ECHOCARDIOGRAPHY IN CHILDREN • Doppler Echocardiographic Features of Pulmonary Vein Stenosis in Ex-Preterm Children

BACKGROUND

Pulmonary vein stenosis (PVS) is a rare disease in children, primarily affecting premature infants and children with congenital heart disease. PVS causes significant morbidity and mortality (estimated 3-year survival rate of 43%).

While diagnosis can be made with computed tomography, magnetic resonance imaging, cardiac catheterization, and echocardiography, only echocardiography which has the advantage of low cost and good safety profile, is convenient for use as a screening tool.

With echocardiography, PVS is suspected when increased pulmonary vein blood flow velocities are measured using Doppler echocardiographic (DE) imaging. The velocity of flow through a stenosed vessel is dependent on the degree of stenosis but also on the flow in that vessel. The more severe the stenosis and the greater the blood flow, the higher the blood flow velocity in the vessel. In severe PVS, however, there is redirection of segmental pulmonary blood flow away from the stenosed venous segments, thus causing a paradoxical decrease in blood flow velocity in these veins. This is likely one of the reasons why the suspicion for PVS is delayed on echocardiography in ex-preterm infants.

Despite routine use of DE imaging at many centers to screen for PVS, the utility of DE imaging for diagnosing PVS has not been studied. The aim of this study was to compare the different DE characteristics of PVS reported in the literature and to identify those most likely to be associated with a diagnosis of PVS confirmed by cardiac catheterization in young children born preterm. The authors hypothesized that in expreterm children, certain DE characteristics of pulmonary venous flow would discriminate between patients with and those without PVS confirmed by cardiac catheterization.

METHODS

This was a single-center retrospective review of spectral Doppler interrogations, invasive hemodynamic measurements, and angiograms of pulmonary veins of children with histories of prematurity was performed. All ex-preterm children <3 years old undergoing cardiac catheterization between January 2014 and January 2019 were included in the study.

Prematurity was defined as birth before 37 weeks’ gestation.

Excluded patients:

1. Patients with single-ventricle physiology because of markedly different pulmonary flow dynamics compared with subjects with two-ventricle physiology.
2. Patients where direct measurement of the pressure difference between the pulmonary vein of interest and the left atrium (pulmonary vein mean pressure gradient) was not performed during cardiac catheterization and PVS was not reported in their medical records.
3. Patients where no echocardiogram was obtained in the 3 months before cardiac catheterization.
4. Patients with congenital diaphragmatic hernia.

The study included 18 children with 25 stenosed pulmonary veins and 29 children with 78 nonstenosed pulmonary veins. The echocardiogram obtained before and closest to the date of cardiac catheterization was reviewed. Spectral Doppler evaluation of the pulmonary veins (both color and spectral Doppler interrogation of all pulmonary veins) was analyzed. If more than one acquisition of spectral Doppler velocity in a pulmonary vein was performed in a study, the one with the highest Doppler velocity was used. All Doppler velocity measurements were averaged over three cardiac cycles. The authors calculated the mean pulmonary vein pressure gradient using the velocity-time integral of the area under the Doppler peak velocity over time curve.

To quantify phasic flow in the pulmonary veins, the authors calculated the absolute difference between the peak systolic velocity and the peak diastolic velocity as well as the pulsatility index. The pulsatility index was calculated as (peak systolic velocity/peak diastolic velocity)/mean velocity during the cardiac cycle. Return of Doppler velocity to baseline was defined as a decrease of velocity to <0.20 m/sec at any time during the cardiac cycle. The echocardiograms were also evaluated for the presence of pulmonary hypertension and right ventricular dysfunction.

RESULTS

Most of the cohort had pulmonary hypertension (defined as mean pulmonary artery pressure > 20 mm Hg and total indexed pulmonary vascular resistance > 3 Wood units • m²), as well as a diagnosis of chronic lung disease in their medical charts.

Invasive hemodynamic data obtained during cardiac catheterization showed a wide range of pulmonary vein gradients in patients with PVS, reflecting the range of degree of PVS in the cohort. Pulmonary hypertension (by both mean pulmonary artery pressure and total pulmonary resistance) was present in patients with and without PVS, though it was more severe in patients with PVS.

Angiographic evidence of PVS did not always correlate with a pulmonary vein mean pressure gradient of >3 mm Hg. Analysis of the echocardiograms showed no significant differences between patients with and those without PVS regarding the frequency of diagnosis of pulmonary hypertension, measures of pulmonary flow dynamics, and measures of right ventricular function. Results of the DE analysis of the pulmonary veins revealed significantly higher peak systolic velocity (‘‘s’’ wave), peak diastolic velocity (‘‘d’’ wave), and mean and peak pulmonary vein gradients in patients with PVS.  Phasic flow was far more common in the absence of PVS. This is also reflected in the absolute difference between the systolic and diastolic flow peak velocities and the pulsatility index. There was a trend toward Doppler velocity’s reaching baseline more often in patients without PVS.

Areas under the ROC curve demonstrated strong accuracy for all measurements, except for the absolute difference between the systolic and diastolic peak velocities and the pulsatility index. The authors report the values at which sensitivity and specificity were each 90%, optimal cutoff values that maximize the percentage correctly classified, and the sensitivity and specificity associated with those cutoffs.

DISCUSSION

The authors commented that this study describes the pulmonary vein DE features – peak systolic and diastolic velocities and the peak and mean pulmonary vein Doppler-derived gradients – associated with PVS in ex-preterm children and suggests cutoff values that have high sensitivity and specificity for the identification of PVS in the studied population.

Absence of phasic flow in the pulmonary vein was also strongly associated with the diagnosis of PVS.

Echocardiography is used as a screening tool for identification of PVS. The identification of PVS by echocardiography is dependent on identification of increased pulmonary vein flow velocity as a proxy for increased pressure gradient across the pulmonary vein. The normal peak systolic velocity of flow in the pulmonary veins in adults is 0.48 to 0.59 m/sec. However, no similar study has been performed in children. Therefore, the authors believe that values used to make the diagnosis of PVS are arbitrary; the cutoff values for pulmonary vein Doppler velocities reported range from 1.5 to 2 m/sec, and the mean pulmonary vein gradients measured by DE imaging range from >2 to >5 mm Hg.

In this study, the authors identified cutoff values for systolic and diastolic Doppler velocities and Doppler-derived mean gradients in ex-preterm children that are significantly lower than the velocities commonly cited in the PVS literature but still have high sensitivity and specificity. For example, the peak systolic velocity and peak diastolic velocity cutoff values for the identification of PVS in this study were both about 0.7 m/sec, compared to a cutoff value > 1.5 m/sec, as described previously (corresponded to 100% specificity but very low sensitivity of 11% to 29% in this cohort.)

Another point discussed was the value of echocardiography in assessing mean pressure gradient across the pulmonary veins. In this study echocardiography performed poorly in predicting the mean pressure gradient across the pulmonary vein by cardiac catheterization. This is because the mean pressure gradient and Doppler velocities across the pulmonary vein are flow dependent and thus overall, not good markers of PVS severity. Therefore, the use of Doppler-derived gradients in severity scoring systems of PVS may need to be reevaluated based on this study results.

CONCLUSIONS

The authors concluded that Pulmonary vein Doppler systolic and diastolic velocities and the Doppler waveform can discriminate between ex-preterm children with and without PVS.

They added that lower cutoff values than currently used should be considered to avoid missed and delayed diagnosis and recommended of setting up a large multicenter prospective study of echocardiographic screening for PVS in ex-preterm children using these lower cutoff pulmonary vein Doppler systolic and diastolic velocities.