Background
Several adult studies have shown that sickle cell disease is associated with cardiac abnormalities and premature death. The aim of this study was to use speckle-tracking strain, a relatively load independent parameter, to evaluate systolic left ventricular (LV) and right ventricular (RV) function in a pediatric sickle cell disease population.
Methods
Twenty-eight patients with sickle cell disease (mean age, 10.0 ± 3.6 years; mean body surface area, 1.14 ± 0.27 m 2 ) and 29 controls matched for age and body surface area were compared. Cardiac output, LV dimension, wall thickness and circumferential strain, LV and RV longitudinal systolic strain, conventional and tissue Doppler parameters, and pulmonary pressure were assessed.
Results
LV cardiac output was significantly higher in patients, as were indexed LV systolic diameter, indexed LV mass, and E/E′ septal ratio. Indexed LV diastolic diameter, wall thickness, LV shortening fraction, and global LV longitudinal and circumferential strains were similar in patients and controls. However, their global RV longitudinal strain was significantly lower, although tricuspid annular plane systolic excursion and color-coded tricuspid S-wave velocity were similar. Among patients, 21% had tricuspid regurgitation velocities > 2.5 m/sec, but none had tricuspid regurgitation velocities > 3 m/sec. Indexed LV diastolic dimension and systolic pulmonary artery pressure were significantly higher in patients whose hemoglobin was <80 g/L, but parameters of systolic and diastolic LV function were similar.
Conclusions
In children with sickle cell disease, LV diastolic function is significantly altered, although LV systolic function, evaluated by global longitudinal strain, is normal. In addition, cardiac output is increased, and elevated tricuspid regurgitation velocity is common, whereas it is never found in controls. Most importantly, global RV longitudinal systolic strain is significantly altered.
Sickle cell disease, a frequent hemoglobinopathy affecting an estimated 30 million people worldwide, is associated with long-standing severe hemolytic anemia and elevated cardiac output, in which the heart begins to dilate in childhood and gradually hypertrophies.
Erythrocyte sickling can lead to a state of resistance to nitric oxide bioactivity, small vessel obstruction, ischemia-reperfusion injury, and infarction in the myocardium and the lung. Sickle cell disease is associated with premature death. Although pulmonary hypertension and sudden death have been found to be major causes, other cardiac illnesses, such as myocardial infarction, heart failure, and cardiac arrhythmia, can account for up to 13% of deaths in adults.
Diastolic dysfunction and elevated pulmonary pressure have been found to be strong predictors of early and high mortality in the adult sickle cell disease population. Three recent studies confirmed that there is diastolic dysfunction and a high prevalence of mild pulmonary hypertension in the pediatric sickle cell disease population. However, because the investigators did not apply the commonly accepted definition of pulmonary hypertension (mean pulmonary arterial pressure > 25 mm Hg) and may have overestimated its importance and prevalence, we prefer to use the term “elevated pulmonary pressure.”
Systolic strain, a Doppler echocardiographic parameter that is independent of ventricular geometry and to a lesser extent of ventricular load, may be particularly useful to assess right ventricular (RV) function, potentially impaired by elevated pulmonary pressure, as well as by left ventricular (LV) diastolic dysfunction. In this study, we used systolic speckle-tracking strain to compare LV and RV function in an asymptomatic group of children and adolescents with sickle cell disease and in a control group. We hypothesized that systolic LV strain and/or RV strain could be altered in children with sickle cell disease. We also sought to link these cardiac parameters with the severity of anemia.
Methods
Study Population
In the sickle cell group, 28 consecutive patients ranging in age from 5 to 18 years (mean age, 10.0 ± 3.6 years) were recruited during their routine outpatient visits. Their disease, which was documented by hemoglobin electrophoresis, included homozygous SS ( n = 23), hemoglobin SC ( n = 3), hemoglobin Sβ+ thalassemia ( n = 1), and Sβ0 thalassemia ( n = 1). All but one of the patients were of Afro-Caribbean origin. None were in crisis at the time of echocardiography. Three patients were on chronic transfusion protocol, and three were on therapy with hydroxyurea. Hemoglobin levels were recorded the same day for all but three patients. Clinical records of the patients with sickle cell disease were reviewed for frequency and duration of exchange transfusion, and all underwent transcranial Doppler examinations for stroke screening.
