Abstract
In this study, results of a functional in vitro study of 2 newly developed valved stents for transcatheter mitral valve implantation are presented.
Two novel stent designs, an oval-shaped and a D-shaped stent with a strut fixation system were developed. The fixation force of the novel stents were tested in vitro in porcine hearts with a tensile test set-up. In further experiments, the stents were equipped with a circular valved stent, and the valve performances were investigated in a pulsatile heart valve tester.
Sufficient mean stent fixation forces in the range of 24.2 ± 0.9 N to 28.6 ± 1.9 N were measured for the different stent models. The novel valved stents showed good performance in an in vitro pulsatile heart valve tester. A sufficient opening area and low opening pressures were measured for all tested mitral valved stents. Compared with an established reference valve, the D-shaped stent and the oval-shaped valved stent showed a lower systolic transvalvular pressure gradient, which indicates slightly greater extent of valvular leakage of the closed valved stents. However, the mitral nitinol valved stents demonstrated adequate durability.
This study indicates a sufficient annular fixation force of the tested transcatheter mitral valve implantation valved stent prototypes. Therefore, these mitral valved stents demonstrate a new type of mitral valved stent design.
INTRODUCTION
Mitral regurgitation is one of the most common heart valve diseases [1]. For severe functional mitral regurgitation, replacement of the valve is the recommended treatment. Surgical replacement is still the gold standard, but according to the Euro Heart Survey, about 49% of the patients are considered to be high-risk patients and are denied open-heart surgical intervention [2]. Transcatheter aortic valve implantation has become increasingly popular in the past 10 years, achieving results comparable to those of surgical alternatives. Transcatheter mitral valve implantation (TMVI) is a promising alternative to surgical mitral valve implantation and represents a treatment option for a large cohort of patients that cannot be referred for open-heart surgery. Nevertheless, TMVI is challenging due to the complex anatomy of the mitral apparatus, the high ventricular pressures and relative movement of the mitral annulus.
Several TMVI devices are in preclinical or in early clinical evaluation [3]. The early clinical results are promising to a certain extend. During the development process, systolic fixation of the TMVI device is a critical design challenge, and different TMVI prototypes use different techniques that include struts, chords and clamps. In previous studies, our group achieved very good results with an apical fixation system [4]. The developed Lutter valve was a precursor model of the successful Tendyne valve [5].
In this study, the in vitro evaluation of 2 novel valved stents with an anatomical mitral shape for TMVI is presented. The first design is an oval-shaped stent, and the second design is a D-shaped stent. Both mitral stents were fabricated with angular fixation struts that protrude into the mitral annulus to achieve secure systolic and diastolic fixation. To determine the optimal design parameters for these novel valved stents, mechanical and functional in vitro testing were conducted. The stent fixation forces were measured in specifically designed tensile tests. In a further experiment, the novel stents were additionally equipped with circular valved stents with a valve made from porcine pericardium. The performance and position stability of the mitral valved stents were investigated in a pulsatile heart valve tester.
METHODS
Stent design
Two different anatomically adapted models of a self-expanding nitinol stent for an atrioventricular valve replacement were developed (Fig. 1). The stents were fabricated from a nitinol tube with a wall thickness of 0.46 mm (R.T.M. Rainer Trapp Medizintechnik GmbH, Graben-Neudorf, Germany). The valve size was determined using the transoesophageal echocardiographic data of our porcine hearts from previous animal studies. The stent sizes were adapted to the mitral annulus size during diastole or filling phase of the left ventricle. Therefore, the stent size was adapted to the maximal annular size obtained during the heart cycle, because the annular size is slightly larger during diastole. The first design had a slightly oval shape with a maximum axis length of 32 mm. The second design had a D-shape of the annular stent body with a maximum diameter of 36 mm. Both stent models were fabricated with 20 annular struts to achieve secure systolic annular fixation and to prevent migration of the valve into the left atrium. The angular struts of the stents protrude at 45° (D-shape and oval) and 60° (oval stent) to the stent. Furthermore, a D-shaped stent and an oval stent with no struts to the aortic side of the stent (15 struts in total) were fabricated (Fig. 2). The height of the stent on the ventricular side is 11 mm. This is the minimal height to enable fixation of a mitral valve inside. Minimization of the annular stent body height is a critical design criteria to maximally reduce the risk of an left ventricular outflow tract obstruction. An assymetrical collar with 10 mm protrusions at 115° on the posterior side and 15 mm protrusions at 140° on the aortic side allows a proper fit to the mitral annulus and prevents migration of the stent towards the ventricular side during diastole. The increased angle of 140° on the aortic side improves the optimal stent alignment and decreases the impact on the aorta (ideally no pressure towards the aortic wall in A2) and the risk of stent fractures.
