Jump to
Abstract
Systemic arterial hypertension predisposes children to cardiovascular risk in childhood and adult life. Despite extensive study of left ventricular (LV) hypertrophy, detailed 3-dimensional strain analysis of cardiac function in hypertensive children has not been reported. The aim of this study was to evaluate LV mechanics (strain, twist, and torsion) in young patients with hypertension compared with a healthy control group and assess factors associated with functional measurements. Sixty-three patients (26 hypertension and 37 normotensive) were enrolled (mean age, 14.3 and 11.4 years; 54% men and 41% men, respectively). All children underwent clinical evaluation and echocardiographic examination, including 3-dimensional strain. There was no difference in LV volumes and ejection fraction between the groups. Myocardial deformation was significantly reduced in those with hypertension compared with controls. For hypertensive and normotensive groups, respectively, global longitudinal strain was −15.1±2.3 versus −18.5±1.9 (P<0.0001), global circumferential strain −15.2±3 versus −19.9±3.1 (<0.0001), global radial strain +44.0±11.3 versus 63.4±10.5 (P<0.0001), and global 3-dimensional strain −26.1±3.8 versus −31.5±3.8 (P<0.0001). Basal clockwise rotation, apical counterclockwise rotation, twist, and torsion were not significantly different. After multivariate regression analyses blood pressure, body mass index and LV mass maintained a significant relationship with measures of LV strain. Similar ventricular volumes and ejection fraction were observed in hypertensive and normotensive children, but children with hypertension had significantly lower strain indices. Whether reduced strain might predict future cardiovascular risk merits further longitudinal study.
Introduction
See Editorial Commentary, pp 1089–1090
Children with high blood pressure (BP) may develop left ventricular (LV) hypertrophy (LVH), which is the most commonly reported cardiac measurement for evaluating target organ effect in this population.1–9 More recently, functional parameters of the left ventricle, such as tissue Doppler imaging, have been found to be more sensitive markers of systolic and diastolic dysfunction.10–13 Tissue Doppler imaging is a 1-dimensional technique, however, and does not provide information on variables such as myocardial rotation, twist, and torsion. Children are well suited to techniques based on 3-dimensional (3D) speckle tracking echocardiography because of excellent acoustic windows, and reference data for these techniques are now available in children.14–16
This study aimed to measure LV strain, twist, and torsion in young hypertensive patients compared with normotensive children. Additionally, we sought to assess whether factors such as BP and mass correlated with indices of LV strain.
Methods
This study was approved by the local research ethics committee, and informed consent was obtained from parents and (where appropriate) children in the study. All procedures were followed in accordance with institutional guidelines.
Study Population
Patients aged <18 years with a diagnosis of primary arterial hypertension were included in the hypertension group. Hypertension was defined as average systolic (SBP) or diastolic BP (DBP) ≥95th percentile (equivalent to z score ≥1.645) or if the patient was being treated for hypertension using the Fourth Report definitions.17 Patients with abnormal estimated glomerular filtration rates using the modified Schwartz formula18–20 were excluded, as were patients who were uncooperative with echocardiography. Treatment for hypertension was recorded (yes/no), but we did not investigate the use of different antihypertensive medications in this study.
Patients participating in the study had clinic BP measurements using a calibrated aneroid instrument and manual BP by a trained investigator using a size appropriate cuff. All patients had 3 BP measurements performed when sitting at rest in clinic, which were averaged to give the BP value used for analysis. Anthropometric (height, weight, and body mass index [BMI]), clinical, and laboratory measures were obtained on the day of the research investigations. Echocardiography was performed by experienced staff blinded to the patients’ BP measurements and treatment. Healthy normotensive children were recruited from the local population.
Echocardiographic Methods
Echocardiographic examinations were performed using the Philips IE33 ultrasound system (Philips, Inc, Andover, MA) with a S5-1MHz transducer and X3-1, X5-1, or X7-2 matrix probes for 3D data acquisition as appropriate for patient size. All studies were stored digitally on a Philips Xcelera image archive (Philips, Inc) and analyzed off-line by 1 author (S.N.) who was blinded to the medical diagnosis, BP, and medical therapy.
Standard 2-Dimensional Echocardiographic Examination
All 2-dimensional echocardiographic values were obtained as the average value of 3 consecutive cardiac cycles. LV end-systolic and end-diastolic diameters, posterior wall thickness, and interventricular septum thickness in diastole were measured in M mode in a parasternal long-axis view. LV mass (LVM) was calculated according to the formula of Devereux.21 When LVH was present, the geometry was assessed by relative wall thickness: 2× LV posterior wall thickness in diastole/LV end-diastolic diameter.22
Doppler Echocardiography
Pulsed-wave Doppler measurement of E and A waves of the mitral valve inflow were obtained in the apical 4-chamber view as recommended.23 Pulsed tissue Doppler was used to measure myocardial velocities at the lateral mitral valve annulus in early (e′), late (a′) diastole, and in systole (s′). The E/A and E/e′ ratios were calculated in each patient.
