- Email: [email protected]
Cardiac Involvement in the Churg-Strauss Syndrome
Cardiac Involvement in the Churg-Strauss Syndrome Giovanna Pelà, PhD*, Giovanni Tirabassi, MD, Paolo Pattoneri, MD, Laura Pavone, MD, Giovanni Garini, MD, and Giacomo Bruschi, MD Churg-Strauss syndrome (CSS) is a rare systemic disease characterized by necrotizing vasculitis and peripheral eosinophilia. Cardiac involvement is considered common and is given a high rank among the causes of morbidity and mortality. The aim of this study was an update on the cardiac manifestations of this syndrome using a noninvasive approach. Sixteen patients with CSS were compared with a gender- and age-matched group of 20 healthy subjects. All patients but 1 were receiving treatment (steroids and/or immunosuppressive drugs). According to the Birmingham vasculitis activity score, 12 patients were in an active phase, and 4 were in drug-induced remission. All subjects underwent M-B-mode echocardiography and Doppler tissue echocardiography. Heart failure, lifethreatening arrhythmias, and other prominent manifestations of heart disease were not observed. No differences were found in left ventricular diameter, volume, mass, or ejection fraction. The 2 groups did not differ in right ventricular diameter and pulmonary pressure. Few and nonspecific changes were detected by 2-dimensional echocardiography, including subclinical pericardial effusion and mitral regurgitation, in fewer than half the subjects. Subjects with CSS showed an impairment of ventricular relaxation. Changes were more prominent in the right ventricle. The peak velocity (PV) of early diastolic tricuspid inflow (E) was about 8% less than in controls, and the velocity of late diastolic inflow (A) was 35% greater. The E/APV ratio was, on average, 33% less. In the left ventricle, EPV was 11% less and APV 11% greater. The E/A ratio was decreased by 22%. Doppler analysis of tissue kinetics confirmed these indications. In the right ventricle, EPV was decreased by 10% and APV was increased by 20% in the patient group. The E/APV ratio was decreased by 29%. In the left ventricle, in which different sites were sampled, the average changes were ⴚ15%, ⴙ1%, and ⴚ23%, respectively. In the left ventricle, the velocity of systolic contraction was also decreased by 12%. Because of the small group size, only some of these differences were statistically significant. In conclusion, these moderate changes, devoid of clinical correlates, contrast with early reports emphasizing cardiac morbidity and poor prognosis in this syndrome. © 2006 Elsevier Inc. All rights reserved. (Am J Cardiol 2006;97:1519 –1524)
Churg-Strauss syndrome (CSS) is a rare necrotizing vasculitis of unknown origin that affects small-sized blood vessels.1,2 Cardiac involvement has been described as an eosinophilic vasculitis, pericarditis, and myocarditis, potentially resulting in heart failure and sometimes leading to death through cardiac tamponade, myocardial infarction, and arrest.3–9 However, several discrepancies appear in published research concerning the prevalence of cardiomyopathy in CSS and its clinical severity. For example, the frequency of cardiac involvement ranges from 17% to 92%.1,10 –15 In some reports, myocardial involvement was believed to be progressive,11–14 whereas in others, it was fully reversible.16,17 An explanation of these variations could lie either in the method used to assess the carditis (e.g., clinical vs instrumental) or in different intensities of treatment. It is now commonly accepted that patients with CSS greatly benefit Department of Internal Medicine, Nephrology and Prevention Sciences, University of Parma, Parma, Italy. Manuscript received August 10, 2005; revised manuscript received and accepted November 29, 2005. * Corresponding author: Tel: 390521033192; fax: 390521033185. E-mail address: [email protected] (G. Pelà). 0002-9149/06/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2005.11.088
from steroid and/or steroid-associated immunosuppressive drugs, but this was not equally clear in old reports. Sixteen patients were compared with 20 age- and gendermatched controls. All patients but 1 were receiving treatment.
Methods Subjects: Sixteen patients with CSS (9 men, 7 women) were enrolled in our study. The patients were evaluated by the interdisciplinary Secondary and Primary Vasculitides study group of the University of Parma from 1998 to 2005. All the patients were hospitalized and underwent physical examinations, standard blood and urinary tests, spirograms (14 of 16 patients) and chest x-rays (14 of 16 patients) before echocardiography. In all patients, CSS was diagnosed on the basis of the classification criteria of the American College of Rheumatology.18 Briefly, 4 of the following 6 findings were required: (1) a history of asthma, (2) eosinophilia ⬎10%, (3) mononeuropathy or polyneuropathy, (4) pulmonary nonfixed infiltrates, (5) paranasal sinus abnormality, and (6) biopsy evidence of extravascular eosinophil infiltration. Actuwww.AJConline.org
1520
The American Journal of Cardiology (www.AJConline.org)
Table 1 Subjects and their basic echocardiographic data Parameter Age (yrs) Body surface area (m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Heart rate (beats/min) Body mass index (kg/m2) LV end-diastolic diameter (cm) LV end-systolic diameter (cm) Septal wall thickness (mm) Posterior wall thickness (mm) LV mass index (g/m2) Relative wall thickness LV fractional shortening (%) Ejection fraction (%) RV end-diastolic diameter (cm) RV long-axis diastolic diameter (cm)
Controls
Patients With CSS
49 ⫾ 19 1.83 ⫾ 0.16 121 ⫾ 13 77 ⫾ 9 71 ⫾ 8 23.9 ⫾ 3.1 4.6 ⫾ 0.4 2.7 ⫾ 0.3 9.7 ⫾ 1.3 8.5 ⫾ 1.2 94 ⫾ 19 0.40 ⫾ 0.05 40 ⫾ 5 71 ⫾ 5 2.8 ⫾ 0.4 6.5 ⫾ 0.8
49 ⫾ 10 1.74 ⫾ 0.19 124 ⫾ 17 81 ⫾ 13 84 ⫾ 14* 24.6 ⫾ 2.6 4.6 ⫾ 0.5 2.9 ⫾ 0.5 9.2 ⫾ 1.5 8.8 ⫾ 1.5 100 ⫾ 24 0.39 ⫾ 0.06 40 ⫾ 6 68 ⫾ 7 3.0 ⫾ 0.2 6.7 ⫾ 0.7
Data are expressed as mean ⫾ SD. * p ⬍0.001, patients versus controls. LV ⫽ left ventricular; RV ⫽ right ventricular.
ally, only skin biopsies were performed in a sizable portion of patients, and a history of asthma was present in all. Organ system involvement was assessed using the Birmingham vasculitis activity score.19 Subjects with scores of 0 were considered in remission; all others were assumed to be in an active phase. A control group of 20 age-matched healthy subjects (10 men, 10 women) was also studied. The 2 groups had similar body surface areas (Table 1). Informed written consent was obtained from all participants. The research protocol was approved by the university’s ethics committee. All procedures conformed to the Declaration of Helsinki. All patients but 1 were treated at the time of the study. The treatment of active disease began with an intravenous bolus of methylprednisolone 1 g for 3 days, followed by oral prednisone 1 mg/kg/day. If remission was observed, the dose was tapered to 0.25 mg/kg/day after 1 year. An immunosuppressive drug, usually cyclophosphamide (2 patients were randomly assigned to methotrexate and 1 to azathioprine), was added to prednisone when (1) remission had not begun after 2 weeks of treatment; (2) the involvement of ⬎3 organ systems was found at presentation; and (3) polyneuropathy, which in our experience is particularly sensitive to cyclophosphamide, was present. The dose was adjusted to the clinical response and the leukocyte count. A count of 3,000/ mm3 was a threshold for decreasing the drug, and a count of 1,500/mm3 was a threshold for discontinuing the drug. Standard echocardiography: The subjects were examined by an experienced sonographer in the left lateral decubitus position using an Aspen instrument (Acuson Corporation, Mountain View, California) equipped with a multihertz sector probe (2 to 4 MHz).
