Heart involvement in Churg–Strauss syndrome: Retrospective study in French Burgundy population in past 10years

Results From a cohort of 31 patients, we found 20 with heart lesions. When heart lesions were present, we noted fewer initial symptoms of digestive disorders ( p < 0.05), lower levels of lung infiltrates and fewer anti-MPO pANCA ( p < 0.05). Heart lesions were linked to CSS in 75% of cases. Their patients were thus younger than those in the other cardiac patients ( p < 0.05), were more likely to have clinical manifestations of heart involvement at diagnosis, were less likely to have lung infiltrates on the X-ray at diagnosis and during flare-ups and less likely to have lung abnormalities on X-rays during flare-ups ( p < 0.05) and higher level of leucocytes and eosinophils at diagnosis. Conclusion Heart lesions directly attributable to CSS are frequent, severe and probably underestimated. A specific physiopathology that is not mediated by ANCA seems to be involved in the genesis of CSS-related heart lesions. Keywords Vasculitis Churg–Strauss syndrome Epidemiology Prevalence Incidence Heart 1 Introduction Churg–Strauss syndrome (CSS) is vasculitis associated with pANCA (anti-neutrophil cytoplasmic antibodies) and is very often complicated by heart involvement [1] . In their series of autopsies, Churg and Strauss found more than 50% of cadavers with heart lesions [2] . The three layers of the heart may be affected. The clinical manifestations are therefore varied and are represented in the Five Factor Score (FFS), which evaluates the prognosis in ANCA-related vasculitis [3] . An FFS of 1 or 2 is an indicator of a poor outcome [4] and often justifies intensification of the treatment [1,3,5–7] . Heart involvement is considered a leading cause of morbidity in CSS [3] , and may be the leading cause of death [8] . Heart involvement was found in 64% of autopsied cases [2] and death was attributed to heart involvement in 27% and 48% in two post mortem studies [2,9] . Recurrent myocardial vasculitis has been described in patients with CSS who benefitted from a heart transplant [10] . Diagnosis is often established late; apart from pericarditis, lesions associated with CSS are varied, and, in the early stages of the disease, difficult to distinguish from those linked to other cardiovascular risk factors. Though heart involvement is a frequent finding in CSS, it is important for the clinician to determine early on what heart anomalies are directly attributable to the vasculitis. If these lesions are detected early, treatments can be introduced to manage the disease before irreversible myocardial lesions occur and thus reduce morbi-mortality associated with this specific heart condition. The aim of our work was to record and analyze cases of heart involvement in CSS in our region of Burgundy, France, in order to determine the prevalence, to compare the clinical and paraclinical characteristics of patients according to the presence or absence of heart involvement and to determine whether the heart lesions were due to CSS or not. 2 Patients and methods From the computer coded hospitalization records, we conducted a retrospective study of all of the patients who were hospitalized at least once in a health care establishment in Burgundy, France, between January 1998 and May 2008 for care related to CSS or one of its complications. For each case, the medical file was examined, the ACR 1990 diagnostic criteria were checked and the clinical and biological manifestations as well as results of cardiac examinations for periods before the diagnosis and during the follow-up were recorded. We noted all of the manifestations of vasculitis including pulmonary manifestations (infiltrates, nodules, exacerbation of asthma, hemoptysis, pleural effusion and chest pain), digestive symptoms (hemorrhage, bowel pain, eosinophilic cholecystitis, vasculitis and ischemia), neurological symptoms (neuropathy, stroke, cognition failure and mononeuropathy), skin symptoms (nodules, purpura, ulcers and digital ischemia), rheumatism symptoms (arthritis and myalgia), renal symptoms (renal failure, nephritis, and crises of arterial hypertension, …), ENT symptoms (nasal polyposis, unexplained chronic sinusitis or chronic