Valve-in-valve outcome: design impact of a pre-existing bioprosthesis on the hydrodynamics of an Edwards Sapien XT valve

Abstract

INTRODUCTION

Since the first transcatheter aortic valve implantation (TAVI) was performed by Cribier et al. in 2002 [1], more than 100 000 TAVI procedures have been conducted worldwide [2] as the procedure was swiftly embraced in clinical practice after regulatory approval. Currently, TAVI is recommended for patients with severe symptomatic aortic stenosis who are deemed inoperable or at prohibitively high risk for surgical aortic valve replacement (SAVR) [3, 4].

In the wake of TAVI’s success, failing surgical bioprostheses have also been treated by transcatheter valve-in-valve implantation (ViV-TAVI). Currently, ViV-TAVI is an alternative to SAVR in patients with severe aortic stenosis deemed by a heart team as inoperable or at high risk [4]. The Medtronic CoreValve (Medtronic, Minneapolis, MN, USA) and the Edwards Sapien XT (Edwards Lifesciences, Irving, CA, USA) transcatheter aortic valve prostheses both received the CE mark and premarket approval by the U.S. Food and Drug Administration for use in degenerated bioprostheses [5–7]. Despite good outcomes, technical difficulties such as device malpositioning, coronary obstruction or elevated post-procedural gradients have been reported more frequently with ViV-TAVI than with TAVI in native aortic valves [8].

The proportion of surgically implanted heart valves that are bioprosthetic has increased. Moreover, this trend also extends to valves implanted in younger patients in order to avoid long-term anticoagulation medications [9]. Consequently, the potential failure of the bioprosthesis followed by the need for repeated aortic valve replacements should be considered as part of a long-term strategy prior to the implantation of a bioprosthesis in a relatively young patient.

The current study analysed the effects of the design of the surgical bioprosthesis on the hydrodynamic function of an Edwards Sapien XT valve implanted using a ViV procedure in four commonly used bioprostheses, namely the Vascutek Aspire (Vascutek, Inchinnan, Scotland, UK), the Edwards Perimount, the Medtronic Mosaic and the St. Jude Medical Trifecta (St. Jude Medical, Saint Paul, MN, USA) valves in order to provide physicians with additional information when choosing the ‘first valve’.

MATERIALS AND METHODS

Surgical aortic valve bioprostheses

The 23-mm size is the most commonly used surgical bioprosthesis according to the latest data from the valve-in-valve international database (VIVID) [8] and was chosen for this study. The Edwards Lifesciences Perimount 23 mm represents 35.2% of failing bioprostheses that are treated with a ViV procedure [8]. The Perimount has bovine pericardial leaflets that are fixed on the inside of a cobalt-nickel stent [10, 11]. The Medtronic Mosaic 23 mm and Vascutek Aspire 23 mm are manufactured using native porcine aortic valves also fixed on the inside of a polymer stent [10, 11]. A total of 11.7% of the patients in the VIVID have a failing mosaic bioprosthesis [8]. In contrast to the aforementioned surgical aortic valves (SAVs), the St. Jude Medical Trifecta 23 mm has bovine pericardial leaflets that are mounted externally to the stent frame, which is made of polyester-covered titanium [10]. The fabric of the suture rings of the bioprosthesis was sealed using silicone to simulate intimal ingrowth and prevent paravalvular leakage outside the bioprosthesis.

Valve-in-valve implantation

The selected 23-mm bioprostheses have an inner diameter (ID) of 18–21 mm [12]. Therefore, a Sapien XT 23 mm was chosen for the ViV implantation as recommended for this diameter range by the manufacturer’s instructions. The Sapien XT 23-mm prosthesis was crimped and subsequently expanded into the bioprosthesis using the Edwards Ascendra+ System. Axial positioning was performed in accordance with the recommendations given by the ViV Aortic App [13]; the ViV pairs are depicted in Fig. 1A–D. Radial alignment of the commissures of the bioprostheses and the Sapien XT was chosen to ensure comparable conditions for all four bioprostheses.

