The tricuspid valve separates the right atrium from the right ventricle and is composed of the anterior, posterior, and septal leaflets that are attached to the fibrous annulus at different levels. Chordae tendineae connect the leaflets to the three papillary muscles at the right ventricular wall.1,2 Stenosis of the tricuspid valve is rare, whereas tricuspid regurgitation (TR) is diagnosed more often.1,3 Tricuspid regurgitation can be congenital or acquired (secondary) and organic or functional. Most commonly, TR is caused by pulmonary hypertension or enlargement of the right ventricle, usually secondary to other cardiac pathologies such as mitral and aortic valvular disease (left side of the heart) and is therefore referred to as functional regurgitation. Dilation of the annulus1,4,5 from mean values of 10.8 cm and 11.4 cm (annular circumference) in female and male hearts, respectively,2 to more than 22 cm results in incomplete coaptation of the valve leaflets and consequent regurgitation.1,6
Therapy of tricuspid valve regurgitation depends on the severity of the lesion, its repairability, and its form. In 1962, the surgeons Barnard and Schrire were the first to successfully replace the tricuspid valve with a prosthesis in a patient with Ebstein disease (congenital).7 Adopting the methods of mitral valve repair, Kay8 performed annuloplasties of the tricuspid valve in the 1960s, followed by new techniques by Carpentier et al.9 and De Vega.10 All these techniques require open heart valve surgery. The first successful percutaneous heart valve implantation in a human was documented by Bonhoeffer et al.11 in 2000 for the pulmonary position. Since then, a vast amount of experience has been gained in the field of transcatheter valve replacement for the aortic and pulmonary position. However, little experience was achieved for the percutaneous implantation of tricuspid heart valve prostheses. Reported series of percutaneous tricuspid valve replacements were valve-in-valve replacements, in which the prosthesis could use the already implanted prosthesis to provide secure anchoring.12–14 In 2005, Boudjemline et al.15 developed a self-expandable stent made from Nitinol wire for minimal-invasive replacement of the tricuspid valve, which reduces the diameter of the annulus providing the opportunity to apply commercially available heart valves. Bai et al.16 developed a similar design. In 2010, Lauten et al.17 presented a new approach for a percutaneous tricuspid valve therapy consisting of two valved stents that are implanted into the inferior and superior vena cava (SVC) to substitute tricuspid valve function. Recently, Iino et al.18 made a new approach for a self-expandable stent (diameter of 27–29 mm) using a superabsorbent polymer that is deployed in a doughnut-shaped pouch around the atrial region of the stent and expands when in contact with blood to fill the gap between the valved stent and the annulus.
It is estimated that, in North America alone, approximately 500,000 people have clinically significant heart failure caused by functional mitral regurgitation. For a majority of these patients, there would be an associated need for percutaneous tricuspid annuloplasty, repair, or replacement after mitral valve surgery. Unlike in the aortic or pulmonary position, however, there are no convenient adjacent defined structures for secure device anchoring.4,19 To overcome the abovementioned disadvantages, the anchoring forces could be transferred from the native annulus to both the SVC and the tricuspid valve annulus (TVA) by a new design in which anchoring forces and elastic longitudinal applied forces together provide the necessary stability to the annulus within the valve prosthesis. This may represent an effective approach to the orthotopic therapy of tricuspid pathology.
This article is organized as follows: First, the approach for a new anatomically customized stent design is presented. Second, the generation of the anatomical three-dimensional (3D) model is described. Then, the design process and the results of the finite element analysis and their validations are depicted. The results of first in vitro tests are presented. Finally, these findings are discussed and conclusions are drawn.
The new approach for TR therapy consists of two tubular anchoring components, i.e., one inside the tricuspid annulus and the other inside the SVC, which are connected by elastic elements to provide stability to hold the device in position while allowing some motion of the heart. Figure 1 shows a schematic sketch of the novel approach inside a computer-generated model of the right heart. The SVC stent constitutes the main anchorage of the complete device. The TVA stent is primarily held in position by five elastic connector struts that connect both stents. The stent positioned inside the tricuspid annulus is meant to contain a tricuspid heart valve. The complete device is held in position by means of radial forces exerted by both stents to the SVC, the tricuspid annulus, and by different-sized connector struts that distribute longitudinally applied forces throughout the entire structure. The placement of the connector struts needs to leave space to ensure the inflow through the inferior vena cava into the right atrium.
Anatomical 3D Model
The new stent design was developed using anatomical data. The geometry of the right ventricle and right TVA, including the inferior and SVC, was obtained by a computer tomography (CT) scan of a porcine heart (65 kg, female). Using the software Mimics and 3-Matic (Materialise, Leuven, Belgium), the data set was postprocessed (digital geometry reconstruction, segmentation, surface generation, and surface smoothing) and transformed into a 3D model of the inner surface of the abovenamed regions of the right heart (Figure 2). Then, relevant distances and diameters for the development of the new stent design for the tricuspid valve were identified and used.