rhinitis) and general signs (weight loss, fever, sweats and anorexia) before diagnosis, at diagnosis and for each relapse. For all patients, we checked if the following were reported in the follow-up: heart abnormalities and specially those described in CSS such as pericarditis, heart failure, aortic or mitral insufficiency, heart enzyme modifications or electrical modifications, chest pain or abnormalities in ultrasound scan, scintigraphy, magnetic resonance imaging (MRI) or coronarography. We analyzed standard laboratory tests and tests for inflammation, and determined total IgE and NT Pro-BNP and troponin I. Atherosclerotic lesions and cardiovascular risk factors were also noted. The impact of treatments and particularly of immunosuppressive drugs on the heart manifestations was evaluated. The heart abnormalities were then stratified by CSS experts into groups according to the likelihood of being due to Churg–Strauss vasculitis. The cardiac CSS group comprised cardiac lesions already reported and imputed to CSS in the literature and those unexplained by another etiology (history of infectious valve disease, atherosclerosis, chronic hypertension, diabetes, smoking, hereditary…) in each patient's history. The group also included those concurrent with a flare-up of the disease (neurological manifestations, fever, pulmonary infiltrates…) which improved or stabilized with high dose corticoids or immunosuppressants or were proven histologically. Lesions with no link to CSS (non-cardiac CSS group) comprised cardiac lesions for which the etiology could be explained by the presence of other cardiovascular risk factors and diseases: atherosclerosis, coronary artery disease, heart failure due to chronic hypertension, diabetes, post streptococci aortic or mitral insufficiency.... For the statistical analysis, it was decided to choose a type-1 error of 5%. Qualitative variables were analyzed using Fisher's exact test; and quantitative variables were analyzed parametrically using Student's t test and non-parametrically using the Wilcoxon test. 3 Results We found 31 different patients followed for CSS. The cardio-vascular risk factors included arterial hypertension in 38.7% of patients (IC = 22–56%), dyslipidemia in 25.8% (IC = 10–41%), former smoking in 19.4% (IC = 5–33%), current smoking in 9.7% (IC = 0–20%), diabetes in 9.7% (IC = 0–20%), stroke in 6.5% (IC = 0–15%) and morbid obesity in 3.2% of cases (0–9%). Twenty patients had heart disease (cardiac group) (64%; IC = 48–81%) ( Fig. 1 ). Patients had valvulopathy (52%; IC = 34–69%) (2/3 acute or progressive mitral regurgitation, 1/4 of aortic stenosis), heart failure (48%; IC = 31–66%) 40% of which were diastolic, severe rhythm troubles (32%; IC = 16–49%) (2 patients with ventricular tachycardia), severe conduction troubles (22.6%; IC = 8–37%) (high degree AV block), ischemic heart disease (22.6%; IC = 8–37%), pericarditis (19%; IC = 5–33%) or endocarditis (3%; IC = 0–9%). All patients had an electrocardiogram and an echocardiography and 80% of patients had an ANCA test, 15% coronarography, 10% cardiac scintigraphy and 10% cardiac MRI. Eighteen patients (58%; IC = 41–75%) had an abnormal ECG (rhythm or conduction troubles, ST segment abnormalities) and 16 (52%; IC = 34–69%) had an abnormal heart ultrasound scan (valvulopathies, heart failure, akinesia and endomyocardial fibrosis). 3.1 Cardiac group versus non-cardiac group of patients There was no statistically significant difference in terms of demographics, history or risk factors of cardio-vascular disease between patients presenting with cardiac involvement whatever the etiology (cardiac group) and those with no cardiac involvement (non-cardiac group). We also noted a lower prevalence of atypical digestive abnormalities in the cardiac group before the diagnosis of CSS (10% versus 63%, OR = 0.07; p < 0.05), a greater proportion of heart signs (40% versus 0%, p < 0.05), a lower prevalence of eye involvement at the time of diagnosis (0% versus 27%, p < 0.05), fewer abnormalities in the pulmonary X-ray (17% versus 47%, p < 0.05) and smaller quantities of lung infiltrates during flare-ups (21% versus 53%, p < 0.05) ( Table 1 : only statistically significant data are shown). The pANCA test was positive in only 7% of patients with heart involvement (cardiac group) whereas we found this in 60% of CSS patients with no heart involvement (non-cardiac group) (OR = 15; p < 0.05) ( Fig. 2 ). We found no differences between the different patient groups for troponin I and NT Pro-BNP levels. 