Background
Chagas disease leads to biventricular heart failure, usually with prominent systemic congestion. Although echocardiography is widely used in clinical routine, the utility of echocardiographic parameters to detect right ventricular (RV) systolic dysfunction in patients with Chagas disease is unknown. We sought to study the diagnostic value of echocardiography, including speckle-tracking parameters, to distinguish individuals with RV systolic dysfunction from those with normal RV systolic function in Chagas disease using cardiac magnetic resonance (CMR) as the reference method.
Methods
In this cross-sectional study, 63 individuals with Chagas disease underwent echocardiography and CMR evaluations. Conventional echocardiographic parameters for RV functional evaluation were tricuspid annular plane systolic excursion, RV systolic excursion velocity, fractional area change, and RV index of myocardial performance. Strain and strain rate were obtained by two-dimensional speckle-tracking echocardiography and defined as “RV free wall,” when based only in segments from RV free wall, or “RV free wall and septum,” when segments from both free RV wall and interventricular septum were included. RV systolic dysfunction was defined as RV ejection fraction (RVEF) < 50% by CMR.
Results
Mean age was 56 ± 14 years, and 58.7% of the patients were men. RV systolic dysfunction was detected by CMR in 18 (28.6%) individuals. RV free wall strain showed the highest correlation with RVEF by CMR ( r = −0.62, P < .001), followed by fractional area change ( r = 0.56, P < .001), RV free wall and septum strain ( r = −0.54, P < .001), RV free wall and septum strain rate ( r = −0.47, P < .001), RV free wall strain rate ( r = −0.45, P < .001), and RV systolic excursion velocity ( r = 0.30, P = .016). The RV index of myocardial performance and tricuspid annular plane systolic excursion showed a small and not significant correlation with RVEF ( r = −0.20, P = .320; r = 0.14; P = .289, respectively). Using predefined cutoffs for RV systolic dysfunction, RV free wall strain (>−22.5% for men and >−23.3% for women) exhibited the highest area under the receiver operating characteristic curve (area under the curve = 0.829) to differentiate the presence from the absence of RV systolic dysfunction in Chagas disease, with a sensitivity and specificity of 67% and 83%, respectively.
Conclusions
RV free wall strain is an appropriate and superior echocardiographic variable for evaluating RV systolic function in Chagas disease, and it should be the method of choice for this purpose.
Highlights
- •
Speckle-tracking echocardiography is a valuable technique to assess RV function.
- •
Right ventricular (RV) free wall strain is the method of choice for RV evaluation in Chagas disease.
- •
The sensitivity of tricuspid annular plane systolic excursion, RV systolic excursion velocity, fractional area change, and RV index of myocardial performance to detect RV dysfunction is much lower.
Chagas disease, a parasitic infection caused by the protozoan Trypanosoma cruzi, is a major public health problem, leading to a high global economic burden and social impact. The World Health Organization estimates that around 7 million people are infected worldwide, mostly in the endemic areas of 21 Latin American countries. Population interchange between endemic and nonendemic areas has increased the number of infected individuals living in other countries, such as the United States and European countries.
Roughly 30%–40% of individuals infected with T. cruzi will develop chronic cardiomyopathy, the most serious and frequent manifestation of Chagas disease. Right ventricular (RV) impairment is a common finding, demonstrated even in the early stages of the disease, as previously reported in seminal studies using radionuclide angiography. Heart failure is usually a late manifestation, characterized by the predominance of systemic congestion over signs of pulmonary congestion. The presence of RV dysfunction is associated with poor prognosis at this advanced stage. Despite RV function playing a pivotal role in Chagas disease, accuracy of echocardiography to detect RV impairment in this clinical setting has not been assessed.
Assessment of RV structure and function by echocardiography can be challenging, especially because of the complex morphofunctional characteristics of this cardiac chamber. The right ventricle shows a crescent shape in a cross-sectional view, while it appears triangular when viewed from the side. In addition, the RV cavity encompasses three distinct components: the inlet, the main cavity, and the outflow regions, which are usually difficult to visualize all together in a single two-dimensional plane. The moderator band and prominent trabeculae in the midapical region frequently preclude an appropriate visualization of the interface between the endocardium and the cavity of the right ventricle. Furthermore, a deep subendocardial layer of longitudinal fibers leads to a predominantly base-to-apex contraction, in contrast to the left ventricular (LV) contraction, which is directed primarily towards the LV cavity centroid.