Figure 1:

Edwards Sapien XT valve implanted in Vascutek Aspire (A), Medtronic Mosaic (B), Edwards Lifesciences Perimount (C) and St. Jude Trifecta (D) valves. (E) is a schematic representation of the CVE left heart simulator.

Open in new tabDownload slide

Test setup

The bioprostheses with the implanted Sapien XT valves were fixed in a rigid aortic root chamber with three sinuses, which was integrated into the pulse duplicator system of the Department of Cardiovascular Engineering (CVE) at the Helmholtz Institute (Aachen, Germany). The CVE pulse duplicator is a hydraulic model of the left human circulatory system consisting of a passively filled atrium, a silicone ventricle in a compression chamber, and resistances and compliances (Fig. 1E). It can produce pressure and flow waveforms that approximate conditions over the required physiological range appropriate for the intended device application. The CVE pulse duplicator has been successfully used in several comparable studies [14–18]. The test fluid used was saline solution at 37 °C. Pressure values were measured using PVB DPT 6000 pressure sensors (CODAN Medizinische Geräte GmbH & Co KG, Germany) while flow data was acquired via 20PXL flow sensors (Transonic Instruments, NY, USA). High-speed video recordings (Fastcam 1024 PCI Model 100K, Photron USA Inc, San Diego, CA, USA) of the tests were taken at 3000 frames per second at a resolution of 512 × 512 pixels.

The hydrodynamic tests were performed at three characteristic circulatory conditions (Table 1). Condition 1 resembles a patient’s state after the intervention, which is characterized by increased heart rate and low cardiac output. After recovery, the patient’s parameters return to a regular state with normotensive conditions (condition 2). During exercise, the heart rate, the aortic pressure and the aortic flow rise as modelled in condition 3.

Data analysis

In this study the mean pressure gradient (MPG), the smallest cross section of the flow—the effective orifice area (EOA)—and the regurgitant fraction were calculated according to ISO 5840-3:2013 from the pressure and flow data to assess valve function [19]. The regurgitant fraction includes closure and the leakage volumes, which are presented as a percentage of the stroke volume. All variables were obtained from 10 cycles, and the results are given as mean ± standard deviation. In addition, statistical analysis of the results was performed using the Welch t-test for a paired comparison of results. Because the Sapien XT maximum orifice area is limited by the SAV’s inner orifice area, the EOA, the geometric opening area (GOA) and the MPG should correlate with the inner orifice area of the SAVs. Therefore, the correlation of these measurements with the orifice area of the SAV derived from the ID as given by Bapat et al. [13] was investigated. A confidence level above 95% (P < 0.05) was considered significant for all results of the statistical analysis.

The high-speed videos were analysed in terms of leaflet dynamics and coaptation as well as possible leaflet stent-frame contact. Additionally, the GOA, defined here as the unobstructed geometric opening, was calculated from the high-speed video recordings.

RESULTS

High-speed video analysis

High-speed analysis of leaflet kinematics showed that the Sapien XT valve closed without visible central leakage in all ViV combinations (Fig. 2A–D). However, the leaflets of the Sapien XT valve implanted in the Aspire and the Mosaic prostheses overlapped and twisted considerably (so-called pin-wheeling) within the coaptation area (Fig. 2A and C, Video 1).

Figure 2:

Still images from the high-speed video recordings at 5 l/min point of operation: Closed valves: Sapien XT in Aspire (A), Sapien XT in Perimount (B), Sapien XT in Mosaic (C) and Sapien XT in Trifecta (D).