3.2 Cardiac CSS group versus non-cardiac CSS group of patients Among the 20 patients presenting with heart involvement (cardiac group), 15 (75%) had lesions that were attributable to CSS (cardiac CSS group) ( Fig. 1 ). Thus, overall, a total of 48% of the 31 patients (IC = 31–66%) had heart lesions directly attributable to Churg–Strauss vasculitis (cardiac CSS group): we found 6 (40%; IC = 15–65%) patients with heart failure with no other etiology, 3 patients (20%; IC = 0–40%) with ischemic lesions on otherwise healthy coronary arteries (normal coronarography), 3 (20%; IC = 0–40%) pericarditis, 3 (20%; IC = 0–40%) aspects of myocardial fibrosis, 3 (20%; IC = 0–40%) evolutive valvulopathies (mitral and aortic regurgitation), and 4 (27%; IC = 4–49%) cases of serious conduction or rhythm troubles (high degree AV block and ventricular tachycardia). The characteristics of the cardiac CSS group are presented in Table 2 . The mean age of patients in the cardiac CSS group was 64 years (median = 66) versus 85 years (median = 85) for those in the non-cardiac CSS group ( p < 0.01). ( Table 3 : only statistically significant data are shown). Even though it was not significant, the disease tended to be diagnosed at a younger age in the cardiac CSS group of patients (median 55 years versus 70 years, p = 0.08), and the time between the first signs and diagnosis appeared shorter (median 54 years versus 69 years; ( p = 0.06)). Moreover, there was a trend towards a lower prevalence of cardiovascular risk factors in the cardiac CSS group than in the non-cardiac CSS group. We were unable to show any demographic differences. Heart involvement due to CSS (cardiac CSS group) was less frequently accompanied by radiologically-apparent infiltrates at the time of diagnosis of the disease ( p < 0.05) ( Table 3 : only statistically significant data are shown). During flare-ups, the absence of abnormalities on chest X-ray ( p < 0.05), of pulmonary infiltrates ( p < 0. 001) and of infiltrated papules ( p < 0.05) also correlated with the presence of heart involvement due to CSS (cardiac CSS group). The mean leucocyte count was higher at the diagnosis in the cardiac CSS group. The number of eosinophils was also higher in the cardiac CSS group, even though the difference did not reach statistical significance. The inflammatory syndrome was more severe in the non-cardiac CSS group; this was seen in the levels of CRP (C-reactive protein) and fibrinogen, but especially for the ESR ( p < 0.05). Rhythm troubles were more frequent in heart abnormalities not linked to CSS (non-cardiac CSS group) ( p < 0.004) ( Table 3 : only statistically significant data are shown). We found no statistically significant difference between the cardiac CSS group and non-cardiac CSS group with regard to the prevalence of abnormalities on the electrocardiogram or the ultrasound scan. No patients underwent cardiac biopsy or catheterization of the right heart. Four patients underwent coronarography, and in three, the results were normal despite electrical and biological signs of myocardial ischemia. Three patients had myocardial scintigraphy and 3 had cardiac MRI; one and three, respectively, revealed perfusion abnormalities. 3.3 Cardiac CSS group versus non-cardiac group of patients We found no statistically significant difference between heart abnormalities linked to CSS (cardiac CSS group) and patients without heart abnormalities (non-cardiac group) with regard to demography, cardiovascular risk factor or medical history. In the cardiac CSS group compared with the non-cardiac group, digestive symptoms were less frequent before diagnosis ( p < 0.01), whereas cardiac signs were more frequent and ocular symptoms were less frequent at diagnosis. Relapses showed more neurological cardiac and renal symptoms in the cardiac CSS group than in the non-cardiac group but less infiltrates in chest radiography ( p < 0.05) ( Table 4 ). We found no difference between patients with