Despite those obstacles, echocardiography can provide several parameters to assess RV structure and function in clinical practice. Recently, speckle-tracking echocardiography has allowed the analysis of RV strain, a measurement of myocardial deformation, which has demonstrated higher accuracy for the detection of RV systolic dysfunction in various clinical settings, such as myocardial infarction and idiopathic dilated cardiomyopathy.
Cardiac magnetic resonance imaging (CMR) permits a highly accurate assessment of the right ventricle and is considered a reference method for this purpose. However, echocardiography is widely available, can be performed at bedside, and is relative inexpensive, thus it is extensively used in the clinical routine.
The primary purpose of this study was to determine the ability of echocardiographic parameters, including speckle-tracking-derived measures, to distinguish the presence from the absence of RV systolic dysfunction in patients with Chagas disease using CMR as the standard reference method.
Methods
Study Population
This prospective cross-sectional study included 63 individuals with chronic Chagas disease, defined by the positivity of at least two distinct serological tests (ELISA, indirect immunofluorescence, indirect hemagglutination) and age ≥ 18 years, who were recruited in the outpatient clinics at the Hospital das Clínicas de Ribeirão Preto, University of São Paulo, Brazil, from July 2012 to August 2014. Exclusion criteria were presence of other cardiomyopathies, obstructive coronary disease, severe cardiac valve disease, or systemic disease with potential effect on the RV function. All participants underwent standard 12-lead electrocardiography (ECG), chest x-ray, two-dimensional echocardiography including speckle-tracking analysis, and CMR. The mean time interval between CMR and echocardiography examinations was 24 ± 16 days, with no changes in clinical or medication status in this period of time.
The institutional research ethics committee approved the study protocol (process no. 4913/2010), which was conducted in accordance with the Helsinki Declaration. Written informed consent was obtained from all participants.
Echocardiography
Transthoracic echocardiography was performed in all individuals using the commercially available ultrasound system Vivid E9 (GE Healthcare, Horten, Norway), with a phased-array transducer of 1.5–4.6 MHz. Acquisitions and analyses were performed by a single certified and experienced physician blinded to CMR data. The exams were recorded in vendor-specific format (raw data). Acquisition and reading protocols followed guidelines previously published. Briefly, individuals were examined in left lateral and dorsal supine decubitus, utilizing the conventional parasternal, apical, and subcostal views, including the apical four-chamber view focused on the right ventricle (AP4-RV). Images were acquired with simultaneous electrocardiographic signal recording, during quiet respiration. At least three entire cardiac cycles were retrospectively recorded, increasing to five entire cardiac cycles in the presence of cardiac arrhythmia. All right chamber parameters were analyzed in the AP4-RV view, with the exception of RV outflow tract dimension, determined at the parasternal long-axis view. RV linear dimensions were measured at end diastole and were described as longitudinal dimension (from the plane of tricuspid annulus to the RV apex) and basal diameter and midcavity diameter (representing the maximal RV short-axis dimension in the basal one third and in the mid-third, respectively). RV systolic function was assessed by the nonvolumetric conventional parameters recommended for the clinical routine: tricuspid annular plane systolic excursion (TAPSE), systolic excursion velocity (RVS′), and fractional area change (FAC), respectively, by M-mode, pulsed tissue Doppler, and two-dimensional echocardiography. RV index of myocardial performance (RIMP), also known as the Tei index, a global estimation of both systolic and diastolic function, was evaluated as well, utilizing the pulsed tissue Doppler method ( Figure 1 ). Right atrium structural analysis included longitudinal and transverse diameters, maximum area, and maximum volume (estimated by disk summation technique). Pulmonary artery systolic pressure (PASP) was estimated when tricuspid regurgitation was present.
Speckle-Tracking Echocardiography
Two-dimensional speckle-tracking analysis was applied to assess RV longitudinal deformation in all 63 participants using the validated EchoPAC software (GE Healthcare), version 112. Images dedicated to speckle-tracking analysis were recorded at a frame rate from 40 to 80 frames/sec. End diastole was automatically defined at the beginning of the QRS complex. End systole was manually determined at the time of the aortic valve closure. Manual RV endocardial border delineation was performed in the AP4-RV view. Thereafter, epicardial borders were automatically defined, creating a region of interest (ROI). If necessary, adjustments in ROI width were made to include the entire RV myocardium region, based on a six-segment model. Finally, myocardium tracking was automatically obtained. Also, a tracking quality control was automatically displayed and a new tracing was performed when at least one segment was poorly tracked. In the presence of a persistent poor tracking of at least one segment, deformation analysis was considered as not feasible. RV deformation was evaluated by longitudinal strain, measured as the peak systolic change in the myocardial length relative to the length at end diastole and represented by percentage of change (%). In addition, longitudinal strain rate, defined as the deformation rate and estimated as the strain temporal derivative, was measured at its systolic peak and expressed as change of strain per unit of time (1/sec). Both RV strain and strain rate were studied using two different approaches: (1) considering all six segments of the RV visualized in the AP4-RV view in order to obtain RV free wall and septum strain and (2) including only the free wall segments (basal, midventricular, and apical; Figure 2 ). Global LV strain was also measured, using the apical four-, three-, and two-chamber views. Longitudinal systolic strain and strain rate reflect shortening, thus better systolic function is represented by more negative values of strain and strain rate.
