Author + information
- Received January 12, 2015
- Revision received August 10, 2015
- Accepted August 13, 2015
- Published online December 21, 2015.
- William M. Wilson, MBBS∗,
- Lee N. Benson, MD†,
- Mark D. Osten, MD∗,
- Ashish Shah, MD∗ and
- Eric M. Horlick, MDCM∗∗ ()
- ∗Toronto Congenital Cardiac Centre for Adults, Peter Munk Cardiac Centre, University Health Network and University of Toronto, Toronto, Ontario, Canada
- †Hospital for SickKids, Toronto, Ontario, Canada
- ↵∗Reprint requests and correspondence:
Dr. Eric Horlick, Toronto Congenital Cardiac Centre for Adults, Toronto General Hospital, 200 Elizabeth Street Room 6E-249, Toronto, Ontario M5G2C4, Canada.
Objectives This study sought to review the outcomes for the Sapien and Sapien XT valves (Edwards Lifesciences, Irvine, California) for percutaneous pulmonary valve implantation (PPVI).
Background PPVI has emerged as a viable alternative to surgery in patients with right ventricular (RV) outflow tract dysfunction. Limited data are available for the Sapien and Sapien XT valves in this setting.
Methods Retrospective analysis was performed for all patients to have undergone PPVI using the Edwards Sapien system at a large quaternary center.
Results Twenty-five patients (70% male, mean age 34 ± 8.9 years) were identified. Primary underlying diagnosis was tetralogy of Fallot (n = 15), Ross procedure (n = 5), and other (n = 5). RV outflow tract characteristics included: biological valve (n = 16) and homograft (n = 9). Technical success was 96%. One patient required elective surgical pulmonary valve replacement for a high residual gradient. Pre-stenting was performed in all cases (52% covered stents). Valve sizes were 23 mm (n = 8), 26 mm (n = 15), and 29 mm (n = 2). Procedural hemodynamics revealed a decrease in the mean RV-to-systemic pressure ratio from 0.64 to 0.36 (p < 0.001) and RV-to-pulmonary artery (PA) gradient from 39 to 9 mm Hg (p < 0.001). No patient had clinically significant pulmonary regurgitation (PR). At a mean follow-up of 3.5 ± 2.1 years (range 0.3 to 7.2 years), there were no deaths. One patient required reintervention (no PR evident immediately post-procedure but severe valvular PR at 1 year requiring a valve-in-valve procedure). There were no episodes of endocarditis and no stent fractures. There was preserved valve function during follow-up with no change in RV-to-PA gradient nor PR severity.
Conclusions The Edwards Sapien system is a viable and durable option for PPVI in this single-center study.
Surgery to reconstruct the right ventricular outflow tract (RVOT) is commonly performed in patients with congenital heart disease. These procedures are not curative, and most patients will require further procedures later in life to address a dysfunctional valve or conduit. Percutaneous pulmonary valve implantation (PPVI) has emerged as an alternative to surgical pulmonary valve replacement in certain patients. Currently available valves suitable for use in this setting include the Melody transcatheter pulmonary valve (Medtronic, Minneapolis, Minnesota) and the Edwards Sapien transcatheter heart valve system (Edwards Lifesciences, Irvine, California). Multiple studies (totaling 464 patients with 1- to 5-year follow-up) have demonstrated efficacy of the Melody valve for PPVI, for RVOTs smaller than 24 mm (few native outflow tracts) (1–4). Initial issues with stent fractures have been resolved with routine pre-stenting. By contrast, published experience with the Sapien system in this setting is limited to 1 prospective, nonrandomized, multicenter study (5) and 3 small case series (6–8), with a total of 72 patients (Table 1). Use of the Sapien valve has afforded treatment of larger conduits and, in select cases, native outflow tracts (9,10). There are little long-term data, however, with most studies reporting up to 6 months follow-up only. We describe our experience with the Sapien system in the pulmonary position.
Between October 2007 and October 2014, 25 patients underwent percutaneous pulmonary valve implantation utilizing the Edwards-Sapien system at the Toronto Congenital Cardiac Centre for Adults (representing one-third of all PPVI procedures performed in the same time period). Patients were considered for PPVI at Toronto Congenital Cardiac Centre for Adults if they had congenital heart disease requiring previous RVOT surgery, with an indication for pulmonary valve replacement according to guidelines (11) (RVOT obstruction with peak instantaneous echocardiography gradient >50 mm Hg or catheter-derived RV/left ventricular pressure ratio >0.7 or significant pulmonary regurgitation with an enlarged RV more than 150 ml/m2). An RVOT diameter between 21 mm and 28 mm (after pre-stenting) was considered appropriate for the Sapien prosthesis. Patients were not considered for the procedure if: weight <30 kg, pregnant, evidence of active infection, or unfavorable RVOT morphology (including the potential for left main coronary compression or occlusion of branch pulmonary arteries).