heart abnormalities linked to CSS (cardiac CSS group) and all others patients ( n = 16), for demography, cardiovascular risk factors or medical history ( Table 5 ). Sonography abnormalities, valvulopathies and heart failure were more frequent in the first group ( p < 0.05). We found a significant difference for clinical events: in the cardiac CSS group there were fewer digestive symptoms before diagnosis and more cardiac events at diagnosis, during relapses and during the follow-up, and less pulmonary infiltrates and chest abnormalities at diagnosis and in relapses ( p < 0.05) ( Table 5 ). When done, positive pANCA was noted in 7% of patients in the cardiac CSS group and in 45% of all patients ( p = 0.05). We found no significant difference between the two groups and between patients with or without pulmonary infiltrates for eosinophil blood count. 4 Discussion Even though the study was retrospective, we revealed that in Burgundy 64% of our patients with Churg–Strauss vasculitis suffered from heart disease during their follow-up and that in 48% of cases the heart involvement was due to CSS. Patients with CSS-specific heart involvement had different clinical and lab characteristics. The prevalence of heart abnormalities in our Burgundy series was slightly higher than those reported in the literature, but a recent study on cardiac MRI showed 62% of cardiac abnormalities in CSS [11] . One of the limits of our work is its retrospective nature and the lack of histological data concerning the heart lesions, but our figures are in agreement with autopsy findings. In the literature, the prevalence of cardiac involvement in CSS varies from 15 to 56% (39% in the Sable et al. series of 112 patients [12] ). It was evaluated in a number of retrospective studies with [5,12–14] or without [8,9] histological confirmation of the disease. Many types of heart involvement have been described in CSS: endomyocardial fibrosis [17,18] , eosinophilic myocarditis [19,20] , coronaritis [21–23] , isolated tachycardia [14] , conduction abnormalities or arrhythmia [24–26] , valvulopathy [15,27] and pericarditis [8,28,29] . Eosinophilic myocarditis generally develops with low levels of cardiac murmur and is a frequent cause of rapidly evolving chronic heart failure [13] . Chest pain is a frequent symptom as was the case in our series. Angina is rather rare [23] , but may resemble Prinzmetal angor [22] , as was the case with one of our patients, or unstable angina [21] . We also noticed a lower prevalence of digestive manifestations and of X-ray abnormalities and infiltrates in cases of heart involvement. We have no explanations concerning the strong association between the absence of pulmonary infiltrates at the diagnosis and heart disease directly linked to CSS. Eosinophil blood counts were not different. Eosinophil counts in the blood do not reflect tissue eosinophilia. In our study, electrocardiogram and echocardiography data contributed little to help distinguish between disease that was specifically linked to CSS and that due to other causes. Anomalies on the electrocardiogram were found in 50% of cases, and 15 to 25% of patients developed congestive heart failure during the follow-up [8,13] . Echocardiography is an essential examination in the context of vasculitis. According to Vignaux et al. this examination should be conducted systematically even in the absence of clinical manifestations [19] . Two studies [15,16] showed that heart involvement with abnormal echocardiography results is far more frequent in patients without ANCA. MRI may reveal infraclinical lesions, inflammatory infiltration of the myocardium, and pericardial effusion. There are now many case reports and some series about cardiac MRI in CSS patients. MRI is clearly a helpful diagnostic method in a given patient and may thus provide information to determine the prognosis and evaluate the efficacy of treatment on cardiac lesions [19,30–32] . To be certain of heart involvement, histological examination is necessary, but this examination carries a considerable iatrogenic risk and does not always contribute to the diagnosis [33] . Indeed, histological study of a biopsy confirms the disease in only 30% of patients presenting typical clinical myocarditis