I. INTRODUCTION The era of percutaneous valve procedures in the pediatric population was ushered in by the first in-man transcatheter implantation of a pulmonary valve in a 12-year-old patient with tetralogy of Fallot (TOF) and a dysfunctional right ventricle (RV) to pulmonary artery (PA) conduit, reported by Bonhoeffer et al. in 2000. The device utilized in that procedure was a bovine jugular venous valve sewn to an expandable stent. After several modifications, the device was marketed as the Melody valve (Medtronic Inc., Minneapolis, MN) (Fig. 13.1), was approved by the Food and Drug Administration (FDA) initially under Humanitarian Device Exemption guidelines in 2010, and received premarket FDA approval in 2015. At the time of this writing, there are two valves approved by the FDA for transcatheter pulmonary valve replacement (percutaneous pulmonary valve replacement [PPVR]), the Melody valve and the Edwards SAPIEN XT valve (Edwards Lifesciences, Irvine, CA), both specifically approved for patients with dysfunctional RV to PA conduits.

Approximately 20% of patients with congenital heart disease have a lesion affecting the RV outflow tract (RVOT), and a subset of these patients require a surgically implanted conduit, often early in childhood. These conduits have a limited life span, either because of somatic growth of the patient or, in fully grown patients, because of acquired stenosis or regurgitation of the valved conduit. The availability of a percutaneous approach will serve to decrease the number of open heart procedures these patients will face over the course of their lifetime, and is one of the most exciting developments in the field of pediatric cardiac intervention in the past decade.

The majority of patients who undergo PPVR at this time have had surgical placement of an RV to PA conduit to treat their underlying congenital heart disease, most commonly TOF and, in smaller numbers, other conotruncal defects such as truncus arteriosus. Another important group of patients benefiting from this technology includes those who have had a Ross procedure to treat their aortic valve disease and are also at risk of developing RV to PA conduit dysfunction. Although approval of the Melody valve was predicated on placement within an existing RV to PA conduit, its use has been extended to patients with dysfunctional bioprosthetic and also native pulmonary valves, provided the dimensions of the RVOT do not exceed the limit required for stable implantation of the valve. Implantation in a native RVOT requires creation of a “landing zone” by placement of one or more stents (“prestenting”); the valve is then deployed within the stented region. Whether the valve is deployed within a previously existing conduit or a stented native RVOT, the indications for intervention are pulmonary stenosis (PS), pulmonary regurgitation (PR), or a combination of both.

A. Pulmonary stenosis. PPVR is recommended when the RV systolic pressure is ≥75 mm Hg, or RV:left ventricular (LV) systolic pressure ratio is >0.7, but may be
considered at lower pressures if RV function is diminished. Generally, this corresponds to a peak instantaneous gradient of ≥50 to 60 mm Hg, and mean gradient of ≥35 mm Hg by echocardiogram.

B. Pulmonary regurgitation. Indications for PPVR for PR are less well defined than for PS and are in fact evolving in part owing to availability of a less invasive option than surgical replacement. The following considerations have moved the pendulum toward earlier replacement than previously recommended:

1. Using an RV end-diastolic volume of ≥170 mL/m², previously the threshold to recommend surgical PVR, RV size does not return to normal following PVR in the majority of patients.

2. RV function is considered normal if RV ejection fraction (RVEF) is ≥45% measured by cardiac MRI. When RV dysfunction is already present at the time of PVR in patients with predominant PR (in contrast to PS) despite evidence of symptomatic improvement and decreased RV volumes after PVR, RVEF frequently remains unchanged.

3. The detrimental effects of chronic PR and irreversible injury to the RV must be weighed against the risk of multiple operations, and the lack of clear evidence to date that PVR improves arrhythmia burden or survival. With a percutaneous alternative, the threshold for intervention might be justifiably lower.

4. Taking the above-mentioned considerations into account, and adapting Geva’s recommendations for surgical PVR to percutaneous PVR, PPVR should be recommended when there is moderate to severe PR (≥25% regurgitant fraction) and:

a. Symptoms, including right heart failure or progressive decrease in exercise tolerance or functional class, or:

b. No symptoms and at least two of the following:

i. RV end-diastolic volume index >150 mL/m²

ii. RV end-systolic volume >80 mL/m²

iii. RVEF <47%

iv. LV ejection fraction <55%

v. QRS duration > 140 ms

vi. Sustained tachyarrhythmia related to right heart volume load (unless a Maze procedure is thought to be indicated, in which case surgical PVR should be performed simultaneously)

vii. Increasing tricuspid regurgitation (TR) on serial evaluations

viii. RVOT obstruction with RV systolic pressure ≥ 2/3 systemic

ix. No associated lesions requiring surgical intervention, such as residual ventricular septal defect (VSD) or severe aortic regurgitation

C. Mixed PS and PR. Many patients with RV to PA conduit dysfunction have coexisting stenosis and regurgitation, resulting in both a pressure and a volume load on the RV. The threshold for intervention should be dictated by the predominant lesion, but may be somewhat lower when both present a significant hemodynamic burden.