(A and B) D-shaped stent design with fixation struts (red circle). (C and D) Oval-shaped mitral stent model with struts (red circle).
(A) D-shaped stent without struts at the aortic side. The angle between the flange and the ventricular part of the stent is α = 140° on the aortic side and β = 115° on the posterior side. The height of the ventricular part is 11 mm. (B) Oval-shaped stent without struts on the aortic side. The 4 cords for fixation testing were mounted at the centre of the ventricular part.
Fixation testing
Porcine hearts from the butcher were prepared for stent deployment and fixation tests. The stents were transapically deployed into the mitral valve position of a porcine heart with a 14-Fr size applicator catheter. Correct positioning after deployment of the stent was ensured by visual inspection. Additionally, the stents were equipped with 4 subannular cords to measure the fixation force with a tensile testing apparatus (Zwicki, Zwick GmbH & Co Kg, Ulm, Germany) (Fig. 3). This is exactly the part where the valve of the stent is fixed. This is the closest possible simulation of the forces of the valve acting on the carrier stent, which we could achieve with our experimental set-up. Then, the porcine heart was placed in the specifically designed fixation box with apical access to the heart to simulate systolic loading of the closed mitral valve. The force applied to the stent was slowly increased. Because the ventricular pressures can be greater than 200 mmHg, we chose 30 N, which is the force on the valve at approximately 300 mmHg, as a cut-off for the maximal force. This is to ensure the entire fixation for a short-term period with very high pressures.
Test set-up to enable tensile test to analyse the fixation force of the novel transcatheter mitral valve implantation prototypes. The heart is positioned in a specifically designed fixation box with apical access to simulate systolic loading.
Pulsatile testing
In further tests, the stents were equipped with an additional circular valved stent. Thereto the nitinol stent body was covered with a polytetrafluoroethylene membrane (Zeus Inc., Orangenburg, SC, USA), and a commercially available bioprosthetic heart valve was sutured into the annular stent body. The function of the valved stent was measured in a hydrodynamic heart valve tester (Fig. 4) [6]. The mitral annulus was modelled from silicon to simulate the mitral valve annulus anatomy. The silicon annulus was adapted to each form of the 3 tested specimens (D-shape, oval and circular) to minimize the effect of transvalvular regurgitation during testing. The valves were tested in this adapted silicon sample holder to evaluate their function, valve fixation and position stability under pulsatile loading. Systolic and diastolic mean pressure differences were evaluated and compared with an established reference valve (CoreValve) to determine the efficiency, sufficiency and opening and closing properties of the valves and to estimate the paravalvular regurgitation.
Image of the test set-up for the hydrodynamic valved stent testing (left) and transcatheter mitral valve implantation prototype within the specifically designed silicon model to enable functional testing of the valved stent in the heart valve tester (right).
Maximal opening area of the tested valves in pulsatile experiments at a heart rate of 70 beats/min (top) and tested valves in closed state (bottom). Oval mitral valved stent (left), D-shaped stent (middle) and CoreValve reference aortic valve (right).
RESULTS
Transapical deployment of the stent prototype was executed in vitro in 5 porcine hearts for each stent model. The deployment of the stent into the correct position in the annular plane was achieved in 14 of the 15 tests. Eight of the 15 stents showed a minor uncritical rotation (<10°). Without the aortic struts, 8 of the 10 newly designed stents were implanted at the optimal annular position. A small uncritical rotation (<10°) was observed in 9 of the 10 cases. Despite a minor and non-critical lateral rotational displacement of 5°–10° in 76% of the cases, correct annular positioning and complete opening of the self-expanding stents were observed in nearly all the tests (Table 1).