3D Echocardiographic Examination for LV Volume, 3D Speckle Tracking, and Twist and Torsion
A full-volume acquisition of the LV was obtained from an apical 4-chamber view. The 3D volume was acquired during 4 cardiac cycles in an end-expiratory breath-holding state. All data sets were stored digitally and analyzed off-line. For analysis of LV volume and ejection fraction, a commercially available software (Cardiac 3D Quantification Advanced [3DQ ADV], QLab 9, Philips) was used. The software automatically identifies the LV border after manually placing points at the mitral valve annulus and the apex in diastole and systole. Papillary muscles are included in the LV cavity. After automatic detection of the LV borders at end diastole and end systole, the LV border was adjusted manually when necessary.
3D global deformation parameters were measured using a semiautomated analysis package (4D LV analysis 3; TomTec, Inc, Munich, Germany). From the 3D full-volume set, the software automatically extracts the apical 4-, 2-, and 3-chamber and the short-axis views of the LV. The LV contours are generated automatically, but the user retains the ability to able to adjust the center of the mitral valve and apex after the 3D endocardial surface is automatically reconstructed and to modify the contour. The papillary muscles were included in the LV cavity. The LV was automatically divided into 16 segments using standard segmentation schemes.24 The software then provided 3D speckle tracking analysis to produce both global (GLS) and segmental longitudinal, circumferential (GCS), radial (GRS), and 3D strain (G3DS) time curves. Each of these strain modalities was derived from 6 segments at the base of the heart, 6 segments at the midventricular location, and 4 segments at the apical level of the heart. From these 3 levels of the heart, averaged peak strains at the 3 levels were determined and peak GLS, GCS, GRS, and G3DS were calculated. 3D Strain was calculated using the vector sum of the longitudinal and circumferential strain without the radial component and described the tangential deformation. The same software also measured LV basal rotation, apical rotation, and LV twist through the cardiac cycle. Peak basal and apical rotation, peak LV twist, and peak LV torsion were automatically calculated. Peak LV torsion was defined as peak LV twist divided by the distance between the basal and apical slices at end diastole.14
Use of z Scores
To account for age and size, we used z scores for SBP and DBP as per the Fourth Report.17 For echocardiographic parameters, z scores were calculated from the data of Kampmann25 (M mode) and Eidem26 (tissue Doppler imaging), respectively. LV mass was indexed to height to the power of 2.7.27,28
Statistical Methods
Data are expressed as the mean±SD. Comparison of variables between the hypertensive patients and the normotensive controls was performed by using the unpaired t test.
Multivariable linear regression analyses were performed to assess the relationship between independent variables, including mean arterial pressure, antihypertensive treatment (yes/no), age, BMI, and LVM with dependent variables, including GLS, GCS, GRS, and G3DS and twist and torsion measures. R2 and the P value for the slope being nonzero were listed as results. A P value <0.05 was considered statistically significant. Statistical analysis was performed using GraphPad Prism 6 and IBM SPSS Statistics 20.
Results
Demographic Characteristics
Sixty-three children were studied comprising of 26 hypertensive patients attending a regional hypertension service and 37 healthy normotensive children (Table 1). SBP and DBP were significantly higher in the hypertension group than normotensive children. Sixteen of 26 patients (62%) in the hypertension group were on antihypertensive treatment but with no difference in BP levels between those with and without treatment in the hypertension group. Hypertension patients were 3 years older on average and had higher BMI than normotensive children (Table 1).
Demographic Characteristics and Clinical Parameters of Study Population
The exact duration of hypertension before clinic presentation is not known, but participants in the study had a median duration of follow-up of 1 year (range, 0–8.5 years). Only 8 patients had hypertension for >3 years.
Two-Dimensional and 3D Parameters
Indexed LVM was significantly higher in the hypertensive group compared with those with normotension with elevated LVM to LV volume ratio (Table 2). Two patients in the hypertension group had a LV mass z score >1.65 (>95th centile) with eccentric LVH.
Two-Dimensional (M Mode) and 3D Echocardiographic Parameters
There was no significant difference in the ventricular volume or ejection fraction between the hypertensive and normotensive groups. GLS, GCS, GRS, and G3DS were significantly lower in hypertensive patients (Figure). There was no correlation between GLS and LV mass indexed to height2.7.