B. RVOT dimensions too small or too large for PPVR. The Melody valve is designed to be expanded to 18, 20, or 22 mm. Expansion to 24 mm is possible, with preliminary data showing good performance of the valve in midterm follow-up. The nominal length of the valve is 28 mm, and it shortens to 26, 24, and 23 mm when expanded to 18, 20, and 22 mm, respectively.

1. Too small: The recommended diameter of the surgically implanted conduit receiving the Melody PPV is ≥16 mm. Regardless of the original conduit size, balloon sizing of the conduit must be performed to identify the narrowest diameter, determined by a visible “waist” on the balloon at relatively low inflation pressure. If after adequate preparation of the landing zone, including high-pressure balloon dilation and/or prestenting, the minimal diameter remains <14 mm, the Melody valve should not be implanted.

2. Too large: The outer diameter of the Melody valve is approximately 2 mm larger than the diameter it is expanded to. Following strict manufacturer’s guidelines, a waist no larger than 20 mm on balloon sizing is required to implant the valve at 22 mm, but in reality it can be implanted up to a waist of 22 mm, particularly after prestenting. Dilating the valve to 24 mm (outer diameter 26 mm) allows implantation in slightly larger RVOTs, up to a waist of 23 to 24 mm. Implantation in larger RVOTs is contraindicated because the valve would not have an anchoring site. The Edwards SAPIEN XT valve can be expanded to 24, 26, and 29 mm, and has allowed implantation up to a minimal diameter of 27 to 28 mm on balloon sizing.

C. Native RVOT. The approved indication for the Melody valve is implantation within a surgically implanted RV to PA conduit; however, successful off-label implantation in the native RVOT (most often following TOF repair) has been reported by several investigators and is being commonly performed. Implantation is contraindicated when the native RVOT is too large to meet sizing criteria (as outlined earlier), which occurs in the large majority of postoperative TOF patients.

D. Coronary artery compression. In some patients with a surgically implanted RV to PA conduit, the conduit lies above or in close proximity to a coronary artery. One of the most important and potentially life-threatening complications of PPVR is coronary artery compression, which has been found to occur in approximately 5% of patients. The underlying mechanism is a change in the size or geometry of the RV to PA conduit as a result of placement of the stented percutaneous valve, resulting in coronary artery compression. The only foolproof technique to rule out this possibility is aortic root or selective coronary angiography (or both if necessary) while simultaneously inflating a balloon (as much as possible to the same diameter as the intended valve diameter) at the site of intended valve implantation. Patients at highest risk for coronary artery compression are those with coronary artery anomalies, most commonly the left anterior descending (LAD) arising from the right sinus of Valsalva in patients with TOF. Patients with transposition of the great arteries (TGA) are also at increased risk, again
with the majority having an anomalous coronary artery congenitally (typically these are patients status post Rastelli procedure for TGA, VSD, and PS).

E. Bacterial endocarditis. Active endocarditis is a contraindication to PPVR.

Consideration of PPVR in a patient with pulmonary valve dysfunction begins with a thorough history and physical examination to ascertain the presence of symptoms, murmurs of PS and/or regurgitation, and signs of right heart failure if present. Cardiopulmonary stress testing, particularly if performed serially, can be very useful in identifying clinical deterioration. The presence of atrial and/or ventricular arrhythmias should be determined. Echocardiography can identify patients with significant enough RVOT obstruction to render them candidates for PPVR, but to evaluate the RVOT anatomy in detail, cardiac MRI is the gold standard. Cardiac MRI is also the gold standard to measure RV volumes and RVEF in patients with PR as the primary indication. Lastly, the detrimental effect of RV dilation and dysfunction on LV systolic and diastolic function is being increasingly recognized, making evaluation of LV performance also important in this patient population.

Patients with moderate to severe PR typically remain asymptomatic for long periods of time but ultimately develop progressive fatigue and activity intolerance, and if left untreated right heart failure will occur. Progressive RV enlargement due to PR is a known risk factor for ventricular arrhythmias and sudden cardiac death late after repair of TOF.