In the control group, 29 children aged 4 to 13 years (mean age, 8.8 ± 2.9 years) and matched for age and body surface area (BSA) were recruited from the general population.
None of the 57 subjects had significant cardiovascular disease, according to clinical history, physical examination, electrocardiography, and complete two-dimensional echocardiography. None of them was taking cardiovascular medication at time of enrollment. The study was approved by the institutional review board.
Echocardiographic Study
Doppler Echocardiography
Doppler echocardiography with color flow was performed using a commercially available system (Vivid 7; GE Vingmed Ultrasound AS, Horten, Norway). Data were acquired from standard parasternal views and from apical four-chamber view. When children were >5 years old and able to cooperate, standard M-mode and two-dimensional images were obtained at end-expiratory apnea and stored in cine loop format for three consecutive beats. All echocardiographic measurements were made from digitally saved images, and observers were blinded to the subjects’ disease status. M-mode measurements were performed according to recommendations of the American Society of Echocardiography. All measurements were indexed to BSA. LV mass was calculated by using the American Society of Echocardiography cube method according to the Deveureux modification and normalized to BSA. Tricuspid annular plane systolic excursion was measured in the four-chamber view using M-mode echocardiography. Left atrial (LA) diameter was measured on M-mode imaging in the parasternal view from the posterior aortic wall to the posterior LA wall.
Mitral flow velocity was recorded in the apical four-chamber view, by placing the sample volume of pulsed-wave Doppler between the tips of the mitral leaflets. Then, peak A and E waves and E deceleration time were measured.
LV stroke volume was calculated by multiplying the LV outflow tract velocity-time integral by aortic valve cross-sectional area, as described previously. The LV outflow tract velocity-time integral, a pulsed Doppler value, was measured in centimeters by placing the sample volume in the LV outflow tract adjacent to the aortic valve and parallel to the blood flow direction. LV cardiac output was calculated by multiplying LV stroke volume by heart rate.
Valvular regurgitation was investigated and graded as mild, moderate, or severe. Systolic pulmonary artery pressure, on the basis of the velocity of the tricuspid regurgitation jet, was calculated using a modified Bernoulli equation and an estimation of right atrial pressure. Right atrial pressure did not vary between patients and controls, because none presented with dilated inferior venae cavae, so we estimated right atrial pressure at 5 mm Hg according to common practice.
Mean pulmonary artery pressure was calculated using the early diastolic velocity of pulmonary regurgitation according to a modified Bernoulli formula, with RV pressure considered to be 0 mm Hg in early diastole.
Speckle-Tracking Strain Analysis
Standard grayscale two-dimensional images were obtained in apical four-chamber and parasternal short-axis views, at the level of the papillary muscle. All images were recorded with a frame rate of 70 to 100 Hz and were digitally saved in cine loop format. Offline speckle-tracking analysis was performed using software for echocardiographic quantification (EchoPAC version 6.1.0; GE Vingmed Ultrasound AS). Endomyocardial borders of the left ventricle were manually traced within the end-systolic frame. The epicardial tracing was automatically generated by the software algorithm and, when necessary, manually adjusted to cover the whole myocardial wall. The tracking algorithm then followed the myocardial speckles during the cardiac cycle. Tracking was accepted only if both visual inspection and EchoPAC software analyses confirmed that it was adequate. Generally, we analyzed three heart cycles, but in a few cases in which cycle length and quality were too different, we assessed only two.
Myocardial longitudinal and circumferential strain values were obtained using this technique. Global strain was determined according to the average of strain values for the six segments taken in four-chamber and in short-axis views. Strain values are expressed as percentages. Negative strain values indicate shortening, and positive strain values indicate lengthening. Longitudinal strain for the left ventricle was assessed in the lateral wall, apex, and septum, whereas only the free wall was traced for the right ventricle ( Figure 1 ). All offline measurements with EchoPAC were performed by a single observer. Interobserver and intraobserver variability was determined by measurement of LV and RV strain parameters in 21 randomly selected subjects (14 patients and seven controls). To assess intraobserver variability in avoiding recall bias, the same observer reassessed the LV and RV segments 6 months later. To assess interobserver variability, a second observer, blinded to the results of the first, performed strain measurements at the same evaluation.