Figure 1. Top panel, an apical 4-chamber view with some of the sampling sites of Doppler tissue measurements. Arrows, the direction of tissue motion. In the RV, the descent of the tricuspid annulus has been explored. On the left, the analogous motion of the mitral annulus was monitored. From the parasternal view, not shown here, the kinetics of the posterior wall were explored. Full black circles denote the sampling sites: lateral and posteroseptal for the LV and lateral for the RV. Bottom panel, a tricuspid annulus velocity pattern in a control subject (left) and in a patient with CSS (right). LA ⫽ left atrium; RA ⫽ right atrium.
The criteria for measuring wall thickness, cavity dimensions, volumes, and ventricular mass conformed to established conventions and are reported in detail elsewhere.20 Blood flow across the mitral and tricuspid valves was also determined by standard techniques. In the diastolic flow profile, the E and A waves were considered, and the peak velocity (PV) and the time–velocity integral (TVI) of either wave were extracted. The PV of the tricuspid regurgitant jet was used to estimate pulmonary artery systolic pressure, where pulmonary artery systolic pressure ⫽ 4V2 (where V is the peak tricuspid regurgitant velocity) ⫹ 10 (a conventional estimate for right atrial pressure). The acceleration time of pulmonary artery systolic flow was also used to obtain an indirect assessment of pulmonary arterial pressure. Pulsed Doppler tissue echocardiography: Doppler tissue echocardiography extracts a velocity tracing from the
1.56 ⫾ 0.5 1.04 ⫾ 0.3 ⬍0.01 4.0 ⫾ 1.2 6.0 ⫾ 1.9 ⬍0.001 33 ⫾ 7 45 ⫾ 10 ⬍0.001 9.2 ⫾ 2.3 8.7 ⫾ 2.9 NS 49 ⫾ 10 45 ⫾ 12 NS 1.77 ⫾ 0.8 1.5 ⫾ 0.6 NS 1.25 ⫾ 0.5 0.97 ⫾ 0.3 0.07 7.4 ⫾ 2.7 7.2 ⫾ 2.0 NS Data are expressed as mean ⫾ SD.
11.6 ⫾ 2.3 9.8 ⫾ 2.3 ⬍0.05 Controls Patients with CSS p Value
69 ⫾ 13 61 ⫾ 11 0.07
60 ⫾ 18 67 ⫾ 15 NS
ATVI (cm) EPV (cm/s) ETVI/ATVI (cm) EPV/APV (cm/s) ATVI (cm) ETVI (cm) EPV (cm/s)
APV (cm/s)
Mitral flow Variable
Cross-sectional echocardiography: In terms of basic echocardiography, no significant group differences were observed in the cavity dimensions of the left ventricle (LV) and the right ventricle (RV). The thickness of the septum and
Table 2 Mitral and tricuspid flow patterns
Results Baseline clinical features: According to the Birmingham vasculitis activity score, 4 patients were in complete remission and 12 were in an acute phase of the illness (Birmingham vasculitis activity score ranges from 5 to 27, mean 15.8). Heart rates were significantly greater in patients with CSS. The means of systolic and diastolic blood pressure were similar in the 2 groups (Table 1), but 8 patients were treated with antihypertensive drugs. Twelve-lead electrocardiograms showed normal patterns in 11 patients and repolarization alterations in 5 patients. One patient had a right bundle branch block. Spirograms indicated airway obstructions in 6 of 14 patients (mild in 5, severe in 1) and normal tracings in 7 of 14 patients; 1 patient had a mild impairment of the diffusing capacity of carbon monoxide. Fibrotic pulmonary nodules and interlobular septal thickening were detected in 7 of 14 patients on standard chest x-rays. At the time of admission, inflammatory markers were altered in 12 subjects with CSS: the erythrocyte sedimentation rate (ESR) in 5 (29.4 ⫾ 33.1 mm/hour) and C-reactive protein (CRP) in 5 (31.3 ⫾ 55.8 mg/dl). Eosinophilia was present in 10 subjects (mean 15.6 ⫾ 13.6% of the leukocyte count; total eosinophils 1,978 ⫾ 2,489/l). One patient had diabetes mellitus requiring oral hypoglycemic treatment.
ETVI (cm)
Statistical analysis: Data are expressed as mean ⫾ SD. The comparison between the group means was made by the 2-tailed Student’s t test, after ascertaining the compatibility of the data with the normal distribution (the KolmogorovSmirnov procedure). For comparison, nonparametric statistics were also assayed, including the Mann-Whitney U statistic, Wald-Wolfowitz runs, and Tukey’s quick tests. The bibliographic basis was the Encyclopedia of Statistical Sciences.22 SPSS software packages (SPSS, Inc., Chicago, Illinois) were used extensively.
APV (cm/s)
Tricuspid Flow
EPV/APV (cm/s)
ETVI/ATVI (cm)
oscillatory motion of myocardial walls during the heart cycle (Figure 1). It is usual to refer to the “waves” of this tracing, which are the main peaks protruding above or below the baseline. The 3 main waves are those of systolic contraction (S), diastolic early filling (E), and atrial contraction (A). The cardiac cycle also presents many minor displacements of the heart walls, such as during isovolumetric contraction and relaxation and diastasis (Figure 1).21 The displacement of the ventricular wall in a given time is derived by integrating the velocity profile over that time. For example, to ascertain the displacement of the mitral ring during systolic contraction, it is sufficient to integrate the S-wave velocity tracing over its entire profile. The positioning of sample volumes at selected sites was described by Pelà et al,20 who also provided an account of measurement variability (Figure 1).