and in 30 to 60% of patients presenting with dilated cardiomyopathy [34,35] . To date, there are no algorithms that make it possible to define the pertinence of heart biopsy in the diagnosis of heart disease due to CSS. Scintigraphy has also contributed little to the diagnosis. [36] . No study has evaluated the usefulness of scintigraphy and coronarography in CSS. A normal coronarography does not eliminate the diagnosis of coronary artery involvement, as the lesions are very often distal [19] . The main coronary trunks, however, may show an aspect of arteritis in the form of dissections [37] or spasms [22] . We have shown a clear decrease in the prevalence of pANCA in Churg–Strauss patients with heart involvement. This finding has been confirmed by other authors, who found pericardial involvement in 36% of ANCA-negative patients and in only 7% of ANCA-positive patients [12] . The presence of ANCA may define a sub-group of patients with a greater prevalence of renal involvement (extracapillary glomerulonephritis), purpura, diffuse intra-alveolar hemorrhage and involvement of the central nervous system [12,38] . In contrast, heart lesions may be more frequent in patients without ANCA [12,39] . In their study, Kümpfer et al. showed that elevated endothelial, leukocytic and platelet-derived microparticles decreased during follow-up and closely paralleled vasculitic activity. Endothelial microparticles might be an additional tool to diagnose and monitor cases of suspected vasculitis-related cardiac involvement in CSS [40] . Moreover, we noted a trend towards a lower prevalence of inflammatory syndromes at the diagnosis in cases of heart involvement directly linked to vasculitis. Agard et al. reported lower levels of CRP in this population [41] . In our study, we also noted higher levels of leucocytes at the diagnosis in patients who went on to develop heart lesions associated with vasculitis than in those who developed heart lesions due to other causes. We found no significant difference with regard to the level of eosinophils even though there was a trend similar to the one above; this may be because our study lacks power because of its small size. However, the eosinophil count does not reflect tissular eosinophilia. Neumann et al. are recently investigating cardiac involvement in Churg–Strauss Syndrome [42] . In this study of 49 patients, a high eosinophil count was associated with specific cardiac lesions due to CSS. These data may provide support to the hypothesis that eosinophils play a preponderant role compared to ANCA in the development of heart lesions. This is corroborated histologically by the presence of large numbers of eosinophils in heart tissues and the direct pathogenic role of eosinophils in tissues in general [43,44] . The eosinophil-rich interstitial inflammatory infiltrate, the vasculitis-related ischemic coronary lesions and the extravascular granulomas may contribute to the physiopathology of heart disease in CSS. Finally, these elements provide support to the hypothesis that in CSS there are two nosological sub-entities with distinct physiopathological mechanisms depending on the presence or absence of heart involvement. The therapeutic management of heart lesions in CSS consists mainly of corticoids in the case of pericarditis [29] and of the addition of cyclophosphamide in the case of myocardial involvement [3] . Management is empirical and is based on the same immunosuppressive therapies as those found in severe visceral involvement, which themselves are based on the therapeutic approaches used in other forms of ANCA-related vasculitis. The preponderant role of eosinophils may encourage the use of targeted molecules such as anti-interleukin 5 monoclonal antibodies, which are currently being tested in phase II/III in CSS [45] or omalizumab (an anti-IgE antibody) [45] used in essential hyper eosinophil syndrome [46] . We also mention that omalizumab may be associated with the onset of CSS in a few patients [47] . Heart disease due to CSS is frequent, heterogeneous, probably underestimated and associated with a poor prognosis. The diagnosis remains difficult and is often late and even retrospective. Patients with heart involvement seem to form a sub-group with a specific physiopathology in the development of heart lesions that predominantly involves eosinophils. 