The focus of the design process is on both stents that are subsequently described in detail. The corresponding stent design, i.e., for the SVC stent and the TVA stent was created with their specific function in mind. Both stents were designed as 3D models with an outer diameter of 10 mm due to the manufacturing process. On the basis of the anatomical data of the previous CT scan, the SVC stent requires an oversized outer diameter of 18 mm to provide sufficient anchoring and the TVA stent one of 40 mm in its expanded state. The SVC stent needs to provide the main fixation for the framework, i.e., a sufficient holding force needs to be applied to the SVC and was therefore designed with a closer mesh. A 3D model was created with repeating unit geometry of simple diamonds (Figure 3A). The adjusted length of the stent due to the expansion process was initially estimated by the Pythagorean theorem and then identified by means of simulation and compared. The length is approximately 36 mm in the expanded state. This method of predicting the changes of geometry worked very well for simple diamond structures and no further iteration was implemented. Meanwhile, the design for the TVA stent consists of a more open structure with five anchoring points (struts bended outward) that provide a fixation inside the annulus. The anchoring points separate the stent into the atrial and the ventricular region. Due to the anatomical geometry, this stent is shorter with an overall length of 24 mm with the ventricular region not exceeding 16 mm to prevent too much extension into the ventricle. The development of this stent was an iterative process of designing a 3D model for a tube diameter of 10 mm and of analyzing the expansion to a diameter of 40 mm with regard to the geometrical displacement of the struts and to stresses after running the finite element analysis. Changes to the initial geometry, such as different radii or lengths of struts, had a direct impact on the development of stresses inside the material and on the resulting geometry. Iteration steps are shown in Figure 3 (B1) and (B2). The initially designed stent geometry (B1) seen in Figure 3 exceeds the required length after expansion and the atrial-ventricular ratio is incorrect. In addition, small radii provoke high stresses inside the material. Iteration B2 possesses correct dimensions for the ventricular region, but the atrial region is still longer than intended. The final geometry (B3) is shown in Figure 3.
The connector struts have elastic elements to account for the motion of the heart after implantation. In addition, the elasticity is required for the crimping process of the complete framework to load the curved stent construct onto a straight catheter for minimal-invasive implantation. The design of the two longest connector struts causes difficulties (high resulting stresses) because they need to be crimped from an expanded length of 50 mm to the length of the shortest connector strut (20 mm length). The idea for resolving the problem is to additionally expand the shortest connector strut during crimping, which results in less crimping distance and lower stresses for the longest connector struts.
Finite Element Analysis
Nitinol was used to manufacture the self-expanding stents and connector struts. Therefore, the material model of superelastic Nitinol incorporated in the simulation software was used to perform the simulations in Abaqus/CAE (Simulia, Providence, RI) with an austenite and martensite elasticity of 40 GPa and 32 GPa, respectively, an austenite and martensite Poisson’s ratio of 0.33, and a process temperature of estimated 22°C. To investigate the expansion of the stents and the connectors, they were modeled with 3D eight-node brick reduced integration elements (C3D8R). To keep the analysis run-time reasonable, a symmetrical part of the TVA and SVC stent, i.e., 36° and 72° was used for the simulation. The mesh size was chosen with regard to the smallest geometrical part of the stent to cover each side along the stent-wall thickness direction with mostly five to seven bricks. The mesh was generated with variably sized elements, i.e., radii were meshed with finer elements than straight struts.
Both stents were simulated separately using a finite element model, consisting of a cylinder as forming tool, and two rigid walls in between the stents were positioned. Using a cylindrical coordinate system, the stents are angularly fixed on both cut sides and expanded to their final outer diameter of 18 mm or 40 mm by the outward motion of the cylinder. The stent design analysis was performed in Abaqus/Explicit using mass scaling by means of the target time increment for accelerated calculation, ensuring that the order of mass scaling and stable time did not differ by more than factor 100 to obtain reasonable results.20 The energy output was evaluated for quasistatic conditions and physically adequate results, i.e., the kinetic energy should not exceed a small fraction, typically 5% to 10%, of the internal energy throughout the majority of the simulation process.21
Simulation Results and Validation of the Anatomically Adapted Design
Figures 4 and 5 (left side) show the geometries of the expanded manufactured SVC and TVA stents, respectively. At the right side, an overlay of the geometry of the manufactured stent and of the geometry calculated by the finite element analysis is shown. The simulation results fit the real stent geometry very well in both cases. Small deviations are identified at the diamond structures in which the simulation results show straighter struts for the SVC stent. The deviations at the rear of the stent are due to optical distortions. The struts of the biggest diamond structures near the eyelets seem to be most deviated. When comparing parts of the stent with the simulation results, it is apparent that the stent was not expanded uniformly, which resulted in minor deviations.
