There were 17 subjects in the HES group and 24 patients in the group of healthy controls. HES patients were recruited between May 2012 and December 2021 prospectively in the Division of Haematology at the University of Szeged. Two patients were excluded from study in the HES group due to suboptimal image quality with 3D echocardiography. From the 15 patients participating in the study, 14 patients had idiopathic HES and 1 patient had HES associated with acute T-lymphoma (mean age: 61.7 $\pm$ 11.2 years, 10 males). The 24 healthy subjects were volunteers (mean age: 55.2 $\pm$ 7.9 years, 12 males), who had no symptoms, did not have any known cardiovascular risk factors or disorders or were taking any medications. None of the HES patients had any hypereosinophilia-related symptoms at baseline. Only some subjects had previous cardiovascular events including non-ST-elevation myocardial infarction (NSTEMI) (n = 1), deep vein thrombosis in the lower extremity (n = 1) and bilateral iliofemoral embolism (n = 1). All patients in the HES and in the control groups were in sinus rhythm without pulmonary embolism, chronic obstructive pulmonary disease or tumors in the past medical history. Regarding the classic cardiovascular risk factors in the HES patients, 1 patient was treated with diabetes mellitus, 8 patients had hypertension and 4 patients were diagnosed with hypercholesterolaemia. HES patients showed several extracardiac manifestations as well including eosinophilic asthma bronchiale (n = 1), eosinophilic dermatitis (n = 1), duodenal eosinophilia (n = 1), granulomatous necrotizing vasculitis and sensory-motor neuropathy with pulmonary involvement (n = 1), tissue (pulmonary) eosinophilia (n = 1), sole skin involvement (n = 1), and acute T-lymphoma associated hypereosinophilia with gastrointestinal involvement (n = 1). Two-dimensional (2D) and 3D echocardiography extended with Doppler assessments were completed in all subjects. The present substudy served as a part of Motion Analysis of the heart and Great vessels by three-dimensional speckle-tracking echocardiography in Pathological cases (MAGYAR-Path) Study evaluating echocardiographic parameters, including valvular annular parameters using 3D (speckle-tracking) echocardiography in various disorders. Ethical approval was provided by the Institutional and Regional Human Biomedical Research Committee of University of Szeged, Hungary (No.: 71/2011 and updated versions) and our study conducted in accordance with the ethical guidelines of the Declaration of Helsinki (1975 and updated versions). All HES patients and healthy controls gave an informed consent.

2D gray-scale images were performed using a broadband PST-30BT phased-array transducer (1–5 MHz) with a Toshiba Artida® (Toshiba Medical Systems, Tokyo, Japan) echocardiographic tool. The following echocardiographic parameters were measured in all patients: diameter of the left atrium, diameter and volume of the LV respecting the cardiac cycle, thickness of the interventricular septum and thickness of the LV posterior wall, LV ejection fraction measured using the Simpson’s method [11]. With pulsed Doppler transmitral flow velocities were measured in diastole and their ratio (E/A) were calculated. With tissue Doppler imaging E/E’ was measured as a ratio of early diastolic transmitral flow velocity (E) and mitral annular velocity (E’). With Doppler echocardiography tricuspid regurgitation pressure gradient was estimated, as well.

3D data acquisitions were completed by a Toshiba Artida${}^{\text{TM}}$ (Toshiba Medical Systems, Tokyo, Japan) cardiac ultrasound tool with a PST-25SX matrix-array transducer. From apical window, 6 subvolumes were collected in 6 heart cycles within a single breath-hold, then 3D Wall Motion Tracking software version 2.7 (Toshiba Medical Systems, Tokyo, Japan) was used for offline image analysis on the automatically created full volume 3D dataset [12, 13, 14]. Left atrium (LA) and LV dimensions were assessed using the parasternal long-axis view. LA volumes were evaluated from apical 2-chamber and 4-chamber longitudinal views, and the measured values were indexed to body surface area.

The end-diastole was considered when peak R wave was on electrocardiogram. The end-systole was considered as the first frame when the aortic valve was closed (at the end of T wave on electrocardiogram). MA/TA assessments were made using optimized image planes on the endpoints of the MA/TA on apical two- and four-chamber views and on C7 short-axis view. Several MA/TA measures and features of their function were calculated at end-diastole and at end-systole (Fig. 1) [15, 16]:

MA/TA dimensions:

$\bullet$ MA/TA diameter: perpendicular line connecting the peak of MA/TA curvature and the middle of the straight MA/TA border,

$\bullet$ MA/TA area: measured by planimetry,

$\bullet$ MA/TA perimeter: measured by planimetry.

MA/TA functional properties:

$\bullet$ MA/TA fractional shortening (MAFS/TAFS) = [end-diastolic MA/TA diameter – end-systolic MA/TA diameter]/end-diastolic MA/TA diameter $\times$ 100,

$\bullet$ MA/TA fractional area change (MAFAC/TAFAC) = [end-diastolic MA/TA area – end-systolic MA/TA area]/end-diastolic MA/TA area $\times$ 100.

Fig. 1.