1521 2.56 ⫾ 1.2 1.70 ⫾ 1.1 ⬍0.05
Cardiomyopathy/The Heart in the Churg-Strauss Syndrome
1522
The American Journal of Cardiology (www.AJConline.org)
the posterior wall as well left ventricular mass were similar in patients with CSS and controls. The ejection fraction (range 54% to 78%) and the fractional shortening of the LV were superimposable in the 2 groups (Table 1). Three patients showed alterations in regional wall motion. In 2, hypokinesia of the anterolateral apex was present; in the third, akinesia of the basal segment of the inferior wall was observed. Barely visible pericardial detachments were present in 6 patients. Mild mitral insufficiency was detected in 10 patients, and in 2 patients, slight aortic insufficiencies were found. Diastolic function: Doppler flow measurements: The absolute and relative velocities of diastolic E and A inflow waves are an established index of ventricular diastolic performance.23 Altered relaxation or ventricular stiffening is associated with a relatively slower E velocity and a relatively faster A velocity. We extended the measurements to the TVIs of the flow-wave profiles, which are a function of the mean rather than the peak velocity of either wave. The tricuspid inflow patterns in patients with CSS showed significant decreases in the EPV/APV and ETVI/ATVI ratios (Table 2). Using the 5th and 95th percentiles of the control distribution as a threshold for “abnormal” values, 5 and 4 patients had abnormal EPV/APV and ETVI/ATVI ratios, respectively. The mitral flow patterns showed significant decreases in ETVI (3 patients were in the ⬍5th percentile) and nonsignificant decreases in EPV and the EPV/APV and ETVI/ATVI ratios (Table 2). Doppler flow measurements were also the basis for estimating pulmonary pressure. The tricuspid regurgitant jet (patients with CSS: pulmonary artery systolic pressure 24.6 ⫾ 3.7 mm Hg, n ⫽ 5; controls: pulmonary artery systolic pressure 25.2 ⫾ 3.5 mm Hg, n ⫽ 11) and the acceleration time of systolic pulmonary flow (patients with CSS: acceleration time 128 ⫾ 16 ms, n ⫽ 12; controls: acceleration time 145 ⫾ 26 ms, n ⫽ 8) provided similar estimates of pulmonary pressure in the 2 groups. Pulsed Doppler tissue echocardiography: Changes in the motion pattern of the right ventricular lateral wall in diastole were similar to those of diastolic tricuspid inflow. When tissue motion is considered, the TVIs assume the simple meaning of the displacement of that wall; put another way, they represent the mean velocity times the duration of that movement. The EPV/APV and ETVI/ATVI ratios were significantly decreased in patients with CSS (Figure 2). Five patients had abnormal EPV/APV ratios, and 6 patients had abnormal ETVI/ATVI ratios. Figure 1 shows an example of an extremely altered Doppler tissue echocardiographic pattern. In the LV, a nonsignificant reduction of the EPV/APV and ETVI/ATVI ratios was observed in each of the 3 sites examined in subjects with CSS (Figure 2). Changes in systolic contraction were found only in the LV (Figure 3). STVI was significantly decreased in the
Figure 2. The E/A tissue ratio, in the 2 groups, evaluated as PV (top) and TVI (bottom) in the posterior wall, posteroseptum, and lateral wall of the LV and in the RV. *p ⫽ 0.08, †p ⬍0.01 (patients [white bars] vs controls [hatched bars]).
posterior wall and in the septum; a nonsignificant reduction was observed in the lateral wall. SPV was slightly reduced in each site. In the RV, SPV and STVI were unchanged (Figure 3). Correlations: Ultrasound diastolic indexes were tested for correlation with hematologic signs of disease (ESR, CRP, and eosinophil count). For the 2 diastolic waves and their ratio, the PV and the TVI were considered (EPV, ETVI, APV, ATVI, EPV/APV, and ETVI/ATVI). These 6 measurements yielded 6 ⫻ 3 possible correlations for each ultrasound sampling. In the RV, 18 correlations were possible with regard to tricuspid inflow. No correlation was significant. Eighteen correlations were also examined with regard to tissue motion. With a significance level of p ⬍0.05, 6 correlations were significant: ESR and APV, ESR and ATVI, ESR and the EPV/APV ratio (inverse), CRP and APV, CRP and ATVI, and eosinophil count and ETVI (inverse).
Cardiomyopathy/The Heart in the Churg-Strauss Syndrome
1523
Discussion
Figure 3. Comparison of controls and patients with CSS with regard to systolic velocities (top) and shortenings (bottom) in the LV, assessed in the posterior wall, posteroseptum, and lateral wall, and in the RV, explored at the lateral site of tricuspid annulus. *p ⫽ 0.08, †p ⬍0.05, ‡p ⬍0.01 (patients [white bars] vs controls [hatched bars]).
In the LV, 18 correlations were possible with regard to mitral inflow. None was significant. Fifty-four (18 ⫻ 3) matches were examined for the 3 samples of tissue motion. Six coefficients were significant: for the septum: ESR and APV, ESR and ATVI, CRP and APV, and CRP and ATVI; for the lateral wall: ESR and EPV and CRP and EPV. On the whole, 12 of 108 correlations were significant. This is slightly better than expected as a result of chance, but correlation coefficients never exceeded 0.6. Correlations between echocardiographic parameters and the clinical activity score offered little enlightenment. The same was true for correlations between ultrasound measurements and the duration of treatment.