Stent positioning and injuries caused by the stent identified upon optical inspection of the heart after fixation testing
| D-shape | D-shaped no aortic struts | Oval (45°) | Oval (45°) no aortic struts | Oval (60°) | |
|---|---|---|---|---|---|
| No migration at required fixation forces | 3 | 2 | 5 | 4 | 5 |
| Stent in required annular position | 5 | 3 | 5 | 5 | 5 |
| Stent in required rotational position | 1 | 0 | 2 | 1 | 2 |
| Injuries due to struts | 5 | 5 | 5 | 5 | 5 |
| Aortic interference | 5 | 0 | 5 | 0 | 5 |
| D-shape | D-shaped no aortic struts | Oval (45°) | Oval (45°) no aortic struts | Oval (60°) | |
|---|---|---|---|---|---|
| No migration at required fixation forces | 3 | 2 | 5 | 4 | 5 |
| Stent in required annular position | 5 | 3 | 5 | 5 | 5 |
| Stent in required rotational position | 1 | 0 | 2 | 1 | 2 |
| Injuries due to struts | 5 | 5 | 5 | 5 | 5 |
| Aortic interference | 5 | 0 | 5 | 0 | 5 |
Each stent criterion was tested in 5 hearts (n = 5) for each stent design. The desired annular position was achieved with nearly all stents. A slight rotational displacement was observed in most of the tested stents, but the rotation was less than 10° and uncritical in all tests. As expected, the struts penetrated into the tissue of the mitral annulus causing minor injuries. An interference with the aorta was observed for all stents with struts on the aortic side. Stents without struts at the aortic side did not interfere with the aorta.
Stent positioning and injuries caused by the stent identified upon optical inspection of the heart after fixation testing
| D-shape | D-shaped no aortic struts | Oval (45°) | Oval (45°) no aortic struts | Oval (60°) | |
|---|---|---|---|---|---|
| No migration at required fixation forces | 3 | 2 | 5 | 4 | 5 |
| Stent in required annular position | 5 | 3 | 5 | 5 | 5 |
| Stent in required rotational position | 1 | 0 | 2 | 1 | 2 |
| Injuries due to struts | 5 | 5 | 5 | 5 | 5 |
| Aortic interference | 5 | 0 | 5 | 0 | 5 |
| D-shape | D-shaped no aortic struts | Oval (45°) | Oval (45°) no aortic struts | Oval (60°) | |
|---|---|---|---|---|---|
| No migration at required fixation forces | 3 | 2 | 5 | 4 | 5 |
| Stent in required annular position | 5 | 3 | 5 | 5 | 5 |
| Stent in required rotational position | 1 | 0 | 2 | 1 | 2 |
| Injuries due to struts | 5 | 5 | 5 | 5 | 5 |
| Aortic interference | 5 | 0 | 5 | 0 | 5 |
Each stent criterion was tested in 5 hearts (n = 5) for each stent design. The desired annular position was achieved with nearly all stents. A slight rotational displacement was observed in most of the tested stents, but the rotation was less than 10° and uncritical in all tests. As expected, the struts penetrated into the tissue of the mitral annulus causing minor injuries. An interference with the aorta was observed for all stents with struts on the aortic side. Stents without struts at the aortic side did not interfere with the aorta.
During systole, the ventricle contracts and pumps the blood through the aortic valve into the aorta. This generates high and maximum ventricular pressures of up to 200 mmHg, compared to the atrial pressures of 0–15 mmHg, meaning that the fixation must withstand a pressure difference of approximately 200 mmHg. The shape-dependent area of the valves is determined to be 7.3 cm2 and 8.4 cm2 for the oval-shaped and the D-shaped stents, respectively. This results in a maximum systolic force from the high ventricular blood pressure of 19.4 N for the oval-shaped stent and 22.4 N for the D-shaped stent.