All patients underwent a comprehensive pre-procedural work-up, including transthoracic echocardiogram to estimate RV systolic pressure from the tricuspid regurgitant jet, to calculate RV-to-pulmonary artery (PA) gradient, to assess pulmonary regurgitation severity, and estimate RV size/function; and magnetic resonance imaging or computed tomography where appropriate to accurately define the RVOT anatomy and accurately assess RV size and systolic function. Catheterization was performed to define hemodynamics and assess the RVOT with angiograms acquired in multiple projections. Selective coronary angiography was performed to define proximity to the RVOT. Repeat selective coronary angiography was performed at the time of the procedure during high-pressure expansion of a noncompliant balloon in the RVOT if warranted.
All procedures were performed under general anesthesia with fluoroscopic guidance. Transesophageal echocardiography was not used. Prophylactic antibiotics were administered at the time of procedure before valve implantation. Large sheath access site was via the femoral vein in 22 patients and jugular vein in 3 patients. Anticoagulation was achieved with intravenous heparin (100 U/kg to achieve and activated clotting time of longer than 250 s). Pre-stenting was performed in all cases and mostly (96%) at the time of valve implantation. The Edwards Sapien transcatheter heart valve was available in 2 sizes (23 and 26 mm), and was used with Retroflex I or III catheters (Edwards Lifesciences) for delivery. The Sapien XT is the newer iteration of the valve and was available in Canada from 2010. It is available in 3 sizes (23, 26, and 29 mm) and was used in 4 cases. The XT valve is crimped and loaded on the NovaFlex delivery system (Edwards Lifesciences) catheter shaft outside the patient. The crimped valve requires repositioning in the inferior vena cava after exiting the sheath, so as to position it on the balloon before delivery and deployment in the RVOT. Rapid pacing was required in 2 patients only, in whom excessive movement was noted during balloon sizing or pre-stenting.
Hemodynamics were reassessed post-valve deployment and angiography performed to assess for degree of pulmonary regurgitation. Closure devices (ProStar or Perclose ProGlide, both Abbott Vascular, Santa Clara, California) were used in all cases for venous hemostasis.
Clinical follow-up was performed at 3 and 12 months and then annually thereafter. Chest x-rays and echocardiography were performed at the time of clinical review. Cardiopulmonary stress testing and cardiac magnetic resonance imaging were performed at 12 months if appropriate.
Unless otherwise indicated, the mean ± SD was calculated for continuous variables, whereas number and percentage were calculated for categorical variables.
Technical success was defined as deployment of the valve to the target area with removal of the delivery catheter from the body without significant residual RVOT dysfunction (defined as mean RV-to-PA gradient <15 mm Hg and pulmonary regurgitation severity mild or less).
Need for reintervention during follow-up was ascertained. All continuous variables were tested for normality before data analysis, and paired t test was used for comparison of mean values pre- versus immediate post-procedure. Due to the differences in follow-up duration, the annual rate of change was calculated and analyzed using 1-sample test. Values of p < 0.05 were considered statistically significant for all tests.
PPVI using the Sapien system was attempted in 25 patients. Baseline characteristics are presented in Table 2.
Technical success was achieved in 24 of 25 patients (96%). A high residual gradient was present in 1 patient, who had a 26-mm Sapien valve implanted into a stenotic 27-mm Ionescu-Shiley bovine pericardial bioprosthetic valve. In retrospect, this bioprosthetic valve had an internal diameter of only 19 mm (confirmed by balloon sizing); the high gradient was inevitable and related to the small size of the bioprosthesis rather than to a procedural issue. This case highlights the importance of confirming gradient reduction after pre-stenting and before valve implantation. In fairness, this case was performed early in the percutaneous pulmonary valve experience when little information was available about true internal diameters of biological valves. The patient required nonurgent surgical pulmonary valve replacement.