Twenty-five percent of our patients were in full remission, whereas the rest (75%) had occurrences or recurrences of disease activity. These figures may not represent the general CSS population,17 because hospitalization may have produced some selection of patients with large activity scores. Despite this, the clinical presentation was milder in our group than in many previous publications, and the recognition of “cardiac involvement” depended entirely on the technique involved. Arrhythmias were rare, episodes of congestive heart failure were not recorded, and clinical pericaridits was not detected in any of the patients. Routine methods such as electrocardiography yielded amorphous data. Even bi-dimensional and Mmode echocardiography were not instructive. Minimal pericardial effusion was detected in 6 of 16 subjects. Data on contractility, as illustrated by the ejection fraction, showed a substantial patient– control overlap. Mild mitral regurgitation was present in 10 of 16 patients and 5 of 20 controls. It was mainly with Doppler quantitative echocardiography that some differences between groups were recognized. The flow pattern through the mitral and tricuspid valves indicated that diastolic kinetics were perturbed in patients with CSS. Relaxation in early diastole had, on average, a slower velocity compared with the end-diastolic dilation associated with atrial contraction. A smaller fraction of ventricular filling occurred in early than in late diastole. These changes were more evident in the RV than in the LV. They suggest a restrictive pathology, possibly linked to myocardial fibrosis, a feature of the hypereosinophilic syndromes.24,25 Doppler measurements of wall motion also reflect ventricular kinetics, with the proviso that specific portions of the myocardium are selected, whereas flow patterns reflect global kinetics. Three sites were sampled in the LV (lateral, septum, and posterior) and only 1 by the side of the tricuspid valve. The site-to-site differences of the LV raise a problem of interpretation, especially considering that statistical significance was associated with some sampling sites and not with others. A look at Figure 2 shows that the differences between the group means had the same sign for all the measurements around the mitral annulus. However, statistical significance (at the conventional level of p ⬍0.05) occurred where the dispersion (SD) of data was least. The same reasoning can be applied in Figure 3 for systolic parameters of the LV. It is again apparent that a small patient– control difference was present at the 3 sampled sites, but significance was associated with measurements having the least SDs (posterior and septum). One may add that resorting to nonparametric statistics, such as the MannWhitney U statistic or Tukey’s quick test, somewhat changed the assignment of p ⬍0.05 significance. For example, the first test assigned a significance also to the posterior sampling data in Figure 2, and the second test to the lateral sampling columns in Figure 3. In the end, we chose to illustrate the results of the Student’s t test because all data were compatible with the
1524
The American Journal of Cardiology (www.AJConline.org)
normal distribution and because this is the most popular statistical procedure in the biologic sciences. Poor correlations were found between ultrasound measurements and inflammatory markers, eosinophil count, or clinical activity score. All these variables were poorly correlated with one another. Della Rossa et al26 similarly reported that clinical and serologic parameters did not correlate in a group of 19 patients. Different explanations can be proposed. First, sensitive echocardiographic measurements may detect a subclinical phase of cardiomyopathy. Second, the vasculitis at a given moment may strike other organs while sparing the heart and vice versa. Third, we have noted a different time course in the response of hematologic versus tissue manifestations of the disease. For example, 1 single injection of prednisone reduces the eosinophil count dramatically but is ineffective on myocardial performance. Only 1 published echocardiographic study included a sizable (n ⫽ 12) number of patients with CSS, and it dates back to 1989.25 The main finding was a high frequency (6 of 12) of severe mitral regurgitation, which in 2 of 6 patients required surgery. Four of 12 patients had left ventricular dilation, with 3 of 4 having depressed shortening fractions. These findings are clearly different from ours. The investigators used continuous Doppler techniques to diagnose mitral regurgitation but not to evaluate diastolic function. A generalized increase of myocardial echocardiographic amplitude suggested myocardial fibrosis. Unfortunately, that report contains no mention of treatment and its effects. It is now common opinion that steroid-immunodepressant protocols suppress many manifestations of this syndrome, including left ventricular failure.2,3,16 We think that a different intensity of treatment explains the difference between many early studies and the present findings. In other respects, this is the first report investigating diastolic function in CSS and the first Doppler tissue echocardiographic study of this condition. For the sake of completeness, we must cite 1 single-case report in which diastolic dysfunction was inferred by means of cardiac catheterization; the measurement in question was left ventricular filling pressure.24 1. Churg J, Strauss L. Allergic granulomatosis, allergic angiitis, and periarteritis nodosa. Am J Pathol 1951;27:277–331. 2. Conron M, Beynon HL. Churg-Strauss syndrome. Thorax 2000;55: 870 – 877. 3. Noth I, Strek ME, Leff AR. Churg-Strauss syndrome. Lancet 2003; 361:587–594. 4. Hasley PB, Follansbee WP, Coulehan JL. Cardiac manifestations of Churg-Strauss syndrome: report of a case and review of the literature. Am Heart J 1990;120:996 –999. 5. Sharma A, De Varennes B, Sniderman AD. Churg-Strauss syndrome presenting with marked eosinophilia and pericardial effusion. Can J Cardiol 1993;9:329 –330. 6. Alvarez-Sala R, Prados C, Armada E, Del Arco A, Villamor J. Congestive cardiomyopathy and endobronchial granulomas as manifestations of Churg-Strauss syndrome. Postgrad Med J 1995;71:365–366.
7. Kozak M, Gill EA, Green LS. The Churg-Strauss syndrome. A case report with angiographically documented coronary involvement and a review of the literature. Chest 1995;107:578 –580. 8. Hellemans S, Dens J, Knockaert D. Coronary involvement in the Churg-Strauss syndrome. Heart 1997;77:576 –578. 9. Val-Bernal JF, Mayorga M, Garcia-Alberdi E, Pozueta JA. Churg-Strauss syndrome and sudden cardiac death. Cardiovasc Pathol 2003;12:94 –97. 10. Abu-Shakra M, Smythe H, Lewtas J, Badley E, Weber D, Keystone E. Outcome of polyarteritis nodosa and Churg-Strauss syndrome. An analysis of twenty-five patients. Arthritis Rheum 1994;37:1798 –1803. 11. Chumbley LC, Harrison EG, DeRemee RA. Allergic granulomatosis and angiitis (Churg-Strauss syndrome). Report and analysis of 30 cases. Mayo Clin Proc 1977;52:477– 484. 12. Gaskin G, Clutterbuck EJ, Pusey CD. Renal disease in the ChurgStrauss syndrome. Contrib Nephrol 1991;94:58 – 65. 13. Haas C, G ¸ eneau C, Odinor JM, De Jaeger C, Lavner M, Lowenstein W, Choubrac P, Durand H. L’angéite allergique avec granulomatose: syndrome de Churg et Strauss. Ann Med Interne 1991;142:335–342. 14. Lanham JG, Elkon KB, Pusey CD, Hughes G. Systemic vasculitis with asthma and eosinophilia: a clinical approach to the Churg-Strauss syndrome. Medicine 1984;63:65– 81. 15. Solans R, Bosch JA, Perez-Bocanegra C, Selva A, Huguet P, Alijotas J, Orriols R, Armadans L, Vilardell M. Churg-Strauss syndrome: outcome and long-term follow-up of 32 patients. Rheumatology 2001;40:763–771. 16. Balestrieri GP, Valentini U, Cerudelli B, Spandrio S, Renaldini E. Reversible myocardial impairment in the Churg-Strauss syndrome: report of a case. Clin Exp Rheumatol 1992;10:75–77. 17. Guillevin L, Cohen P, Gayraud M, Lhote F, Jarrousse B, Casassus P. Churg-Strauss syndrome. Clinical study and long term follow-up of 96 patients. Medicine 1999;78:26 –37. 18. Masi AT, Hunder GG, Lie JT, Michel BA, Bloch DA, Arend WP, Calabrese LH, Edworthy SM, Fauci AS, Leavitt RY, et al. The American College of Rheumatology 1990 criteria for the classification of Churg-Strauss syndrome (allergic granulomatosis and angiitis). Arthritis Rheum 1990;33:1094 –1100. 19. Luqmani RA, Bacon PA, Moots RJ, Janssen BA, Pall A, Emery P, Savage C, Adu D. Birmingham vasculitis activity score (BVAS) in systemic necrotising vasculitis. Q J Med 1994;87:671– 678. 20. Pelà G, Bruschi G, Montagna L, Manara M, Manca C. Left and right ventricular adaptation assessed by Doppler tissue echocardiography in athletes. J Am Soc Echo 2004;17:205–211. 21. Palka P, Lange A, Fleming AD, Sutherland R, Fenn LN, McDicken WN. Doppler Tissue imaging: myocardial wall motion velocities in normal subjects. J Am Soc Echocardiogr 1995;8:659 – 668. 22. Encyclopedia of Statistical Sciences. New York, New York: John Wiley, 1983. 23. Appleton CP, Hatle LK, Popp RL. Relation of transmitral flow velocity patterns to left ventricular diastolic function: new insights from combined hemodynamic and Doppler echocardiographic study. J Am Coll Cardiol 1988;12:426 – 440. 24. Frustaci A, Gentiloni N, Chimenti C, Natale L, Gasbarrini G, Maseri A. Necrotizing myocardial vasculitis in Churg-Strauss syndrome: clinicohistologic evaluation of steroids and immunosuppressive therapy. Chest 1998;114:1484 –1489. 25. Morgan JM, Raposo L, Gibson DG. Cardiac involvement in ChurgStrauss syndrome shown by echocardiography. Br Heart J 1989;62: 462– 466. 26. Della Rossa A, Baldini C, Tavoni A, Tognetti A, Neglia D, Sambuceti G, Puccini R, Colangelo C, Bombardieri S. Churg-Strauss syndrome: clinical and serological features of 19 patients from a single Italian centre. Rheumatology 2002;41:1286 –1294.