The oval stent with 45° angle struts had a maximum fixation force of 29.4 ± 0.9 N (Table 2). In all experiments, the stent withstood forces greater than the required 19.4 N (at 200 mmHg). The oval stent with 60° struts showed a mean fixation force of 28.6 ± 1.9 N. In all experiments, the stent withstood forces greater than the required 19.4 N (at 200 mmHg).
Maximal Ffix under tensile loading and Δpsys, EOA and a Δpdias determined during the hydrodynamic pulsatile tests at a heart rate of 70 beats/min and a stroke volume of 75 ml
| Ffix (N) | Δpdias (mmHg) | EOA (cm2) | Δpsys (mmHg) | |
|---|---|---|---|---|
| CoreValve | 5.8 ± 0.1 | 1.98 | 50.3 ± 0.6 | |
| D-shape | 24.2 ± 6 | 8.8 ± 0.2 | 1.32 | 26.4 ± 0.5 |
| D-shape (no aortic struts) | 22.6 ± 5.2 | |||
| Oval (45°) | 29.4 ± 0.9 | 10.9 ± 0.5 | 1.05 | 38.4 ± 0.1 |
| Oval (no aortic struts) | 25 ± 5.2 | |||
| Oval (60°) | 28.6 ± 1.9 |
| Ffix (N) | Δpdias (mmHg) | EOA (cm2) | Δpsys (mmHg) | |
|---|---|---|---|---|
| CoreValve | 5.8 ± 0.1 | 1.98 | 50.3 ± 0.6 | |
| D-shape | 24.2 ± 6 | 8.8 ± 0.2 | 1.32 | 26.4 ± 0.5 |
| D-shape (no aortic struts) | 22.6 ± 5.2 | |||
| Oval (45°) | 29.4 ± 0.9 | 10.9 ± 0.5 | 1.05 | 38.4 ± 0.1 |
| Oval (no aortic struts) | 25 ± 5.2 | |||
| Oval (60°) | 28.6 ± 1.9 |
The fixation force was measured in 5 porcine hearts (n = 5) for each stent model. For the pulsatile tests, 6 consecutive simulated heart cycles were evaluated to determine the opening area and the diastolic and systolic pressure differences for each valve.
EOA: effective orifice area; Ffix: fixation force; Δpdias: mean diastolic transvalvular pressure difference; Δpsys: mean systolic transvalvular pressure difference.
Maximal Ffix under tensile loading and Δpsys, EOA and a Δpdias determined during the hydrodynamic pulsatile tests at a heart rate of 70 beats/min and a stroke volume of 75 ml
| Ffix (N) | Δpdias (mmHg) | EOA (cm2) | Δpsys (mmHg) | |
|---|---|---|---|---|
| CoreValve | 5.8 ± 0.1 | 1.98 | 50.3 ± 0.6 | |
| D-shape | 24.2 ± 6 | 8.8 ± 0.2 | 1.32 | 26.4 ± 0.5 |
| D-shape (no aortic struts) | 22.6 ± 5.2 | |||
| Oval (45°) | 29.4 ± 0.9 | 10.9 ± 0.5 | 1.05 | 38.4 ± 0.1 |
| Oval (no aortic struts) | 25 ± 5.2 | |||
| Oval (60°) | 28.6 ± 1.9 |
| Ffix (N) | Δpdias (mmHg) | EOA (cm2) | Δpsys (mmHg) | |
|---|---|---|---|---|
| CoreValve | 5.8 ± 0.1 | 1.98 | 50.3 ± 0.6 | |
| D-shape | 24.2 ± 6 | 8.8 ± 0.2 | 1.32 | 26.4 ± 0.5 |
| D-shape (no aortic struts) | 22.6 ± 5.2 | |||
| Oval (45°) | 29.4 ± 0.9 | 10.9 ± 0.5 | 1.05 | 38.4 ± 0.1 |
| Oval (no aortic struts) | 25 ± 5.2 | |||
| Oval (60°) | 28.6 ± 1.9 |
The fixation force was measured in 5 porcine hearts (n = 5) for each stent model. For the pulsatile tests, 6 consecutive simulated heart cycles were evaluated to determine the opening area and the diastolic and systolic pressure differences for each valve.