B. Physical examination. Increased RV impulse can be felt with significant pressure and/or volume loading of the RV. The typical murmur of PS is a systolic ejection murmur maximal at the upper left sternal border, radiating to the back. A thrill may also be palpable. Generally, the intensity of the murmur increases with the severity of obstruction, but patients with RV to PA conduits that are directly under the chest wall may have relatively loud murmurs with mild obstruction. The length of the murmur increases with increasing stenosis, and the pulmonary closure sound becomes inaudible. The murmur of PR is a low-pitched diastolic murmur also best heard at the left upper sternal border. Even with severe PR, a grade >3 murmur is unusual. Almost always in patients with PR, a “to-and-fro” murmur is heard as the increased flow through the valve causes a murmur of “relative” PS. A fixed and widely split S₂ is common owing to the almost universally present right bundle branch block (RBBB) in repaired TOF and truncus arteriosus patients. Patients with right heart failure have the typical findings of elevated jugular venous pressure, hepatomegaly, and peripheral edema. A right-sided S₄ may be present. In severe cases, splenomegaly from chronic passive congestion and ascites may also be present.

C. Electrocardiography. The majority of patients undergoing PPVR are repaired TOF patients, and characteristically the ECG in these patients has an RBBB pattern. In postoperative TOF patients with PR, a QRS duration ≥180 ms, which correlates with chronic RV volume overload and marked RV enlargement, has been identified as a risk factor for life-threatening ventricular arrhythmias and sudden cardiac death. Symptoms suggestive of ventricular arrhythmias in this subset of patients should be thoroughly evaluated, including electrophysiologic study depending on the clinical scenario, and it is one of the factors taken into account when deciding the timing of PPVR.

1. Pulmonary stenosis: The echocardiogram is an excellent screening test for patients with RVOT dysfunction, and in the majority, it can give a good estimate of the severity of obstruction by Doppler interrogation of the RVOT. A peak instantaneous gradient ≥50 to 60 mm Hg and mean gradient of ≥35 mm Hg would typically correlate with an RV pressure ≥65% to 75% systemic. RV function can be qualitatively assessed. In the presence of significant dysfunction and low cardiac output, a lower gradient may be measured despite significant stenosis because of inability of the RV to generate a higher pressure and less than normal flow across the pulmonary valve. In patients with a long, tubular stenosis, as can be seen with a diffusely narrow RV to PA conduit, the echo-estimated gradient can overestimate the severity of obstruction. If there is enough TR to obtain a reliable Doppler signal, the RV pressure can be estimated by measuring the TR velocity (V) and applying the Bernoulli equation (RV pressure = 4V² + right atrial pressure). In the absence of PA hypertension, which is not typically present in this patient population, a TR velocity of ≥4 m/s would indicate at least moderate to severe obstruction. When using the TR velocity to estimate RV pressure, it must be kept in mind that a subset of patients, particularly those with repaired TOF, may have branch pulmonary artery stenosis (PAS), which if present bilaterally could also result in RV hypertension with or without RVOT obstruction. Similarly, repaired TOF patients with RVOT obstruction who have not undergone prior surgical PVR may have subvalvar PS because of inadequate infundibular resection at the time of their initial repair. Such patients may require surgical RVOT reconstruction instead of PPVR. Detailed understanding of the entire RVOT and PA anatomy is imperative when planning PPVR, and echocardiographic imaging, especially in larger patients, must be complemented by other imaging modalities for adequate visualization of all the relevant structures.

2. Pulmonary regurgitation: The degree of PR can be assessed qualitatively by color Doppler interrogation of the pulmonary valve. The width of the PR jet, degree of extension into the RV cavity, and flow pattern in the PAs can differentiate between mild and moderate or severe PR. Flow reversal in the distal PAs indicates severe PR. More important than grading the severity of PR is assessment of the effect of PR-induced volume loading on the RV when evaluating a patient for PPVR. The degree of RV enlargement and RV contractility can be ascertained from multiple views, including parasternal long- and short-axis and apical four-chamber views. Because of the purely qualitative nature of the echocardiogram when imaging the RV, when there appears to be significant enough RV enlargement and/or dysfunction to consider PPVR, cardiac MRI should be performed to quantitate RV volumes and RVEF. In patients unable to undergo MRI owing to the presence of a pacemaker, cardiac CT scan may be considered, although limited data have shown some overestimation of RV volumes and underestimation of PR fraction by CT when compared to MRI in patients with repaired TOF. It is also important to note that progression of TR over time in an individual patient may signal progressive RV dilation and dysfunction. Changes in echocardiographic parameters of LV systolic and diastolic function should also be noted, as ventriculo-ventricular mechanical interactions may affect LV performance.