Global longitudinal, circumferential, radial, and 3-dimensional (3D) strain in hypertensive and normotensive children. Hypertensive (HTN) patients, normotensive children. ****P<0.0001.
With respect to longitudinal, circumferential, radial, and 3D strain values at the apical, mid, and basal levels, in almost all 3 segments, hypertensive patients had significantly lower strain compared with the normotensive group (Figures S1 through S4 in the online-only Data Supplement). The least difference between the hypertensive and the normotensive group was in the apical region. Circumferential apical strain was the only segment without a significantly lower strain in hypertensive patients compared with those with normotension.
Twist and Torsion of the Left Ventricle
Basal clockwise rotation (−) and apical counterclockwise rotation (+), as well as LV twist and torsion showed no statistical difference between the 2 groups.
Doppler Parameters
All mean Doppler and tissue Doppler z scores were within the normal range. There was a significantly higher A-wave z score in the hypertensive group (Table 3). The lateral mitral valve annulus z score for e′ was lower than in those with normotension, and in consequence, the E/e′ z score was significantly higher in the hypertensive patients (although still within normal limits). s′ z scores were lower in hypertensive than normotensive patients but again within normal limits. There was a weak but significant positive correlation between e′ z score and LV mass/height2.7 in the hypertensive group (P=0.021; Pearson, 0.460).
Doppler Echocardiographic Parameters
A positive correlation was observed between the e′ velocity z score and the GLS (P<0.0028; R2, 0.17), but there was no correlation between the s′ velocity z score (P<0.07; R2, 0.07) or the E/e′ ratio with the GLS (P<0.31; R2, 0.02; Figure S5).
SBP and DBP z Score Correlations
If normotensive and hypertensive patients are all included, SBP z score had a significant positive correlation with GLS (P<0.005; Pearson, 0.473), LV E/e′ z score (P<0.05; Pearson, 0.295), and the septal a′ z score (P<0.005; Pearson, 0.402). SBP z score had a negative correlation with LV e′ z score (P<0.05; Pearson, −0.297) and septal e′ z score (P<0.05; Pearson, −0.365).
DBP z score correlated positively with GLS (P<0.05; Pearson, 0.314) and septal a′ z score (P<0.05; Pearson, 0.277) and negatively with septal e′ z score (P<0.05; Pearson, −0.366).
Discussion
In this study, we have compared measures of cardiac function in hypertensive and normotensive children. As expected, BP and indexed LV mass were significantly higher in the hypertensive patients. However, LV diastolic volume, systolic volume, and ejection fraction were not significantly different between the study groups. Use of advanced and more sensitive echocardiographic techniques, such as tissue Doppler imaging and strain analysis, unmasked significant differences in systolic and diastolic ventricular function between the hypertensive patients and healthy normotensive children.
By tissue Doppler, early myocardial relaxation velocity (e′) and systolic velocity (s′) were significantly reduced in hypertension compared with the normotensive group. The E/e′ ratio—a surrogate for filling pressure—was significantly higher in the hypertension group who also had a higher mitral valve inflow A-wave velocity. Our data are consistent with the results of Zamojska et al10 and Agu et al12 who showed that in hypertensive children, z scores for lateral mitral valve annulus tissue Doppler velocities for e′ and s′ are significantly lower, and E/e′ ratio is increased compared with those with normotension. These tissue Doppler findings suggest impaired LV relaxation and longitudinal systolic function in young hypertensive patients. In adult patients, including those with hypertension and LV hypertrophy, the early diastolic mitral annular velocity is a predictor of cardiac mortality,29,30 but the prognostic significance of this finding in younger patients remains unknown.
Additionally, there was a correlation between higher systolic and diastolic pressures and parameters of diastolic dysfunction (higher E/e′ and a′ z scores and lower e′ z scores).
GLS, GCS, GRS, and G3DS were all significantly lower in the hypertension group compared with those with normotension confirming the use of these measurements for early detection of functional cardiac impairment consistent with recently published comparable results in adult patients.31,32
LV ejection fraction was no different in hypertensive patients compared with normotensive children. This measure takes no account of when ventricular ejection occurs. Previous data from our group has shown that in adult hypertensive patients with preserved LV ejection fraction, more severe hypertension was associated with delayed peak myocardial wall stress and impaired diastolic function.33 In the current study, tonometry was not undertaken to calculate the timing of peak myocardial wall stress.
