Background
Doppler tissue imaging may help identify children with dyssynchrony who could benefit from resynchronization therapy. However, few studies have quantified dyssynchrony measures in children; no study has investigated the relationship among age, heart rate, and dyssynchrony measures in children; and no study has quantified cross-correlation delay in children. The aim of this study was to test the hypotheses that measures of left ventricular dyssynchrony would correlate with age, primarily because of the correlation between heart rate and age, and that children with cardiomyopathy would have left ventricular dyssynchrony.
Methods
Sixty healthy children and 11 children with dilated cardiomyopathy were prospectively enrolled. Seven dyssynchrony measures were quantified: septal-to-lateral delay, peak velocity difference, the standard deviations of times to peak in 12 segments in systole and diastole, and cross-correlation delay in systole, diastole, and the whole cycle.
Results
The seven dyssynchrony measures were either not correlated with age or only weakly correlated with age after correcting for heart rate using Bazett’s formula. Septal-to-lateral delay, peak velocity difference, and the standard deviation of times to peak in 12 segments in systole showed dyssynchrony in 57% to 85% of normal controls, compared with 20% for cross-correlation delay in the whole cycle and 3% for the standard deviation of times to peak in 12 segments in diastole. Cross-correlation delay in systole, cross-correlation delay in diastole, cross-correlation delay in the whole cycle, and the standard deviation of times to peak in 12 segments in diastole were elevated in children with dilated cardiomyopathy compared with controls.
Conclusions
Echocardiographic dyssynchrony measures should be corrected for heart rate using Bazett’s formula in children. Time-to-peak Doppler tissue imaging dyssynchrony measures classify many healthy children as having abnormalities with the timing of left ventricular contraction, which suggests that the methodology is not accurate in children. In preliminary studies, cross-correlation dyssynchrony measures show elevated systolic and diastolic measures of dyssynchrony in children with dilated cardiomyopathy compared with controls, which deserves further investigation to help identify children who may benefit from resynchronization therapy.
Cardiac resynchronization therapy (CRT) is a class 1A treatment for adults with stable, New York Heart Association class III or IV heart failure with prolonged QRS duration and low left ventricular (LV) ejection fraction despite optimal medical therapy. The success of CRT in adults has led to its use in children with heart failure. However, criteria for selecting children who will most likely benefit from CRT are not yet well defined. The few pediatric patients who have received and benefited from CRT differ substantially from adults who receive CRT in their etiologies (71% had congenital heart disease), symptoms (only 38% were in New York Heart Association class III or IV), and objective ventricular measures (only 54% had an ejection fraction ≤35% and QRS duration ≥120 msec). Thus, adult criteria for CRT implantation are not likely to apply to pediatric patients.
Doppler tissue imaging (DTI) has been used in adult patients with heart failure to diagnose LV dyssynchrony, which is thought to be the primary therapeutic target of CRT. Thus, DTI may likewise play an important role in the development of criteria to select children for CRT. In fact, DTI has already been used in multiple studies to identify dyssynchrony in children with heart disease.
However, only two studies have reported normal ranges of DTI dyssynchrony measures in large, prospective groups of normal children. Studies using DTI in adults have largely ignored the effects of heart rate on DTI dyssynchrony measures, and no study has investigated the relationship among age, heart rate, and measures of LV dyssynchrony in normal children. Additionally, no study has reported normal ranges of cross-correlation delay (XCD) in healthy children, which is a new DTI dyssynchrony measure that has shown promise in adults and therefore should be evaluated in children. Moreover, recent reports have shown that DTI dyssynchrony measures may be falsely abnormal in a large proportion of normal adults and have poor reproducibility and controversial utility in predicting response to CRT, which may limit their utility in children.
Thus, our objectives were to (1) investigate the relationship among age, heart rate, and measures of LV dyssynchrony in a large, prospective cohort of normal children; (2) report normal values, including estimates of reproducibility for XCD parameters in normal children; and (3) quantify measures of dyssynchrony in a group of children with dilated cardiomyopathy for comparison with the normal subjects. We hypothesized that (1) measures of LV dyssynchrony would correlate with age, primarily because of the correlation between heart rate and age in children, and (2) children with dilated cardiomyopathy have LV dyssynchrony as quantified with XCD.
Methods
The study was approved by the institutional review boards of Emory University and Children’s Healthcare of Atlanta. All subjects and their parents provided informed consent or assent.
Normal Controls
Sixty healthy children ranging in age from 0 to 18 years were prospectively recruited by our research nurses through word of mouth and recruitment flyers and were not referred for routine echocardiography. Inclusion criteria were (1) age 0 to 18 years, (2) QRS duration less than the 95th percentile on standard 12-lead electrocardiography, and (3) no history of heart disease. Exclusion criteria were (1) structural abnormality or abnormal function on echocardiography; (2) history of systemic hypertension or metabolic, genetic, neuromuscular, or lung disorders; and (3) history of any rhythm abnormality.
Patients with Dilated Cardiomyopathy
Eleven children with dilated cardiomyopathy were prospectively recruited to participate in the study for comparison with the normal controls. Inclusion criteria were (1) age 0 to 18 years and (2) diagnosis of dilated cardiomyopathy by the patient’s attending physician. Patients were excluded if they had histories of heart surgery or congenital heart disease. We defined cardiomyopathy as an enlarged ventricle ( Z score > 2) with decreased ventricular function (fractional shortening < 28%) without anatomic or other known causes, such as coarctation, aortic stenosis, history of congenital heart disease or heart surgery, or suspected myocarditis.