Open in new tabDownload slide

The GOA of the Sapien XT valve in the Perimount (Fig. 3D–F) and Trifecta (Fig. 3J–L) valves was close to circular in the open position, whereas the open leaflets of the Sapien XT valve paired with the Aspire (Fig. 3A–C) and the Mosaic (Fig. 3G–I) valves displayed a more triangular shape. In the Sapien XT matched with the Aspire, this triangular shape was also slightly tilted towards the lower right side, resulting in systolic contact between the lower leaflet and the stent frame (marked by black arrow in Fig. 3A–C). In addition, the Sapien XT-Aspire configuration showed an uneven closing that was most evident at measurement condition 2. One leaflet reached the closed position before the other leaflets started closing. Complete closure of the Sapien XT in the Aspire occurred about 80 ms after the first leaflet reached the closed position (Video 1). Such an effect was not seen in any other combination.

Figure 3:

Still images from the high-speed video recordings: open valves at all points of operation. White arrows mark areas of axial overlap.

Open in new tabDownload slide

Hydrodynamic results

The EOA and the MPG increased significantly with increased cardiac output in all ViV combinations (P < 0.001) with the exception of the EOA in the Mosaic valve between measurement points 1 and 2 (1.50 ± 0.01 cm² vs 1.51 ± 0.01 cm², P = 0.26) (Fig. 4A and B, Table 2). The EOA correlated with the area of the ID of the SAV given by Bapat et al. [13] (Table 3). The correlation was statistically significant for measurement conditions 2 and 3 (P < 0.05). Also, the MPG correlated negatively with the growing ID of the surgical valve (Table 4), which was statistically significant at measurement condition 1 (P < 0.05).

Figure 4:

Results of the hydrodynamic tests: EOA in cm² (A), MPG in mmHg (B) and regurgitation as percentage of the forward flow (C) for all four Sapien XT–SAV pairs at the three measurement points. All values given as mean ± SD. Where the comparison of two values using Welch’s t-test resulted in a significant difference, the confidence level is denoted by * for P < 0.05, ** for P < 0.01 and *** for P < 0.001. EOA: effective oriface area; MPG: mean pressure gradient.

Open in new tabDownload slide

Table 2:

Open in new tab

Results of the hydrodynamic tests

Sapien XT valve in .	Measurement condition .	MPG (mmHg) .	EOA (cm2) .	Closure volume (% of SV) .	Leakage volume (% of SV) .	Total regurgitation (% of SV) .
Aspire	1	5.7  ± 0.1	1.3  ± 0.0	5.5  ± 0.3	0.6  ± 0.5	6.1  ± 0.7
	2	13.2  ± 0.2	1.6  ± 0.0	1.9  ± 0.3	1.3  ± 0.7	3.2  ± 0.7
	3	20.3  ± 0.4	1.7  ± 0.0	1.2  ± 0.3	0.7  ± 0.9	2.0  ± 1.2
Perimount	1	5.2  ± 0.4	1.4  ± 0.0	6.0  ± 0.3	3.0  ± 0.8	9.0  ± 1.0
	2	11.2  ± 0.1	1.9  ± 0.0	1.9  ± 0.2	3.5  ± 1.3	5.4  ± 1.3
	3	15.4  ± 0.3	2.0  ± 0.0	2.0  ± 0.2	1.8  ± 0.6	3.9  ± 0.6
Mosaic	1	6.3  ± 0.1	1.2  ± 0.0	4.7  ± 0.3	1.0  ± 0.8	5.6  ± 0.7
	2	15.6  ± 0.2	1.5  ± 0.0	1.6  ± 0.5	0.9  ± 1.2	2.4  ± 1.3
	3	22.6  ± 0.5	1.5  ± 0.0	1.3  ± 0.3	1.0  ± 0.8	2.3  ± 0.8
Trifecta	1	5.8  ± 0.1	1.4  ± 0.0	6.6  ± 0.3	7.6  ± 0.6	14.2  ± 0.6
	2	13.0  ± 0.2	1.8  ± 0.0	2.4  ± 0.5	8.9  ± 1.0	11.2  ± 1.4
	3	17.9  ± 0.4	1.8  ± 0.0	2.8  ± 0.4	4.0  ± 0.5	6.8  ± 1.4

