Author + information
- Received June 10, 2015
- Revision received August 11, 2015
- Accepted August 12, 2015
- Published online December 28, 2015.
- Prem A. Midha, MS∗,
- Vrishank Raghav, PhD∗,
- Jose F. Condado, MD, MS†,
- Sivakkumar Arjunon, PhD∗,
- Domingo E. Uceda, BS∗,
- Stamatios Lerakis, MD†,
- Vinod H. Thourani, MD†,
- Vasilis Babaliaros, MD† and
- Ajit P. Yoganathan, PhD∗∗ ()
- ↵∗Reprint requests and correspondence:
Dr. Ajit P. Yoganathan, Georgia Institute of Technology & Emory University, Wallace H. Coulter Department of Biomedical Engineering, Technology Enterprise Park, Suite 200, 387 Technology Circle, Atlanta, Georgia 30313-2412.
Objectives The aim of this study was to investigate the hemodynamic performance of a transcatheter heart valve (THV) deployed at different valve-in-valve positions in an in vitro model using a small surgical bioprosthesis.
Background Patients at high surgical risk with failing 19-mm surgical aortic bioprostheses are not candidates for valve-in-valve transcatheter aortic valve replacement, because of risk for high transvalvular pressure gradients (TVPGs) and patient-prosthesis mismatch.
Methods A 19-mm stented aortic bioprosthesis was mounted into the aortic chamber of a pulse duplicator, and a 23-mm low-profile balloon-expandable THV was deployed (valve-in-valve) in 4 positions: normal (bottom of the THV stent aligned with the bottom of the surgical bioprosthesis sewing ring) and 3, 6, and 8 mm above the normal position. Under controlled hemodynamic status, the effect of these THV positions on valve performance (mean TVPG, geometric orifice area, and effective orifice area), thrombotic potential (sinus shear stress), and migration risk (pullout force and embolization flow rate) were assessed.
Results Compared with normal implantation, a progressive reduction of mean TVPG was observed with each supra-annular THV position (normal: 33.10 mm Hg; 3 mm: 24.69 mm Hg; 6 mm: 19.16 mm Hg; and 8 mm: 12.98 mm Hg; p < 0.001). Simultaneously, we observed increases in geometric orifice area (normal: 0.83 cm2; 8 mm: 1.60 cm2; p < 0.001) and effective orifice area (normal: 0.80 cm2; 8 mm: 1.28 cm2; p < 0.001) and reductions in sinus shear stresses (normal: 153 dyne/cm2; 8 mm: 40 dyne/cm2; p < 0.001), pullout forces (normal: 1.55 N; 8 mm: 0.68 N; p < 0.05), and embolization flow rates (normal: 32.91 l/min; 8 mm: 26.06 l/min; p < 0.01).
Conclusions Supra-annular implantation of a THV in a small surgical bioprosthesis reduces mean TVPG but may increase the risk for leaflet thrombosis and valve migration. A 3- to 6-mm supra-annular deployment could be an optimal position in these cases.
Valve-in-valve (VIV) transcatheter aortic valve replacement (TAVR) is a feasible treatment for patients with failing aortic bioprostheses (1). Although VIV implantation may restore valve function and improve symptoms, adverse events such as elevated post-procedural gradients (28.4%), coronary obstruction (3.5%), device malpositioning (15.0%), and valve leaflet thrombosis (4%) have been reported (2–5). Current transcatheter heart valves (THVs) were not designed for deployment into semirigid bioprostheses. Patients with small surgical bioprostheses (i.e., 19- or 21-mm valves) are generally excluded from VIV TAVR because of the risk for patient-prosthesis mismatch. In these patients, some have experimented with alternative techniques, such as supra-annular deployment, in an attempt to bypass the geometric constraints imposed by the bioprosthesis frame and improve post-procedural gradients (6–8). However, supra-annular deployment could increase the risk for coronary obstruction (9), leaflet thrombosis from flow stagnation within the sinus region (10–14), and THV migration (15). In this in vitro study, we attempted to bridge a gap in VIV TAVR knowledge by assessing hemodynamic and potential safety parameters using a low-profile balloon-expandable THV at different levels of implantation.