5 Learning points • Cardiac involvement in CSS is very frequent and underestimated. • Heart involvement in CSS is difficult and the diagnosis is often late and retrospective. • Heart involvement in CSS is associated with a lower positivity of pANCA anti-MPO. • Patients with heart manifestations linked to CSS have different clinical presentation with less pulmonary infiltrates at diagnosis and during relapses. • Patients with heart manifestations linked to CSS had less history of gastrointestinal manifestations before de diagnosis. • This data suggest two sub-entities in Churg–Strauss syndromes with different physiopathologies. References [1] P. Seo J.H. Stone The antineutrophil cytoplasmic antibody-associated vasculitides Am J Med 117 2004 39 50 [2] J. Churg L. Strauss Allergic granulomatosis, allergic angiitis, and periarteritis nodosa Am J Pathol 27 1951 277 301 [3] L. Guillevin F. Lhote M. Gayraud P. Cohen B. Jarrousse O. Lortholary Prognostic factors in polyarteritis nodosa and Churg–Strauss syndrome. A prospective study in 342 patients Medicine (Baltimore) 75 1996 17 28 [4] L. Guillevin C. Pagnoux L. Mouthon Churg–Strauss syndrome Semin Respir Crit Care Med 25 2004 535 545 [5] R. Solans J.A. Bosch C. Perez-Bocanegra A. Selva P. Huguet J. Alijotas Churg–Strauss syndrome: outcome and long-term follow-up of 32 patients Rheumatology (Oxford) 40 2001 763 771 [6] C. Ribi P. Cohen C. Pagnoux A. Mahr J.P. Arene D. Lauque Treatment of Churg–Strauss syndrome without poor-prognosis factors: a multicenter, prospective, randomized, open-label study of seventy-two patients Arthritis Rheum 58 2008 586 594 [7] P. Cohen C. Pagnoux A. Mahr J.P. Arene L. Mouthon V. Le Guern Churg–Strauss syndrome with poor-prognosis factors: a prospective multicenter trial comparing glucocorticoids and six or twelve cyclophosphamide pulses in forty-eight patients Arthritis Rheum 57 2007 686 693 [8] J.G. Lanham K.B. Elkon C.D. Pusey G.R. Hughes Systemic vasculitis with asthma and eosinophilia: a clinical approach to the Churg–Strauss syndrome Medicine (Baltimore) 63 1984 65 81 [9] L.C. Chumbley E.G. Harrison Jr. R.A. DeRemee Allergic granulomatosis and angiitis (Churg–Strauss syndrome). Report and analysis of 30 cases Mayo Clin Proc 52 1977 477 484 [10] R.A. Henderson P. Hasleton B.N. Hamid Recurrence of Churg Strauss vasculitis in a transplanted heart Br Heart J 70 1993 553 [11] Dennert RM, van Paassen P, Schalla S, Kuznetsova T, Alzand BS, Staessen JA, et al. Cardiac involvement in Churg–Strauss syndrome. Arthritis Rheum 2010; 62:627-34. [12] R. Sable-Fourtassou P. Cohen A. Mahr C. Pagnoux L. Mouthon D. Jayne Antineutrophil cytoplasmic antibodies and the Churg–Strauss syndrome Ann Intern Med 143 2005 632 638 [13] L. Guillevin P. Cohen M. Gayraud F. Lhote B. Jarrousse P. Casassus Churg–Strauss syndrome. Clinical study and long-term follow-up of 96 patients Medicine (Baltimore) 78 1999 26 37 [14] C. Haas C. Le Jeunne P. Choubrac H. Durand F.C. Hugues Churg–Strauss syndrome. Retrospective study of 20 cases Bull Acad Natl Med 185 2001 1113 1130 discussion 30-3 [15] J.M. Morgan L. Raposo D.G. Gibson Cardiac involvement in Churg–Strauss syndrome shown by echocardiography Br Heart J 62 1989 462 466 [16] G. Pela G. Tirabassi P. Pattoneri L. Pavone G. Garini G. Bruschi Cardiac involvement in the Churg–Strauss syndrome Am J Cardiol 97 2006 1519 1524 [17] C.R. McGavin A.J. Marshall C.T. Lewis Churg–Strauss syndrome with critical endomyocardial fibrosis: 10 year survival after combined surgical and medical management Heart 87 2002 E5 [18] S. Hayashi S. Furuya H. Imamura Fulminant eosinophilic endomyocarditis in an asthmatic patient treated with pranlukast after corticosteroid withdrawal Heart 86 2001 E7 [19] O. Vignaux J. Marmursztejn P. Cohen E. Bruguiere D. Duboc L. Guillevin Cardiac imaging in ANCA-associated vasculitis Presse Méd 36 2007 902 906 [20] N.G. Uren P.J. Hammond Myopericarditis in Churg–Strauss syndrome Tex Heart Inst J 18 1991 127 131 [21] A.D. Wagner G.P. Meyer M. Rihl A. Rathmann U. Wittkop H. Zeidler Acute coronary syndrome associated with Churg–Strauss syndrome Vasc Health Risk Manag 3 2007 775 779 [22] P. Petrakopoulou