Sapien XT valve in .	Measurement condition .	MPG (mmHg) .	EOA (cm2) .	Closure volume (% of SV) .	Leakage volume (% of SV) .	Total regurgitation (% of SV) .
Aspire	1	5.7  ± 0.1	1.3  ± 0.0	5.5  ± 0.3	0.6  ± 0.5	6.1  ± 0.7
	2	13.2  ± 0.2	1.6  ± 0.0	1.9  ± 0.3	1.3  ± 0.7	3.2  ± 0.7
	3	20.3  ± 0.4	1.7  ± 0.0	1.2  ± 0.3	0.7  ± 0.9	2.0  ± 1.2
Perimount	1	5.2  ± 0.4	1.4  ± 0.0	6.0  ± 0.3	3.0  ± 0.8	9.0  ± 1.0
	2	11.2  ± 0.1	1.9  ± 0.0	1.9  ± 0.2	3.5  ± 1.3	5.4  ± 1.3
	3	15.4  ± 0.3	2.0  ± 0.0	2.0  ± 0.2	1.8  ± 0.6	3.9  ± 0.6
Mosaic	1	6.3  ± 0.1	1.2  ± 0.0	4.7  ± 0.3	1.0  ± 0.8	5.6  ± 0.7
	2	15.6  ± 0.2	1.5  ± 0.0	1.6  ± 0.5	0.9  ± 1.2	2.4  ± 1.3
	3	22.6  ± 0.5	1.5  ± 0.0	1.3  ± 0.3	1.0  ± 0.8	2.3  ± 0.8
Trifecta	1	5.8  ± 0.1	1.4  ± 0.0	6.6  ± 0.3	7.6  ± 0.6	14.2  ± 0.6
	2	13.0  ± 0.2	1.8  ± 0.0	2.4  ± 0.5	8.9  ± 1.0	11.2  ± 1.4
	3	17.9  ± 0.4	1.8  ± 0.0	2.8  ± 0.4	4.0  ± 0.5	6.8  ± 1.4

EOA according to the ISO 5840 in cm², MPG in mmHg, and closure volume, leakage volume, and total regurgitation as percentage of the forward flow for all four Sapien XT–SAV pairs at the three measurement points. All values are given as mean ± SD.

EOA: effective orifice area; MPG: mean pressure gradient; SV: stroke volume.

Table 2:

Open in new tab

Results of the hydrodynamic tests

Sapien XT valve in .	Measurement condition .	MPG (mmHg) .	EOA (cm2) .	Closure volume (% of SV) .	Leakage volume (% of SV) .	Total regurgitation (% of SV) .
Aspire	1	5.7  ± 0.1	1.3  ± 0.0	5.5  ± 0.3	0.6  ± 0.5	6.1  ± 0.7
	2	13.2  ± 0.2	1.6  ± 0.0	1.9  ± 0.3	1.3  ± 0.7	3.2  ± 0.7
	3	20.3  ± 0.4	1.7  ± 0.0	1.2  ± 0.3	0.7  ± 0.9	2.0  ± 1.2
Perimount	1	5.2  ± 0.4	1.4  ± 0.0	6.0  ± 0.3	3.0  ± 0.8	9.0  ± 1.0
	2	11.2  ± 0.1	1.9  ± 0.0	1.9  ± 0.2	3.5  ± 1.3	5.4  ± 1.3
	3	15.4  ± 0.3	2.0  ± 0.0	2.0  ± 0.2	1.8  ± 0.6	3.9  ± 0.6
Mosaic	1	6.3  ± 0.1	1.2  ± 0.0	4.7  ± 0.3	1.0  ± 0.8	5.6  ± 0.7
	2	15.6  ± 0.2	1.5  ± 0.0	1.6  ± 0.5	0.9  ± 1.2	2.4  ± 1.3
	3	22.6  ± 0.5	1.5  ± 0.0	1.3  ± 0.3	1.0  ± 0.8	2.3  ± 0.8
Trifecta	1	5.8  ± 0.1	1.4  ± 0.0	6.6  ± 0.3	7.6  ± 0.6	14.2  ± 0.6
	2	13.0  ± 0.2	1.8  ± 0.0	2.4  ± 0.5	8.9  ± 1.0	11.2  ± 1.4
	3	17.9  ± 0.4	1.8  ± 0.0	2.8  ± 0.4	4.0  ± 0.5	6.8  ± 1.4

