Leaflet Mechanical Stress in Different Designs and Generations of Transcatheter Aortic Valves: An in Vitro Study

Abstract

Background

It is unknown whether bioprostheses used for transcatheter aortic valve implantation will have similar long-term durability as those used for surgical aortic valve replacement. Repetitive mechanical stress applied to the valve leaflets, particularly during diastole, is the main determinant of structural valve deterioration. Leaflet mechanical stress cannot be measured in vivo. The objective of this in vitro/in silico study was thus to compare the magnitude and regional distribution of leaflet mechanical stress in old vs new generations of self-expanding (SE) vs balloon expandable (BE) transcatheter heart valves (THVs).

Methods

A double activation simulator was used for in vitro testing of two generations of SE THV (Medtronic CoreValve 26 mm and EVOLUT PRO 26 mm) and two generations of BE THV (Edwards SAPIEN 23 mm vs SAPIEN-3 23 mm). These THVs were implanted within a 21-mm aortic annulus. A noncontact system based on stereophotogammetry and digital image correlation with high spatial and temporal resolution (2000 img/sec) was used to visualize the valve leaflet motion and perform the three-dimensional analysis. A finite element model of the valve was developed, and the leaflet deformation obtained from the digital image correlation analysis was applied to the finite element model to calculate local leaflet mechanical stress during diastole.

Results

The maximum von Mises leaflet stress was higher in early vs new THV generation (p < 0.05) and in BE vs SE THV (p < 0.05): early generation BE: 2.48 vs SE: 1.40 MPa; new generation BE: 1.68 vs SE: 1.07 MPa. For both types of THV, the highest values of leaflet stress were primarily observed in the upper leaflet edge near the commissures and to a lesser extent in the mid-portion of the leaflet body, which is the area where structural leaflet deterioration most often occurs in vivo.

Conclusions

The results of this in vitro/in silico study suggest that: i) Newer generations of THVs have ∼30% lower leaflet mechanical stress than the early generations; ii) For a given generation, SE THVs have lower leaflet mechanical stress than BE THVs. Further studies are needed to determine if these differences between new vs early THV generations and between SE vs BE THVs will translate into significant differences in long-term valve durability in vivo.

Introduction

Aortic valve stenosis is the third most frequent cardiovascular disease in high-income countries after hypertension and coronary artery disease and is associated with increased risk of heart failure and mortality. There is currently no pharmacotherapy to slow or halt the progression of aortic stenosis, and aortic valve replacement with a prosthetic valve is the only available option for the treatment of this disease. Aortic stenosis is responsible for ∼125,000 deaths and ∼350,000 aortic valve replacement procedures per year worldwide, with an associated loss of 1.8 million disability-adjusted life years per year.¹

Surgical aortic valve replacement (SAVR) with a bioprosthesis or a mechanical valve has been the standard of care for the past 60 years. However, transcatheter aortic valve implantation (TAVI) has become a valuable less invasive alternative to SAVR.2, 3, 4 The indication for TAVI has been expanded from high to low surgical risk populations, and the long-term durability of the valves has become a key subject of interest. The main limitation of bioprosthetic valves that are used for SAVR or TAVI is their limited durability due to structural valve deterioration, which implies a risk for a future valve reintervention, especially in younger populations with long life expectancy. Furthermore, the long-term durability of TAVI valves is largely unknown due to the absence of long-term follow-up data in a large number of patients.⁵

The structural valve deterioration and failure of bioprosthetic heart valves is a complex and multifaceted process that involves valve-related and patient-related factors.⁵ Increased and repetitive leaflet mechanical stress appears to be a key determinant of structural valve deterioration.⁶ In silico and in vitro studies provide a unique opportunity to assess and compare the leaflet mechanical stresses of the different models and sizes of bioprosthetic valves in various anatomic and hemodynamic conditions. These data may help to rapidly inform on the potential durability of new generations of bioprosthetic valves.