CMR Imaging
Cine magnetic resonance imaging was acquired using steady-state free precession pulse sequence gated by ECG, obtained with a 1.5 Tesla scanner Achieva (Phillips, the Netherlands), following a predefined protocol: repetition time, 3.8 msec; echo time, 6 msec; flip angle, 45°; 30 acquisition phases; matrix, 256 × 160; and field of view, 360–400 mm. Left and right ventricles were analyzed in short-axis slices (thickness, 8 mm; gap between slices, 2 mm). All exams were recorded in the Digital Imaging and Communication in Medicine pattern and analyzed by a single certified and experienced reader (CMR reader, G.J.V.; cechocardiography reader, H.T.M.; second echocardiography reader for the reproducibility analysis, L.G.G.), blinded to echocardiographic data, utilizing MASS software (Leiden University, Leiden, the Netherlands). Simpson’s disk summation technique was used to determine RV ejection fraction (RVEF) and LV ejection fraction (LVEF) from short-axis slices. The first basal slice was defined as the one immediately adjacent to the atrioventricular junction. RV outflow tract was not included in the analysis of RV volumes. The largest and the smallest volumes throughout the cardiac cycle were used to define end-diastolic and end-systolic volumes, respectively. Endocardial borders were manually delineated. Papillary muscles and trabeculae were considered as part of ventricular cavity. RV systolic dysfunction was defined as RVEF < 50%, as described elsewhere.
Reproducibility
Reproducibility analysis for echocardiographic parameters of RV functional assessment was performed in 20 subjects randomly selected. For intraobserver reproducibility, a rereading was done at least 30 days after the first reading, blinded to the previous results. For interobserver analysis, the second independent reader was blinded to the analysis of the first reader. Inter- and intrareader reproducibility analyses were evaluated using an intraclass correlation coefficient to assess absolute agreement and a coefficient of variation, determined by the standard deviation divided by the mean and expressed in percentage.
Statistical Analysis
Normality of continuous data was assessed by histograms and the Shapiro-Wilk test. Continuous data are expressed as mean ± standard deviation if normally distributed, or as median [interquartile range] if not normally distributed. Categorical data are presented as absolute values and percentages. Student’s t test and Wilcoxon rank-sum test were used to compare clinical and echocardiographic characteristics between two subgroups defined according to the presence or absence of RV systolic dysfunction. Correlation of RVEF (measured by CMR) with echocardiographic parameters of RV functional assessment was verified by Pearson’s correlation coefficient. Receiver operating characteristic (ROC) analysis was performed to evaluate the ability of echocardiographic parameters to distinguish the presence from the absence of RV systolic dysfunction, defined as reduced RVEF by CMR as described above, in patients with Chagas disease. Sensitivity and specificity of each echocardiographic parameter were tested using cutoffs previously specified for identification of RV systolic dysfunction: TAPSE < 17 mm, RVS′ < 9.5 cm/sec, FAC < 35%, and tissue-Doppler derived RIMP > .54. For RV free wall longitudinal strain, RV systolic dysfunction was previously determined as >−22.5% for men and >−23.3% for women, whereas for RV free wall and septum strain, RV systolic dysfunction was >−20.0% and >−20.3% for men and women, respectively. Cutoffs to identify RV systolic dysfunction using strain rate parameters are not well established in the literature, thus sensitivity and specificity for these parameters were not tested in this study. P < .05 was considered statistically significant. Statistical analysis was performed using Stata 14.0 (StataCorp, College Station, TX).