LV twist and LV torsion play an important role in LV systolic and diastolic mechanics. To our knowledge, no previous study has investigated twist and torsion dynamics in hypertensive children. In adult hypertensive patients, there are several investigations showing that LV twist and torsion are preserved in the high-normal BP subjects and significantly impaired in the hypertensive patients.31,34–36 In our study population, twist and torsion were no different in hypertensive patients compared with those with normotension. A potential explanation could be that in our cohort of hypertensive children, the disease state had not been present for sufficiently long to have had an impact on the myocardium to impair LV twist and torsion. Despite increased LV mass, LV volume and LV ejection fraction were similar in hypertensive and normotensive groups.
Our data shows a strong negative correlation of GLS, GCS, GRS, and the G3DS with BP (Table 4). Those correlations are mirrored by weaker correlations between strain measures and left ventricular mass index and BMI in multivariate analyses. Several of these independent relationships varied by different types of strain (Table 4). The G3DS only has BMI (and not mean arterial pressure) as a significant relationship.
Multiple Linear Regression Analysis of Left Ventricular Strain With Brachial Mean Arterial Pressure in Hypertension Group
It is known that there is a strong association between obesity and hypertension in children and adolescents,37,38 which is consistent with our study finding of a higher BMI in the hypertension versus the healthy normotensive group. Equally, there is clear evidence of hypertension leading to LVH.6,39 In children and adolescents with systemic arterial hypertension, the presence of obesity has been associated with marked LVH. The relative impact of excess weight and elevated BP will need longitudinal evaluation, and our data can suggest association only.
Study Limitations
This is a cross-sectional study, hence the impact of change in variables including BMI or BP has not been assessed. The hypertensive patients were on average 3 years older with a significantly higher BMI than the normotensive controls. 3D analysis was performed on a single analysis platform for consistency, but our results may not be able to be extrapolated to different analysis software. We have made extensive use of indexed variables or z scores based on published normal data but accept that the interpretation of cardiovascular variables in children remains difficult because variables may be influenced by patient age, size, and heart rate.
Perspectives
The observation of reduction of strain in hypertensive asymptomatic young patients may provide more information on global and regional myocardial function in hypertension that might improve pathophysiologic understanding of progression of myocardial dysfunction. The combination of reduced strain and abnormal diastolic LV filling may play a role in the development of chronic cardiac dysfunction in hypertensive patients. Further prospective studies are required to answer whether reduced strain can predict future cardiovascular risk and whether strain can be normalized by interventions such as BP control.
Acknowledgments
We would like to thank the echocardiography staff of the Evelina London Children’s Hospital.
Sources of Funding
The authors (M.D. Sinha and J.M. Simpson) acknowledge support by the Department of Health through the National Institute for Health Research comprehensive Biomedical Research Centre award to Guy’s and St Thomas’ National Health Service Foundation Trust in partnership with King’s College London and King’s College Hospital National Health Service Foundation Trust. The Division of Imaging Sciences also receives support as the Centre of Excellence in Medical Engineering (funded by the Wellcome Trust and Engineering and Physical Sciences Research Council; grant number WT 088641/Z/09/Z) and the British Heart Foundation Centre of Excellence (British Heart Foundation award RE/08/03). Dr Susanne Navarini received financial support from the Swiss National Science Foundation.
Disclosures
None.
Footnotes
The online-only Data Supplement is available with this article at http://hyper.ahajournals.org/lookup/suppl/doi:10.1161/HYPERTENSIONAHA.117.09574/-/DC1.
- Received April 29, 2017.
- Revision received May 23, 2017.
- Accepted October 3, 2017.
- © 2017 American Heart Association, Inc.
References
Novelty and Significance
What Is New?
Children with hypertension have significantly reduced global circumferential, longitudinal, and radial strain.
The reduction in strain is associated with a higher mean arterial pressure.
Torsion and twist in this age group do not seem to be different from controls.
What Is Relevant?
Use of tissue Doppler and 3-dimensional strain techniques shows differences in hypertensive patients, which are not evident, using measures such as ejection fraction or fractional shortening.
Measurement of 2-dimensional strain and tissue Doppler should form part of the assessment of hypertensive children and adolescents.
Torsion and twist are not affected in this age group but reduced later in life so there may be an opportunity for medical intervention before adverse remodelling.
Summary
Hypertensive children have reduced strain with preserved ejection fraction and fractional shortening. Further longitudinal studies from childhood to adulthood are required to ascertain whether this is a risk factor for later symptomatic cardiac dysfunction.
This Issue
Jump to
Article Tools
Share this Article
- Myocardial Deformation Measured by 3-Dimensional Speckle Tracking in Children and Adolescents With Systemic Arterial HypertensionNovelty and Significance