The two most commonly (by far) used transcatheter heart valve (THV) devices are the Medtronic self-expanding (SE) and the Edwards Lifesciences balloon expandable (BE) THVs. The fundamental principles behind these two valves are similar; that is, they are composed of an expandable stent supporting three pericardium leaflets deployed within a calcified aortic valve. However, these 2 THVs present major differences in their design: SE vs BE stent, and supra-annular vs intra-annular position of the leaflets within the stent. The objective of this in vitro/in silico study was to perform a side-by-side comparison of the early vs new generations of SE and BE THVs with regards to the magnitude and distribution of leaflet stress and strain.

Methods

A double activation simulator (Supplementary Figure 1) was used for in vitro testing of two generations of SE THV (Medtronic CoreValve 26 mm and EVOLUT PRO 26 mm) and two generations of BE THV (Edwards SAPIEN 23 mm vs SAPIEN-3 23 mm). The early-generation THVs were used once before the previous study, whereas the new-generation THVs were new and used for the first time in this study. A biaxial stretch testing was performed on planar biaxial tensile testing device to determine the material properties of the THVs. Leaflet thickness was measured using Smartscope (model MVP, OGP Inc, Switzerland) with a resolution of <5 μm. Table 1 presents the characteristics of the design and materials of these THVs. The THVs were implanted within a standardized 21-mm aortic annulus according to the THV sizing charts recommended by the manufacturers (Medtronic Inc and Edwards Lifesciences).

Table 1.

Design, biomechanical characteristics, and leaflet mechanical stress of early and new generations of BE vs SE THVs

	Data
	Self-expanding THV		Balloon expandable THV
	Early generation CoreValve-26 mm	New generation EVOLUT PRO 26 mm	Early generation Edwards SAPIEN 23 mm	New generation SAPIEN-3 23 mm
Stent material (mm)	Nitinol	Nitinol	Stainless steel	Cobalt-Chromium
Stent thickness (mm)	0.40	0.38	0.65	0.55
Leaflet material	Porcine pericardium	Porcine pericardium	Bovine pericardium	Bovine pericardium
Leaflet thickness (mm)	0.35 ± 0.03	0.28 ± 0.03	0.55 ± 0.04	0.34 ± 0.03
Leaflet angle (º)	55.0	66.0	81.0	80.0
Maximum stent flexibility (mm)		1.07 ± 0.045†		0.05 ± 0.001†
Pinwheeling index	4.69 ± 1.69	3.55 ± 0.98	8.54 ± 2.16†	8.59 ± 0.63†
Maximum principal stress (MPa)	1.40 ± 0.03	1.07 ± 0.02	2.48 ± 0.05	1.68 ± 0.03
Maximum global leaflet stress (MPa)	0.81 ± 0.32†	0.64 ± 0.26†	1.40 ± 0.60∗†	0.99 ± 0.44∗†

Abbreviations: BE, balloon expandable; SE, self-expanding; THV, transcatheter heart valve.

^∗p < 0.05 between early vs new generation within the same type of THV.

^†p < 0.05 between BE vs SE THV within the same generation.

In Vitro Cardiovascular Simulation

The double activation left heart duplicator system⁷ used for this study is able to accurately simulate the human blood circulation from the pulmonary veins to the peripheral systemic capillaries by reproducing physiological pressure and flow waveforms. The circulating fluid was a mixed saline glycerol solution with a viscosity of 3.8 cP similar to blood viscosity and was maintained at a constant temperature of 37 °C. The system includes: i) anatomically shaped, deformable silicone molds of left heart cavities, which are compressed or stretched in order to mimic left heart contraction and relaxation; and ii) a glass model of the aortic root and ascending aorta that enables optimal visualization of valve leaflets.⁸ Contraction of the left ventricle and left atrium were obtained using two ViVitro piston pumps (ViVitro Inc, Victoria, Canada). Pressures in both the left ventricle and the aortic root were recorded by micro-tip pressure catheters (Millar catheter and signal conditioning unit, Millar Instruments, Houston, TX, accuracy ±0.5% maximum full scale). Transvalvular flow was measured with an electromagnetic flowmeter (Model 501, Carolina Medical Electronics Inc, East Bend, USA, accuracy ±1% maximum full scale) positioned immediately below the bioprosthesis. Pump activation and signal acquisition were controlled with LabVIEW8.2 (National Instruments, Texas, USA) to achieve physiological flow as recommended in the ISO standards for heart valve testing in normal flow conditions (ISO 5840-3).