Churg-Strauss syndrome (CSS) is a rare necrotizing vasculitis of unknown origin that affects small-sized blood vessels.1,2 Cardiac involvement has been described as an eosinophilic vasculitis, pericarditis, and myocarditis, potentially resulting in heart failure and sometimes leading to death through cardiac tamponade, myocardial infarction, and arrest.3–9 However, several discrepancies appear in published research concerning the prevalence of cardiomyopathy in CSS and its clinical severity. For example, the frequency of cardiac involvement ranges from 17% to 92%.1,10 –15 In some reports, myocardial involvement was believed to be progressive,11–14 whereas in others, it was fully reversible.16,17 An explanation of these variations could lie either in the method used to assess the carditis (e.g., clinical vs instrumental) or in different intensities of treatment. It is now commonly accepted that patients with CSS greatly benefit Department of Internal Medicine, Nephrology and Prevention Sciences, University of Parma, Parma, Italy. Manuscript received August 10, 2005; revised manuscript received and accepted November 29, 2005. * Corresponding author: Tel: 390521033192; fax: 390521033185. E-mail address: [email protected] (G. Pelà). 0002-9149/06/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2005.11.088
from steroid and/or steroid-associated immunosuppressive drugs, but this was not equally clear in old reports. Sixteen patients were compared with 20 age- and gendermatched controls. All patients but 1 were receiving treatment.
Methods Subjects: Sixteen patients with CSS (9 men, 7 women) were enrolled in our study. The patients were evaluated by the interdisciplinary Secondary and Primary Vasculitides study group of the University of Parma from 1998 to 2005. All the patients were hospitalized and underwent physical examinations, standard blood and urinary tests, spirograms (14 of 16 patients) and chest x-rays (14 of 16 patients) before echocardiography. In all patients, CSS was diagnosed on the basis of the classification criteria of the American College of Rheumatology.18 Briefly, 4 of the following 6 findings were required: (1) a history of asthma, (2) eosinophilia ⬎10%, (3) mononeuropathy or polyneuropathy, (4) pulmonary nonfixed infiltrates, (5) paranasal sinus abnormality, and (6) biopsy evidence of extravascular eosinophil infiltration. Actuwww.AJConline.org
1520
The American Journal of Cardiology (www.AJConline.org)
Table 1 Subjects and their basic echocardiographic data Parameter Age (yrs) Body surface area (m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Heart rate (beats/min) Body mass index (kg/m2) LV end-diastolic diameter (cm) LV end-systolic diameter (cm) Septal wall thickness (mm) Posterior wall thickness (mm) LV mass index (g/m2) Relative wall thickness LV fractional shortening (%) Ejection fraction (%) RV end-diastolic diameter (cm) RV long-axis diastolic diameter (cm)
Controls
Patients With CSS
49 ⫾ 19 1.83 ⫾ 0.16 121 ⫾ 13 77 ⫾ 9 71 ⫾ 8 23.9 ⫾ 3.1 4.6 ⫾ 0.4 2.7 ⫾ 0.3 9.7 ⫾ 1.3 8.5 ⫾ 1.2 94 ⫾ 19 0.40 ⫾ 0.05 40 ⫾ 5 71 ⫾ 5 2.8 ⫾ 0.4 6.5 ⫾ 0.8
49 ⫾ 10 1.74 ⫾ 0.19 124 ⫾ 17 81 ⫾ 13 84 ⫾ 14* 24.6 ⫾ 2.6 4.6 ⫾ 0.5 2.9 ⫾ 0.5 9.2 ⫾ 1.5 8.8 ⫾ 1.5 100 ⫾ 24 0.39 ⫾ 0.06 40 ⫾ 6 68 ⫾ 7 3.0 ⫾ 0.2 6.7 ⫾ 0.7
Data are expressed as mean ⫾ SD. * p ⬍0.001, patients versus controls. LV ⫽ left ventricular; RV ⫽ right ventricular.
ally, only skin biopsies were performed in a sizable portion of patients, and a history of asthma was present in all. Organ system involvement was assessed using the Birmingham vasculitis activity score.19 Subjects with scores of 0 were considered in remission; all others were assumed to be in an active phase. A control group of 20 age-matched healthy subjects (10 men, 10 women) was also studied. The 2 groups had similar body surface areas (Table 1). Informed written consent was obtained from all participants. The research protocol was approved by the university’s ethics committee. All procedures conformed to the Declaration of Helsinki. All patients but 1 were treated at the time of the study. The treatment of active disease began with an intravenous bolus of methylprednisolone 1 g for 3 days, followed by oral prednisone 1 mg/kg/day. If remission was observed, the dose was tapered to 0.25 mg/kg/day after 1 year. An immunosuppressive drug, usually cyclophosphamide (2 patients were randomly assigned to methotrexate and 1 to azathioprine), was added to prednisone when (1) remission had not begun after 2 weeks of treatment; (2) the involvement of ⬎3 organ systems was found at presentation; and (3) polyneuropathy, which in our experience is particularly sensitive to cyclophosphamide, was present. The dose was adjusted to the clinical response and the leukocyte count. A count of 3,000/ mm3 was a threshold for decreasing the drug, and a count of 1,500/mm3 was a threshold for discontinuing the drug. Standard echocardiography: The subjects were examined by an experienced sonographer in the left lateral decubitus position using an Aspen instrument (Acuson Corporation, Mountain View, California) equipped with a multihertz sector probe (2 to 4 MHz).
