Elevated ventricular filling pressure is a marker of diastolic dysfunction and a known risk factor for failure of single-ventricle surgical palliation. Doppler echocardiography has been shown valuable in identifying patients with elevated ventricular end-diastolic pressure (VEDP) in other settings, but its utility in evaluating pediatric patients with single ventricle is unclear. The aim of this study was to compare Doppler parameters to simultaneous catheter measurements of VEDP in children with single ventricle.
All consecutive patients (age < 18 years) with single ventricle who underwent simultaneous echocardiography and catheterization in 2009 and 2010 were included in this prospective study. Data regarding diagnosis, morphology of the “single” ventricle, prior surgeries, Doppler tissue imaging (DTI), atrioventricular valve inflow and pulmonary vein Doppler, and myocardial performance index (MPI) were collected. Ventricular Doppler echocardiography was performed from the dominant ventricle. Simultaneous Doppler and catheter measurements of systolic and diastolic function and VEDP were obtained. Correlation of continuous variables was examined using linear regression analysis. Receiver operating characteristic curves, two-sample Student’s t tests, χ 2 analyses, and Fisher’s exact tests were used as appropriate.
A total of 32 patients (15 male; mean age, 30.2 ± 22 months) were studied (nine post-Fontan, 15 post-Glenn, and eight pre-Glenn). Mean systemic arterial saturation was 81 ± 10%, mean VEDP was 11 ± 3 mm Hg, and mean echocardiographically estimated ejection fraction was 55 ± 7%. VEDP was correlated positively with E/E′ ratio ( r = 0.44, P < .01), pulmonary vein atrial reversal duration ( r = 0.77, P < .001), and E′ ( r = 0.49, P < .01). Receiver operating characteristic curve analysis using an E/E′ cutoff of 12 showed sensitivity of 90% (95% confidence interval, 54.1%–99.5%) and specificity of 75.0% (95% confidence interval, 47.4%–91.7%) for identifying VEDP > 10 mm Hg. Single left ventricles had higher DTI S′ and E′ velocities and lower E/E′ ratios and MPIs compared with single right ventricles. S′ velocity correlated positively with ejection fraction ( r = 0.77, P < .001) and negatively with single left ventricular MPI ( r = −0.46, P < .01).
In patients with single-ventricle physiology, DTI and pulmonary vein Doppler echocardiographic parameters correlated modestly with direct measurement of VEDP and may be helpful in identifying patients with elevated filling pressures. In addition, single left ventricles had better systolic and diastolic performance than single right ventricles. DTI systolic velocities and left ventricular MPI correlated well with ventricular ejection fraction.
The Glenn (cavopulmonary anastomosis, stage 2) and Fontan (total cavopulmonary anastomosis, stage 3) operations are currently part of “standard” surgical palliation in patients with functional single ventricle. After the Glenn and Fontan procedures, the venous blood must traverse the pulmonary vascular bed without a ventricular pump. Since the inception of single-ventricle surgical palliation, cardiac catheterization has been considered mandatory for detailed hemodynamic assessment and angiographic evaluation of the extracardiac thoracic vasculature and to evaluate ventricular filling pressures. Elevated ventricular end-diastolic pressure (VEDP), and the resultant increase in atrial pressure can impede the passive flow of the blood through the pulmonary vasculature in a Glenn or Fontan circuit. Elevated VEDP is a well-established marker of diastolic dysfunction and a risk factor for Glenn and Fontan failure. Advances in noninvasive diagnostic techniques such as echocardiography, magnetic resonance imaging, and computed tomography have made it possible to effectively evaluate the extracardiac thoracic vasculature. However, cardiac catheterization continues to be used routinely at many centers in the assessment of children who are referred for Glenn or Fontan operations to perform thorough hemodynamic assessment for the possibility of elevated pulmonary artery pressure or increased VEDP. There is increasing evidence that in patients with normal ventricular function and other low-risk subjects, the omission of routine cardiac catheterization would not impair the identification of inoperable subjects or those at risk for adverse early postoperative outcomes. Doppler echocardiographic parameters correlate with VEDP in other diseases. However, many recent studies have demonstrated discrepancies and modest correlations between Doppler estimated filling pressure and cardiac catheterization–derived filling pressures in a variety of diseases. It is unclear whether these findings apply to pediatric patients with functional single ventricle. Although there are studies evaluating Doppler parameters and inherent diastolic dysfunction in single-ventricle patients, there are no data correlating these findings with simultaneous invasive measurements of ventricular filling pressure. The objective of this study was to compare Doppler echocardiographic parameters with simultaneous invasive measurements of VEDP in pediatric patients with functional single ventricle.