Hemodynamic Testing

In order to obtain the hemodynamic data for the THVs, the tested heart rate was set to 70 bpm, mean aortic pressure to 100 mmHg, and stroke volume to 70 ml. Doppler echocardiographic measurements were performed using the General Electric Vivid 7 system (GE Health Medical, Horten, Norway), with a phased array 3.5 MHz probe. The mean transaortic gradient and aortic velocity-time integral were measured by continuous-wave Doppler on ten beats and averaged. Aortic and ventricular pressures and stroke volume were measured and averaged over 100 cycles.

Valve effective orifice area (EOA) was calculated using the continuity equation by dividing the stroke volume measured with the electromagnetic flowmeter by the echocardiographic aortic velocity-time integral. The geometric orifice area was obtained using a high-speed camera (SA3 Fastcam, Photron, Japan) by acquiring an en-face view of the valve during the cardiac cycle at a frame rate of 1000 images per second. Postprocessing of the recorded images was done using MATLAB (MATLAB R2015a, The MathWorks, Inc, Natick, Massachusetts, United States). The inward displacement of commissures during diastole was calculated using custom-coded tracking system using MATLAB (Supplementary Figure 2). The inward displacement at each commissure was analyzed separately and then averaged for each THV. The maximum inward displacement was calculated as the summation of the values of maximum expansion and maximum deflection of the commissures. The degree of pinwheeling index was assessed by the relative difference between the contour of the actual leaflet coaptation line (L_actual) and the ideal straight coaptation line (L_ideal) (Supplementary Figure 3)⁹:

$$Pinwheelingindex=\frac{L_{actual}-L_{ideal}}{L_{ideal}}$$

Measurement of Leaflet Strain by Stereophotogammetry and Digital Image Correlation

A noncontact system based on stereophotogammetry and digital image correlation (DIC) with high spatial and temporal resolution enabled video capture and performance of frame-by-frame analysis of the valve leaflet motion, three-dimensional (3D) analysis, and strain measurements. To obtain a high-contrast stochastic pattern for the use of DIC,^10,11 the leaflet surface was applied with a fine speckle pattern using a black tissue dye (Shandon Tissue Marking DYE, Thermo Scientific) (Figure 1).^8,13 Leaflet motion was recorded with two high-speed cameras (Fastcam SA3, Photron, Inc, San Diego) equipped with 105 mm lenses (EF 24 Reflex lenses, Sigma, Tokio). Each camera recorded 24-bit true color value images with a resolution of 1024 x 1024 pixels at a frame rate of 2000 frames per second. To ensure simultaneous recording, both cameras were triggered with one shutter release. Before each image acquisition, the camera setup was calibrated. Measurement of displacement was based on 3D DIC conducted with commercial system VIC3D (Correlated Solutions, Inc). Postprocessing of the experimental data of the valve leaflet was realized using the VIC3D 8 postprocessing tool. Since the area of interest needs to be visualized by both cameras, computation of the complete 3D structure of the 3 leaflets during opening and closing was not possible due to complicated THV geometry (stent). Leaflet displacement, stress and strain were computed only during diastole (i.e., when the valve is closed). The principal strain is the extreme values of the normal strain possible in the material. These are the maximum normal strain (major principal strain) and the minimum normal strain (minor principal strain) possible for a specific point on a structural element. Positive signs in strain describe tension, and negative signs describe compression. The accuracy for the DIC was 0.05 pixels for both BE THVs and both SE THVs.

Figure 1.

The THVs (a) before and (b) after the speckle pattern application: I. BE THV-Early Generation (SAPIEN 23mm), II. BE THV-New Generation (SAPIEN-3) 23 mm, III. SE THV-Early Generation (CoreValve 26 mm), IV. SE THV-New Generation (EVOLUT-PRO 26 mm).Approximately 1500 speckle points are applied on the surface of the leaflets.