EOA: effective orifice area; Ffix: fixation force; Δpdias: mean diastolic transvalvular pressure difference; Δpsys: mean systolic transvalvular pressure difference.
The D-shaped stent had a mean fixation force of 24.2 ± 6 N. In 3 experiments, a maximum fixation force greater than the required 22.4 N (transvalvular pressure difference of 200 mmHg) were achieved. In the other 2 measurements, the fixation forces were 16 N and 21 N.
The oval stent model without struts on the aortic side showed a fixation force of 25 ± 5.2 N. In 4 experiments, the fixation force was higher than the required 19.4 N. In 1 test, the maximum fixation force was 18 N. The fixation force of the D-shaped stent model with no struts on the aortic side was determined to be 22.6 ± 6.9 N. In 2 tests, the test limit of 30 N was achieved, and in 3 tests, the maximum fixation force was lower than the desired 22.4 N (15 N, 19 N and 19 N).
An optical inspection was performed after the fixation testing to determine the damaging effect of the struts. As expected, the struts penetrated into the tissue of the mitral annulus causing minor injuries. These injuries were small perforations and cuts that mainly affected the surface of the tissue. An interference with the aorta was observed for all stents with struts on the aortic side. Stents models without struts at the aortic side did not interfere with the aorta (Table 1).
For further investigation, the oval stent with 45° struts and the D-shaped 45° struts were equipped with circular valved stents with leaflets from porcine pericardium to test their performance under hydrodynamic loading. During hydrodynamic simulated diastole at a heart rate of 70 beats/min and a stroke volume of 75 ml, the oval valved stent exhibited a transvalvular pressure difference of 10.9 ± 0.5 mmHg together with a maximal effective orifice area of 1.05 cm2 (Table 2, Fig. 5). The D-shaped valved stent showed a lower diastolic pressure difference of 8.8 ± 0.2 mmHg and a larger orifice area of 1.32 cm2. The lowest diastolic pressure and hence the highest opening area for the CoreValve Evolut R 29 mm reference valve was 5.8 ± 0.1 mmHg and the effective orifice area was 1.98 cm2. The determined mean systolic transvalvular pressure difference during the closing phase of the valve was 26.4 ± 0.5 mmHg for the D-shaped valve and 38.4 ± 0.1 mmHg for the oval-shaped valve. The CoreValve reference valve showed the highest systolic pressure difference of 50.3 ± 0.6 mmHg, which indicates a superior leak tightness of the CoreValve Evolut R compared with the novel mitral valves.
In addition, a change in the annular size of the valved stent during the heart cycle was observed. The opening and closing sequences of the valves demonstrate a slight change in size of the A-P direction: The oval stent exhibits a diastolic length of 27.3 ± 0.3 mm and a systolic length of 26.8 ± 0.4 mm. The D-shaped stent showed a diastolic length of 27.5 ± 0.3 mm and a systolic length of 26.5 ± 0.3 mm. No change of dimensions could be measured for the CoreValve reference valve. The change of lateral dimension was less than 0.3 mm; therefore, no major change in the lateral dimension was observed. Furthermore, no areas of central regurgitation of the valve were detected. In contrast, very small areas of paravalvular leakage occurred between the outer margin of the valved stent and the added outer specifically formed (oval-shaped and D-shaped) stent.
DISCUSSION
The development and testing of mitral valved stents for transcatheter implantation into the beating heart is the focus of recent research. Different groups worldwide have developed several prototypes currently undergoing preclinical and early clinical investigation at different stages [7–11]. The need for such a device is high, as almost half of all patients with severe symptomatic mitral regurgitation cannot be referred for the gold standard of open-heart surgery [2]. During the design process, the development team particularly focuses on fixation of the TMVI device during systole. During systole, the high pressure in the left ventricle acts on the closed mitral valve and a secure fixation within the complex anatomy of the mitral apparatus must be achieved. A migration into the left atrium must be prevented. Because of the different interests, access to experimental data of the various prostheses is very limited, which slows down the design progress as well as the comparability of the devices.