Figure 1. Top panel, an apical 4-chamber view with some of the sampling sites of Doppler tissue measurements. Arrows, the direction of tissue motion. In the RV, the descent of the tricuspid annulus has been explored. On the left, the analogous motion of the mitral annulus was monitored. From the parasternal view, not shown here, the kinetics of the posterior wall were explored. Full black circles denote the sampling sites: lateral and posteroseptal for the LV and lateral for the RV. Bottom panel, a tricuspid annulus velocity pattern in a control subject (left) and in a patient with CSS (right). LA ⫽ left atrium; RA ⫽ right atrium.
The criteria for measuring wall thickness, cavity dimensions, volumes, and ventricular mass conformed to established conventions and are reported in detail elsewhere.20 Blood flow across the mitral and tricuspid valves was also determined by standard techniques. In the diastolic flow profile, the E and A waves were considered, and the peak velocity (PV) and the time–velocity integral (TVI) of either wave were extracted. The PV of the tricuspid regurgitant jet was used to estimate pulmonary artery systolic pressure, where pulmonary artery systolic pressure ⫽ 4V2 (where V is the peak tricuspid regurgitant velocity) ⫹ 10 (a conventional estimate for right atrial pressure). The acceleration time of pulmonary artery systolic flow was also used to obtain an indirect assessment of pulmonary arterial pressure. Pulsed Doppler tissue echocardiography: Doppler tissue echocardiography extracts a velocity tracing from the
1.56 ⫾ 0.5 1.04 ⫾ 0.3 ⬍0.01 4.0 ⫾ 1.2 6.0 ⫾ 1.9 ⬍0.001 33 ⫾ 7 45 ⫾ 10 ⬍0.001 9.2 ⫾ 2.3 8.7 ⫾ 2.9 NS 49 ⫾ 10 45 ⫾ 12 NS 1.77 ⫾ 0.8 1.5 ⫾ 0.6 NS 1.25 ⫾ 0.5 0.97 ⫾ 0.3 0.07 7.4 ⫾ 2.7 7.2 ⫾ 2.0 NS Data are expressed as mean ⫾ SD.
11.6 ⫾ 2.3 9.8 ⫾ 2.3 ⬍0.05 Controls Patients with CSS p Value
69 ⫾ 13 61 ⫾ 11 0.07
60 ⫾ 18 67 ⫾ 15 NS
ATVI (cm) EPV (cm/s) ETVI/ATVI (cm) EPV/APV (cm/s) ATVI (cm) ETVI (cm) EPV (cm/s)
APV (cm/s)
Mitral flow Variable
Cross-sectional echocardiography: In terms of basic echocardiography, no significant group differences were observed in the cavity dimensions of the left ventricle (LV) and the right ventricle (RV). The thickness of the septum and
Table 2 Mitral and tricuspid flow patterns
Results Baseline clinical features: According to the Birmingham vasculitis activity score, 4 patients were in complete remission and 12 were in an acute phase of the illness (Birmingham vasculitis activity score ranges from 5 to 27, mean 15.8). Heart rates were significantly greater in patients with CSS. The means of systolic and diastolic blood pressure were similar in the 2 groups (Table 1), but 8 patients were treated with antihypertensive drugs. Twelve-lead electrocardiograms showed normal patterns in 11 patients and repolarization alterations in 5 patients. One patient had a right bundle branch block. Spirograms indicated airway obstructions in 6 of 14 patients (mild in 5, severe in 1) and normal tracings in 7 of 14 patients; 1 patient had a mild impairment of the diffusing capacity of carbon monoxide. Fibrotic pulmonary nodules and interlobular septal thickening were detected in 7 of 14 patients on standard chest x-rays. At the time of admission, inflammatory markers were altered in 12 subjects with CSS: the erythrocyte sedimentation rate (ESR) in 5 (29.4 ⫾ 33.1 mm/hour) and C-reactive protein (CRP) in 5 (31.3 ⫾ 55.8 mg/dl). Eosinophilia was present in 10 subjects (mean 15.6 ⫾ 13.6% of the leukocyte count; total eosinophils 1,978 ⫾ 2,489/l). One patient had diabetes mellitus requiring oral hypoglycemic treatment.
ETVI (cm)
Statistical analysis: Data are expressed as mean ⫾ SD. The comparison between the group means was made by the 2-tailed Student’s t test, after ascertaining the compatibility of the data with the normal distribution (the KolmogorovSmirnov procedure). For comparison, nonparametric statistics were also assayed, including the Mann-Whitney U statistic, Wald-Wolfowitz runs, and Tukey’s quick tests. The bibliographic basis was the Encyclopedia of Statistical Sciences.22 SPSS software packages (SPSS, Inc., Chicago, Illinois) were used extensively.
APV (cm/s)
Tricuspid Flow
EPV/APV (cm/s)
ETVI/ATVI (cm)
oscillatory motion of myocardial walls during the heart cycle (Figure 1). It is usual to refer to the “waves” of this tracing, which are the main peaks protruding above or below the baseline. The 3 main waves are those of systolic contraction (S), diastolic early filling (E), and atrial contraction (A). The cardiac cycle also presents many minor displacements of the heart walls, such as during isovolumetric contraction and relaxation and diastasis (Figure 1).21 The displacement of the ventricular wall in a given time is derived by integrating the velocity profile over that time. For example, to ascertain the displacement of the mitral ring during systolic contraction, it is sufficient to integrate the S-wave velocity tracing over its entire profile. The positioning of sample volumes at selected sites was described by Pelà et al,20 who also provided an account of measurement variability (Figure 1).