Abbreviations: BE, balloon expandable; SE, self-expanding; THV, transcatheter heart valve.

Computation of Leaflet Stress by Finite Element Analysis

A finite element model of the valve was then developed using a micro-computed tomography scan of the tested valves, recreating a 3D valve geometry. Two generations of SE THV and two generations of BE THV were imaged with a desktop micro-computed tomography scanner (NanoScan PET-CT, Mediso) in different orientations and intensities to distinguish leaflet and stent geometries. The high-resolution Digital Imaging and Communications in Medicine images were used to create a 3D valve model in format “.stl”. The finalized surface of both tissue material and stent was imported into Hypermesh (Altair HyperWorks, Altair Engineering, USA) to generate a surface mesh (Figure 2) with accurate size ex vivo. Leaflet rheological properties were based on experimental data acquired from bi-axial tensile testing of both bovine and porcine tissue. Poisson's ratio was set to 0.45 to account for near incompressibility of soft tissues, and density was set to 1100 kg/m³. Stent geometry was considered as a rigid body. Leaflet geometry was modeled using 2D shell elements with a thickness of 0.55 vs 0.34 and 0.35 vs 0.28 mm for early vs new BE THVs and early vs new SE THVs, respectively, to mimic the leaflet tissue thickness. Leaflet nodes were tied with stent nodes (as if it was a suture) and for leaflet-to-leaflet contact, the symmetric Type 7 (surface-to-surface auto-impact) contact definition was chosen to avoid leaflet penetration.

Figure 2.

Example of FE model of (a) balloon expandable and (b) self-expanding THV. FE model of two types of THVs used in the study; (a) balloon expandable, (b) self-expanding.

Abbreviations: FE, finite element; THV, transcatheter heart valve.

A custom-coded MATLAB program was used to perform the data alignment by matching the point cloud extracted from the first moment of analyzed VIC3D data (closed leaflets) to a numerical surface mesh exported from Hypermesh. The displacement obtained from the DIC analysis was applied in the finite element model using imposed displacement function, which allowed a realistic motion of the leaflets during diastole. In order to calculate the local mechanical stress (Supplementary Figure 4), FE simulations were performed using Radioss (Altair Engineering, USA).^8,13 Implementation of experimental results to FE model is further explained in online supplementary methods.

The von Mises stress data generated by FE model was examined during diastole at 4 key time points (Figure 3a): coaptation-initial reference (T1), leaflet closure with maximum displacement/stress (T2), mid-diastole (T3) and end-diastole (T4). Four main regions of the leaflets were analyzed: the commissures, the free edge, the belly of the leaflet, and the lower belly/subcommissure (Figure 3b).

Figure 3.

Schematic representation of the analyzed timing and regions of leaflets. Panel A: Four key timings of the valve leaflet kinetics during the cycle were analyzed: leaflet coaptation—initial reference (T1), leaflet closure with maximum displacement/stress (T2), mid-diastole (T3), and end-diastole (T4). Panel B: Four main regions of the leaflets were analyzed: commissures, free edge, belly of the leaflet, and lower belly/subcommissure.

Statistical Analysis

Continuous variables are presented as peak values or mean values ± SD. Mean values of leaflet strain and stress were averaged over the 3 valve leaflets in each of the 4 regions of the leaflet during each of the 4 time points of the cardiac cycle selected. Mean values of leaflet stress and strain were compared according to type and generation of valve using Student t-test. The statistical analyses were performed using the R Project for Statistical Computing, and a p-value <0.05 was considered statistically significant.

Results

Valve Hemodynamic Performance

Table 2 shows valve hemodynamic performance according to the type (BE vs SE) and generation (early vs new) of the THV. Overall EOAs were larger, and gradients were lower with the new generation vs the early generation of THVs for both the BE and SE THVs. Furthermore, for a given generation, the EOAs were larger and gradients lower with the SE vs BE THV.

Table 2.

