Author + information
- Received November 6, 2004
- Revision received December 21, 2004
- Accepted January 11, 2005
- Published online July 19, 2005.
- Younes Boudjemline, MD⁎,†,⁎ (, )
- Gabriella Agnoletti, MD⁎,
- Damien Bonnet, MD⁎,†,
- Luc Behr, DVM‡,
- Nicolas Borenstein, DVM‡,
- Daniel Sidi, MD⁎,† and
- Philipp Bonhoeffer, MD§
- ↵⁎Reprint requests and correspondence:
Dr. Younes Boudjemline, Service de Cardiologie Pédiatrique, Hôpital Necker Enfants Malades, 149 Rue de Sévres, 75015 Paris Cedex, France
Objectives The goal of this study was to develop a device for percutaneous replacement of the tricuspid valve in animals.
Background Percutaneous valve replacement has recently been introduced, and early clinical experience has been reported. To date, this technique is limited to the replacement of pulmonary and aortic valves in selected patients.
Methods A newly designed nitinol stent, forming two large disks separated by a cylinder with a diameter of 18 mm, was specially designed for the purpose of this study. An 18-mm bovine valve was mounted in the central part of the stent, and a polytetrafluoroethylene membrane was sutured onto the ventricular disk. Eight ewes were equally divided into two groups (group 1, acute study; group 2, killed at one month).
Results Seven of eight devices were successfully delivered in the desired position. In one animal, the device was trapped in tricuspid cordae, leading to its incomplete opening. A significant paravalvular leak was noticed in one animal of group 2. Mean right atrial pressure increased from 5 to 7 mm Hg and did not change during the follow-up. At autopsy, examination confirmed the good position of devices in successfully implanted animals.
Conclusions Implantation of a semi-lunar valve in the tricuspid position is possible in ewes through a transcatheter approach. A disk-based nitinol stent is needed to allow valve implantation in the atrioventricular position. These studies open new perspectives into tricuspid as well as mitral valve replacement.
Percutaneous valve implantation has been introduced recently with promising results (1–6). Experience is presently limited to aortic and pulmonary valves. Extensive work has been undertaken to extend indications to the whole spectrum of aortic and pulmonary valve diseases as well as to atrioventricular cardiac valves. Various approaches have been reported in animals to repair the mitral valve, namely the edge-to-edge or Alfieri repair using an intervascular clip, and the annuloplasty through the coronary sinus (7,8). However, to date, no transcatheter technique has been described to replace mitral or tricuspid valves. To broaden the indications of percutaneous approach to atrioventricular valves, we developed a self-expandable stent that allowed: 1) downsizing of the diameter of the annulus to the available valve diameter, and 2) device fixation to the annulus. We report here our experience with this new device in ewes.
We designed a self-expandable stent constructed from a 0.22-mm nitinol wire (AMF, Reuilly, France). It is a symmetrical device with an overall length when deployed of 15 mm. It is formed by two flat disks and a tubular portion. The disks and the central part had a spontaneous diameter of 40 mm and 18 mm, respectively (Fig. 1A).It is braided using a single wire, making all parts physically interconnected. When deployed, the disks tended to join in the middle of the tubular part (Fig. 1B). Because of this design and the alloy properties, the tendency of apposition of these disks created forces that fixed the device around the annulus with one disk being deployed in the right ventricle (RV) and one in the right atrium (RA). Additionally, the tubular part interconnected between the disks acted as a supporting structure for the valve to be implanted.
A naturally valved venous segment, harvested from the bovine jugular vein (Contegra, Medtronic Inc., Minneapolis, Minnesota) was prepared as previously reported (1,2) and mounted into the tubular part of the self-expandable stent. To guarantee the sealing of the device, we sutured a polytetrafluoroethylene (PTFE) membrane, usually used for covered stents (Zeus Inc., Orangeburg, South Carolina), on the outside of the ventricular disk (Fig. 1C). The atrial disk was not covered to limit the risk of coronary sinus occlusion. All devices were stored in glutaraldehyde solution until their use.
The delivery system
The delivery system consisted of a “homemade” front-loading 18-F long sheath (Cook Inc., Charenton le Pont, France). For the purpose of the study, the distal tip of a dilator was cut off. A piece of catheter was fixed to this part to liberate space for stent placement. The length of this space (i.e., piece of catheter) was 5.5 cm, which was the length of the device when in the constrained position. At the tip of the catheter, a 1-cm-long dilator was attached to allow for a smooth transition between the tip and the sheath and to facilitate the tracking of the delivery system during its course. The sheath could freely slide over the device. No balloon was necessary to deliver the nitinol stent that spontaneously deployed at the time of uncovering.