Three-dimensional (3D) echocardiographic assessment of mitral and tricuspid annuli in a patient with hypereosinophilic syndrome. (A) apical four-chamber view, (B) apical two-chamber view and cross-sectional view (C7) of the mitral and tricuspid annuli optimalised on mitral and tricuspid annular images. Yellow arrows indicate plane of the mitral and tricuspid annuli. Abbreviations: Area, mitral/tricuspid annular area; Circ, mitral/tricuspid annular perimeter; Dist, mitral/tricuspid annular diameter; LA, left atrium; LV, left ventricle; RA, right atrium; RV, right ventricle.

While categorical data were expressed in counts and percentages (%), continuous variables were demonstrated in mean $\pm$ standard deviation format. Levene’s test was used for testing homogeneity of variances. Student’s t-test was used in case of normally distributed data, while non-normally distributed data were analyzed with Mann-Whitney Wilcoxon test. Chi-squared test and Fisher’s exact test were performed for statistical analysis of categorical variables. Pearson’s correlation coefficients were used to evaluate correlations. For intraobserver and interobserver correlations, intraclass correlation coefficients (ICCs) were calculated [17]. The Bland–Altman method was used to determine intraobserver and interobserver agreements [18]. Multivariable regression analysis with covariable age, hypertension, hyperlipidemia and presence of HES was used for assessment of independent predictors of reduced MAFAC and MAFS. All statistical tests were two-sided. Statistical significance was considered in case of p 0.05. Statistical analyses were performed using the MedCalc software package (MedCalc, Inc., Mariakerke, Belgium).

The incidence of hypertension and hyperlipidaemia was higher in the HES group, other demographic parameters were similar between the groups examined (Table 1).

Absolute eosinophil count (7.4 $\pm$ 4.4 $\times$ 10${}^9$/L vs. 0.2 $\pm$ 0.2 $\times$ 10${}^9$/L, p = 0.03), eosinophil ratio (45.2 $\pm$ 16.5% vs. 3.1 $\pm$ 2.1%, p = 0.01) and white blood cell count (16.2 $\pm$ 6.1 $\times$ 10${}^9$/L vs. 6.3 $\pm$ 0.9 $\times$ 10${}^9$/L, p = 0.02) were significantly increased in HES patients compared to that of controls. Other laboratory findings including red blood cell count (4.1 $\pm$ 0.6 T/L vs. 4.2 $\pm$ 0.3 T/L, p = 0.81), hemoglobin (125.0 $\pm$ 16.9 g/L vs. 129.1 $\pm$ 9.1 g/L, p = 0.80), platelet count (272.3 $\pm$ 156.1 $\times$ 10${}^9$/L vs. 271.9 $\pm$ 155.2 $\times$ 10${}^9$/L, p = 0.85) and hematocrit (36.5 $\pm$ 4.1% vs. 36.9 $\pm$ 5.3%, p = 0.88) did not differ between the groups.

Table 1 was used for demonstration of 2D echocardiographic parameters of HES patients and controls. Only interventricular septum was significantly thickened in patients with HES, no other parameter differed significantly between the groups examined. Only one HES patient showed grade 2 mitral regurgitation, other HES patients and controls did not show larger than grade 1 valvular insufficiency or had significant valvular stenoses.

Increased end-diastolic and end-systolic MA diameters, areas and perimeters together with reduced MAFAC and MAFS could be detected in HES patients as compared to those of controls. From TA morphological parameters, only end-diastolic TA area and end-systolic TA diameter were significantly increased in HES patients who had preserved TA functional parameters (Tables 2,3). Comparative analysis of MA and TA parameters was done between HES cases with vs. without hypertension without significant differences between these parameters (Table 4).

No correlations were found either between 2D and 3D echocardiography-derived parameters, or with any of the laboratory findings in HES patients. The logistic regression model identified presence of HES as an independent predictor of reduced MAFAC (hazard ratio (HR) 1.80, 95% CI of HR: 1.21 to 3.45, p 0.05) and MAFS (HR 1.76, 95% CI of HR: 1.20 to 3.32, p 0.05).

3D echocardiography-derived end-diastolic and end-systolic MA/TA dimensions were measured twice by the same observer (intraobserver agreement) and by two independent observers (interobserver agreement), the values were expressed as mean $\pm$ SD together with corresponding ICCs, the results are presented in Table 5.

Two out of 17 HES patients (12%) were excluded from the study as image quality was poor (inadequate for visual qualitative analysis with or without artifacts). The overall feasibility of MA/TA measurements proved to be 88%.

The following important limitations were present:

– Assessing any strains, rotational or dyssynchrony parameters of any chambers by 3D (speckle-tracking) echocardiography was not the aim of this study [12, 13, 14].

– HES is a rare disease, therefore we were able to collect clinical and 3D echocardiography-derived data of only relatively few HES patients [1, 2, 3, 4, 5, 6].

– There was a higher ratio of hypertension and hyperlipidemia in HES patients, which could affect result. However, comparative analysis between HES patients with vs. without hypertension did not find any differences in MA and TA data.

– In addition to the currently available technical development, lower temporal and spatial resolution features 3D echocardiography as compared to 2D echocardiography. This methodologic limitation could affect measurements [12, 13, 14].