1521 2.56 ⫾ 1.2 1.70 ⫾ 1.1 ⬍0.05
Cardiomyopathy/The Heart in the Churg-Strauss Syndrome
1522
The American Journal of Cardiology (www.AJConline.org)
the posterior wall as well left ventricular mass were similar in patients with CSS and controls. The ejection fraction (range 54% to 78%) and the fractional shortening of the LV were superimposable in the 2 groups (Table 1). Three patients showed alterations in regional wall motion. In 2, hypokinesia of the anterolateral apex was present; in the third, akinesia of the basal segment of the inferior wall was observed. Barely visible pericardial detachments were present in 6 patients. Mild mitral insufficiency was detected in 10 patients, and in 2 patients, slight aortic insufficiencies were found. Diastolic function: Doppler flow measurements: The absolute and relative velocities of diastolic E and A inflow waves are an established index of ventricular diastolic performance.23 Altered relaxation or ventricular stiffening is associated with a relatively slower E velocity and a relatively faster A velocity. We extended the measurements to the TVIs of the flow-wave profiles, which are a function of the mean rather than the peak velocity of either wave. The tricuspid inflow patterns in patients with CSS showed significant decreases in the EPV/APV and ETVI/ATVI ratios (Table 2). Using the 5th and 95th percentiles of the control distribution as a threshold for “abnormal” values, 5 and 4 patients had abnormal EPV/APV and ETVI/ATVI ratios, respectively. The mitral flow patterns showed significant decreases in ETVI (3 patients were in the ⬍5th percentile) and nonsignificant decreases in EPV and the EPV/APV and ETVI/ATVI ratios (Table 2). Doppler flow measurements were also the basis for estimating pulmonary pressure. The tricuspid regurgitant jet (patients with CSS: pulmonary artery systolic pressure 24.6 ⫾ 3.7 mm Hg, n ⫽ 5; controls: pulmonary artery systolic pressure 25.2 ⫾ 3.5 mm Hg, n ⫽ 11) and the acceleration time of systolic pulmonary flow (patients with CSS: acceleration time 128 ⫾ 16 ms, n ⫽ 12; controls: acceleration time 145 ⫾ 26 ms, n ⫽ 8) provided similar estimates of pulmonary pressure in the 2 groups. Pulsed Doppler tissue echocardiography: Changes in the motion pattern of the right ventricular lateral wall in diastole were similar to those of diastolic tricuspid inflow. When tissue motion is considered, the TVIs assume the simple meaning of the displacement of that wall; put another way, they represent the mean velocity times the duration of that movement. The EPV/APV and ETVI/ATVI ratios were significantly decreased in patients with CSS (Figure 2). Five patients had abnormal EPV/APV ratios, and 6 patients had abnormal ETVI/ATVI ratios. Figure 1 shows an example of an extremely altered Doppler tissue echocardiographic pattern. In the LV, a nonsignificant reduction of the EPV/APV and ETVI/ATVI ratios was observed in each of the 3 sites examined in subjects with CSS (Figure 2). Changes in systolic contraction were found only in the LV (Figure 3). STVI was significantly decreased in the
Figure 2. The E/A tissue ratio, in the 2 groups, evaluated as PV (top) and TVI (bottom) in the posterior wall, posteroseptum, and lateral wall of the LV and in the RV. *p ⫽ 0.08, †p ⬍0.01 (patients [white bars] vs controls [hatched bars]).
posterior wall and in the septum; a nonsignificant reduction was observed in the lateral wall. SPV was slightly reduced in each site. In the RV, SPV and STVI were unchanged (Figure 3). Correlations: Ultrasound diastolic indexes were tested for correlation with hematologic signs of disease (ESR, CRP, and eosinophil count). For the 2 diastolic waves and their ratio, the PV and the TVI were considered (EPV, ETVI, APV, ATVI, EPV/APV, and ETVI/ATVI). These 6 measurements yielded 6 ⫻ 3 possible correlations for each ultrasound sampling. In the RV, 18 correlations were possible with regard to tricuspid inflow. No correlation was significant. Eighteen correlations were also examined with regard to tissue motion. With a significance level of p ⬍0.05, 6 correlations were significant: ESR and APV, ESR and ATVI, ESR and the EPV/APV ratio (inverse), CRP and APV, CRP and ATVI, and eosinophil count and ETVI (inverse).
Cardiomyopathy/The Heart in the Churg-Strauss Syndrome
1523
Discussion
Figure 3. Comparison of controls and patients with CSS with regard to systolic velocities (top) and shortenings (bottom) in the LV, assessed in the posterior wall, posteroseptum, and lateral wall, and in the RV, explored at the lateral site of tricuspid annulus. *p ⫽ 0.08, †p ⬍0.05, ‡p ⬍0.01 (patients [white bars] vs controls [hatched bars]).
In the LV, 18 correlations were possible with regard to mitral inflow. None was significant. Fifty-four (18 ⫻ 3) matches were examined for the 3 samples of tissue motion. Six coefficients were significant: for the septum: ESR and APV, ESR and ATVI, CRP and APV, and CRP and ATVI; for the lateral wall: ESR and EPV and CRP and EPV. On the whole, 12 of 108 correlations were significant. This is slightly better than expected as a result of chance, but correlation coefficients never exceeded 0.6. Correlations between echocardiographic parameters and the clinical activity score offered little enlightenment. The same was true for correlations between ultrasound measurements and the duration of treatment.
Twenty-five percent of our patients were in full remission, whereas the rest (75%) had occurrences or recurrences of disease activity. These figures may not represent the general CSS population,17 because hospitalization may have produced some selection of patients with large activity scores. Despite this, the clinical presentation was milder in our group than in many previous publications, and the recognition of “cardiac involvement” depended entirely on the technique involved. Arrhythmias were rare, episodes of congestive heart failure were not recorded, and clinical pericaridits was not detected in any of the patients. Routine methods such as electrocardiography yielded amorphous data. Even bi-dimensional and Mmode echocardiography were not instructive. Minimal pericardial effusion was detected in 6 of 16 subjects. Data on contractility, as illustrated by the ejection fraction, showed a substantial patient– control overlap. Mild mitral regurgitation was present in 10 of 16 patients and 5 of 20 controls. It was mainly with Doppler quantitative echocardiography that some differences between groups were recognized. The flow pattern through the mitral and tricuspid valves indicated that diastolic kinetics were perturbed in patients with CSS. Relaxation in early diastole had, on average, a slower velocity compared with the end-diastolic dilation associated with atrial contraction. A smaller fraction of ventricular filling occurred in early than in late diastole. These changes were more evident in the RV than in the LV. They suggest a restrictive pathology, possibly linked to myocardial fibrosis, a feature of the hypereosinophilic syndromes.24,25 Doppler measurements of wall motion also reflect ventricular kinetics, with the proviso that specific portions of the myocardium are selected, whereas flow patterns reflect global kinetics. Three sites were sampled in the LV (lateral, septum, and posterior) and only 1 by the side of the tricuspid valve. The site-to-site differences of the LV raise a problem of interpretation, especially considering that statistical significance was associated with some sampling sites and not with others. A look at Figure 2 shows that the differences between the group means had the same sign for all the measurements around the mitral annulus. However, statistical significance (at the conventional level of p ⬍0.05) occurred where the dispersion (SD) of data was least. The same reasoning can be applied in Figure 3 for systolic parameters of the LV. It is again apparent that a small patient– control difference was present at the 3 sampled sites, but significance was associated with measurements having the least SDs (posterior and septum). One may add that resorting to nonparametric statistics, such as the MannWhitney U statistic or Tukey’s quick test, somewhat changed the assignment of p ⬍0.05 significance. For example, the first test assigned a significance also to the posterior sampling data in Figure 2, and the second test to the lateral sampling columns in Figure 3. In the end, we chose to illustrate the results of the Student’s t test because all data were compatible with the
