that patients with cardiac involvement have distinct clinical characteristics from those without cardiac involvement in this cohort, which may provide some meaningful information for physicians during clinical practice. What's more, our study also suggests that cardiac involvement predicts poor prognosis and higher disease activity in EGPA patients. Comprehensive cardiac examinations and appropriate treatment are essential to make early diagnosis and improve the outcome of cardiac involvement of EGPA patients in our study.
However, there are several limitations of our study. First of all, this is a cross-sectional and retrospective study that may bring in some biases, which may have an influence on the results. Lack of endomyocardial biopsy to confirm the diagnosis of cardiac involvement is another limitation of our study.

Conclusion
In conclusion, cardiac involvement was common in this EGPA cohort and mainly presented with chest pain. Compared with those without cardiac involvement, patients with cardiac involvement tended to have association with younger onset age, higher eosinophil count, higher disease activity and poorer prognosis. Comprehensive cardiac examinations are important to make early diagnosis. High-dose glucocorticoids combined with CTX might be the preferred treatment which can prevent the acceleration of cardiac involvement effectively.

Ethics approval and consent to participate
The study was approved by the Ethics Committee of Peking Union Medical College Hospital. Written informed consent couldn't be obtained due to the retrospective nature of this study.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.