Sapien XT valve in .	Measurement condition .	MPG (mmHg) .	EOA (cm2) .	Closure volume (% of SV) .	Leakage volume (% of SV) .	Total regurgitation (% of SV) .
Aspire	1	5.7  ± 0.1	1.3  ± 0.0	5.5  ± 0.3	0.6  ± 0.5	6.1  ± 0.7
	2	13.2  ± 0.2	1.6  ± 0.0	1.9  ± 0.3	1.3  ± 0.7	3.2  ± 0.7
	3	20.3  ± 0.4	1.7  ± 0.0	1.2  ± 0.3	0.7  ± 0.9	2.0  ± 1.2
Perimount	1	5.2  ± 0.4	1.4  ± 0.0	6.0  ± 0.3	3.0  ± 0.8	9.0  ± 1.0
	2	11.2  ± 0.1	1.9  ± 0.0	1.9  ± 0.2	3.5  ± 1.3	5.4  ± 1.3
	3	15.4  ± 0.3	2.0  ± 0.0	2.0  ± 0.2	1.8  ± 0.6	3.9  ± 0.6
Mosaic	1	6.3  ± 0.1	1.2  ± 0.0	4.7  ± 0.3	1.0  ± 0.8	5.6  ± 0.7
	2	15.6  ± 0.2	1.5  ± 0.0	1.6  ± 0.5	0.9  ± 1.2	2.4  ± 1.3
	3	22.6  ± 0.5	1.5  ± 0.0	1.3  ± 0.3	1.0  ± 0.8	2.3  ± 0.8
Trifecta	1	5.8  ± 0.1	1.4  ± 0.0	6.6  ± 0.3	7.6  ± 0.6	14.2  ± 0.6
	2	13.0  ± 0.2	1.8  ± 0.0	2.4  ± 0.5	8.9  ± 1.0	11.2  ± 1.4
	3	17.9  ± 0.4	1.8  ± 0.0	2.8  ± 0.4	4.0  ± 0.5	6.8  ± 1.4

EOA according to the ISO 5840 in cm², MPG in mmHg, and closure volume, leakage volume, and total regurgitation as percentage of the forward flow for all four Sapien XT–SAV pairs at the three measurement points. All values are given as mean ± SD.

EOA: effective orifice area; MPG: mean pressure gradient; SV: stroke volume.

Table 3:

Open in new tab

Measurements of the GOA from high-speed video stills and the ‘true ID’ as given by Bapat et al. [13]

Sapien XT valve in .	Measurement condition .	GOA (mm2) .	‘True ID’ (mm) .
Aspire	1	2.3	19.0
	2	2.5
	3	2.6
Perimount	1	2.6	21.0
	2	2.9
	3	3.1
Mosaic	1	2.2	18.5
	2	2.2
	3	2.3
Trifecta	1	2.6	20.5
	2	2.8
	3	2.8

Sapien XT valve in .	Measurement condition .	GOA (mm2) .	‘True ID’ (mm) .
Aspire	1	2.3	19.0
	2	2.5
	3	2.6
Perimount	1	2.6	21.0
	2	2.9
	3	3.1
Mosaic	1	2.2	18.5
	2	2.2
	3	2.3
Trifecta	1	2.6	20.5
	2	2.8
	3	2.8

GOA: geometric orifice area, ID: inner diameter.