The study was conducted in the Georgia Tech Left Heart Simulator (Figure 1), a validated pulsatile flow loop that simulates physiological and pathophysiological conditions of the heart (16,17). The surgical bioprosthesis was mounted in the aortic chamber, which is an idealized rigid acrylic chamber designed to simulate the aortic sinus and ascending aorta (Figure 2). The chamber dimensions were based on published average anatomic measurements (18,19). The flow rate through the valve is adjusted through a LabVIEW version 12 (National Instruments Corporation, Austin, Texas) triggered solenoid system that controls a bladder pump and is measured through an electromagnetic flow probe (600 series, Carolina Medical Electronics, East Bend, North Carolina). Aortic and ventricular pressure waveforms (Figure 1) are tuned through lumped systemic resistance and compliance and are measured with pressure transducers (Deltran DPT-200, Utah Medical Products, Inc., Midvale, Utah) on either side of the valve annulus.
Valve models and deployment
A 23-mm balloon-expandable SAPIEN XT (Edwards Lifesciences, Irvine, California) was deployed within a 19-mm PERIMOUNT (17-mm inside diameter; Edwards Lifesciences) in 4 positions: normal (bottom of the THV stent aligned with the bottom of the surgical bioprosthesis sewing ring) and 3, 6, and 8 mm above the normal position. Please note that the SAPIEN XT is considered a low-profile TAVR device. In all supra-annular positions, a second balloon inflation was performed to flare the aortic end of THV (“flower pot” geometry; Figure 3).
Mean transvalvular pressure gradients (TVPG) were measured and used as surrogates of VIV performance. The working fluid was a 3.5-cSt saline-glycerin solution (approximately 36% glycerin by volume in 0.9% NaCl) to match the kinematic viscosity of blood. The flow rate was tuned for a mean cardiac output of 5 l/min and a systolic duration of 35%. The resistance and compliance were then adjusted to ensure a diastolic aortic pressure of 80 mm Hg and a systolic pressure of 120 mm Hg (Figure 1). Two hundred consecutive cardiac cycles of hemodynamic data (aortic pressure, ventricular pressure, and flow rate) were collected for each test condition at a 1-kHz data-sampling rate using a custom LabVIEW Virtual Instrument.
Aortic valve orifice area
Geometric orifice area (GOA) was determined through en face high-speed imaging using a Karl Storz 88630 SF borescope (Karl Storz, Tuttlingen, Germany) connected to a Basler A504k high-speed charge-coupled device camera (Basler Corporation, Exton, Pennsylvania). Four cardiac cycles (n = 4) of data were collected for each implantation position. The images were manually segmented and scaled to determine peak systolic GOA (Figure 4). Effective orifice area (EOA) was computed using the Gorlin equation (20):[Equation 1]
High-speed particle image velocimetry experiments were conducted using the flow system described previously. The purpose of these experiments was to obtain time-resolved measurements of the flow velocities in the sinus (Figure 2) as a means of assessing relative risk for stagnation-induced thrombosis. A diode-pumped solid-state laser (2 W, 532 nm, Shenzhen optlaser Technologies, Shenzhen, China) was used as the light source, a custom LaVision scanner (LaVision, Göttingen, Germany) was used to convert the continuous beam into a high-frequency pulse, and a complementary metal-oxide semiconductor camera (Phantom Miro M/R/LC123, 1,920 × 1,200 pixels, 730 frames/s; Vision Research, Wayne, New Jersey) was used to image the particles in a single plane. Fluorescent polymeric rhodamine-B particles of diameters 1 to 20 μm were used as seeding particles to visualize the flow field under laser illumination. The resultant velocity field in the sinus was used to calculate the viscous shear stress (VSS) field throughout the cardiac cycle and is given by the following expression:[Equation 2]
The VSS data presented in this work were computed over 9 cardiac cycles. This velocity field–dependent measure is critically linked to thrombotic potential of the leaflets of the valve (21,22).
