The use of a surgically placed, valved conduit between the subpulmonary ventricle and the pulmonary artery in patients with congenital heart defects was first reported in the mid 1960s.1 Such conduits made possible the repair of many complex congenital cardiac lesions involving atresia or hypoplasia of the right ventricular outflow tract (RVOT). Patients with multiple types of congenital cardiac lesions benefited from reparative surgery involving RVOT conduits and, as technical advances in surgery occurred, these procedures were performed in younger (and smaller) patients.2-11 Although initial results were exciting, it soon became apparent that conduit failure in the form of progressive valve/conduit stenosis from calcification or intimal proliferation, and/or valvular degeneration causing progressive pulmonary regurgitation (PR), was inevitable for all types of RVOT conduits.12-21 In the growing child, somatic outgrowth of the conduit(s) occurs. Irrespective of the cause of conduit failure, the need for multiple surgical procedures for conduit revisions was to become the rule for these patients. Furthermore, some found that second and subsequent conduits had shorter freedom from failure than the originally placed conduits.15

Transcatheter therapies, such as balloon valvuloplasty/ angioplasty and/or bare-metal stenting, have been instituted to provide palliative relief for conduit stenosis in hopes of extending the functional life span of conduits and have been shown to be somewhat beneficial in prolonging interval(s) between conduit replacement( s).22-25 However, especially in the case of conduit stenting, patients are left with significant PR. Multiple long-term follow-up studies have suggested that chronic volume overload of the RV can lead to impaired systolic and diastolic function of the RV, dilation of the tricuspid valve annulus leading to significant tricuspid regurgitation, impaired exercise capacity, atrial and ventricular arrhythmias, heart failure, and an increased risk of sudden death.25-30 Furthermore, restoration of pulmonary valve competence at an appropriate time has resulted in improvement in RV function, exercise tolerance, and incidence of arrhythmias.28,30,31 Until recently, there has been no transcatheter or less-invasive treatment for PR and therefore, these patients have either been managed medically or referred back for repeat surgery.

For nearly 20 years, investigators have attempted to come up with a nonsurgical treatment for failed conduits and valves.32-34 Dr. Philipp Bonhoeffer was the first (in 2000) to perform percutaneous replacement of the pulmonary valve in a right ventricular–pulmonary artery (RV-PA) prosthetic conduit of a human with valve dysfunction. 35 Since then, rapid advancements in the development of percutaneous catheter-based therapies for cardiac valve repair have occurred.36-41

Currently, in the United States, there are only two transcatheter heart valves that have been used in the pulmonary position for RVOT dysfunction: the Medtronic Melody transcatheter pulmonary valve (Medtronic, Inc., Minneapolis, MN) and the Edwards Sapien transcatheter heart valve (THV) (Edwards Lifesciences, Irvine, CA).42-46 The basic concept of both devices involves intravascular stent and bioprosthetic valve technologies.

This article provides an update on percutaneous valves used in the treatment of RVOT conduit dysfunction. It also reviews recommended criteria for use, special procedural considerations for transcatheter valve placement in the RVOT position, and recently published outcomes regarding the use of these valves for RVOT dysfunction.

The current Melody valve is a modification of the initial valve used by Dr. Bonhoeffer and consists of a platinum- iridium stent with a valved segment of natural bovine jugular vein.34,39 The Melody valve has an initial stent length of 28 mm and an initial diameter of 18 mm (Figure 1A). The venous segment is sutured to the stent around its entire circumference at inflow and outflow and at each stent node using 5-0 polypropylene sutures. The suture is blue at the outflow to signify the outflow end of the valve for appropriate alignment of the valve onto the delivery system and at implant (Figure 1B). The valve can be crimped down to 6 mm for loading and reexpanded up to 22 mm in diameter, depending on which size delivery system is used.

Delivery System
The Melody valve is implanted using the Ensemble catheter delivery system (Medtronic, Inc). The Ensemble system is composed of a balloon-in-balloon deployment catheter, which comes in three different sizes (18, 20, and 22 mm) based on the outer balloon diameter of the balloon- in-balloon catheter onto which the Melody valve is front-loaded and hand crimped. There is an outer retractable sheath that covers the valve during delivery and is pulled back just before deployment. The sheath has a side arm for flushing or delivering contrast. The entire system has a 22-F crossing profile no matter which outer balloon diameter is used (Figure 2).