Doppler Tissue Imaging
Doppler tissue imaging (DTI) was evaluated in the four-chamber apical view using color-coded DTI, placing the sample volume at the septal mitral annulus to assess the E′-wave velocity and at the tricuspid annulus to measure S-wave velocity. Three consecutive cycles were recorded.
Offline Analysis
Digital images were obtained digital cine loop format for offline analysis in EchoPAC version BT06 (GE Vingmed Ultrasound AS). Global longitudinal LV and RV strain, as well as global LV circumferential strain, were measured using two-dimensional strain software.
Objectives
The primary objective was to use myocardial strain with two-dimensional speckle tracking echocardiography to evaluate systolic RV and LV function in a pediatric population with sickle cell disease compared with that in a control group. A secondary objective was to compare LV and RV function in the two groups according to more conventional criteria. Another was to compare functional parameters in the patients with sickle cell disease according to the severity of their anemia using an 80 g/L cutoff value for hemoglobin, a recognized factor for poor prognosis.
Statistical Analysis
Data are expressed as mean ± SD or as percentages. Because the number of subjects in the study was small and some of the variables did not have normal distributions, nonparametric Mann-Whitney U tests were used to assess differences in the two groups. Pearson’s correlation coefficients were used to assess the univariate relations between cardiac and selected clinical variables (age and hemoglobin) and between some dependent cardiac variables among themselves. P values < .05 were considered to indicate statistical significance. SAS (SAS Institute Inc., Cary, NC) was used for calculation.
Intraobserver and interobserver variability for the different parameters of strain were assessed in 21 randomly chosen study patients (14 with sickle cell disease and seven controls) using the Bland-Altman approach, including the calculation of mean bias (average difference between measurements) and the lower and upper limits of agreement (95% limits of agreement of mean bias). In addition, the coefficient of variation (standard deviation of the difference of paired samples divided by the average of the paired samples) and the intraclass correlation coefficient were assessed. Observers were blinded to the results. For the intraobserver and interobserver comparisons, the delay between the reviews was 6 months. P values < .05 were considered statistically significant.
Results
Clinical Characteristics
Patients’ clinical characteristics compared with those of controls are shown in Table 1 . Heart rate was significantly higher in patients than in controls ( Table 1 ).
| Variable | Controls ( n = 29) | Patients ( n = 28) | P |
|---|---|---|---|
| Male gender | 21 (75%) | 16 (57%) | .22 |
| Age (y) | 8.8 ± 2.9 | 10.0 ± 3.6 | .24 |
| BSA (m 2 ) | 1.07 ± 0.27 | 1.14 ± 0.27 | .36 |
| Heart rate (beats/min) | 79.1 ± 11.3 | 85.9 ± 8.6 | .024 |
Of the 28 patients, 57% were male. This study was representative of the prevalence of different genotypes of sickle cell disease in the French population, of which 80% are homozygous SS, 15% SC, and 5% Sβ thalassemia. Indeed, 82% of the patients we studied were SS, 10.5% were SC, and 6% were Sβ thalassemia. The mean hemoglobin level was 86.1 ± 16.5 g/L (rang, 66–129 g/L), and in 11 patients (39.3%), hemoglobin was <80 g/L. All patients were in New York Heart Association class I both at inclusion and at 2-year follow-up. Clinical history showed thoracic vaso-occlusive events in four patients (14%) at inclusion and in eight patients (28%) at 2-year follow-up. At 2-year follow-up, children who had had tricuspid regurgitation velocities (TRVs) > 2.5 m/sec at inclusion reported more episodes of acute chest syndrome than those who had not (50% vs 22.7%). Severe neurologic findings, defined as a history of stroke or pathology on transcranial Doppler, were found in four patients (14%) at inclusion; their frequency did not increase at 2-year follow-up.