Procedural data are presented in Table 3. Valve sizes used were 23 mm (n = 8), 26 mm (n = 15), and 29 mm (n = 2). The Sapien XT valve was used in 4 cases, the first being delivered via a Glenn shunt (Figure 1). The mean number of stents used for pre-stenting was 1.1 ± 0.2 (range 1 to 2). Fifty-two percent of the stents used were covered. Stent types used were Cheatham-Platinum bare and covered stents (NuMed, Hopkinton, New York), Palmaz XL series (J&J Interventional Systems Co., Warren, New Jersey), Genesis (J&J Interventional Systems Co.), and Andrastent (Andramed, Reutlingen, Germany). There appeared to be a reduction in procedural time and radiation exposure over the course of the study; median values for the first 12 versus the remainder of cases were 195 versus 125 min and 27,696 versus 21,050 cGycm2, respectively.
Complications occurred in 3 patients. Valve malposition contributed to a significant paravalvular leak in 1 patient, who required insertion of a second transcatheter valve. In this case, there existed significant regurgitation after implantation of a 26-mm Sapien valve within a bare stent, which was placed in a failing 29-mm Hancock II prosthesis (true internal diameter 24 mm) (Figure 2). Insertion of a covered stent and second 26-mm Sapien valve successfully eliminated pulmonary regurgitation. Loss of wire position (Lunderquist Extrastiff wire, Cook Medical, Bloomington, Indiana) occurred in another patient when attempting to deliver a 26-mm Sapien valve with subsequent need for vascular surgery to remove the valve from the body. The procedure was completed successfully with another Sapien valve. Mild hemoptysis occurred in another patient, which did not require treatment.
Hemodynamic data pre and post PPVI are presented in Table 4.
Mean follow-up was 3.5 ± 2.1 years (range 0.3 to 7.2 years). There were no deaths during follow-up. Endocarditis was not reported nor were stent fractures. There was 1 reintervention, which occurred in a 24-year-old man with tetralogy of Fallot who underwent successful PPVI for a failing 25-mm Perimount prosthesis (true internal diameter 23-mm) with a 26-mm Sapien valve. No pulmonary regurgitation was evident initially post-procedure; however, the patient returned at 1 year with recurrent symptoms and severe valvular regurgitation of unclear cause. A successful valve-in-valve procedure was performed with a 22-mm Melody valve (Medtronic), but at 6-month routine follow-up, there existed severe valvular pulmonary insufficiency. The etiology remains unclear with no evidence for endocarditis nor stent fracture. The patient is being considered for surgical pulmonary valve replacement.
There was a sustained improvement in New York Heart Association (NYHA) functional class post-PPVI (median [interquartile range]: NYHA functional class 2 [2 to 3] pre-PPVI versus 1 [1 to 1] as at last follow-up, p = 0.031). At last follow-up, all patients were NYHA functional class 1 or 2 (78% NYHA functional class 1). Ten patients had paired cardiopulmonary studies with a trend to an improved functional capacity (Vo2 peak % predicted 53 ± 11% pre-PPVI and 60 ± 13% post-PPVI with an average annual percent change of 4.5% [−0.8% to 9.8%, p = 0.09]).
There was no evidence of progressive valve dysfunction on serial echocardiography. Mean RV systolic pressure and pulmonary valve peak gradient by echocardiography were 48 ± 8 mm Hg and 20 ± 9 mm Hg at the initial study post-procedure, and there was no significant rise in these values over the follow-up duration with the average annual percent change of 1.7% (95% confidence interval: −0.8% to 4.2%; p = 0.178) for RV systolic pressure and 7.4% (95% confidence interval: −2.2% to 17.0%; p = 0.124) for pulmonary valve peak gradient, respectively (Figures 3A and 3B). All patients had nil or mild pulmonary regurgitation at the initial echocardiogram post-PPVI (80% nil or trace, 20% mild). Ninety-six percent had nil or mild pulmonary regurgitation at last follow-up (1 patient had severe pulmonary regurgitation as previously discussed) (Figure 3C).
Procedural success in this series was high (96%), which compares favorably with published series using the Sapien (5–8) or Melody (1–4) valves in the pulmonic position. The Sapien prosthesis was successfully deployed in all cases of this series, and the single “failure” case related to a high residual gradient after valve insertion, which reflected a small surgical prosthesis rather than a problem with the Sapien system itself. This case highlights the need to have a good understanding of surgical valves (in particular, their internal diameters) and more importantly, the need to invasively measure RV-to-PA gradient post-stent insertion before valve implantation to ensure complete abolition of a gradient. Balloon compliance testing can be useful in this instance to accurately define the waist diameter of a surgical prosthesis.