Preparation of the animals
Animals were treated according to the European regulations (9), and the protocol was approved by the institutional ethics committee. Eight ewes weighing 60 to 70 kg were included. We intended to implant in the tricuspid position a device sheltering an 18-mm valve as a one-step procedure. Animals were divided into two equal groups according to the killing time points. General anesthesia was induced with 10 mg/kg of thiopental and maintained with isoflurane. Right jugular and femoral veins were prepared for catheterization.
Percutaneous replacement of the tricuspid valve
Through the right jugular vein, a 5-F right Judkins coronary catheter (Cordis, Issy les Moulineaux, France) was advanced in the distal right pulmonary artery (PA). Through this catheter, a 0.035-inch extra-stiff guidewire (Amplatzer, Golden Valley, Minnesota) was positioned distally. The valved device was loaded into the delivery system, inserted over the previously positioned wire, and advanced into the RV. As with devices for closure of atrial septal defects, the distal disk was deployed in the RV by pulling on the external sheath while maintaining the dilator in position. This disk was then applied to the tricuspid annulus by pulling on the external sheath and dilator. After deployment of the tubular part containing the valve, the second disk was delivered similarly in the RA (Fig. 2).The two disks sandwiched the annulus, with one disk laying into the RV and the proximal one in the RA.
Epicardial echocardiography imaging and cardiac catheterization
The RV and RA pressures were measured before and after device implantation. Angiographic evaluation consisted of an atrial injection and a right ventriculography. A small left thoracotomy was performed in all animals to allow for an epicardial echocardiography. Angiograms and echocardiography were initially performed to define the anatomy of the area of interest and to measure the maximum diameter of the tricuspid annulus. Studies were also repeated after implantation and before killing to confirm the appropriate position and sealing of the device and to verify the function of the implanted valves. In animals with tricuspid regurgitation, a RV angiography was performed through the RV because the regurgitation could be enhanced or created by the position of the catheter through the implanted valve.
Grafts were electively explanted one hour (group 1) and one month (group 2) after valve implantation. The subvalvular area was examined to determine the relationship between the cordae and the device and the position of the implanted device in relation to the tricuspid valve annulus. After cutting down the interventricular septum, the device was harvested with the RA and RV free wall and rinsed to remove excess intraluminal blood. Valvular competency was grossly tested by passing fluid in the graft. The RA was finally dissected and inspected macroscopically to look for injuries and for the position of the proximal disk.
The mean maximum diameter of the tricuspid annulus was 30 mm, ranging from 27 to 35 mm. The mean RA pressure increased from 5 to 7 mm Hg after valve implantation (range 4 to 8 mm Hg) (Tables 1 and 2).⇓⇓Unsustained ventricular and atrial ectopic beats occurred during wire placement and device deployment in all animals. No sustained or hemodynamically relevant arrhythmias were recorded during the study.
Short-term evaluation (group 1)
In group 1, three of four devices were successfully implanted with good function of valves (Table 2). In one animal, it was impossible to completely deploy the valve despite attempts to dilate the device with a balloon catheter. In this animal, the device was not aligned with the tricuspid annulus. Systemic blood pressure subsequently decreased, and the QRS enlarged with ST-segment depression. Angiographic and echocardiographic evaluations showed a severe paravalvular leak. At autopsy, the ventricular disk was trapped in the tricuspid cordae, explaining its incomplete deployment. In one animal, the dilator glued on the tip of the delivery system embolized in the PA. In another ewe, echographic data acquired during deployment showed that the ventricular disk was incorrectly opened in the RA. Because the device was not fully deployed, it was possible to reload it in the RA by pushing on the Mullins sheath while maintaining the dilator and the wire positions. After reloading, the delivery system was re-advanced in the RV and the device was delivered properly.
One-month evaluation (group 2)
In group 2, all devices were successfully implanted. The mean RA pressure did not significantly change when comparing acute and chronic evaluations (7 vs. 7.3 mm Hg). There was no early or late stent migration (Tables 2 and 3).⇓Evaluations showed that implants were in the desired position and confirmed the sealing of the device, showing no significant leak in three of four animals (Figs. 2 and 3).⇓In one animal, a significant paravalvular leak was found at the one-month evaluation. At autopsy, a pericardial effusion was found and the PTFE was torn beside a weld fracture. Elsewhere, valves were competent and no other stent fractures were found. At autopsy, valve leaflets were thin and mobile in all animals. All devices were sitting in the area of the tricuspid annulus. Devices were partially covered by a fibrous tissue, making devices impossible to retrieve without structural damage (Fig. 4).No macroscopic damage was noted when inspecting the right cavities. As expected, the proximal and distal disks were respectively in the RA and the RV. The tricuspid native valve was completely inactivated by the stent and partially retracted.