In the context of this study, 2 novel TMVI devices with anatomical mitral designs were developed and mechanical and functional evaluation was conducted. In contrast to previously developed transcatheter mitral valve devices, the newly developed stents combine an oval-shaped or D-shaped design specifically adapted to the anatomy of the mitral annulus with annular fixation struts, which protrude into the mitral annulus to achieve a secure systolic fixation. Although one of the previous valved stents had a D-shape (Tendyne), these newly developed stent prototypes in this study were mainly designed without neocords (tethers) for secure fixation of the valved stent. To define the optimal design parameters for these novel stents, mechanical and functional in vitro testing was conducted. The stent fixation forces were measured in specifically designed tensile tests, which proved to be well suited for the specific setting of these tests. In a further experiment, the novel stents were equipped with a circular valved stent including a porcine pericardium valve. The performance and position stability of the valve were investigated in a pulsatile heart valve tester.
All stents could be deployed and positioned in vitro in porcine hearts. A challenge was the positioning of the stent without interference of the valve from the anatomically natural anteromedial leaflet, the posterolateral leaflet and, especially, the chordae tendineae. However, all valved stents were deployed without destruction of the natural mitral apparatus. As expected, minor injuries from strut penetration into the mitral annulus were observed in all tested hearts after testing the fixation force and subsequent removal of the stents. Furthermore, because of the small ventricular height of the stent models (11 mm), no obstruction of the left ventricular outflow tract was observed. Only the stents with struts on the aortic side compressed the aorta and in rare cases even penetrated the aorta. This problem did not occur with the stents without fixation struts at the aortic side.
The oval stents had the highest fixation forces. There were no significant differences observed between 45° and 60° struts. Regarding injuries to the aorta or the coronary sinus, 45° struts might be favourable due to smaller lateral penetration depth. The D-shaped design showed slightly lower but still sufficient fixation forces in 3 of the 5 cases. The stents were slightly oversized (small hearts from the butcher), which led to minor deformation of the D-shaped stents before fixation testing. Because of this, the 2 valved stents slightly migrated under lower forces than the oval valved stents. The newly developed stents showed a very good fixation and maintained a good overall anchorage. Nevertheless, an additional apical fixation with neochords could further improve durability and anchorage of the device.
During hydrodynamic tests, the fixation struts provided very good position stability without migration. A rotation of the valved stents, which occurred with D-shaped valved stents with an apical fixation in a previous study, was prevented with struts in this study [12]. The asymmetric atrial flange prevented the stents from migrating into the ventricle during diastole.
The transvalvular pressure difference is related to the orifice area. The CoreValve Evolut R reference valved stent exhibits the highest orifice area and therefore the lowest diastolic transvalvular pressures. The mean diastolic pressure differences are slightly higher for the oval and the D-shaped valved stents compared with the CoreValve Evolut R reference valved stent. Their orifice area is therefore smaller but still large enough to assure sufficient blood flow into the ventricle.
The systolic transvalvular pressure difference is related to the closing properties and the regurgitation of the valve. The pressure difference of the reference CoreValve Evolut R is slightly higher than the oval-shaped and D-shaped valved stent. This is attributed to the stent in stent procedure for the oval-shaped and D-shaped models. For this method, a circular valve-carrying stent is inserted into an oval-shaped or D-shaped fixation stent. Possible leaks may have occurred between these 2 stents, which may decrease the systolic pressure difference. To accomplish better leak tightness of the valved stent, the suture of the valve leaflets and the polytetrafluoroethylene membrane lining between valved stent and the fixation stent have to be improved in additional studies.
Limitations
Additional studies will overcome the current limitations of the preliminary in vitro studies: the experimental setting, the in vitro heart filling conditions and the juvenile mitral annulus of porcine hearts.
CONCLUSION
The first in vitro prototype testing of the novel valved stent models for TMVI were successful, as all tested mitral valved stent models met the criteria for positioning and deployment. The mitral valved stents showed a sufficient fixation with angular struts. This novel mitral valved stent with strut fixation showed great potential for TMVI to treat severe mitral valve disease in the high-risk patients.
Funding
This work was supported by the German Center of Cardiovascular Research (DZHK).
Conflict of interest: Georg Lutter is a consultant of Tendyne Holding Inc. All other authors have nothing to declare.