Melody Outcomes and Follow-Up Data
Initial outcomes data involving use of the Melody valve in the pulmonary position in human subjects with RVOT dysfunction are from Dr. Bonhoeffer's group in Europe.35,36,39,47 In September 2006, the first prospective, nonrandomized, multicenter trial involving the Melody valve was initiated in the United States under an investigational device exemption using a standardized protocol. The objectives of the trial were to determine safety, procedural success, and short-term effectiveness of the Melody valve by assessing 6-month outcomes. Inclusion criteria included patients aged more than 5 years, weight > 30 kg, presence of a surgically placed conduit with an initial implant diameter of > 16 mm, and evidence of conduit dysfunction defined as either: 1) New York Heart Association (NYHA) class II, III, IV with RVOT mean Doppler gradient by echo > 35 mm Hg or moderate-to-severe PR; or 2) NYHA class I with RVOT mean gradient > 40 mm Hg or severe PR with RV dilation and dysfunction (RVFS < 40%) by echocardiography. The initial short-term results involving the first 34 patients enrolled in the study were reported by Zahn et al.42 The short- and medium-term outcomes of 136 patients enrolled through August 2009 were recently published by McElhinney et al.43 In the updated report of these patients by McElhinney et al, the procedural success rate was high at 98%, with only 6% experiencing serious procedural adverse events, which included uncontained conduit rupture in two patients, guidewireinduced pulmonary artery perforation in two patients, and femoral vein thrombosis in one patient. There was one death after implantation that was not device related. This patient died 20 days after implantation after experiencing complications due to coronary dissection with diagnostic coronary angiography for coronary assessment before valve implantation. Although only 15% of the patients were categorized as NYHA class I before implantation, at every follow-up interval after implantation at least 75% were categorized as NYHA class I, with no patients considered class III or IV. Before implantation, more than 80% of the patients had moderate to severe PR; however, at each follow-up interval out to 2 years, more than 90% of the patients had no/trivial PR and no patients had more than mild PR (Figure 3). Paired preimplantation and 6-month postimplantation cardiac MRI data were consistent with a significant decrease in RV end-diastolic volumes, RV mass, and PR fractions at follow- up. Stent fractures were observed in 26% of the patients, with the majority occurring in patients who did not undergo prestenting of their conduit before valve implantation. Most of these patients had been enrolled early on in the study when concomitant procedures (such as prestenting of the conduit) were not allowed. McElhinney et al did report that 10 of 122 patients (8%) required reintervention during follow-up, with all reinterventions performed due to increased RVOT gradients. All but one of these patients were noted to have stent fractures at the time of reintervention. Initial reintervention was balloon angioplasty in two patients and placement of a second Melody valve in eight patients. Only one patient required surgical conduit replacement at follow-up, and late mortality was zero.

The Edwards Sapien (formerly known as the Cribier- Edwards) THV was initially developed as a percutaneous valve option for aortic valve disease and was first successfully placed in a human for calcified aortic stenosis by Cribier.37 In 2005, a previous version of the Sapien valve was used under a compassionate use protocol for percutaneous placement in an obstructed RV-PA homograft of a patient who had previously undergone a Ross procedure. 44 Most recently, the Edwards Sapien THV was used in the treatment of seven patients with RV-PA conduit failure.45 The Sapien valve also combines balloon-expandable stent and bioprosthetic valve technology. The frame is a radiopaque, stainless steel, expandable support structure with an integrated, unidirectional, trileaflet, tissue valve fabricated from three equal sections of bovine pericardium. The valve leaflets are hand-sewn to the stent frame (Figure 4). The valve is available in 23-mm- and 26-mm-diameter sizes, with heights of 14.5 and 16 mm, respectively.45

Delivery System
The Sapien valve has been delivered via a transfemoral route using the RetroFlex 3 delivery system (Edwards Lifesciences). The RetroFlex 3 system consists of the introducer sheath set, a dilator kit, and the RetroFlex balloon catheters. The delivery system is an articulating “flex” catheter with a handle that provides a rotational grip for articulation of the distal portion of the catheter, a tapered tip at the distal end of the delivery system to facilitate tracking of the Sapien valve into the RVOT, and a balloon for deployment of the valve. This valve requires a specialized crimper that symmetrically reduces the overall diameter of the valve from its expanded size to its collapsed (mounting) size on the delivery balloon catheter.

Coronary Artery Testing
A few comments should be made regarding the risk of coronary artery compression with stent and/or valve placement in the RVOT position, because this is the most common reason why patients who would otherwise meet criteria for a percutaneous valve do not receive one. As has been discussed in previous studies, it is imperative that assessment for possible coronary compression be performed before placing a stent and/or valve in the RVOT conduit.24,42,48,49 In the patient population requiring RVOT conduits, the respective cardiac defects may either have anomalous coronary artery anatomy, such as a single coronary giving rise to significant branches that cross over the front of the heart, and/or the aorta may be in such an anterior or rotated position that even normal origins of the coronaries may still be relatively displaced. Still in other patients, surgery for their congenital cardiac defects may include reimplantation of the coronary arteries, such as in an arterial switch operation or a Ross procedure. The reimplanted coronaries, particularly the left coronary artery, may be at risk of compression during RVOT conduit stenting or valve placement, depending on the postsurgical anatomic relationship between the more anteriorly placed RV-PA conduit and the coronaries. Often, it is not possible to predict the risk of coronary compression with noninvasive testing; therefore, the risk of coronary compression must be evaluated at catheterization before bare-metal stenting and/or valve placement. Initial screening is done by aortic root or selective coronary angiography with a wire or catheter in the RV-PA conduit. If there is any concern of a close anatomic relationship between the conduit and a significant coronary branch, simultaneous inflation of a balloon within the conduit and aortic root and/or selective coronary angiography should be performed (Figure 5). A few important technical considerations include: (1) be sure the appropriate balloon size and inflation pressure is used to best approximate the resultant diameter of the conduit after stent/valve implantation; (2) multiple imaging projections may be necessary to provide the best assessment of the conduit-coronary relationship; (3) remember that selective coronary angiography may mask ostial stenosis or very proximal obstruction during simultaneous conduit balloon dilation; (4) use very dilute contrast to inflate the balloon because a more concentrated mixture may make it difficult to see the course of the coronary arteries behind or underneath the balloon; (5) make sure the widest part of the balloon being used to dilate the conduit is actually positioned over the region or coronary in question; and (6) remember that ringed conduits or bioprosthetic valve rings do not remove the risk of coronary compression.

