Author + information
- Received October 2, 2012
- Revision received January 1, 2013
- Accepted February 2, 2013
- Published online June 1, 2013.
- Michael W. Cullen, MD∗,
- Allison K. Cabalka, MD†,
- Oluseun O. Alli, MBBS‡,
- Sorin V. Pislaru, MD, PhD∗,
- Paul Sorajja, MD∗,
- Vuyisile T. Nkomo, MD, MPH∗,
- Joseph F. Malouf, MD∗,
- Frank Cetta, MD∗,†,
- Donald J. Hagler, MD∗,† and
- Charanjit S. Rihal, MD, MBA∗∗ ()
- ∗Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
- †Division of Pediatric Cardiology, Mayo Clinic, Rochester, Minnesota
- ‡Division of Cardiovascular Disease, University of Alabama–Birmingham, Birmingham, Alabama
- ↵∗Reprint requests and correspondence:
Dr. Charanjit S. Rihal, Division of Cardiovascular Diseases, Department of Internal Medicine, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905.
Objectives The purpose of this study was to report the results of percutaneous valve-in-valve therapy using the Melody valve (Medtronic, Minneapolis, Minnesota) for patients with degenerated mitral and tricuspid bioprosthetic valves.
Background Open surgery for replacement of degenerated bioprosthetic valves is associated with morbidity and mortality.
Methods Nineteen patients (median age 65 years, range 10 to 88 years; 7 males) with degenerated mitral (n = 9) or tricuspid (n = 10) bioprosthetic valves underwent transvenous valve-in-valve implantation of the Melody valve.
Results In the mitral patients, the mean Society of Thoracic Surgeons mortality score was 13.3 ± 5.6%. All patients had a prosthetic valve mean diastolic inflow gradient ≥5 mm Hg. Moderate or worse regurgitation was present in 7 of 9 mitral and 7 of 10 tricuspid patients. Implantation of a Melody valve was successful in all. Among the mitral patients, mean diastolic gradient decreased from 12.3 ± 4.6 mm Hg to 5.2 ± 2 mm Hg (p < 0.01). Residual regurgitation was trivial to mild in 6, mild to moderate in 2, and moderate in 1 patient. Among the tricuspid patients, mean diastolic gradient decreased from 10.0 ± 4.3 mm Hg to 5.6 ± 2.5 mm Hg (p < 0.01). Residual regurgitation was trivial to mild in 9 and mild to moderate in 1 patient. New York Heart Association functional class improved in 17 of 19 patients (p < 0.01). No periprocedural deaths, myocardial infarctions, strokes, or valve embolizations occurred. Vascular access site complications occurred in 4 patients.
Conclusions Percutaneous valve-in-valve implantation of the Melody valve in the mitral or tricuspid position for treatment of bioprosthetic valve dysfunction is feasible and can lead to significant symptomatic improvement in carefully selected high-risk patients.
Bioprosthetic valves are frequently used for patients with acquired and congenital cardiac disease despite the valves’ tendency toward earlier degeneration and valvular dysfunction (1). For patients who require treatment of degenerated bioprosthetic valves, repeat sternotomy and open surgery carry higher risk, especially in older patients with comorbid conditions (2,3). Transcatheter implantation of percutaneously delivered valves into degenerated bioprostheses (i.e., valve-in-valve therapy) has emerged as an alternative to open surgery. Reports have described valve-in-valve therapy in all 4 native valve positions (4–17), but the greatest application has been for aortic bioprosthetic degeneration.
The Melody valve (Medtronic, Minneapolis, Minnesota) is a bovine jugular venous valve designed and approved for percutaneous implantation in the pulmonary position in patients with dysfunctional right ventricular to pulmonary artery conduits and pulmonary bioprosthetic valves or homografts (18,19). Valve-in-valve Melody implantation has been reported for treatment of degenerated tricuspid bioprosthesis (5,20,21), but not for the treatment of left-sided degenerated bioprosthetic valves.
This series describes the clinical feasibility and efficacy of percutaneous implantation of the Melody valve in patients with degenerated bioprosthetic mitral or tricuspid valves.
The study was approved by the Mayo Clinic Institutional Review Board. From July 2011 through November 2012, 19 patients underwent percutaneous implantation of the Melody valve into the mitral (n = 9) or tricuspid (n = 10) position. Patients were considered candidates for the procedure if they had significant bioprosthetic mitral or tricuspid valve dysfunction (either stenosis, regurgitation, or both) with comorbid conditions that would preclude a repeat sternotomy and valve replacement. Consultation with a cardiac surgeon occurred before proceeding with percutaneous valve-in-valve therapy. All patients or parents received detailed instruction on the potential risks of the procedure, including the off-label use of the Melody valve. Alternatives, including repeat open surgery and medical therapy, were carefully discussed. All patients or parents provided informed consent for the procedure.