The new stent design was connected and mounted inside a silicone model of the extracted anatomical geometry of the CT data and into a porcine heart to verify the anatomical fit, see Figure 6 left and right, respectively. The stent construct was crimped by hand to introduce the device into both objects. The SVC stent and the TVA stent align and adapt to the SVC and the TVA of the silicone model as well as to the structure of the porcine heart. It seems that inside the silicone model the anchoring is mostly provided by the SVC stent as it was planned (Figure 6, left). Attempts at pulling the device out of the model were without success. The introduction of the stent device into the porcine heart confirmed the general anatomical fit to the unstructured wall geometry of the TVA. The TVA stent adapts to the shape of the annulus of the native tricuspid valve, which is pushed against the annulus wall. The fixation of the SVC stent could not be demonstrated inside the porcine model due to a missing SVC but was verified inside the silicone model.
Initial in vitro tests were performed to examine the function of the new concept. A customized trileaflet heart valve was sewn into the TVA stent using bovine and porcine pericardium for the leaflets and the sealing skirt, respectively, to prevent paravalvular leakage. A mock loop simulating physiological parameters of the right heart was mounted and the stent construct tested inside the anatomically adapted silicone model. A pressure difference of approximately 5/45 mm Hg was applied for pulsatile flow with 70 bpm. High-speed video recordings showed that the stent construct adapts to the surrounding structure in motion at each point of the cardiac cycle. No migration was visible over the duration of the testing (several hours to half a day) and both stents stayed at their initial position. The configuration of the struts leaves enough space to ensure the inflow through the inferior vena cava; moreover, they are placed to both sides of the vessel at the atrial wall, see Figure 7 left and right for top and side view, respectively.
Discussion and Conclusion
A new concept for designing stents for the tricuspid valve was developed and validated. It was shown that the design is anatomically adapted inside the right heart, including the SVC, right atrium, and TVA. The primary anchoring to the right heart is provided by the SVC stent inside the SVC and supported by the TVA stent inside the TVA.
Comparing this approach with other previously developed percutaneous concepts for the tricuspid position showed that the majority of devices had the disadvantage of using the unstructured tricuspid annulus as the sole anchoring point.15,16,18 In addition, the design by Iino et al.18 might provoke embolization and toxic reactions due to the material used. Lauten et al.17 do not use the annulus for anchoring but introduce two heart valve prostheses into the superior and inferior vena cava. This might lead to obstruction of the hepatic veins and to a ventricularization of the right atrium.
The approach presented here displaces the anchoring toward the SVC to prevent the apparent disadvantages of concepts by other research groups. Initial in vitro tests showed the feasibility of the approach, i.e., no migration occurred and the stents and struts adapted to the surrounding geometry. The connector struts exerted not only enough strength to ensure the stability of the device, but also enough flexibility to allow for the pulsatile motion of the heart. Concern arose that the struts inside the right atrium may lead to arrhythmias or thrombosis, but previously performed animal trials with a similar concept showed no lesion or injury of the surrounding tissue and of the electrical activity of the heart, especially the sinoatrial node or atrioventricular node. In addition, thrombosis was not seen and could be prevented by covering the struts and stents with medication or hemocompatible coating. Further knowledge in this field needs to be gained in planned animal trials with the presented design.
The healthy heart is an organ always in motion, possessing a mean rotation in apical circumferential direction of 10.9 ± 4.8 degrees counterclockwise for the left ventricle and 1.1 ± 5.8 degrees clockwise for the right ventricle at aortic valve closure. Moreover, the right ventricle differs from the left ventricle and possesses almost no rotation, rather a “tightening belt” motion.22 In patients with TR secondary to left heart diseases, the function of the right heart is reduced and minimal motion is left. A study by Ring et al.23 found that the tricuspid annulus of patients with dilated right hearts is nearly circular and the change in mean tricuspid annular area does not reach statistical significance, i.e., 1598 mm2 diastolic vs. 1534 mm2 systolic, which results in a change of tricuspid radius of 0.45 mm. It can be concluded that there is not much twist and motion of the right heart left, which could lead to fatigue of the Nitinol stents and struts. Nevertheless, durability tests and animal trials will be performed to gain knowledge about failure conditions of the stent construct.
The developed customized heart valve consisting of three leaflets was sewn symmetrically into the TVA stent. The valve opened and closed completely but needs further development of leaflet design due to the large diameter of 40 mm, to optimize fluid dynamical properties and reach a small crimping diameter. The connection between stents and struts needs to be optimized because the nuts and bolts used here are likely to loosen, which may result in failure of the whole concept. Moreover, a customized crimping technique will be developed to load the arched device onto a straight catheter. Further in vitro and in vivo tests will be performed to study the stent design and its fixation including holding forces for further optimization.