M-Mode Parameters and Echocardiographic Findings
M-mode measurements in patients and controls are shown in Table 2 . Indexed systolic LV diameter was significantly greater in patients than in controls ( P = .017; Table 2 ). Indexed diastolic LV diameter also tended to be greater, but this did not reach statistical significance ( P = .057). Indexed LV mass was significantly greater in patients than in controls ( P < .0001; Table 2 ). According to their LV mass-for-height centile curves, eight patients with sickle cell disease (28.5%) had LV hypertrophy. Mild or discrete pulmonary regurgitation was found in 22 of the 28 patients, which allowed us to measure mean pulmonary artery pressure (12.4 ± 2.8 mm Hg). No patient had a mean pulmonary artery pressure > 25 mm Hg, which is the definition of pulmonary hypertension according to European guidelines. Mild tricuspid regurgitation was observed in 24 of the 28 patients (86%). Systolic pulmonary pressure was measurable in 18 of the 28 patients (64%) (mean, 25.5 ± 4.9 mm Hg). TRVs were <2.5 m/sec in 12 patients and between 2.5 and 2.9 m/sec in six patients.
| Variable | Controls ( n = 29) | Patients ( n = 28) | P |
|---|---|---|---|
| Indexed LVd (mm/m 2 ) | 38.8 ± 7.3 | 42.4 ± 6.7 | .057 |
| Indexed LVs (mm/m 2 ) | 24.2 ± 5.0 | 27.5 ± 4.7 | .017 |
| Indexed IVSd (mm/m 2 ) | 6.5 ± 1.5 | 6.9 ± 1.5 | . 21 |
| Indexed PWd (mm/m 2 ) | 5.5 ± 1.6 | 6.1 ± 1.6 | .08 |
| Indexed LV mass (g/m 2 ) | 57.5 ± 14.6 | 94.4 ± 25.9 | <.0001 |
| SF (%) | 37.7 ± 4.3 | 35.1 ± 4.1 | .11 |
| Cardiac output (L/min) | 3.45 ± 0.95 | 5.34 ± 1.28 | <.0001 |
| E wave (cm/sec) | 96.1 ± 20.3 | 122.3 ± 19.8 ( n = 27) | <.0001 |
| A wave (cm/sec) | 51.8 ± 12.4 | 57.2 ± 17.0 ( n = 27) | .19 |
| E/A ratio | 1.91 ± 0.5 | 2.25 ± 0.6 ( n = 27) | .03 |
| DT (msec) | 192.4 ± 59.3 ( n = 27) | 175.3 ± 37.5 ( n = 26) | .5 |
| E′ septal (cm/sec) | 10.3 ± 1.5 ( n = 29) | 10.9 ± 1.4 ( n = 26) | .33 |
| E/E′ | 9.4 ± 2 | 11.4 ± 2.1 ( n = 25) | .0013 |
| LA diameter (mm) | 23.2 ± 2.8 | 28.4 ± 3.1 | <.0001 |
| Systolic pulmonary pressure (mm Hg) | 20 ± 5.0 ( n = 2) | 26.5 ± 4.9 ( n = 18) | |
| Mean pulmonary pressure (mm Hg) | 10.5 ( n = 1) | 12.4 ± 2.8 ( n = 22) | |
| Global LV longitudinal strain (%) | −22.2 ± 2.6 | −22.5 ± 5.0 ( n = 25) | .48 |
| Global LV circumferential strain (%) | −17.9 ± 3.6 ( n = 25) | −19.2 ± 3.8 ( n = 24) | .6 |
| Global RV longitudinal strain (%) | −34.1 ± 5.0 ( n = 26) | −28.5 ± 5.6 ( n = 21) | .0003 |
| Tricuspid S wave (cm/sec) | 10.4 ± 1.4 ( n = 29) | 11.1 ± 1.6 ( n = 21) | .14 |
| TAPSE (mm) | 22.3 ± 3.4 ( n = 25) | 24.2 ± 4.2 ( n = 26) | .2 |
Because tricuspid regurgitation was absent or very mild in 27 of the 29 controls, pulmonary systolic pressure could not be measured. There was a positive relation between systolic pulmonary pressure and indexed LV mass in the sickle cell disease group ( r = 0.49, P = .037).
LV Diastolic Function
Functional diastolic Doppler echocardiographic parameters are shown in Table 2 . E waves in patients were higher than in controls ( P < .0001), as were E/E′ ratios ( P = .00125; Table 2 ). A waves in patients were higher than in controls, but this was not statistically significant ( P = .19). E′ color-coded DTI waves were similar in patients and in controls ( P = .33; Table 2 ). None of the patient showed a restrictive filling pattern when a deceleration time of 110 msec was used as a cutoff value for diagnosis. LA diameter, a potential marker of diastolic dysfunction, was found to be significantly higher in patients ( Table 2 ).