Mitral valve-in-valve procedure
The mitral valve-in-valve procedure was performed in the cardiac catheterization laboratory (Fig. 1). Patients were placed under general endotracheal anesthesia. Intraprocedural imaging was performed with transesophageal echocardiography (TEE). Transseptal puncture was performed using standard techniques. The atrial septum was sequentially dilated with 14-F and 21-F dilators. A 20-F Dry Seal sheath (Gore Medical, Flagstaff, Arizona) was introduced into the right femoral vein, and an 8.5-F medium curve Agilis sheath (St. Jude Medical, St. Paul, Minnesota) was placed in the left atrium over a Torayguide guidewire (Toray Industries, Tokyo, Japan). Unfractionated heparin (100 U/kg) was administered to ensure adequate systemic anticoagulation, and the activated clotting time was monitored regularly to maintain a level >250 s.
Coronary angiography was performed to delineate the course of the left anterior descending coronary artery and its major branches before left ventricular puncture. The left ventricular apical puncture was performed with an 18-gauge needle or a One-Step Centesis Catheter (Merit Medical Systems, South Jordan, Utah) under fluoroscopic and transthoracic echocardiography guidance. A 6-F sheath was then advanced into the left ventricle over a 0.038-inch wire (22).
An exchange length 0.035-inch angled extra-support glide wire (Terumo, Somerset, New Jersey) was introduced through the Agilis sheath into the left atrium and advanced across the mitral bioprosthesis into the left ventricle. The glide wire was snared in the left ventricle and exteriorized through the left ventricular apical sheath, creating a wire rail between the right femoral vein and the left ventricular apex (Fig. 1C). In selected cases, a 5-F pacing catheter was advanced into the right ventricle via the left femoral vein. The internal diameter of the dysfunctional valve was measured with TEE, and balloon sizing performed with a compliant balloon in selected cases. Balloon sizing was not performed if it was apparent the internal diameter of the existing bioprosthetic valve would support the Melody valve. In all cases, the Melody valve was mounted onto the 22-mm Ensemble delivery system (Medtronic) and delivered antegrade via the right femoral vein, across the atrial septum, and into the dysfunctional prosthesis over the arteriovenous rail. The valve was carefully positioned across the bioprosthesis and deployed under rapid ventricular pacing using fluoroscopy and TEE guidance. The left ventricular apical puncture site was subsequently closed with a 6-mm Amplatzer Vascular Plug II (AVP II, St. Jude Medical) that expanded to fill the left ventricular apical defect. Anticoagulation was reversed with protamine, and the venous sheath site was closed with a figure-of-eight suture (23).
Tricuspid valve-in-valve procedure
The tricuspid valve-in-valve procedure was performed in the cardiac catheterization laboratory under general anesthesia for 9 patients and conscious sedation for 1 patient (Fig. 2). Imaging was performed with intracardiac echocardiography (ICE) in 7 patients and TEE in 3 patients. The valve was delivered either through the right internal jugular vein or the right femoral vein. An extra-stiff 0.035-inch exchange length Amplatz wire was advanced into a distal pulmonary artery through a balloon wedge catheter. Balloon sizing of the prosthesis was then performed with a 22-mm Z-Med II balloon (NuMed, Hopkinton, New York). The Melody valve was mounted onto the 22-mm Ensemble delivery system, advanced over the exchange wire, and maneuvered across the dysfunctional tricuspid prosthesis. Position was confirmed, and the valve was deployed under fluoroscopic and echocardiographic imaging. The venous access site was subsequently closed with a figure-of-eight suture (23).
Society of Thoracic Surgeons (STS) risk score was calculated for the mitral valve patients using the mitral valve replacement algorithm. Clinical events for myocardial infarction, stroke, emergency surgery, bleeding, and vascular complications were defined using standard criteria (24–26). Continuous variables were expressed as mean ± SD if normally distributed or median with interquartile range if skewed. Normal distribution was tested with the Shapiro-Wilk statistic, and logarithmic transformations were performed when appropriate. Paired t tests and Wilcoxon signed rank tests compared pre- and post-procedure variables within patients. We defined surveillance period as the time between the procedure and the last clinical contact with the patient. Analyses were performed using JMP statistical software, version 9 (SAS Institute, Cary, North Carolina).