Hemodynamic performance of early and new generations of BE vs SE THVs

	Data
	Self-expanding THV		Balloon expandable THV
	Early generation CoreValve-26 mm	New generation EVOLUT PRO 26 mm	Early generation Edwards SAPIEN 23 mm	New generation SAPIEN-3 23 mm
MTPG (mmHg)	9.60 ± 0.50∗†	8.70 ± 0.30∗†	11.96 ± 0.10∗†	10.50 ± 0.29∗†
EOA (cm2)	1.48 ± 0.02∗†	1.70 ± 0.02∗†	1.31 ± 0.01∗†	1.57 ± 0.05∗†
GOA (cm2)	2.19 ± 0.08†	2.18 ± 0.08	2.06 ± 0.001∗†	2.14 ± 0.001

Abbreviations: BE, balloon expandable; EOA, effective orifice area; GOA, geometric orifice area; MTPG, mean transprosthetic gradient; SE, self-expanding; THV, transcatheter heart valve.

^∗p < 0.05 between early vs new generation within the same type of THV.

^†p < 0.05 between BE vs SE THV within the same generation.

Inward Displacement of Commissures During the Diastole

The inward displacement of commissures during the diastole for new generations of BE and SE THV are shown in Figure 4 and Table 1. For the BE and SE THV, the summation of the inward displacement of commissures during diastole was 0.05 ± 0.001 mm and 1.07 ± 0.045 mm, respectively. The stent of SE THVs had higher (p < 0.001) inward displacement of commissures compared to BE THV.

Figure 4.

Comparison of the inward displacement of commissures during the diastole according to type of THV. Inward displacement of commissures as a function of time is represented for BE THV-New Generation (SAPIEN 23 mm) in red and SE THV-New Generation (EVOLUT-PRO 26 mm) in light blue.

Abbreviations: BE, balloon expandable; SE, self-expanding; THV, transcatheter heart valve.

Leaflet Pinwheeling Index

The results of the pinwheeling index are presented in Table 1. There was no significant difference between early vs new generation within the same type of THV. Both early and new generations of BE THVs had higher pinwheeling index compared to same generation of SE THVs (p < 0.05).

Leaflet Strain Measured by DIC

The major (ε1) and minor (ε2) principal strains were analyzed during diastole at leaflet coaptation (reference) and fully closed position (Figure 5). During the leaflet closing, the peak values of the major principal strain were observed at the free edge near the leaflet commissures for both the BE THVs and SE THVs.

Figure 5.

Comparison of leaflet strain during diastole according to type and generation of THV. Early-generation THVs are represented in Figures IA-IVB and new generations in Figures VA-VIIIB. Initial reference images at the time of leaflet coaptation (T1) are represented in Panels A and leaflet closure with maximum displacement (T2) are represented in Panels B. Major and minor principal strains for BE THV-Early Generation are represented in Panels I and II, respectively. Major and minor principal strains for SE THV-Early Generation are represented in Panels III and IV respectively. Major and minor principal strains for BE THV-New Generation are represented in Panels V and VI, respectively. Major and minor principal strains for SE THV-New Generation are represented in Panels VII and VIII, respectively. Positive signs in strain describe tension, and negative signs describe compression.

Abbreviations: BE, balloon expandable; SE, self-expanding; THV, transcatheter heart valve.

Leaflet Mechanical Stress Computed by Finite Element Analysis

The maximum values of the leaflet stress generally occur in early diastole at the time of leaflet closure (T2). The oscillations observed on the curves of leaflet stress during the diastolic phase for the SE THVs may be explained by the motion of the leaflets during and following leaflet closure.

At the time of leaflet closure, relatively low mechanical stresses (1.05 vs 0.75 MPa for early vs new BE THV and 0.43 vs 0.46 MPa for early vs new SE THV) were observed at the belly of the leaflet for each THV. The maximum Von Mises stress reached 2.48 vs 1.68 MPa for early vs new BE THV and 1.40 vs 1.07 MPa for early vs new SE THV in the region near the commissures when the valve was fully closed (T2). The maximum von Mises leaflet stress was higher in early vs new THV generation THVs (p < 0.01) and in BE vs SE THVs (p < 0.01) (Figure 6). Newer generations of THVs showed 30% lower peak mechanical stress than the early-generation THVs. SE THVs showed 40% lower peak mechanical stress than BE THVs (Table 1). The stress distribution/patterns were similar for both early and new generations of both THVs with the highest stress values occurring near the commissure regions.