Use of the Sapien system affords treatment of larger diameter RVOTs (up to 28 mm after pre-stenting) than does the Melody valve (up to 24 mm after pre-stenting). Implantation of larger Sapien valves may, therefore, allow for more valve-in-valve procedures in the future in the event of valve dysfunction.
The Melody valve is a bovine jugular valve sutured to a 34-mm-long Cheatham-Platinum stent. Significant recoil with a high stent fracture rate was encountered in the early experience with this valve, leading to a change in practice in preparation of the RVOT; routine pre-stenting with 1 or multiple stents has significantly reduced the incidence of stent fracture/recoil (4). Pre-stenting is also usually performed when using the Sapien prosthesis in the pulmonic position (86% of cases in published series), not necessarily to prevent stent fractures (given it has a high radial strength), but rather to ensure appropriate preparation and elimination of RVOT gradient, which may not otherwise be achieved by the short stent frame of the Sapien valve. We advocate high-pressure post-dilation of the stent before valve deployment to address any residual narrowing/gradient, and we caution against using valve deployment to treat residual stenosis. Bare stents were predominantly used in previous series of the Sapien valve in the pulmonic position. Pre-stenting was performed in all cases in our series, with a trend toward the use of covered stents in more recent cases, primarily to mitigate against paravalvular leak. All patients except 2 required a single stent only for pre-stenting, and of the 13 patients in whom covered stents were used, recoil was encountered in only 1 (for which deployment of a J&J 5014 stent was effective in providing sufficient radial strength before valve insertion). One patient in our series had significant paravalvular leak post-insertion of a 26-mm Sapien valve after pre-stenting with a bare-metal stent in a bioprosthesis. This was likely related to malposition of the valve (implanted slightly on the RVOT side of the bioprosthesis within a bare stent) such that the covered region of the Sapien prosthesis did not achieve a 360° seal at the level of the previous bioprosthesis ring (Figure 2). Leakage thus occurred around the valve, through the upper uncovered struts of the Sapien prosthesis. Covered stent use in this setting within a bioprosthesis effectively elongates the landing zone for the valve where a 360° seal can be more easily achieved. There are advantages and disadvantages to covered stent use in this setting that need to be considered. Firstly, covered stent use can reduce the risk of malposition. There is a large global experience with valve-in-valve implantation with the Sapien system in the aortic position (12), where it is relatively easy to reposition a partially deployed valve, however, tortuosity encountered in the pulmonary position means precise repositioning is difficult making malposition a greater issue. This phenomenon of “geographic miss” is less of an issue within homografts than within bioprostheses, however, covered stent use in homografts may be important to mitigate against the uncommon event of conduit rupture in calcified conduits. In large, irregular conduits, where rupture is less of an issue, the placement of a covered stent may similarly reduce the risk of paravalvular leak. Covered stent use, however, mandates use of a larger French sheath for delivery, which may make sheath delivery more difficult through tortuous anatomy. Another potential issue is the contribution of covered stents themselves to a higher gradient across the RVOT, either through creation of a smaller inner diameter or if there is protrusion into the RVOT. Furthermore, migration or malposition of a covered stent may result in occlusion of a branch pulmonary artery. Clearly, there are risks and benefits of covered stent use to be balanced, but we feel advantages (protection against rupture and elimination of pulmonary valve leak in particular) are reasonably compelling and consider covered stent use appropriate within homografts, large irregular conduits, and bioprosthetic valves.
Use of the newer Sapien XT valve in the pulmonic position has not been previously published. It was used successfully in 4 patients in this series, the first being deployed successfully via a Glenn shunt in a patient with pulmonary atresia with intact septum who underwent a 1.5 ventricle repair in childhood. The valve orientation in this case was the same as is used in the aortic position, but the realignment maneuver had to be performed within the eSheath, which was more difficult to achieve than realignment within the inferior vena cava. When delivered via the RVOT (from the jugular or femoral veins), the Sapien XT valve needs to be mounted in the opposite direction, and a theoretical concern existed initially regarding leaflet damage due to involution during the realignment maneuver. Subsequent to our first case via the Glenn shunt, the manufacturer was able to confirm through bench testing that this maneuver would not damage the valve prosthesis and, thus, the NovaFlex delivery system could be used via the jugular or femoral vein. Indeed, implantation of XT valves from the femoral vein was subsequently performed in 3 patients without issue. Both the Melody and Sapien valves can be difficult to deliver in tortuous anatomy. We did not feel that delivery of the 29-mm Sapien XT valve was more difficult than the smaller diameter valves or the Melody valve, and we encountered no issues with valve delivery from the internal jugular approach. Techniques to facilitate delivery of the Sapien prosthesis in tortuous anatomy include separation of the valve from the pusher and use of a stiff guidewire (we would usually start with a Lunderquist Extrastiff wire (Cook Medical) but consider use of a Meier wire (Boston Scientific, Natick, Massachusetts) if delivery is anticipated to be a significant issue) to guide the softer balloon shaft and valve through the conduit. A later stage maneuver if difficulty is encountered would include making a reverse loop in the right atrium; however, in our experience, this approach may increase the risk of loss of wire position.