Table 3:

Open in new tab

Measurements of the GOA from high-speed video stills and the ‘true ID’ as given by Bapat et al. [13]

Sapien XT valve in .	Measurement condition .	GOA (mm2) .	‘True ID’ (mm) .
Aspire	1	2.3	19.0
	2	2.5
	3	2.6
Perimount	1	2.6	21.0
	2	2.9
	3	3.1
Mosaic	1	2.2	18.5
	2	2.2
	3	2.3
Trifecta	1	2.6	20.5
	2	2.8
	3	2.8

Sapien XT valve in .	Measurement condition .	GOA (mm2) .	‘True ID’ (mm) .
Aspire	1	2.3	19.0
	2	2.5
	3	2.6
Perimount	1	2.6	21.0
	2	2.9
	3	3.1
Mosaic	1	2.2	18.5
	2	2.2
	3	2.3
Trifecta	1	2.6	20.5
	2	2.8
	3	2.8

GOA: geometric orifice area, ID: inner diameter.

Table 4:

Open in new tab

Results of the linear regression analyses of the EOA, GOA and MPG in relation to the area of the surgical aortic valves derived from the inner diameter as given by Bapat et al. [13]

.	EOA .		GOA .		MPG .
.	R2 .	P .	R2 .	P .	R2 .	P .
3l/min	0.81	0.102	0.95	0.024	0.63	0.204
5l/min	0.93	0.035	0.95	0.025	0.76	0.128
8l/min	0.93	0.038	0.95	0.026	0.95	0.026

.	EOA .		GOA .		MPG .
.	R2 .	P .	R2 .	P .	R2 .	P .
3l/min	0.81	0.102	0.95	0.024	0.63	0.204
5l/min	0.93	0.035	0.95	0.025	0.76	0.128
8l/min	0.93	0.038	0.95	0.026	0.95	0.026

EOA: effective orifice area; GOA: geometric orifice area; MPG: mean pressure gradient.

Table 4:

Open in new tab

Results of the linear regression analyses of the EOA, GOA and MPG in relation to the area of the surgical aortic valves derived from the inner diameter as given by Bapat et al. [13]

.	EOA .		GOA .		MPG .
.	R2 .	P .	R2 .	P .	R2 .	P .
3l/min	0.81	0.102	0.95	0.024	0.63	0.204
5l/min	0.93	0.035	0.95	0.025	0.76	0.128
8l/min	0.93	0.038	0.95	0.026	0.95	0.026

.	EOA .		GOA .		MPG .
.	R2 .	P .	R2 .	P .	R2 .	P .
3l/min	0.81	0.102	0.95	0.024	0.63	0.204
5l/min	0.93	0.035	0.95	0.025	0.76	0.128
8l/min	0.93	0.038	0.95	0.026	0.95	0.026

EOA: effective orifice area; GOA: geometric orifice area; MPG: mean pressure gradient.

Similar to those of EOA and MPG, regurgitation measurements (Fig. 4C and Table 2) showed considerable differences between the ViV combinations. Analysis of the high-speed video recordings showed complete trileaflet coaptation for each ViV pair (Fig. 2). A complete coaptation of all leaflets corresponded to no transvalvular regurgitation, so leakage after the closure of the valve was attributed to paravalvular leakage.

The closure volume as a percentage of the forward flow was largest at measurement condition 1 in all ViV combinations (Fig. 4C). For all ViV combinations, the closure volume significantly decreased from measurement condition 1 to condition 2 (P < 0.001). However, the decrease in the closure volume between measurement conditions 2 and 3 was only significant in the Sapien XT in the Aspire (P <0.001). In the Aspire and in the Mosaic a significant difference in the closure volume of the Sapien XT occurred only at measurement condition 1 (5.5 ± 0.3 vs 4.7 ± 0.3% of SV, P < 0.001), whereas the uneven leaflet motion of the Sapien XT in the Aspire was more pronounced in measurement conditions 2 and 3.

Although the leakage of the Sapien XT when implanted in the Aspire and the Mosaic valves showed no significant difference, a significant increase in the leakage was recorded in all measurement conditions when the Sapien XT was matched with the Perimount (P < 0.01 for all pairs with corresponding measurement conditions). In the Sapien XT-Trifecta pair, the leakage was largest and constituted another significant increase in comparison to the Sapien XT in the Perimount (P < 0.001 for all measurement conditions).