Table 1 outlines the baseline clinical characteristics of the study patients. Seven of 19 (37%) were male. The mean age was 75 ± 11 years for the mitral patients and 42 ± 24 years for the tricuspid patients. Three of the 10 tricuspid valve implant patients were ≤18 years of age. Median age of the dysfunctional valve was 6 (4 to 11) years. The size of the dysfunctional valve outer sewing ring ranged from 27 to 29 mm for the mitral valve patients and from 25 to 33 mm for the tricuspid valve patients. Measured internal diameter of the dysfunctional valve was 18 to 22 mm in the mitral patients by TEE and 17 to 23 mm in the tricuspid patients by balloon sizing. According to the manufacturers’ specifications, the reported internal diameter of the dysfunctional valves was 24 to 28 mm in the mitral patients and 20.5 to 31 mm in the tricuspid patients. Mean STS risk score for mortality risk in the mitral valve patients was 13.3 ± 5.6%. Despite high surgical risk, 2 of the 9 mitral patients and 4 of the 10 tricuspid patients were potentially candidates for either percutaneous valve-in-valve implantation or open surgical repair. After a multidisciplinary “heart team” discussion, the patients elected the percutaneous procedure. The other 7 mitral and 6 tricuspid cases were too high risk for conventional valve surgery and underwent percutaneous valve-in-valve replacement as an alternative to medical therapy alone. Four of the 10 tricuspid valve patients had congenital heart disease, including 3 with Ebstein’s anomaly and 1 with pulmonary valve stenosis and tricuspid valve dysplasia. Three of the 10 tricuspid patients had significant ascites.
At baseline, mean diastolic inflow gradient was 12.3 ± 4.6 mm Hg among the mitral patients. Seven of the 9 mitral patients had moderate or worse bioprosthetic mitral valve regurgitation. Mean baseline diastolic inflow gradient was 10 ± 4.3 mm Hg among the tricuspid patients. Seven of the 10 tricuspid patients had moderate or worse tricuspid valve regurgitation (Table 2).
All patients underwent successful implantation of a 22-mm Melody valve into the existing dysfunctional bioprosthetic valve. Table 3 displays the procedure characteristics according to valve-in-valve implantation position.
No periprocedural death, myocardial infarction, stroke, stent fracture, or valve embolization occurred in either the mitral or tricuspid patients. Vascular complications occurred in 4 patients. These included a right iliac artery dissection treated endovascularly, a right thigh hematoma treated endovascularly, a left femoral artery pseudoaneurysm treated surgically, and bleeding at the femoral access site that responded to conservative therapy. Two patients developed a left hemothorax requiring chest tube draining, likely related to the left ventricular apical puncture. One mitral valve patient required closure of the transseptal puncture site with an Amplatzer septal occluder (St. Jude Medical) due to significant bidirectional shunting across the transseptal puncture site.
Mean diastolic inflow gradient decreased from 12.3 ± 4.6 mm Hg to 5.2 ± 2 mm Hg for the mitral valve patients (Fig. 3B) (p < 0.01) and from 10 ± 4.3 mm Hg to 5.6 ± 2.5 mm Hg for the tricuspid valve implants (Fig. 3B) (p < 0.01). Degree of valve regurgitation improved after valve-in-valve implantation. In the mitral valve patients, post-procedure regurgitation was undetectable in 2 patients, trivial in 2 patients, mild in 2 patients, mild to moderate in 2 patients, and moderate in 1 patient (Table 2). In the tricuspid valve patients, post-procedure regurgitation was trivial in 5 patients, mild in 4 patients, and mild to moderate in 1 patient (Table 2).
No patients died during hospitalization. Median duration of hospitalization was 5 (1 to 14) days for all patients, 9 (5 to 15) days for the mitral valve patients, and 2 (1 to 9) days for the tricuspid valve patients.
Median follow-up was 41 (11 to 209) days for all patients. Three tricuspid valve patients were readmitted within 30 days of their procedure: 1 patient had a femoral artery pseudoaneurysm requiring surgical repair and 1 patient had fever due to phlebitis. An 11-year-old congenital patient with previous early bioprosthetic failure experienced Melody valve thrombosis 18 days after his procedure. He required surgical tricuspid valve re-replacement with a homograft valve. Pathological examination of the excised Melody valve revealed an occlusive thrombus. Despite anticoagulation, acute failure of the homograft tricuspid prosthesis occurred within a few days of redo surgical tricuspid valve replacement. Thrombophilia evaluation revealed heparin-induced thrombocytopenia. The patient was listed for cardiac transplantation and remains clinically stable despite bioprosthetic valve stenosis.