DTI Data Acquisition
Apical two-chamber, three-chamber, and four-chamber DTI of the myocardium was performed using a 5S or an M3S probe (Vivid 7; GE Vingmed Ultrasound AS, Horten, Norway) as appropriate for patient size. The myocardial walls were aligned parallel to the Doppler beam to minimize the angle of insonation, and the frame rate was optimized to a minimum of 100 Hz. Pulsed Doppler images of the aortic outflow tract were acquired to enable the definition of systole during postprocessing. All studies were done without the use of sedation, and images were acquired during end-expiration when possible.
DTI Dyssynchrony Measures
Postprocessing software (EchoPAC PC version 6.0.0; GE Vingmed Ultrasound AS) was used to export velocity curves from the DTI data for the calculation of seven published dyssynchrony measures.
Systolic dyssynchrony measures were (1) the basal septal-to-lateral delay in time to peak systolic velocity (SLD) ; (2) the maximum difference in times to peak systolic velocity between any two of the basal septal, lateral, anterior, and inferior LV segments (MaxDiff) ; (3) the standard deviation of times to peak systolic velocity in the 12 basal and midwall segments of the left ventricle (Ts-SD) ; and (4) systolic XCD (XCD systole ).
Diastolic dyssynchrony measures were (1) the standard deviation of times to peak diastolic velocity in the 12 basal and midwall segments of the left ventricle (Te-SD) and (2) diastolic XCD (XCD diastole ).
The whole-cycle dyssynchrony measure was (1) whole-cycle XCD (XCD whole ).
Quantification of XCD whole , XCD systole , and XCD diastole has been described previously. Figure 1 shows an example calculation of these cross-correlation parameters for a set of velocity curves from a child with dilated cardiomyopathy. Briefly, a cross-correlation function was used to calculate the time delay that resulted in the maximum correlation between velocity curves from opposing ventricular segments. One velocity curve was shifted relative to the other curve, and the normalized cross-correlation value was computed for each time shift (the normalized correlation coefficient ranges from 1 = perfect synchrony to −1 = perfect dyssynchrony). The time shift between the two curves resulting in the maximum correlation value was defined as the XCD between the two curves. The entire velocity curves were used to calculate XCD whole , while the systolic (from mitral valve closure to aortic valve closure) and diastolic (from aortic valve closure to mitral valve closure) portions, respectively, were used to calculate XCD systole and XCD diastole .
Regions of interest were sized to encompass the entire LV segment of interest, which has previously been shown to minimize variability. An average velocity curve was generated from three cardiac cycles of velocity data before quantifying the measures of dyssynchrony. This averaging was done with the cine-compound function in EchoPAC to improve the signal-to-noise ratio of the extracted velocity curves. All velocity data were exported using stationary regions of interest. Pulsed Doppler of the aortic outflow tract was used to define systole for the identification of peak systolic velocity.
Times to peak systolic velocities were automatically identified by a computer program written in MATLAB quantitative analysis software version 7.10 (The MathWorks Inc., Natick, MA). This program imported velocity curves and valve timings from the echocardiographic postprocessing software (EchoPAC PC version 6.0.0) and exported the time from the Q wave to the maximum velocity in the ejection phase. Peaks were checked by an expert observer to ensure accuracy.
Relationship among Age, Heart Rate, and Measures of Dyssynchrony
The relationship between each dyssynchrony measure and age was determined using Pearson’s linear correlation coefficient. A P value <.05 was considered to indicate a statistically significant correlation. This correlation was explored using both raw data and data corrected for heart rate according to Bazett’s formula :
Dyssynchrony corrected = Dyssynchrony uncorrected RR ,
Comparison between Normal Controls and Patients with Dilated Cardiomyopathy
Each of the seven DTI dyssynchrony measures was quantified in the 60 normal controls and the 11 patients with dilated cardiomyopathy. The mean value of each of the seven dyssynchrony measures was compared between the normal controls and the patients with dilated cardiomyopathy using an unpaired Student’s t test. P values <.05 were considered statistically significant.
Thresholds to Diagnose Dyssynchrony
Published adult threshold values were investigated to determine the percentages of both normal children and patients with dilated cardiomyopathy who had values for dyssynchrony measures above levels considered to be diagnostic of dyssynchrony in adult patients. The threshold values were defined as follows: (1) SLD > 60 msec, (2) MaxDiff > 65 msec, (3) Ts-SD > 34.4 msec, (4) Te-SD > 34 msec, and (5) XCD whole > 31 msec. Note that XCD systole and XCD diastole do not have published thresholds considered to be diagnostic of dyssynchrony, so these parameters were excluded from this analysis.
Interobserver Reproducibility
Interobserver reproducibility was assessed on 12 randomly selected echocardiograms by two independent observers. Reproducibility was displayed graphically using Bland-Altman plots. A coefficient of variation (CoV) was calculated for each dyssynchrony measure using the following equation:
CoV = σ i ¯ M ,
σ i ¯ = ∑ i = 1 12 D i , 1 + D i , 2 2 12 ,
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