1524
The American Journal of Cardiology (www.AJConline.org)
normal distribution and because this is the most popular statistical procedure in the biologic sciences. Poor correlations were found between ultrasound measurements and inflammatory markers, eosinophil count, or clinical activity score. All these variables were poorly correlated with one another. Della Rossa et al26 similarly reported that clinical and serologic parameters did not correlate in a group of 19 patients. Different explanations can be proposed. First, sensitive echocardiographic measurements may detect a subclinical phase of cardiomyopathy. Second, the vasculitis at a given moment may strike other organs while sparing the heart and vice versa. Third, we have noted a different time course in the response of hematologic versus tissue manifestations of the disease. For example, 1 single injection of prednisone reduces the eosinophil count dramatically but is ineffective on myocardial performance. Only 1 published echocardiographic study included a sizable (n ⫽ 12) number of patients with CSS, and it dates back to 1989.25 The main finding was a high frequency (6 of 12) of severe mitral regurgitation, which in 2 of 6 patients required surgery. Four of 12 patients had left ventricular dilation, with 3 of 4 having depressed shortening fractions. These findings are clearly different from ours. The investigators used continuous Doppler techniques to diagnose mitral regurgitation but not to evaluate diastolic function. A generalized increase of myocardial echocardiographic amplitude suggested myocardial fibrosis. Unfortunately, that report contains no mention of treatment and its effects. It is now common opinion that steroid-immunodepressant protocols suppress many manifestations of this syndrome, including left ventricular failure.2,3,16 We think that a different intensity of treatment explains the difference between many early studies and the present findings. In other respects, this is the first report investigating diastolic function in CSS and the first Doppler tissue echocardiographic study of this condition. For the sake of completeness, we must cite 1 single-case report in which diastolic dysfunction was inferred by means of cardiac catheterization; the measurement in question was left ventricular filling pressure.24 1. Churg J, Strauss L. Allergic granulomatosis, allergic angiitis, and periarteritis nodosa. Am J Pathol 1951;27:277–331. 2. Conron M, Beynon HL. Churg-Strauss syndrome. Thorax 2000;55: 870 – 877. 3. Noth I, Strek ME, Leff AR. Churg-Strauss syndrome. Lancet 2003; 361:587–594. 4. Hasley PB, Follansbee WP, Coulehan JL. Cardiac manifestations of Churg-Strauss syndrome: report of a case and review of the literature. Am Heart J 1990;120:996 –999. 5. Sharma A, De Varennes B, Sniderman AD. Churg-Strauss syndrome presenting with marked eosinophilia and pericardial effusion. Can J Cardiol 1993;9:329 –330. 6. Alvarez-Sala R, Prados C, Armada E, Del Arco A, Villamor J. Congestive cardiomyopathy and endobronchial granulomas as manifestations of Churg-Strauss syndrome. Postgrad Med J 1995;71:365–366.
7. Kozak M, Gill EA, Green LS. The Churg-Strauss syndrome. A case report with angiographically documented coronary involvement and a review of the literature. Chest 1995;107:578 –580. 8. Hellemans S, Dens J, Knockaert D. Coronary involvement in the Churg-Strauss syndrome. Heart 1997;77:576 –578. 9. Val-Bernal JF, Mayorga M, Garcia-Alberdi E, Pozueta JA. Churg-Strauss syndrome and sudden cardiac death. Cardiovasc Pathol 2003;12:94 –97. 10. Abu-Shakra M, Smythe H, Lewtas J, Badley E, Weber D, Keystone E. Outcome of polyarteritis nodosa and Churg-Strauss syndrome. An analysis of twenty-five patients. Arthritis Rheum 1994;37:1798 –1803. 11. Chumbley LC, Harrison EG, DeRemee RA. Allergic granulomatosis and angiitis (Churg-Strauss syndrome). Report and analysis of 30 cases. Mayo Clin Proc 1977;52:477– 484. 12. Gaskin G, Clutterbuck EJ, Pusey CD. Renal disease in the ChurgStrauss syndrome. Contrib Nephrol 1991;94:58 – 65. 13. Haas C, G ¸ eneau C, Odinor JM, De Jaeger C, Lavner M, Lowenstein W, Choubrac P, Durand H. L’angéite allergique avec granulomatose: syndrome de Churg et Strauss. Ann Med Interne 1991;142:335–342. 14. Lanham JG, Elkon KB, Pusey CD, Hughes G. Systemic vasculitis with asthma and eosinophilia: a clinical approach to the Churg-Strauss syndrome. Medicine 1984;63:65– 81. 15. Solans R, Bosch JA, Perez-Bocanegra C, Selva A, Huguet P, Alijotas J, Orriols R, Armadans L, Vilardell M. Churg-Strauss syndrome: outcome and long-term follow-up of 32 patients. Rheumatology 2001;40:763–771. 16. Balestrieri GP, Valentini U, Cerudelli B, Spandrio S, Renaldini E. Reversible myocardial impairment in the Churg-Strauss syndrome: report of a case. Clin Exp Rheumatol 1992;10:75–77. 17. Guillevin L, Cohen P, Gayraud M, Lhote F, Jarrousse B, Casassus P. Churg-Strauss syndrome. Clinical study and long term follow-up of 96 patients. Medicine 1999;78:26 –37. 18. Masi AT, Hunder GG, Lie JT, Michel BA, Bloch DA, Arend WP, Calabrese LH, Edworthy SM, Fauci AS, Leavitt RY, et al. The American College of Rheumatology 1990 criteria for the classification of Churg-Strauss syndrome (allergic granulomatosis and angiitis). Arthritis Rheum 1990;33:1094 –1100. 19. Luqmani RA, Bacon PA, Moots RJ, Janssen BA, Pall A, Emery P, Savage C, Adu D. Birmingham vasculitis activity score (BVAS) in systemic necrotising vasculitis. Q J Med 1994;87:671– 678. 20. Pelà G, Bruschi G, Montagna L, Manara M, Manca C. Left and right ventricular adaptation assessed by Doppler tissue echocardiography in athletes. J Am Soc Echo 2004;17:205–211. 21. Palka P, Lange A, Fleming AD, Sutherland R, Fenn LN, McDicken WN. Doppler Tissue imaging: myocardial wall motion velocities in normal subjects. J Am Soc Echocardiogr 1995;8:659 – 668. 22. Encyclopedia of Statistical Sciences. New York, New York: John Wiley, 1983. 23. Appleton CP, Hatle LK, Popp RL. Relation of transmitral flow velocity patterns to left ventricular diastolic function: new insights from combined hemodynamic and Doppler echocardiographic study. J Am Coll Cardiol 1988;12:426 – 440. 24. Frustaci A, Gentiloni N, Chimenti C, Natale L, Gasbarrini G, Maseri A. Necrotizing myocardial vasculitis in Churg-Strauss syndrome: clinicohistologic evaluation of steroids and immunosuppressive therapy. Chest 1998;114:1484 –1489. 25. Morgan JM, Raposo L, Gibson DG. Cardiac involvement in ChurgStrauss syndrome shown by echocardiography. Br Heart J 1989;62: 462– 466. 26. Della Rossa A, Baldini C, Tavoni A, Tognetti A, Neglia D, Sambuceti G, Puccini R, Colangelo C, Bombardieri S. Churg-Strauss syndrome: clinical and serological features of 19 patients from a single Italian centre. Rheumatology 2002;41:1286 –1294.