W.M. Franz P. Boekstegers M. Weis Vasospastic angina pectoris associated with Churg–Strauss syndrome Nat Clin Pract Cardiovasc Med 2 2005 484 489 quiz 90 [23] M. Kozak L.S. Green E.A. Gill The Churg–Strauss syndrome. A case report with angiographically documented coronary involvement and a review of the literature Chest 107 1995 578 580 [24] C. Fong G. Schmidt N. Cain P. Cranswick A.M. Tonkin Churg–Strauss syndrome, cardiac involvement and life threatening ventricular arrhythmias Aust N Z J Med 22 1992 167 168 [25] M. Rabusin L. Lepore F. Costantinides R. Bussani A child with severe asthma Lancet 351 1998 32 [26] G.P. Balestrieri U. Valentini B. Cerudelli S. Spandrio E. Renaldini Reversible myocardial impairment in the Churg–Strauss syndrome: report of a case Clin Exp Rheumatol 10 1992 75 77 [27] L. Doherty P. Kumar R. Bexton C. Plummer J. Dark F. Black Aortic regurgitation and Churg–Strauss syndrome QJM 98 2005 772 773 [28] I. Noth M.E. Strek A.R. Leff Churg–Strauss syndrome Lancet 361 2003 587 594 [29] C. Agard E. Rendu V. Leguern T. Ponge A. Masseau J.H. Barrier Churg–Strauss syndrome revealed by granulomatous acute pericarditis: two case reports and a review of the literature Semin Arthritis Rheum 36 2007 386 391 [30] C. Silva J.C. Moon A.G. Elkington A.S. John R.H. Mohiaddin D.J. Pennell Myocardial late gadolinium enhancement in specific cardiomyopathies by cardiovascular magnetic resonance: a preliminary experience J Cardiovasc Med (Hagerstown) 8 2007 1076 1079 [31] J.P. Smedema P. van Paassen M.J. van Kroonenburgh G. Snoep H.J. Crijns J.W. Tervaert Cardiac involvement of Churg Strauss syndrome demonstrated by magnetic resonance imaging Clin Exp Rheumatol 22 2004 S75 S78 [32] S. Mavrogeni M.N. Manoussakis T.C. Karagiorga M. Douskou D. Panagiotakos V. Bournia Detection of coronary artery lesions and myocardial necrosis by magnetic resonance in systemic necrotizing vasculitides Arthritis Rheum 61 2009 1121 1129 [33] D.J. Davies Small vessel vasculitis Cardiovasc Pathol 14 2005 335 346 [34] C.C. Marboe J.J. Fenoglio Jr Biopsy diagnosis of myocarditis Cardiovasc Clin 18 1988 137 154 [35] N.S. Peters P.A. Poole-Wilson Myocarditis—a controversial disease J R Soc Med 84 1991 1 2 [36] E. Vallejo C. Mendoza-Gonzalez A. Aranda Churg–Strauss syndrome and myocardial perfusion SPECT imaging J Nucl Cardiol 11 2004 358 360 [37] J.C. Hunsaker III W.N. O'Connor J.T. Lie Spontaneous coronary arterial dissection and isolated eosinophilic coronary arteritis: sudden cardiac death in a patient with a limited variant of Churg–Strauss syndrome Mayo Clin Proc 67 1992 761 766 [38] K.A. Keogh U. Specks Churg–Strauss syndrome: clinical presentation, antineutrophil cytoplasmic antibodies, and leukotriene receptor antagonists Am J Med 115 2003 284 290 [39] P. Guilpain P. Cohen C. Pagnoux D. Duboc O. Vignaux J.-F. Viallard Atteinte cardiaque sévère au cours du syndrome de Churg et Strauss Rev Med Interne 2007 S36 S82 [40] P. Kumpers U. Erdbrugger M. Grossheim G.P. Meyer M. Hiss W. Gwinner Endothelial microparticles as a diagnostic aid in Churg–Strauss vasculitis-induced cardiomyopathy Clin Exp Rheumatol 26 2008 S86 S89 [41] C. Agard E. Rendu N. Piriou S. Pattier C. Sagan A. Masseau Atteinte cardiaque sévère du syndrome de Churg et Strauss: étude rétrospective de 11 observations Rev Med Interne 2007 S36 S82 [42] T. Neumann B. Manger M. Schmid C. Kroegel A. Hansch W.A. Kaiser Cardiac involvement in Churg–Strauss syndrome: impact of endomyocarditis Medicine (Baltimore) 88 2009 236 243 [43] C.B. Pearce J.D. McMeekin T.N. Moyana J. Sibley A case of peripartum eosinophilic myocarditis Can J Cardiol 15 1999 465 468 [44] P.U. Ogbogu D.R. Rosing M.K. Horne III Cardiovascular manifestations of hypereosinophilic syndromes Immunol Allergy Clin North Am 27 2007 457 475 [45] E. Bargagli C. Madioni C. Olivieri F. Penza P. Rottoli Churg–Strauss vasculitis in a patient treated with omalizumab J Asthma 45 2008 115 116 [46] M.E. Rothenberg A.D. Klion F.E. Roufosse J.E. Kahn P.F. Weller H.U. Simon Treatment of patients with the hypereosinophilic syndrome with mepolizumab N Engl J Med 358 2008 1215 1228 [47] X. Puechal P. Rivereau F. Vinchon Churg–Strauss syndrome associated with omalizumab Eur J Intern Med 19 2008 364 366