One mitral patient died of unknown causes 110 days after his valve-in-valve procedure. At the time of this patient’s last clinical contact with our institution, he had New York Heart Association (NYHA) functional class III symptoms, only a slight improvement from his pre-operative status. A second mitral valve patient was dismissed to palliative care due to persistent severe heart failure. He died the day after hospital dismissal. A third mitral valve patient died of unknown causes 19 days after her valve-in-valve procedure and 12 days after hospital discharge. At the time of her last contact with our institution, she reported NYHA functional class II symptoms. Overall, NYHA functional class improved in 8 of the 9 mitral valve patients and 9 of the 10 tricuspid valve patients (Fig. 3A) (p < 0.01).
The present study demonstrates the feasibility and clinical efficacy of percutaneous, transvenous valve-in-valve implantation of the Melody valve for treatment of dysfunctional bioprosthetic mitral and tricuspid valves. This series describes a novel transvenous, transseptal technique utilizing an apical rail to facilitate mitral valve-in-valve implantation. Percutaneous transvenous Melody valve implantation reduced inflow obstruction and prosthetic valve regurgitation. In the majority of patients, it improved clinical symptoms. The procedure was safe and well tolerated, with no periprocedural mortality. These findings support an emerging role for percutaneous, transvenous implantation of the Melody valve in patients with dysfunctional bioprosthetic mitral or tricuspid valves in which repeat sternotomy and open cardiac surgery carry substantial perioperative risk.
The Melody valve has an established role in the treatment of right-sided compared with left-sided valvular disease (18,27). Previous studies have suggested adequate performance under systemic pressures or in the setting of pulmonary hypertension (28). Our series adds to the previous literature supporting the feasibility of Melody valve implantation in higher pressure environments.
Other series of patients undergoing transcatheter valve-in-valve implantation have used the Edwards Sapien valve (Edwards Lifesciences, Irvine, California) (5,6,8). We chose the Melody valve for several reasons. The Melody valve is longer in length than the Edwards Sapien valve (29), facilitating coaxial alignment. The longer length of the Melody valve also allows it to cover the entire dysfunctional bioprosthetic valve, so stenting the deployment site before implantation is not necessary. At the time of our first procedures, the Melody valve was the only transcatheter valve available for off-label use in the United States. Finally, the Melody valve provided a favorable size profile for our patients. Although the reported internal diameter of the dysfunctional bioprosthetic valves in our series was as high as 31 mm, the measured internal diameter ranged from 17 to 23 mm. This decrease in size reflects degeneration of the dysfunctional prostheses. The Melody valve, by contrast, expands up to 24.06 mm, with a small percentage of recoil on the 22-mm Ensemble delivery system (30). Therefore, the Melody valve provided an appropriate size profile to fit the measured internal diameters of the dysfunctional bioprosthetic valves in our series.
The first reported attempt at mitral valve-in-valve implantation was unsuccessful due to an inability to align the percutaneous valve in a coaxial position with the dysfunctional bioprosthetic valve (4). Other series have overcome this problem through implantation via a transapical approach (5,6,8). Our approach uses a continuous rail from the femoral vein, through the atrial septum, with exteriorization via the left ventricular apex. This facilitates coaxial alignment of the percutaneous Melody valve with the dysfunctional bioprosthetic valve, allows delivery from the venous side, and minimizes the sheath size necessary to access the left ventricle. In this series, our approach did not result in any cases of valve embolization, poor valve alignment, or dysfunction of the Melody valves placed in the mitral position.
Vascular complications occurred in 4 of 19 patients (21%) in our series. All 4 vascular complications occurred on the side of the venous sheath that we used for Melody valve delivery. In 3 of the 4 cases, we also obtained arterial access at the same site. It is likely that these vascular complications were related to the large size of the venous sheath used for Melody valve delivery, concurrent arterial access, and patient comorbidities.
The left ventricular puncture carries established risks, with complication rates documented as high as 30% to 40% (22,31). In our series, 2 of 9 patients (22%) undergoing mitral valve-in-valve implantation developed a left hemothorax requiring chest tube drainage. No patients developed pericardial tamponade, arrhythmias, or persistent pain related to the left ventricular apical puncture. Therefore, despite its risks, use of a left ventricular apical puncture as part of the mitral valve-in-valve implantation procedure seems an acceptable method to facilitate alignment of the percutaneous valve.
Difficulty obtaining adequate coaxial alignment also occurs during tricuspid valve-in-valve implantation. Some authors have advocated an open transthoracic procedure to facilitate alignment (4,32,33). However, our series demonstrates that adequate alignment can occur through either a transfemoral or transjugular approach and use of a relatively flexible support wire.