The resulting regurgitation also displayed no significant differences between the Sapien XT in the Aspire or the Mosaic while the increases in regurgitation in the Sapien XT-Perimount pair were significant for all measurement conditions (P < 0.001, Table 2). The regurgitation again increased significantly in the Sapien XT when it was implanted in the Trifecta (P < 0.001, Table 2) because the leakage comprised most of the regurgitation.

Valve area

DISCUSSION

This study reports detailed hydrodynamic tests for ViV procedures with an Edwards Sapien XT 23-mm valve mounted in a Vascutek Aspire, an Edwards Perimount, a Medtronic Mosaic and a St. Jude Medical Trifecta valve, all of 23-mm label size. The results of this study show a significant impact of the ID of the SAV as well as its leaflet material and mounting on the hydrodynamic performance and the leaflet dynamics of a Sapien XT implanted ViV.

For treatment of degenerated surgical bioprostheses, transcatheter valve-in-valve procedures are performed with increasing frequency [8]. Advantages of these procedures are reduction of surgical trauma and avoidance of cardiopulmonary bypass. However, insights into the VIVID reveal an increased risk for device malposition, elevated pressure gradients and impaired hemodynamic function compared to standard TAVI [8].

To minimize the risks of ViV, the underlying causes need to be investigated. Our group recently demonstrated the influence of aortic annulus ovality on transcatheter heart valves [16] and also found that the hydrodynamic performance of the CoreValve and the Sapien XT valves in a ViV setting is significantly dependent on the transcatheter aortic valve and the surgical aortic valve [18]. In addition, we showed that in vitro testing can be successfully integrated into an algorithm of patient-specific, preprocedural planning for ViV [17], demonstrating that in vitro testing can provide clinicians with crucial information to minimize patient risk.

In the current study, the Sapien XT valve showed significantly different hydrodynamic results as well as leaflet dynamics in all four bioprostheses, although it still fulfilled the minimum acceptance criteria of ISO 5840-3:2013 [19]. Also, the Sapien XT valve had a post-procedural MPG below 20 mmHg for flow condition 2 in all ViV combinations (Fig. 4B, Table 2), which is defined as device success in the updated VARC-2 criteria [20].

However, the Sapien XT valve was under expanded in the smallest valves—the Aspire and the Mosaic. The under expansion was evident in the triangular shape of the opening of the prosthesis during systole as well as by the pin-wheeling of the leaflets in the closed position. Moreover, axial overlap occurred more often in the under expanded valves, limiting the increases seen in the GOA and EOA and thereby causing increased MPGs. In addition to the inferior hydrodynamics, the pin-wheeling adds stress to the leaflets, which may reduce the long-term durability. Therefore, a bioprosthesis with an ID in the upper range of the Sapien XT’s requirements is associated with improved heamodynamics and reduced stress.

Because the Sapien XT closes completely in all ViV combinations, it is assumed that no central leakage occurs. At the same time, the sewing rings of the SAVs were sealed with silicone and fixed in the test chamber to exclude paravalvular leakage via the sewing ring and the wall of the test chamber. Therefore, the recorded leakage consisted only of paravalvular leakage between the SAV and the implanted transcatheter heart valve (Sapien XT). However, the ViV pairs exhibited large differences in the recorded leakage volumes. The likely causes for these increases were the differences in the leaflet material and in the mounting of the leaflets. Whereas the bovine pericardium of the Perimount’s leaflets is pushed outward and forms a tube, the porcine leaflets of the Aspire and the Mosaic are less stiff and provide a better seal for the gap between the Sapien XT and the SAV (Fig. 1A and C). In contrast, the Trifecta has externally mounted leaflets that result in the Sapien XT being pressed into the textile lining of the Trifecta, resulting in increased paravalvular leakage in this hydrodynamic study. Thus, the differences in the leakage volumes in this study can be explained by the variations in materials and mountings of the leaflets of the bioprostheses and the stent designs of the SAVs.