In this prospective study, we included all consecutive patients (age < 18 years) with functional single ventricle who underwent simultaneous echocardiography and cardiac catheterization as part of their routine preoperative cardiac evaluation at our institution between 2009 and 2010. Cardiac catheterization for hemodynamic measurements and evaluation of vascular structures was routinely performed before Glenn and Fontan palliation at our center. In all patients, echocardiographic measurements were made after patients were intubated on positive-pressure ventilation and within a few minutes before the cardiac catheterization measurements. Echocardiographic and cardiac catheterization measurements were made under near identical hemodynamic and physiologic states, without breath holding. Patients who were chronically paced or not in sinus rhythm were excluded from the study. Patients with greater than mild atrioventricular valve regurgitation, outflow tract obstruction, branch pulmonary artery stenosis, coarctation of the aorta, or fused atrioventricular valve inflow Doppler were excluded from the study.
For echocardiographic evaluation, views equivalent to a standard apical four-chamber were obtained. Echocardiographic measurements obtained included atrioventricular valve inflow velocities (E and A from the dominant ventricle); E/A ratio (from the dominant ventricle); and pulmonary vein systolic, diastolic, and atrial reversal velocities and duration. Peak Doppler velocities were analyzed to determine early (E) and late (A) diastolic flow across the atrioventricular valve of the dominant ventricle. Doppler tissue imaging (DTI) E′, A′, and S′ velocities were obtained for both the septal and ventricular free walls at the level of the atrioventricular valve annulus using pulsed spectral Doppler sampling in all patients. The E/E′ ratio was defined as the ratio of E′ velocity obtained from the dominant ventricular free wall at the level of the atrioventricular valve annulus to the E-wave velocity of the dominant ventricle at the tip of atrioventricular valve. Myocardial performance index (MPI) was calculated using corresponding DTI values of the free wall using the following equation: MPI = (isovolumic contraction time + isovolumic relaxation time)/ejection time. All echocardiographic measurements were made offline in triplicate from three consecutive waveforms and averaged by a single observer blinded to the results of catheterization. Ventricular echocardiographic measurements were obtained from the dominant ventricle. Ejection fraction was measured using the modified Simpson method by a single investigator.
All measurements were performed under general anesthesia. Patients were allowed clear fluids up to 2 hours before the procedure. Per our routine cardiac catheterization protocol, in an attempt to achieve closer to normal hydration status, fasting patients with metabolic acidosis, low blood pressure, or low venous pressure secondary to dehydration were given a crystalloid bolus of 10 ml/kg before hemodynamic measurements. Patients were maintained on maintenance fluid throughout the case. In patients to whom fluid boluses were administered, VEDP and echocardiographic measurements were made after administration. VEDP was measured at catheterization via a fluid-filled catheter connected to an external pressure transducer zeroed to atmospheric pressure at the patient’s midaxillary line and was taken in millimeters of mercury as the point just before the rapid rise in ventricular pressure corresponding to ventricular systole.
The following analyses were performed: (1) the relationship between noninvasively derived Doppler echocardiographic parameters and invasively obtained VEDP, (2) differences between right ventricular and left ventricular MPIs, (3) variations in regional myocardial velocities using DTI, (4) the relationship between DTI parameters and ejection fraction, and (5) DTI by diagnosis.
Correlation of continuous variables was examined using linear regression analysis. Receiver operating characteristic (ROC) curves, two-sample Student’s t tests, χ 2 analyses, and Fisher’s exact tests were used as appropriate. A VEDP of 10 mm Hg was used as a cutoff value. This cutoff value was based on a previous publication indicating that in an optimal Fontan circuit, the systemic venous pressure should not exceed 18 to 20 mm Hg when the 6 to 8 mm Hg mean transpulmonary gradient is added to the mean left-sided filling pressure. The institutional review boards of the University of Utah and Primary Children’s Medical Center approved this study.
A total of 32 patients (15 male; mean age, 30.2 ± 22 months) were studied. Characteristics of the study population are shown in Table 1 . There were no significant differences in age, gender, and oxygen saturation between the single right ventricle and single left ventricle cohorts ( Table 1 ). The Doppler values obtained are shown in Table 2 . Two patients with fused E and A waves and one patient with a significant outflow tract obstruction were excluded from the study. None of the patients in the study group had more than mild atrioventricular valve regurgitation. The ventricular free wall DTI values for E′ and S′ and inflow E/A ratio were significantly lower in single right ventricles than in single left ventricles ( P < .05 for all). The E/E′ ratio was lower for single left ventricles than for single right ventricles ( P = .02). The MPI was significantly higher for single right ventricles than for single left ventricles ( P < .05). All other Doppler values were similar between the single right and left ventricle cohorts. In this study, tissue Doppler velocities for the free wall were used for comparison with VEDP. Deformation of the ventricular septum and septal myocardial velocities were lower compared with free wall velocities, especially in the presence of a poorly functioning, hypertensive, and rudimentary ventricular cavity, as seen in patients with pulmonary atresia and intact ventricular septum or hypoplastic left-heart syndrome with aortic atresia or stenosis and mitral stenosis ( Figure 1 ).