1. Shemin RJCohn LH. Tricuspid valve disease. Cardiac Surgery in the Adult. 2008 New York McGraw-Hill:pp. 1111–1128
2. Silver MD, Lam JH, Ranganathan N, Wigle ED. Morphology of the human tricuspid valve. Circulation. 1971;58:333–348
3. Roberts WC, Sullivan MF. Combined mitral valve stenosis and tricuspid valve stenosis: Morphologic observations after mitral and tricuspid valve replacements or mitral replacement and tricuspid valve commissurotomy. Am J Cardiol. 1986;58:850–852
4. Anyanwu AC, Chikwe J, Adams DH. Tricuspid valve repair for treatment and prevention of secondary tricuspid regurgitation in patients undergoing mitral valve surgery. Curr Cardiol Rep. 2008;10:110–117
5. Marui A, Mochizuki T, Mitsui N, Koyama T, Horibe M. Isolated tricuspid regurgitation caused by a dilated tricuspid annulus. Ann Thorac Surg. 1998;66:560–562
6. Dreyfus GD, Corbi PJ, Chan KM, Bahrami T. Secondary tricuspid regurgitation or dilatation: Which should be the criteria for surgical repair? Ann Thorac Surg. 2005;79:127–132
7. Charles RG, Barnard CN, Beck W. Tricuspid valve replacement for Ebstein’s anomaly. A 19 year review of the first case. Br Heart J. 1981;46:578–580
8. Kay JH. Surgical treatment of tricuspid regurgitation. Ann Thorac Surg. 1992;53:1132–1133
9. Carpentier A, Deloche A, Dauptain J, et al. A new reconstructive operation for correction of mitral and tricuspid insufficiency. J Thorac Cardiovasc Surg. 1971;61:1–13
10. De Vega NG. [Selective, adjustable and permanent annuloplasty. An original technic for the treatment of tricuspid insufficiency]. Rev Esp Cardiol. 1972;25:555–556
11. Bonhoeffer P, Boudjemline Y, Saliba Z, et al. Percutaneous replacement of pulmonary valve in a right-ventricle to pulmonary-artery prosthetic conduit with valve dysfunction. Lancet. 2000;356:1403–1405
12. Roberts P, Spina R, Vallely M, Wilson M, Bailey B, Celermajer DS. Percutaneous tricuspid valve replacement for a stenosed bioprosthesis. Circ Cardiovasc Interv. 2010;3:e14–e15
13. Roberts PA, Boudjemline Y, Cheatham JP, et al. Percutaneous tricuspid valve replacement in congenital and acquired heart disease. J Am Coll Cardiol. 2011;58:117–122
14. Van Garsse LA, Ter Bekke RM, van Ommen VG. Percutaneous transcatheter valve-in-valve implantation in stenosed tricuspid valve bioprosthesis. Circulation. 2011;123:e219–e221
15. Boudjemline Y, Agnoletti G, Bonnet D, et al. Steps toward the percutaneous replacement of atrioventricular valves an experimental study. J Am Coll Cardiol. 2005;46:360–365
16. Bai Y, Zong GJ, Wang HR, et al. An integrated pericardial valved stent special for percutaneous tricuspid implantation: An animal feasibility study. J Surg Res. 2010;160:215–221
17. Lauten A, Figulla HR, Willich C, et al. Percutaneous caval stent valve implantation: Investigation of an interventional approach for treatment of tricuspid regurgitation. Eur Heart J. 2010;31:1274–1281
18. Iino K, Lozonschi L, Metzner A, et al. Tricuspid valved stent implantation: Novel stent with a self-expandable super-absorbent polymer. Eur J Cardiothorac Surg. 2011;40:503–507
19. Rogers JH, Bolling SF. The tricuspid valve: Current perspective and evolving management of tricuspid regurgitation. Circulation. 2009;119:2718–2725
20. Kainz A, Krimpelstätter K, Zeman K FE-simulation of thin strip and temper rolling processesin Abaqus Austria Users’ Conference. Vienna November 24–25, 2003
21. Pawtucket RIABAQUS, Inc.Abaqus: Getting Started with Abaqus, V6.4. 2003 Hibbitt Karlsson & Sorenson, Inc.
22. Gustafsson U, Lindqvist P, Waldenström A. Apical circumferential motion of the right and the left ventricles in healthy subjects described with speckle tracking. J Am Soc Echocardiogr. 2008;21:1326–1330
23. Ring L, Rana BS, Kydd A, Boyd J, Parker K, Rusk RA. Dynamics of the tricuspid valve annulus in normal and dilated right hearts: A three-dimensional transoesophageal echocardiography study. Eur Heart J Cardiovasc Imaging. 2012;13:756–762