The complication rate in this series was low, in keeping with other published PPVI experience. Surgical cutdown was required in 1 case to remove an undeployed Sapien 26-mm valve after loss of wire position in the branch pulmonary artery. This may be of historical interest given undeployed XT and Sapien 3 valves can be removed via the eSheaths through which they are delivered, given that they have a dynamic expansion mechanism afforded by a longitudinal slit in the sheath, which allows for transient sheath expansion. When removing a valve in this manner, it is preferable that the undeployed valve is positioned on the balloon (i.e., in the post-alignment position), rather than on the shaft, for stability during removal.
Median follow-up was short in prior published studies (5–8); it ranged from 5.7 to 22 months, and was approximately 6 months in 3 of the 4 prior series. Over a mean 3.5-year follow-up in our study, there was a high freedom from reintervention and no evidence for progressive valve dysfunction on serial echocardiograms. Valve failure occurred in 1 patient during follow-up who re-presented with symptomatic severe pulmonary regurgitation at 1 year. The exact cause for failure was unclear, and there was no evidence for endocarditis. The same patient had early failure of a surgical bioprosthesis previously, raising the possibility of an autoimmune phenomenon.
Although there are limited data for the Sapien system in the pulmonic position, there are abundant favorable data supporting use of the Sapien system in the aortic position in high-risk elderly patients with severe aortic stenosis, where valve durability does not appear to be an issue (13,14).
Valve fractures were not seen in our series but were not systemically searched for. There are no reports to date of fractures involving the Sapien prosthesis in the aortic or pulmonic position. There were no confirmed cases of endocarditis in our series.
This is a small, observational study, and larger studies with longer follow-up duration are warranted to confirm the durability and safety of the Sapien system in the pulmonic position. Ideally, a standardized follow-up protocol would be employed to track functional status and valve function.
Use of the Edwards Sapien system in the pulmonic position is a viable and durable option with low valve failure rates during long-term follow up. Pre-stenting with a covered stent can be useful to elongate the landing zone for the valve and prevent paravalvular leaks. Future registry data should reinforce the utility of this valve for PPVI.
WHAT IS KNOWN? Percutaneous pulmonary valve implantation has emerged as a viable alternative to surgery in patients with right ventricular outflow tract dysfunction.
WHAT IS NEW? The Melody Transcatheter Pulmonary Valve has a size limitation (maximum outer diameter of 24 mm), whereas the largest Edwards Sapien valve has a diameter of 29 mm, which affords treatment of patients with larger homografts, biological valves, and selected native outflow tracts. This study adds to the worldwide experience with the Edwards Sapien system in the pulmonic position, for which there are limited data otherwise.
WHAT IS NEXT? Future larger studies should reinforce the safety and durability of this valve for percutaneous pulmonary valve implantation.
Dr. Horlick is a proctor and a consultant for Edwards Lifesciences; and has received research funding from Edwards Lifesciences. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- New York Heart Association
- pulmonary artery
- percutaneous pulmonary valve implantation
- right ventricle/ventricular
- right ventricular outflow tract
- Received January 12, 2015.
- Revision received August 10, 2015.
- Accepted August 13, 2015.
- American College of Cardiology Foundation
- Eicken A.,
- Ewert P.,
- Hager A.,
- et al.
- Lurz P.,
- Coats L.,
- Khambadkone S.,
- et al.
- McElhinney D.B.,
- Hellenbrand W.E.,
- Zahn E.M.,
- et al.
- Kenny D.,
- Hijazi Z.M.,
- Kar S.,
- et al.
- Bertels R.A.,
- Blom N.A.,
- Schalij M.J.
- Guccione P.,
- Milanesi O.,
- Hijazi Z.M.,
- Pongiglione G.
- Warnes C.A.,
- Williams R.G.,
- Bashore T.M.,
- et al.