The pullout force was measured as a means of determining the relative embolization risk for each VIV implantation position. The bioprosthesis was sutured to the bottom of a rigid acrylic chamber filled with saline using 2.0 Ethibond sutures (Ethicon US, Somerville, New Jersey). A single continuous 2.0 Ethibond suture was also attached to the THV stent at 3 equally spaced locations to ensure equally distributed tension (Figure 5). The THV was then deployed into the surgical bioprosthesis at the desired location. The THV harness was attached to a MARK-10 Series 3 digital force gauge (MARK-10, Copiague, New York), which records at 10 Hz. Force was applied gradually until the valve migrated and the force measured by the gauge decreased. Because of the “flower pot” configuration of the THV, the primary concern was antegrade migration into the ascending aorta. Each test condition was repeated 4 times.
Embolization flow rate
In addition to pullout forces, embolization risk was assessed by subjecting each THV deployment to gradually increasing steady antegrade flow. The flow rate at which each THV deployment embolized was recorded. Each test was repeated 4 times.
The data are presented as mean ± SD. Normality of all data was tested using the Anderson-Darling method. One-way analysis of variance was used to analyze independent sample sets with Tukey’s post-hoc test for comparisons between multiple groups. The pullout force data were not normally distributed, and therefore, a Mann-Whitney U test was used instead. Values of p < 0.05 were considered to indicate statistical significance, and the analysis was done using SPSS Statistics for Mac version 20.0 (IBM Corporation, Armonk, New York).
The TVPG is reported as a mean measured over 299 cycles of gathered data (Figure 6). The dashed red line indicates the threshold for device success as defined by Valve Academic Research Consortium-2 (23) and American Heart Association and American College of Cardiology (24) guidelines. As expected, our normal VIV position resulted in a significantly higher mean TVPG than the surgical bioprosthesis alone (33.10 ± 0.37 mm Hg vs. 13.07 ± 0.16 mm Hg; p < 0.001). These TVPGs progressively decreased with each supra-annular implantation to 24.69 ± 0.38 mm Hg, 19.16 ± 0.26 mm Hg, and 12.98 ± 0.25 mm Hg at supra-annular 3, 6, and 8 mm, respectively. At 8 mm above normal implantation, the mean TVPG was similar to the control case and was the lowest TVPG for any VIV position.
The reduction in mean TVPG with increasingly supra-annular VIV deployment was due primarily to increases in GOA (Figure 7) through further flaring of the THV (more exaggerated “flower pot” geometry). The dashed line in the figure indicates the threshold for severe aortic valve stenosis (24). The GOA of the control case was 1.78 ± 0.01 cm2 and dropped to 0.83 ± 0.01 cm2 after deployment in the normal position (p < 0.001). At successively supra-annular positions, the GOA increased from 0.99 ± 0.01 cm2 to 1.2 ± 0.01 cm2 and to 1.6 ± 0.02 cm2 (p < 0.001).
The VSS results are maximum values obtained after spatial integration of the shear stress fields over the area of the sinus at each time point in the cardiac cycle. Figure 8 illustrates the variation of maximum VSS among THV deployments. It was observed that the VSS decreases with increasing deployment height of the THV, ranging from 1.53 ± 0.21 to 0.40 ± 0.10 dyne/cm2 (p < 0.01). Given that human blood has been shown to form aggregates at shear rates <46 s−1 (1.61 dyne/cm2), the stagnation-induced thrombosis safety threshold lies above all VIV deployments (14).
The pullout forces are reported as mean ± SD. With each successive supra-annular deployment, the contact area between the THV and surgical aortic valve replacement reduced, resulting in lower required force for migration. Although the THV did not migrate under physiological pulsatile flow conditions at any of the deployment positions, the measured pullout force reduced drastically between normal and 3-mm supra-annular deployment positions (1.55 vs. 0.9125 N, p = 0.029) and steadily reduced across the 3- to 8-mm positions, as shown in Figure 9 (p < 0.06).