Figure 6.

Comparison of leaflet stress magnitude and distribution obtained by finite element analysis at leaflet closure according to type and generation of THV. This figure shows the magnitude and distribution of leaflet mechanical stress at the time of leaflet closure according to different types and generations of THVs: (a) BE THV-Early Generation (SAPIEN 23 mm), (b) SE THV-Early Generation (CoreValve 26 mm), (c) BE THV-New Generation (SAPIEN-3 23 mm), (d) SE THV-New Generation (EVOLUT-PRO 26 mm). The red and orange areas of pattern represent high values, whereas blue values are low values.

Abbreviations: BE, balloon expandable; SE, self-expanding; THV, transcatheter heart valve.

Discussion

The main findings of this study are: i) The highest values of leaflet mechanical stress occurred during early diastole at the time of leaflet closure and for both types and generations of THVs. These high values were observed at the level of the leaflet free edge near the commissures and to a lesser extent on the leaflet belly; ii) The peak mechanical stress was, on average, ∼30% lower with newer generations vs early generation of THVs, and ∼40% lower in SE vs BE THVs of the same generation.

Clinical relevance and utility of in vitro/in silico studies for the assessment of valve leaflet mechanical stress.

Structural valve deterioration is driven by several factors including: i) biologically driven tissue degeneration and mineralization; and ii) mechanical repetitive mechanical stress imposed on the bioprosthetic valve leaflets.3, 4, 5^,10, 11, 12^,14, 15, 16 The repetitive mechanical loading of the valve leaflets during the cardiac cycle and especially during the diastolic phase, causes damage to the collagen fibers as well as breakdown of the nonfibrous part of the extracellular matrix and contributes to mineralization of the leaflet tissue.^14,17,18 Although many factors related to the patient or the valve are involved in structural valve deterioration, repetitive leaflet mechanical stress is a key determinant of this process.

Structural valve deterioration leads to hemodynamic valve deterioration, which may cause bioprosthetic valve failure and therefore require reintervention. Recent studies revealed an incidence of hemodynamic valve deterioration related to structural valve deterioration >30% and of reintervention >3% at 10 years following biological SAVR.¹² The design, sizing, and positioning of THVs have been improved since the introduction of TAVI to minimize leaflet mechanical stress and optimize the THV hemodynamic performance and durability. The mid-term durability of the recent generations of THVs (5-7 years) are encouraging, but long-term durability data are not yet available.19, 20, 21, 22, 23, 24 By coupling in vitro experiments and numerical simulation, THV hemodynamic performance and leaflet stress can be assessed accurately in different simulated clinical scenarios. Moreover, the magnitude and distribution of the leaflet stress in the different generations and design of THVs may be compared during the different phases of the cardiac cycle in well-standardized conditions.

One of the most valuable optical options to quantitatively assess valve leaflet dynamics is the DIC method that was applied in the present study.¹⁹ Indeed, this method does not physically interact with the leaflets and hence does not interfere with valve dynamics. This method allows measurement of leaflet deformation and strain but does not allow quantitation of leaflet mechanical stress, per se. In this study, we thus corroborated the leaflet stress results predicted by FE model with leaflet strain data obtained in vitro by DIC.

The mechanical testing, and in particular the high-cycle fatigue testing required by international standards and regulatory requirements, do not reflect the actual and physiologic conditions of the human circulatory system, and the 200 million cycles required by 5840 ISO standard only reflect 5 years of human life. Hence, the mechanical testing required in these standards may not be sufficient to identify the risk of structural valve deterioration and failure in the new types and generations of bioprosthetic valves. A patent case example to illustrate the major limitations of the mechanical testing prescribed by ISO and regulatory standards is the Trifecta, a surgical bioprosthetic valve that passed and exceeded the high-cycle fatigue testing experiments but had an unacceptable rate of early failure in patients.25, 26, 27 Interestingly, we previously published a study using the same platform and method as used in the present study, which revealed higher leaflet mechanical stress in the Trifecta valve vs other bioprosthetic valves.⁸ When repeated billions of times during patient’s life, even small differences in leaflet mechanical stress between two valves may translate into significant differences in durability.