In the in vitro implantation, the commissural alignment of the Sapien XT with the SAVs was intended as a worst case scenario for leaflet dynamics, because the SAV stent limits the expansion of the Sapien XT’s commissures. The alignment should have had no impact on the hydrodynamics, as Azadani et al. found no significant influence of the alignment on the hydrodynamics of Sapien XT replicas [21]. Although great care was taken to align the commissures during expansion, commissural alignment was not always fully achieved in the Aspire and the Mosaic (Fig. 2A and C) valves. A slight tilt of the Sapien XT in the Aspire towards the lower right side resulted in the lower leaflet touching the stent during systole (Fig. 3A–C). Because the Aspire and Mosaic valves were at the lower range of the sizing recommendations, we concluded that the ID of the SAVs was probably responsible for the misalignments.

The slight tilt of the Sapien XT within the Aspire valve resulted in asymmetrical closing of the Sapien XT’s leaflets. This situtation should theoretically result in an increase in the closure volume. However, we could not measure a significant difference compared to the Sapien XT in the Mosaic, which has a comparable ID and design. Thus, the slight tilt does not seem to have an impact on the hydrodynamic performance. However, this uneven leaflet motion adds additional stress to the Sapien XT’s leaflets and may well reduce their durability. In a clinical setting, the limited ability to manoeuvre the prosthesis and calcifications within the landing zone likely often results in a slight tilt of a TAV. Consequently, uneven leaflet motion may occur more frequently than anticipated in clinical ViV procedures and limit the durability of TAVs.

With ViV procedures being performed with increasing frequency, a strategy to deal with the possibility of repeated valve replacement in patients with a life expectancy exceeding the predicted durability of the bioprosthesis should take the design of the SAV into consideration. In this study, an SAV with a large valve area provided better conditions for positioning a TAV in subsequent ViV-TAVI and improved the hemodynamic results. Regarding specific SAV designs, the present study suggests that valves with externally mounted leaflets have an increased risk of paravalvular leakage between the Sapien XT and the SAVs. In addition, externally mounted leaflets in SAVs are a risk factor for coronary obstruction in ViV implantations [22–24]. When subsequent ViV is foreseeable, we recommend that the physician avoid using an SAV with externally mounted leaflets and choose a bioprosthesis with the largest possible ID.

LIMITATIONS

For this study, we used non-damaged prostheses provided by the manufacturers. However, calcifications and pannus ingrowth were not modelled in this study but can significantly influence the heamodynamics of ViV implantations. Furthermore, we calculated the EOA from the pressure differences using the Gorlin formula as required by ISO 5840-3:2013. Differences between our values and those obtained with clinical data may occur, because data aquisition uses other methods and calculations are done differently [19, 25].

CONCLUSION

We showed that all ViV pairs met the criteria of the ISO standards and guidelines. However, ViV under expansion due to an SAV prosthesis with a small ID may lead to increased MPG, decreased EOA, pin-wheeling and uneven leaflet motion. Furthermore, the materials and mountings of the leaflets appear to be crucial design features that influence the performance of the ViV because internally fixed leaflets provided a better seal while externally fixed leaflets created substantially more paravalvular leakage. The results of this study may be useful for the treating physician. Particularly for younger patients, a long-term strategy should be developed prior to implantation of the first surgical valve. Prosthesis ID and how the leaflets are mounted seem to be critical issues.

Funding

This work was supported by the ADUMED-Stiftung.

Conflict of interest: Buntaro Fujita: received travel compensation from Edwards Lifesciences™ and Symetis™; Stephan Ensminger: Proctor and consultant for Edwards Lifesciences™; proctor and member of the SAB of JenaValve™; received speaker honoraria from Edwards Lifesciences™ and Symetis™; received travel compensation from Edwards Lifesciences™; Ulrich Steinseifer: Consultant for JenaValve™ and Biotronik™. All other authors declare no conflict of interest.

REFERENCES

Author notes