In this study, migration into the left ventricle was not a concern, because the supra-annular implantation provides additional “geometric” resistance under retrograde flow. Dwyer et al. (25) derived estimates of fluid forces on a transcatheter valve on the basis of pressure gradients, viscous forces, and momentum changes and showed that pressure forces accounted for approximately 75% of the total forces. On the basis of these results, a baseline safety threshold was computed (Figure 9):[Equation 3][Equation 4]where GOASAVR is the maximum surgical bioprosthesis GOA on the basis of the internal diameter of the valve, and GOATAVR is the GOA determined using en face imaging of the deployed THV (Figure 4). This threshold represents the theoretical lower limit of pullout force necessary to avoid embolization. Applying a theoretical additional 10% stenosis by area to this threshold yields the upper safety threshold shown in Figure 9. It should be noted that the 3-mm implantation position does not meet this upper safety threshold.
Embolization flow rate
The embolization flow rates followed a similar trend as the pullout forces, though the separation between deployments was not as distinct (Figure 10). The normal deployment embolized at 32.91 ± 0.85 l/min, while the supra-annular deployments of 3, 6, and 8 mm embolized at 31.57 ± 0.75 l/min, 29.37 ± 0.73 l/min, and 26.06 ± 1.27 l/min, respectively. The safety threshold of 30 l/min was defined by a peak instantaneous systolic flow rate through the aortic valve in a healthy adult with a cardiac output of 5 to 6 l/min (26). Most TAVR patients have some degree of impaired cardiac output, making 30 l/min peak instantaneous flow rate a fairly conservative threshold.
The results of this study (Table 1) suggest that supra-annular THV implantation can lead to lower mean TVPG than “normal” implantation after VIV TAVR in patients with small surgical bioprostheses (∼19 mm). We found that increasingly supra-annular deployment resulted in even lower TVPGs, ultimately reaching similar values to our control (a normal functioning bioprosthetic valve). This improvement of gradients can be explained by a better expansion of the downstream portion of the THV at supra-annular locations, demonstrated by the larger GOA measurements and EOA calculations at these levels. Thus, high-implantation VIV TAVR could be performed with good outcomes in patients at high surgical risk who, to this date, are not considered candidates because of the sizes of their original surgical bioprostheses. Moreover, prior clinical registries have reported that VIV TAVR in patients with small surgical bioprostheses (defined as ≤21 mm) have higher mortality than patients with larger surgical bioprostheses who undergo this same procedure, mostly because the former are associated with an increased rate of patient-prosthesis mismatch and elevated TVPGs (27). Although we focused only on the hemodynamic data of high-implantation VIV TAVR in a 19-mm surgical bioprosthesis, we propose that such a procedure may benefit patients with surgical bioprostheses of other sizes.
The observed benefit of high THV implantation measured by mean TVPG could be eclipsed by an increased risk for leaflet thrombosis. Increasingly supra-annular deployment resulted in reduced sinus velocities and VSS levels, which could increase the risk for leaflet thrombosis from flow stagnation (13,14). Recent evidence suggests that thrombosis is an underreported problem affecting THVs. However, the pathophysiology of this process is poorly understood, with most thrombus deposition occurring on the valve leaflets (5,28–30). Although none of our VIV TAVR deployment positions yielded VSS levels higher than the currently understood safety threshold values for thrombus formation, the exact clinical significance of this flow stagnation remains unclear and needs further investigation. Traditionally, thrombosis is discussed in terms of Virchow’s triad (materials, biochemistry, and fluid flow), and the conditions that influence thrombosis risk are highly patient specific. Our in vitro model did not incorporate factors such as aortic distensibility and coronary flow and does not account for patient-specific anatomy or physiology, which can alter thrombosis risk. Because the intention of our study was to understand the fluid mechanics and hemodynamic implications of an alternative deployment, we focused on inspecting the differences observed in thrombosis risk strictly on the basis of fluid mechanics metrics.