Results
Demographic and clinical characteristics of the participants are described in Table 1 . The mean age was 56 ± 14 years, and 37 (58.7%) patients were men. Most participants were in New York Heart Association functional class I [49 (77.8%)]. Hypertension was the most prevalent comorbidity, present in 29 (46%) individuals. RVEF ranged from 26.3% to 75.0% (mean, 55.7± 11.6%). Reduced RVEF was identified in 18 (28.6%) individuals. LVEF measured by CMR ranged from 11.2% to 71.4% (mean, 48.3% ± 13.3%).
| N = 63 | |
|---|---|
| Age (y) | 56 ± 14 |
| Males | 37 (58.7%) |
| Body mass index (kg/m 2 ) | 22.1 [21.4–23.3] |
| New York Heart Association functional class | |
| I | 49 (77.8%) |
| II | 12 (19.0%) |
| III | 2 (3.2%) |
| Comorbidities | |
| Hypertension | 29 (46%) |
| Dyslipidemia | 23 (36.5%) |
| Diabetes mellitus | 5 (7.9%) |
| Ischemic stroke | 4 (6.3%) |
| History of smoking | 21 (33.3%) |
| Medications | |
| ACEI or ARB | 27 (42.9%) |
| Spironolactone | 9 (14.3%) |
| Beta blocker | 22 (34.9%) |
| Calcium channel blocker | 3 (4.8%) |
| Furosemide | 10 (15.9%) |
| Thiazide diuretic | 11 (17.5%) |
| Amiodarone | 10 (15.9%) |
| Digoxin | 4 (6.3%) |
| ECG | |
| Sinus rhythm | 54 (85.7%) |
| Atrial fibrillation | 6 (9.5%) |
| Right atrial rhythm | 3 (4.8%) |
| RBBB | 21 (33.3%) |
| Incomplete RBBB | 6 (9.5%) |
| Left bundle branch block | 3 (4.8%) |
| Left anterior fascicular block | 22 (34.9%) |
| Low QRS voltage | 10 (15.9%) |
| Chest x-ray | |
| Cardiomegaly | 19 (30.2%) |
Due to poor image quality, TAPSE was not feasible in one individual, whereas RVS′, FAC, and RIMP were not feasible in two participants. RV speckle-tracking analyses were not feasible in three examinations due to persistently poor tracking, despite retracing attempts.
In comparison with the subgroup with preserved RV systolic function, the subgroup with RV systolic dysfunction showed significantly lower LVEF (53% [36%–62%] vs 62% [51%–67%]; P < .001) and lower LV systolic deformation, reflected by less negative values of global LV longitudinal strain (−13.3% ± 4.5% vs −16.0% ± 3.8%; P = .042; Table 2 ). There was no association of RV or right atrial dimensions with RVEF measured by CMR. Of note, RV basal diameter showed a trend toward significance to be higher in the RV systolic dysfunction subgroup (38 ± 6 mm vs 35 ± 6 mm; P = .069). RV longitudinal diameter showed the highest correlation with RV end-diastolic volume by CMR ( r = 0.66, P < .001), followed by RV basal diameter ( r = 0.58, P < .001) and RV midcavity diameter ( r = 0.43, P < .001).
| All, N = 63 | No RV systolic dysfunction, n = 45 (71.4%) | With RV systolic dysfunction, n = 18 (28.6%) | P value | |
|---|---|---|---|---|
| Left chambers | ||||
| LA volume/BSA (mL) | 31 [23–40] | 31 [24–40] | 31 [23–41] | .659 ∗ |
| LV end-diastolic diameter (mm) | 50 ± 8 | 50 ± 8 | 52 ± 10 | .251 † |
| LV mass index (g/m 2 ) | 156 [130–195] | 154 [126–195] | 159 [144–184] | .600 ∗ |
| LVEF (%) | 61 [48–67] | 62 [51–67] | 53 [36–62] | .013 ∗ |
| LV longitudinal peak systolic strain (%) | −15.6 ± 4.1 | −16.0 ± 3.8 | −13.3 ± 4.5 | .020 † |
| LV wall motion score index | 1.1 [1.0–1.6] | 1.1 [1.0–1.4] | 1.1 [1.0–1.6] | .519 ∗ |
| Septal e′ (cm/sec) | 6 [5–8] | 6 [5–8] | 5 [6–8] | .470 ∗ |
| Septal E/e′ (cm/sec) | 11 [8–14] | 11 [8–13] | 10 [9–21] | .575 ∗ |
| Right atrium | ||||
| RA transversal diameter (mm) | 37 [33–42] | 36 [33–42] | 38 [34–42] | .341 ∗ |