The timing of the changes in leaflet stress during the cardiac cycle as well as its regional distribution within the leaflet observed in the present in vitro study are consistent with what is reported in vivo studies.^28,29 The bioprosthetic valve leaflets experience high stresses during valve closure and cyclic flexural and compressive stresses during valve opening. Tissue rupture and leaflet tear are, at least in part, related to flexural fatigue^15,28,30,31 and the effect of compressive stresses on the inner surface of the bent tissue.³² Pathologic examinations of explanted bioprosthetic valves have revealed that tears of the leaflet free edge frequently occur in the areas of high localized mechanical stress, such as near the commissures.³³ Studies have also suggested that leaflet tissue calcification generally starts to occur at the commissures and leaflet attachment region, extending towards the leaflet belly region.^6,34 Numerical simulation studies have reported that the maximum stress in THVs occurs during diastole near the commissures,^35,36 such as in the case of surgical bioprosthetic valves.³⁷

Valve Related Factors Determining Leaflet Mechanical Stress

The magnitude of the mechanical forces applied at the level of leaflet free-edge could be related to these factors: i) the leaflet anisotropy; ii) the angle the leaflet free-edge forms with the stent post (Figure 7); iii) the inward displacement of commissures (Supplementary Figure 2); and iv) the extent of pinwheeling motion of the leaflets (Supplementary Figure 3). In the present study, these factors differed between new vs older generations of THVs and between SE vs BE THVs. However, these associations do not necessarily imply causality and are thus hypothesis-generating. Further studies are needed to determine the independent contribution of each of these factors to the magnitude of leaflet stress. (Table 1).

Figure 7.

Angle between leaflet free-edge and stent post. The angle (⍺) is measured when the valve leaflets are closed. Example of the angle measurement (a) for BE THV and (b) for SE THV.

Abbreviations: BE, balloon expandable; SE, self-expanding; THV, transcatheter heart valve.

Ideally, the design and biomechanical characteristics of a bioprosthetic valve should closely mimic those of a normal native aortic valve, that is, homogeneous leaflet tissue with high stiffness in the circumferential direction and important elasticity in the radial direction.³⁸ The fibers of the aortic valve can be represented as the main structural cables of a suspension bridge direction.³⁸ In this representation, the vertical loads carried by the belly of the valve leaflets are transferred to the commissures, which explains why the highest tensile forces occur within the fibers near the free edge and particularly near the commissures. The analogy with the suspended bridge concept also implies that, theoretically, the smaller the angle between the leaflet free-edge and the stent post, the lower the load imposed on the leaflet (Figure 7).

The flexibility of the supporting stent is also believed to contribute to optimal valve function and durability.^39,40 In the native aortic valve, the leaflets join the aortic wall at the level of the commissures. During systole, the aorta expands, thus pulling the commissures apart and facilitating leaflet opening. During diastole, the inward motion of the commissures may contribute to absorb a portion of the kinetic energy during valve closure and thereby potentially reduce the mechanical stress transmitted to the leaflet. Hence, in bioprosthetic valves, greater flexibility of the stent might theoretically help to reduce the stress on the leaflets.

Leaflet Mechanical Stress According to THV Generation

Each of THVs tested in the present study underwent extensive mechanical testing based on international standards and regulatory requirements in order to achieve the market approval. They were subjected to different tests such as mechanical durability (200 million cycles or more) and failure testing and have demonstrated sufficient mechanical durability within acceptable limits, with a safety factor, to be approved for clinical use.