No data are presently available on percutaneous valve replacement of atrioventricular valves. Before the availability of mitral homograft valves, semi-lunar valves were used for that indication. For percutaneous implantation the use of such valves is possible, but several difficulties must be resolved. First, the discrepancy between the size of the available transcatheter valve and the size of the annulus makes the use of current stent designs impossible. Second, the valve must be anchored to the annulus not to embolize. Third, in the case of percutaneous reduction of the annulus size, the device must ensure the perfect sealing of the gap between the true and reduced annulus diameters. Therefore, it was necessary to develop a device for this indication that fulfilled the previous criteria. For that purpose, we designed a new self-expandable stent formed of two disks separated by a tubular part. The diameter of the two disks was chosen to be slightly larger than the diameter of the tricuspid annulus to allow for anchoring. Mechanical fixation was ensured by trapping the annulus between the two disks. An 18-mm valve was sutured in the tubular part of the device. Finally, the PTFE covering guaranteed the sealing of the device. The implantation of these newly designed stents was feasible in seven of eight ewes. The implantation of this newly designed device permitted the reduction of the annulus diameter to the desired diameter with no significant increase in RA pressure. This hemodynamic finding did not change in any animals with “late” killing time points. We failed to implant one valved device because it was trapped in the tricuspid cordae. This should be avoided by a more careful echographic assessment before complete release of the device. Although it has not seemed to be necessary to time the delivery of the device in coordination with a specific phase of the cardiac cycle, it could be important to avoid entrapment in the tricuspid cordae. Techniques of rapid ventricular pacing or vagal stimulation probably need to be investigated further. No migration occurred early or during the follow-up. A significant paravalvular leak occurred in one animal. At autopsy at one-month follow-up, the PTFE membrane was torn beside a weld fracture.
Study limitations and unanswered questions
First, in the present study, only animals with a normal tricuspid valve were evaluated. In patients with Ebstein anomaly, the deployment of the device we designed could be hampered because no clear annulus is identifiable between the RA and the RV. The placement of this stent needs a patent annulus separating the two cavities with a larger diameter. Further developments are needed to encompass these difficulties. The major concern is the risk of myocardial trauma secondary to the presence of the stent in regard to cyclic contraction of the heart. The stent is also submitted to high stress during cardiac cycle, which can lead to stent fracture more frequently than in the vascular position. During the overall study, no myocardial complication occurred, but a weld fracture was noticed once at the one-month evaluation. This fracture led to the tear of the PTFE membrane, which further led to a paravalvular leak and progressive right heart failure. Long-term studies and bench tests (different from “standard” stents implanted in the vascular position) will be necessary to address these concerns more accurately.
In conclusion, we report here new insights into atrioventricular valve replacement using a transcatheter technique. Further developments and experimental studies are, however, necessary before considering the use of such device in people. This technique could open new perspectives into transcatheter replacement of the mitral valve.
The authors thank Philippe Marx, who developed this new stent in cooperation with Younes Boudjemline. We also thank Numed and Medtronic Inc. for their technical support.
Supported by The Fondation de l’Avenir, Paris, France.
- Abbreviations and Acronyms
- pulmonary artery
- right atrium/atrial
- right ventricle/ventricular
- Received November 6, 2004.
- Revision received December 21, 2004.
- Accepted January 11, 2005.
- American College of Cardiology Foundation
- Bonhoeffer P.,
- Boudjemline Y.,
- Saliba Z.,
- et al.
- Boudjemline Y.,
- Bonhoeffer P.
- Boudjemline Y.,
- Bonhoeffer P.
- Bonhoeffer P.,
- Boudjemline Y.,
- Qureshi S.A.,
- et al.
- Cribier A.,
- Eltchaninoff H.,
- Bash A.,
- et al.
- St. Goar F.G.,
- Fann J.I.,
- Komtebedde J.,
- et al.
- ↵European Convention for the Protection of Animals Used for Experimental or Scientific Works. Official Journal of the European Community. L222/29-L222/37. August 24, 1999.