E. Cardiac MRI. Cardiac MRI is the gold standard for quantification of RV volume, RVEF, and PR and TR regurgitant fraction. It also provides excellent visualization of the entire RVOT and supravalvar anatomy, including the subvalvar region and branch PAs. It is important to tailor the study protocol to obtain all the relevant information, the details of which are beyond the scope of this writing, but well outlined by Geva (see Suggested Readings). MRI is subject to significant artifact when metal-containing prosthetic pulmonary valves or PA stents are present. However, quantification of RV volume is not typically compromised by the presence of these prosthetic materials. LV function should also be evaluated. As discussed earlier, the threshold of RV dimensions
for recommending PVR has trended downward over the past several years. Indications for PPVR based on MRI quantification of RV volume and regurgitant fraction were detailed earlier.

F. Cardiopulmonary stress testing. Functional capacity should be measured serially in patients being followed with PR. With the patient serving as his/her own control, a decrease in exercise capacity would support consideration of PPVR, provided other parameters for PPVR are met.

G. Biomarkers. Pro brain-type natriuretic peptide (proBNP) is being increasingly used in patients with congenital heart disease as a marker of deteriorating cardiac status. However, proBNP elevation is not a prerequisite for PPVR. ProBNP elevation in a patient with documented pulmonary valve dysfunction may support intervention, but waiting for this to occur may be intervening at a later than optimal time.

H. Supravalvar branch PAS. The majority of patients undergoing PPVR are repaired TOF patients. There is a significant incidence of branch PAS in these patients, either native or as a result of prior surgery. PAS may also be present in patients following repair of truncus arteriosus. Detailed assessment of the PA anatomy prior to PPVR is imperative. As mentioned earlier, RV hypertension can result from obstruction at any level along the RVOT and branch PAs, in addition to PA hypertension, and no patient should arrive in the catheterization laboratory for PPVR without an understanding of the possibility for multilevel obstruction. Cardiac MRI (and also cardiac CT) provides the necessary anatomic detail to exclude significant PAS. When unilateral PAS is present, flow distribution to each lung should be calculated by MRI to help evaluate the severity of the stenosis. Pressure recordings and PA angiography during the catheterization may be necessary for more detailed evaluation. In some cases, intervention for PAS in addition to PPVR is indicated and can sometimes be performed during the same procedure.

I. Three-dimensional models. Owing to the extensive variability in RVOT anatomy and size limitations imposed by the current valve technology, patient-specific three-dimensional print models derived from cardiac MRI datasets can be very useful in assessing the candidacy of a patient for PPVR and for planning the procedure. The model can be used to test the RVOT with different-size balloons under fluoroscopy, of course understanding the limitation of the material behaving differently than the actual conduit/tissue. This technology is commercially available, though time consuming, expensive, and not yet widely used in clinical practice, but for a specific patient with complex anatomy it may be considered.

Before describing the technical details of the PPVR procedure, it is important to discuss the institutional and operator requirements recommended for a successful transcatheter pulmonary valve program. In addition to the interventional cardiologist, the program requires the collaboration of multiple subspecialists in a “heart team” model, including cardiac surgeons, noninvasive cardiologists, cardiac radiologists, and cardiac anesthesiologists, all of whom actively care for patients with congenital heart disease. The recommended institutional volume of congenital/structural catheterizations is at least 150 cases per year, 100 of which should be therapeutic. The individual performing the procedure should have performed at least 100 catheterizations, 50 of which should have been interventions. He/she should have experience with stent implantation in the RVOT, PAs, removal of embolized foreign bodies, and assessment of the coronary arteries. A biplane catheterization laboratory is preferred. A surgical program with at least 100 open heart surgeries per year in patients with congenital heart disease and extracorporeal membrane oxygenation (ECMO) availability must exist in the same institution. Participation in a national registry (IMPACT) is also recommended.

1. Vascular access: Implantation of the Melody valve requires a #22 French delivery sheath. The diagnostic portion of the procedure can be performed with a
smaller sheath if placement of the valve is not fairly certain, but from a practical standpoint, starting with a #18 French 30-cm sheath allows easy transition to the #22 French Melody ensemble and enables advancement of other long sheaths for “prestenting” (see later), including long sheaths large enough to deliver a covered stent emergently should a conduit rupture occur. Most operators elect to preclose the vein to minimize the risk of bleeding complications at the groin site. In our experience three ProGlide Perclose devices (Abbott Vascular, Santa Clara, CA) have worked well. If maneuvering large sheaths to the RVOT is difficult, access from the right internal jugular vein (RIJ) may be obtained, and not uncommonly this approach is found to be easier. This is particularly true in smaller hearts where the double curve from the tricuspid valve to the RVOT may be more difficult to negotiate, and advancing from the RIJ may be a smoother route.