Our patients undergoing mitral valve-in-valve implantation were elderly and ill, with multiple comorbid conditions and a mean STS mortality score of 13.3%. This is similar to patients in other valve-in-valve series (4–6,8). Despite their high risk, patients tolerated the procedure well. Nevertheless, valve-in-valve therapy should be reserved for patients whose surgical risk and comorbid medical conditions would preclude surgical valve intervention.
Our group of 10 patients undergoing Melody valve implantation in the tricuspid position included a heterogeneous range of pathology. The youngest 4 patients in our group all had a history of congenital heart disease requiring prior tricuspid valve replacement. The 6 older patients underwent their initial tricuspid valve replacement for tricuspid valve endocarditis in 1 case, carcinoid heart disease in 1 case, rheumatic heart disease in 1 case, and presumed functional tricuspid regurgitation in the other 3 cases. Previous reports describing tricuspid valve-in-valve therapy have occurred in a similarly heterogeneous population, including children with congenital heart disease (20,21,34), adults with congenital heart disease (21,35,36), patients with prior infective endocarditis (21,37), and patients with a history of rheumatic (21,33,38) or carcinoid (32) heart disease. The success of the procedure in our study group coupled with the positive outcomes from prior reports suggest that tricuspid valve-in-valve therapy has the potential to benefit patients with a variety of medical conditions leading to bioprosthetic tricuspid valve failure.
All patients in our series had the benefit of intraprocedural imaging with TEE (n = 12) or ICE (n = 7). Imaging support facilitated Melody valve sizing, Melody valve alignment, and assessment of valve-in-valve function immediately after implantation. We found 3-dimensional echocardiography particularly useful for the assessment of dysfunctional bioprosthetic valve morphology prior to valve-in-valve implantation (Figs. 1A and 2A) and Melody valve geometry after the procedure (Figs. 1E and 2F).
This is a single-institution, retrospective review of patient records after percutaneous valve-in-valve implantation of the Melody valve in the mitral or tricuspid position. Patients were highly selected. Post-procedure care may not have occurred at our institution, potentially limiting our ability to capture all post-procedural events. Finally, the immediate success of valve-in-valve implantation at improving hemodynamics and patient symptoms may not necessarily predict long-term outcomes. Study of larger cohorts with prolonged surveillance will be necessary to address this question. These will become possible as valve-in-valve technology and implantation techniques mature.
We report a series of 19 patients undergoing percutaneous transvenous Melody valve-in-valve implantation for treatment of dysfunctional bioprosthetic mitral or tricuspid valves. We also describe a novel technical approach utilizing an apical rail to facilitate mitral valve-in-valve implantation. Procedures were well tolerated and resulted in improved valve function and heart failure symptoms. Longer surveillance and larger cohorts are necessary to determine the ultimate role of percutaneous valve-in-valve implantation in degenerated bioprosthetic valves.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- intracardiac echocardiography
- New York Heart Association
- Society of Thoracic Surgeons
- transesophageal echocardiography
- Received October 2, 2012.
- Revision received January 1, 2013.
- Accepted February 2, 2013.
- American College of Cardiology Foundation
- Ruel M.,
- Chan V.,
- Bédard P.,
- et al.
- Webb J.G.,
- Wood D.A.,
- Ye J.,
- et al.
- Gurvitch R.,
- Cheung A.,
- Ye J.,
- et al.
- Seiffert M.,
- Conradi L.,
- Baldus S.,
- et al.
- de Weger A.,
- Tavilla G.,
- Ng A.C.T.,
- et al.
- Montorfano M.,
- Latib A.,
- Chieffo A.,
- et al.
- McElhinney D.B.,
- Hellenbrand W.E.,
- Zahn E.M.,
- et al.
- Zahn E.M.,
- Hellenbrand W.E.,
- Lock J.E.,
- McElhinney D.B.
- Riede F.T.,
- Dahnert I.
- Roberts P.A.,
- Boudjemline Y.,
- Cheatham J.P.,
- et al.
- Thygesen K.,
- Alpert J.S.,
- White H.D.,
- et al.
- Mehran R.,
- Rao S.V.,
- Bhatt D.L.,
- et al.
- Adams H.P. Jr..,
- Bendixen B.H.,
- Kappelle L.J.,
- et al.
- Hasan B.S.,
- McElhinney D.B.,
- Brown D.W.,
- et al.
- ↵Medtronic, Inc. Melody Transcatheter Pulmonary Valve and Ensemble Transcatheter Delivery System: Instructions for Use. 2010.
- Gaia D.F.,
- Palma J.H.,
- de Souza J.A.M.,
- Buffolo E.
- Roberts P.,
- Spina R.,
- Vallely M.,
- Wilson M.,
- Bailey B.,
- Celermajer D.S.