|Age (months)||30.2 ± 22 (5–120)|
|Body surface area (m 2 )||0.52 ± 0.31 (0.27–1.3)|
|Oxygen saturation in room air (%)||81 ± 10 (70–89)|
|Single-RV cohort ( n = 17)|
|Age (months)||32.1 ± 25|
|Oxygen saturation in room air (%)||83 ± 13|
|Hypoplastic left-heart syndrome||10|
|Atrioventricular septal defect (right dominant)||2|
|Single-LV cohort ( n = 15)||15|
|Age (months)||28.3 ± 80|
|Oxygen saturation in room air (%)||78 ± 12|
|Atrioventricular septal defect (left dominant)||2|
|Pulmonary atresia with intact ventricular septum||2|
|Variable||Single LV |
( n = 15)
|Single RV |
( n = 17)
|Mitral valve E (cm/sec)||72.6 ± 10.5||—|
|Mitral valve A (cm/sec)||42.2 ± 19.2||—|
|Mitral valve E/A ratio||1.5 ± 0.43||—|
|Tricuspid valve E||—||74.6 ± 14.2|
|Tricuspid valve A||—||53.4 ± 18.4|
|Tricuspid valve E/A ratio||—||1.23 ± 0.4|
|Left ventricular E′ (cm/sec)||10.8 ± 6.6||—|
|Left ventricular S′ (cm/sec)||5.9 ± 1.6||—|
|Left ventricular E/E′ ratio||8.1 ± 4.7||—|
|MPI||0.49 ± 0.1||0.56 ± 0.1|
|Right ventricular E′ (cm/sec)||7.2 ± 2.4|
|Right ventricular S′ (cm/sec)||4.9 ± 1.2|
|Right ventricular E/E′ ratio||—||9.3 ± 7.2|
|PV systolic velocity (cm/sec||45 ± 15||48 ± 18|
|PV diastolic velocity (cm/sec)||58 ± 22||56 ± 19|
|PV atrial reversal velocity (cm/sec)||38 ± 17 ∗||42 ± 22 †|
|PV atrial reversal duration (msec)||98 ± 38 ∗||89 ± 40 †|
All patients in this study cohort had subjectively normal systolic function as reported by the physician reading the echocardiogram. The calculated ejection fraction (modified Simpson’s method) for the entire cohort of single ventricle patients was 55 ± 7% (left ventricular ejection fraction range, 54%-62%; right ventricular ejection fraction range, 48%-56%). For the entire cohort, single ventricular free wall S′ velocity was correlated positively with ejection fraction ( r = 0.77, P < .001) and negatively with single left ventricle MPI ( r = −0.46, P < .01). No other statistically significant relationship was found between ejection fraction and Doppler-derived parameters.
During cardiac catheterization, the mean systemic arterial oxygen saturations was 81 ± 10%, and the mean VEDP was 11 ± 3 mm Hg (range, 7-16 mm Hg). VEDP was 17 mm Hg in one patient, 16 mm Hg in one patient, 15 mm Hg in two patients, 14 mm Hg in two patients, and ≤13 mm Hg in the remainder. One patient in this study received a 10 mL/kg bolus because of low blood pressure and metabolic acidosis. VEDP and echocardiographic measurements were obtained after the fluid bolus under similar hemodynamic and fluid statuses. VEDP obtained during simultaneous echocardiography and catheterization was correlated positively with free wall E/E′ ratio ( r = 0.44, P < .01), pulmonary vein atrial reversal duration ( r = 0.77, P < .001), and ventricular free wall E′ ( r = 0.49, P < .01). Other Doppler echocardiographic parameters were not correlated with VEDP obtained during cardiac catheterization. Figure 2 shows the side-by-side box plots for subjects grouped by E/E′ ratio on the basis of VEDP cutoffs used by Ommen et al . (<8, 8-15, and >15 mm Hg) in relation to the VEDP measured during cardiac catheterization. To study the ability of ventricular free wall E/E′ ratio to predict VEDP > 10 mm Hg, ROC curves were constructed. Using an E/E′ ratio cutoff value of 10 resulted in an area under the ROC curve of 0.74, sensitivity of 66.7% (95% confidence interval [CI], 41.1%–85.6%), and specificity of 75.0% (95% CI, 35.6%–95.5%) ( Figure 3 ). Using an E/E′ cutoff of 12, a second ROC curve was generated ( Figure 3 ), with an area under the ROC curve of 0.756, sensitivity of 90.0% (95% CI, 54.1%–99.5%), and specificity of 75.0% (95% CI, 47.4%–91.7%). The curve using an E/E′ cutoff of 12 had higher sensitivity but similar specificity compared with the E/E′ cutoff of 10. However, the confidence intervals for both curves were wide.