LV Systolic Function Using Conventional Parameters and Speckle-Tracking Strain
Systolic Doppler echocardiographic parameters are shown in Table 2 . Shortening fractions were within normal limits and were similar in both groups ( Table 2 ). Global longitudinal LV strain in patients was not different from that in controls (22.5 ± 5.0% vs 22.2 ± 2.6%, P = .48 ). Global circumferential LV strain was not different in patients than in controls (19.1 ± 4.0% vs 17.9 ± 3.6%, P = .259). LV cardiac output was significantly higher in patients than in controls ( P < .0001; Table 2 ).
RV Systolic Function Using Conventional Parameters and Speckle-Tracking Strain
Tricuspid annular plane systolic excursion was similar in the two groups ( Table 2 ). Color-coded DTI tricuspid annular S-wave velocities were also similar in the two groups ( Table 2 ). Global longitudinal RV strain was significantly lower in patients than in controls (−28.4 ± 5.5% vs −34.8 ± 4.5%, P = .0003). Because control subjects were well matched for age, sex, gender, and BSA, a covariance analysis was performed to test whether RV strain, when adjusted for cardiac output, remained lower in children with sickle cell disease than in controls, and the results were unchanged.
Subanalysis in the Sickle Cell Disease Group
Indexed LV diastolic dimension and systolic pulmonary artery pressure were significantly higher in patients whose hemoglobin was <80 g/L, but parameters of systolic and diastolic LV function were similar (fractional shortening, ejection fraction, global LV longitudinal strain, and global LV circumferential strain; Table 3 ). Accordingly, LV indexed systolic and diastolic diameters had inverse relations to hemoglobin level ( r = −0.438, P = .02, and r = −0.532, P = .04), but there was no correlation between hemoglobin level and E/E′ index ( r = −0.240).
| Variable | Hemoglobin ≤ 80 g/L | Hemoglobin > 80 g/L | P | ||
|---|---|---|---|---|---|
| Value | n | Value | n | ||
| Indexed LVd (mm/m 2 ) | 45.58 ± 4.85 | 11 | 40.28 ± 7.04 | 17 | .03 |
| SF (%) | 0.36 ± 0.04 | 11 | 0.35 ± 0.04 | 17 | .38 |
| EF (%) | 65.15 ± 5.43 | 11 | 63.89 ± 5.14 | 17 | .54 |
| Indexed LVs (mm/m 2 ) | 29.19 ± 3.65 | 11 | 26.4 ± 5.15 | 17 | .07 |
| IVSd (mm) | 7.64 ± 1.64 | 11 | 6.48 ± 1.4 | 17 | .06 |
| PWd (mm) | 6.38 ± 1.31 | 11 | 6.1 ± 1.61 | 17 | .64 |
| Indexed LV mass (g/m 2 ) | 98.98 ± 22.33 | 11 | 91.48 ± 28.2 | 17 | .46 |
| E wave (cm/sec) | 127.36 ± 22.69 | 11 | 118.88 ± 17.37 | 16 | .28 |
| A wave (cm/sec) | 59.82 ± 23.76 | 11 | 55.44 ± 10.86 | 16 | .52 |
| E/A ratio | 2.3 ± 0.65 | 11 | 2.21 ± 0.49 | 16 | .68 |
| DT (msec) | 178.91 ± 36.19 | 11 | 172.73 ± 39.43 | 15 | .69 |
| E′ septal (cm/sec) | 13.8 ± 1.69 | 10 | 14.38 ± 2.96 | 16 | .58 |
| E/E′ ratio | 11.7 ± 2.21 | 10 | 10.46 ± 3.44 | 16 | .41 |
| Systolic pulmonary pressure (mm Hg) | 29.83 ± 1.6 | 6 | 24.75 ± 5.08 | 12 | .01 |
| Mean pulmonary pressure (mm Hg) | 12.43 ± 3.51 | 7 | 12.45 ± 2.57 | 15 | .99 |
| Global LV longitudinal strain (%) | 21.44 ± 3.64 | 10 | 23.27 ± 5.71 | 17 | .64 |
| Global circumferential strain (%) | 19.86 ± 4.75 | 10 | 18.57 ± 3.6 | 12 | .97 |
| Global RV longitudinal strain (%) | 28.47 ± 6.28 | 9 | 28.53 ± 5.32 | 11 | .85 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