Conduit Rupture
It is not uncommon for contained conduit tears to occur with balloon dilation for conduit stenosis. Most often these tears are hemodynamically insignificant. In the United States investigational device exemption trial involving the Melody valve, uncontained conduit ruptures occurred in two patients. In the first instance, the patient was temporarily palliated with a chest tube and then underwent emergent surgical conduit replacement. The second patient was treated under emergent use protocol with a large-diameter, covered stent. Of note, the Melody valve itself functions as a covered stent, with the venous wall segment attached to the stent frame; however, ruptures of the conduit may not be in the same region that optimal valve placement would be and/or the ruptures may be longer than the postimplantation valve length.

Pulmonary Artery Injury
Pulmonary artery injury (perforation) is another serious procedural adverse event noted in the Melody trial. Vessel injury was presumed to have occurred due to wire perforation. Percutaneous valve implants require use of very stiff guidewires and great care should be taken in attaining appropriate and stable wire position before valve deployment to keep the risk of vessel injury to a minimum.

Despite improved stent technologies over the years, stent fractures are a common complication after implanting intravascular stents into cardiovascular-related positions. 23,24,50-53 This has been shown to be true for the Melody valve, as well.43,47,54,55 The etiology of stent fractures is thought to be secondary to the dynamic implant environment and the intrinsic metal characteristics of the stents. With ongoing implantation experience, the Melody investigators have implemented certain practices to prepare conduits for implantation in the hopes of decreasing the incidence of stent fractures. These include: (1) aggressive predilation of calcified and stenotic conduits using high-pressure balloons and a balloon diameter < 110% of the conduit diameter at implantation; (2) prestenting conduits with bare-metal stent(s) until evidence of recoil of the conduit during balloon deflation is gone; and (3) attempting to implant the valve in a position that is not directly under the sternum, when possible. Preliminary evaluation of the Melody data suggests that when these techniques are implemented, the incidence of stent fractures at follow-up is significantly decreased, with 35% of fractures occurring in conduits not prestented compared to only 7% of fractures occurring in prestented conduits. When valve stent fracture occurs and reintervention is indicated, placement of bare-metal stent(s) within the initially placed valve with subsequent placement of a second valve has been shown to be successful.43,47

On January 25, 2010, the United States Food and Drug Administration formally approved the Melody valve for placement in dysfunctional RVOT conduits under a Humanitarian Device Exemption, making it the first transcatheter heart valve to receive United States Food and Drug Administration approval. Enrollment into the United States feasibility trial involving use of the Edwards Sapien THV in the pulmonic position was completed in May 2010. Plans for a pivotal study are in progress.

Percutaneous valve therapies for RVOT conduit dysfunction have been shown to be feasible, safe, and effective, with high procedural success rates and low adverse event rates. Procedural considerations, such as careful assessment for risk of coronary compression, cautious use of stiff guidewires, and awareness of the potential for serious adverse events such as conduit rupture, as well as knowing how to prevent and manage such events, will lead to safe and successful percutaneous valve implantations.

Julie A. Vincent, MD, FACC, FSCAI, is Director of the Cardiac Catheterization Laboratories and Director of the Pediatric Cardiology Fellowship Program at Morgan Stanley Children's Hospital of New York-Presbyterian and the Komansky Center for Children's Health at Weill Cornell Medical Center. She is Assistant Professor of Clinical Pediatrics at Columbia University, College of Physicians & Surgeons in New York, New York. She has disclosed that she is a consultant to and proctor for Medtronic, Inc. Dr. Vincent may be reached at (212) 305-6069;

William E. Hellenbrand, MD, FSCAI, is the Director of the Division of Pediatric Cardiology, a combined program at Morgan Stanley Children's Hospital of New York- Presbyterian and the Komansky Center for Children's Health at Weill Cornell Medical Center. He is Professor of Clinical Pediatrics at Columbia University, College of Physicians & Surgeons in New York, New York; and holds the position of the Welton M. Gersony Professor of Clinical Pediatric Cardiology. He has disclosed that he is a consultant to and proctor for Medtronic, Inc. Dr. Hellenbrand may be reached at (212) 342-0610;