We also observed that supra-annular THV implantation resulted in a reduction of force necessary to dislodge the THV. Interestingly, we found that the largest drop in pullout force occurred between the normal and the 3-mm deployments, possibly because of a reduction of the contact area with the internal surface of the surgical bioprosthesis suture cuff. The coefficient of friction between the THV stent and the surgical bioprosthesis suture cuff is likely substantially higher than between the THV stent and the surgical bioprosthesis leaflets. Therefore, there is a gradual decrease in the pullout force at further supra-annular positions, where the contact area of the THV stent with the suture cuff is minimal. Although these lower forces could translate to an increased risk for valve migration, especially in situations of high cardiac output, the calculated (and conservative) baseline safety threshold for valve migration was never reached under physiologic pulsatile flow conditions, even at the highest THV implantation (8 mm). Furthermore, the commonly seen calcification and fibrosis in failing bioprostheses is likely to increase the amount of force required to dislodge the THV in a patient, suggesting that in a worst-case scenario, though possible, antegrade THV embolization is unlikely to occur.
On the basis of this in vitro evidence and threshold values, we recommend deployment of a SAPIEN XT valve between 3 and 6 mm supra-annular in a patient with a failing 19-mm PERIMOUNT. Although we acknowledge the limitations of this study, we would like to emphasize that clinicians must consider patient-specific anatomic characteristics and carefully weigh the benefit of high THV implantation in reducing post-procedural gradients against the potential risk for valve leaflet thrombosis and device migration in potential candidates for VIV TAVR. Similar in vitro studies using other surgical bioprostheses and TAVR devices, including the Medtronic CoreValve, are critically important to optimize VIV performance and patient outcomes.
WHAT IS KNOWN? Patients with small surgical bioprostheses are generally excluded from VIV TAVR because of the risk for patient-prosthesis mismatch. We performed a risk-benefit analysis of supra-annular deployment of a THV in one such simulated in vitro patient.
WHAT IS NEW? On the basis of the evidence from this work, we recommend VIV deployment of a THV between 3 and 6 mm supra-annular in a patient with a small failing bioprosthesis. However, clinicians are strongly urged to consider patient-specific anatomic characteristics and carefully weigh the benefit of high THV implantation in reducing post-procedural gradients against the potential risk for valve leaflet thrombosis and device migration in potential candidates for VIV TAVR.
WHAT IS NEXT? Similar in vitro studies using other surgical bioprostheses and TAVR devices, including the Medtronic CoreValve, are critically important to optimize VIV performance and patient outcomes.
The authors acknowledge the members of the Cardiovascular Fluid Mechanics Laboratory for their assistance and feedback.
The work at the Cardiovascular Fluid Mechanics Laboratory at the Georgia Institute of Technology was funded through discretionary funds available to Dr. Yoganathan through the Wallace H. Coulter Endowed Chair. Dr. Thourani is a consultant or researcher for Edwards Lifesciences, Medtronic, St. Jude Medical, Sorin Medical, Boston Scientific, Abbott Medical, and DirectFlow Medical. Drs. Babaliaros and Lerakis are consultants or researchers for Edwards Lifesciences and Abbott Medical. Dr. Yoganathan is a consultant or researcher for St. Jude Medical and Boston Scientific. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- effective orifice area
- geometric orifice area
- transcatheter aortic valve replacement
- transcatheter heart valve
- transvalvular pressure gradient
- viscous shear stress
- Received June 10, 2015.
- Revision received August 11, 2015.
- Accepted August 12, 2015.
- American College of Cardiology Foundation
- Azadani A.N.,
- Jaussaud N.,
- Matthews P.B.,
- Ge L.,
- Chuter T.A.M.,
- Tseng E.E.
- Hansson N.,
- Leetmaa T.,
- Leipsic J.A.,
- et al.
- De Marchena E.,
- Mesa J.,
- Pomenti S.,
- et al.
- Chevalier F.,
- Leipsic J.,
- Généreux P.
- Dvir D.,
- Leipsic J.,
- Blanke P.,
- et al.
- Falahatpisheh A.,
- Kheradvar A.
- Gorbet M.B.,
- Sefton M.V.
- Saikrishnan N.,
- Kumar G.,
- Sawaya F.J.,
- Lerakis S.,
- Yoganathan A.P.
- Kappetein A.P.,
- Head S.J.,
- Généreux P.,
- et al.
- Nishimura R.A.,
- Otto C.M.,
- Bonow R.O.,
- et al.
- Boron W.F.,
- Boulpaep E.L.