The present study reveals that leaflet mechanical strain and stress are reduced by approximately 30% with new generations vs early generations of THVs, which is consistent with the valve durability data reported in recent clinical studies. Indeed, an analysis of the PARTNER 2 trial and registry reported that the third generation of BE THV SAPIEN 3 has better mid-term (5-year) durability than the second generation (SAPIEN XT) and similar durability compared to surgical bioprosthetic valves.²³ The factors that could potentially explain the lower leaflet mechanical stress in new vs early generations of THVs include: i) the angle between the leaflet free-edge forms and the stent post; and ii) greater stent flexibility, particularly for the new generation of SE THVs (Table 1). Other factors such as differences in leaflet shape, commissure attachment, and leaflet tissue processing may also contribute to these differences in leaflet mechanical stress between different generations of THVs.

Leaflet Mechanical Stress According to THV Design

The present study reports that leaflet mechanical stress is, on average, 40% lower with SE vs BE THVs of the same generation. The main factors that could explain this finding include the smaller angle between leaflet and stent and the greater inward displacement of commissures/stent flexibility, and the lower pinwheeling motion index in SE vs BE THVs (Figure 4, Figure 7 and Table 1).

The CHOICE trial is, until now, the sole randomized trial that performed head-to-head comparison of valve durability in SE vs BE THV platforms.²⁴ This study reported higher incidence of structural valve deterioration in BE (6.6%) vs SE (0%) THVs at 5-year follow-up.²⁴ However, this study was underpowered to assess structural valve deterioration, and it only included early generations of THVs (CoreValve and SAPIEN XT). Recently, two other randomized trials (SMART and BEST) have been launched to compare the hemodynamic, clinical, and valve durability outcomes of new generations of BE vs SE THVs.

Study Limitations

This is an in vitro study. However, the measurement of leaflet mechanical stress would be difficult or impossible to perform in vivo. The leaflet mechanical strain was not measured directly but derived from high-speed camera imaging and FE model. Direct measurements would imply the installation of sensors directly on the leaflets, which would alter the mechanical behavior of the leaflets. The leaflet displacement obtained in the in vitro experiments was imposed to the FE model in order to perform the leaflet stress analyses, and no fluid was included in the numerical simulation. Leaflets were assumed to be composed of isotropic linear material. In the present study, the THVs were tested under only one set of hemodynamic condition and in rigid model of the root. Only one cardiac cycle was tested because the frequency (2000 img/sec) shortened the camera’s record time. Only one part of the cardiac cycle (diastole) was presented in this study due to the complex geometry of THVs.

The present study was limited to a few models and sizes of THV. However, the methods developed in this study could be applied to a large variety of valve models and sizes, aortic annulus size and shape (circular vs elliptic), percentage of THV oversizing, THV position, etc. in order to determine which factors, and particularly those that are modifiable, influence the leaflet mechanical stress and thus potentially valve durability.

Conclusions

The novel in vitro/in silico methods proposed in the present study provide a useful tool to determine and compare the magnitude and distribution of the leaflet strain and stress during the cardiac cycle in different types of bioprosthetic valves and under representative anatomic and physiological conditions.

The results of this in vitro/in silico study suggest that: i) Newer generations of THVs have ∼30% lower leaflet mechanical stress than the early generations; ii) For a given generation, SE THVs have ∼40% lower leaflet mechanical stress than BE THVs. Further in vivo studies are needed to determine if these differences will translate into significant differences in long-term valve durability.

Ethics Statement

The relevant ethical guidelines have been followed in the research.

Consent Statement

This study did not utilize any data related to patients or participants, or their parents or legal guardians. The tests were performed on the bench, and the study was solely in vitro.

Funding

The authors have no funding to report.

Review Statement

Given his role as an editor, Philippe Pibarot, DVM, PhD, FASE, had no involvement in the peer review of this article and has no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to Guest Editor Ajit Yoganathan, MD.

Disclosure Statement

P. Pibarot reports research grants from Edwards Lifesciences, Medtronic, and Pi-Cardia. R. Rieu reports research grant from Medtronic. J. Rodés-Cabau is a consultant for and reports institutional research grants from Edwards Lifesciences, Medtronic, and Boston Scientific. All other authors have no conflict of interest to disclose.

Guest Editor: Ajit Yoganathan, MD

Supplementary Material

Supplementary Figures 1-

4