| RA longitudinal diameter (mm) | 47 [44–51] | 47 [44–51] | 47 [44–50] | .896 ∗ |
| RA area (m 2 ) | 16 [14–22] | 16 [14–22] | 18 [14–28] | .197 ∗ |
| RA volume/BSA (mL) | 23 [18–34] | 23 [18–34] | 23 [18–31] | .867 ∗ |
| Right ventricle | ||||
| RV outflow tract proximal diameter (mm) | 31 ± 3 | 32 ± 4 | 31 ± 4 | .373 † |
| RV basal diameter (mm) | 36 ± 6 | 35 ± 6 | 38 ± 6 | .069 † |
| RV midcavity diameter (mm) | 24 ± 5 | 24 ± 5 | 24 ± 6 | .661 † |
| RV longitudinal diameter (mm) | 72 ± 10 | 71 ± 9 | 73 ± 10 | .369 † |
| RV lateral e′ (cm/sec) | 10 [8–12] | 10 [8–12] | 10 [8–12] | .597 ∗ |
| RV lateral a′ (cm/sec) | 15 [12–18] | 15 [12–19] | 15 [11–15] | .214 ∗ |
| RV lateral e′/a | 0.65 [0.53; 0.89] | 0.64 [0.50; 0.89] | 0.67 [0.53; 0.85] | .474 ∗ |
| TAPSE (mm) | 19 ± 3 | 20 ± 4 | 19 ± 3 | .329 † |
| RV systolic velocity (RVS′) (cm/sec) | 11.8 ± 2.5 | 12.1 ± 2.6 | 10.9 ± 2.2 | .107 † |
| FAC (%) | 47 ± 11 | 50 ± 11 | 41 ± 9 | .003 † |
| RIMP | 410 [364–500] | 402 [353–500] | 476 [366–533] | .438 ∗ |
| RV free wall and septum strain (%) | −21.9 ± 4.5 | −23.3 ± 3.9 | −18.6 ± 4.3 | <.001 † |
| RV free wall strain (%) | −26.1 ± 5.9 | −28.2 ± 5.0 | −21.2 ± 4.9 | <.001 † |
| RV free wall and septum strain rate (1/sec) | −1.02 ± 0.2 | −1.07 ± 0.20 | −0.92 ± 0.19 | .008 † |
| RV free wall strain rate (1/sec) | −1.45 ± 0.3 | −1.46 ± 0.35 | −1.28 ± 0.34 | .060 † |
Mitral regurgitation was identified in 22 (34.9%) subjects, all of them classified as secondary, with normal valve leaflets and chordae, mostly in mild degree, except for three cases graded as moderate (one in the RV dysfunction subgroup). Tricuspid regurgitation was found in 24 (38.1%) individuals, with only one classified as moderate, in an individual with RV dysfunction.
PASP was estimated in all 24 participants with tricuspid regurgitation. PASP in those with preserved RVEF ( n = 16) ranged from 21 to 59 mmHg (median, 33 [IQR, 28–38]), while in those with reduced RVEF ( n = 8), it varied from 23 to 48 mmHg. In the other 10 individuals with reduced RVEF but absence of tricuspid regurgitation, pulmonary artery acceleration time (PAAcT) was < 100 msec, indicating pulmonary hypertension, in one patient (PAAcT = 88 ms), borderline (PAAcT ≥ 100 and < 130 ms) in other seven individuals (ranging from 104 to 120 msec), and not feasible in two participants due to poor quality in Doppler tracing.
Correlations of LV global strain with RV free wall strain and RV free wall and septum strain were r = 0.42 ( P = .001) and r = 0.55 ( P < .001), respectively.
Assessment of late gadolinium enhancement (LGE) of the right ventricle by CMR can be problematic due to the thin walls of this cardiac chamber. Although evaluation of LGE has not been included in the original design of this study, LGE was performed for clinical purposes. Examining LGE in the right ventricle, we found suggestive signs of focal fibrosis in only four patients.
Echocardiographic Ability to Distinguish the Presence from the Absence of RV Systolic Dysfunction
RV free wall strain exhibited the highest correlation with RVEF measured by CMR ( r = −0.62, P < .001), followed by FAC ( r = 0.56, P < .001), RV free wall and septum strain ( r = −0.54, P < .001), RV free wall and septum strain rate ( r = −0.47, P < .001), RV free wall strain rate ( r = −0.45, P < .001), and RVS′ ( r = 0.30, P = .016; Table 3 ). RIMP and TAPSE showed a small and not significant correlation with RVEF ( r = −0.20, P = .32; r = 0.14; P = .289, respectively).
