Author + information
- Received April 16, 2018
- Revision received June 4, 2018
- Accepted June 13, 2018
- Published online September 3, 2018.
- Janarthanan Sathananthan, MBChB, MPHa,
- Stephanie Sellers, PhDb,
- Aaron Barlow, BSc, PhDb,
- Rob Fraser, MScd,
- Viktória Stanová, Dipl-Inge,
- Anson Cheung, MDa,
- Jian Ye, MDa,
- Abdullah Alenezi, MDa,
- Dale J. Murdoch, MBBSa,
- Mark Hensey, MB BCh BAOa,
- Danny Dvir, MDf,
- Philipp Blanke, MDa,
- Régis Rieu, PhDe,
- David Wood, MDa,
- Phillippe Pibarot, DVM, PhDc,
- Jonathan Leipsic, MDa and
- John Webb, MDa,∗ ()
- aCentre for Heart Valve Innovation, St Paul’s Hospital, University of British Columbia, Vancouver, Canada
- bCentre for Heart Lung Innovation, St Paul's Hospital, Vancouver, Canada
- cQuebec Heart & Lung Institute, Laval University, Quebec, Canada
- dViVitro Labs, Victoria, Canada
- eLBA UMR_T24, IFSTTAR, Aix-Marseille University, Marseille, France
- fDepartment of Cardiology, University of Washington, Seattle, Washington
- ↵∗Address for correspondence:
Dr. John Webb, St. Paul’s Hospital, 1081 Burrard Street, Vancouver, British Columbia V6Z 1Y6, Canada.
Objectives This study assessed the effect of overexpansion beyond labeled size (diameter) of transcatheter heart valves through an ex vivo bench study.
Background Transcatheter heart valves function optimally when expanded to specific dimensions. However, clinicians may sometimes wish to overexpand balloon-expandable valves to address specific clinical challenges. The implications of overexpansion have assumed considerable importance, and objective information to guide practice is limited.
Methods We evaluated SAPIEN 3 transcatheter heart valves (Edwards Lifesciences, Irvine, California). Valves (diameters of 23, 26, and 29 mm) were expanded to nominal dimensions, and then incrementally overexpanded with balloons sized 1-, 2-, and 3-mm larger than the recommended diameter. Valves underwent visual, microcomputed tomography, and hydrodynamic evaluation at various degrees of overexpansion.
Results SAPIEN 3 valves with labeled diameters of 23, 26, and 29 mm could be incrementally overexpanded to midvalve diameters of 26.4, 28.4, and 31.2 mm, respectively. With overexpansion, there was visible restriction of the valve leaflets, which was particularly evident with the smaller valves. After maximal overexpansion of a 26-mm valve a leaflet tear was observed. High-speed video demonstrated impaired leaflet motion of both the 23- and 26-mm valves and hydrodynamic testing documented a regurgitant fraction for the 23- and 26-mm valves above accepted international standards. The maximally overexpanded 29-mm SAPIEN 3 still had relatively normal leaflet motion and excellent hydrodynamic function. Durability was not specifically evaluated.
Conclusions Overexpansion of balloon-expandable valves is possible. However, excessive overexpansion may be associated with impaired hydrodynamic function, acute leaflet failure, and reduced durability. Smaller valves may be at greater risk with overexpansion than larger valves. Overexpansion is best avoided unless clinical circumstances are compelling.
A number of transcatheter heart valves (THVs) are available for the treatment of severe aortic stenosis, each with distinctly different design features (1). The SAPIEN 3 (S3) valve (Edwards Lifesciences, Irvine, California) is currently the most commonly used THV, with 20-, 23-, 26-, and 29-mm labeled diameters available. The manufacturer recommends deployment of this balloon-expandable THV with a specified inflation volume of diluted contrast to achieve the nominal expanded diameter for each size of THV. The manufacturer states that extensive testing has demonstrated that expansion to these exact dimensions allows for optimal valve hydrodynamic function and durability.
However, there are clinical situations where optimal valve function and durability must be balanced against competing concerns, such as annular rupture due to calcification or paravalvular regurgitation due to a THV that is “too small.” Clinicians might desire a THV intermediate in size between the nominal diameters available from the manufacturer. To accomplish this end, underfilling or overfilling of the delivery system balloon outside of the manufacturers recommendations is common (2,3). The terms soft or hard are commonly used. There are also other clinical situations where clinicians may wish to overexpand the THV beyond nominal diameter such as in patients with large annuli or more recently during valve-in-valve transcatheter aortic valve replacement where bioprosthetic valve fracture is also performed (3–7). The consequences of THV overexpansion in the long term remains unknown.
We sought to assess the effects of overexpanding the S3 valve beyond labeled size (diameter) through an ex vivo bench study.
The valves tested were a 23-, 26-, and 29-mm S3 THV. The S3 THV is made of a cobalt-chromium alloy frame, bovine pericardial leaflets and an adaptive polyethylene terephthalate fabric seal at the inflow level of the valve. The 23-, 26-, and 29-mm S3 valves have expanded heights of 18, 20, and 22.5 mm, respectively, when fully expanded as per manufacturer specifications (8).
In this study, we assessed if the S3 THV could be successfully overexpanded to varying degrees beyond nominal size (diameter). We also assessed the effect of overexpansion on valve/leaflet integrity and function. The achieved diameter by underfilling and overfilling the valve delivery balloon by approximately 10% compared with nominal volume (ml) was also assessed. Nominal size was defined as the labeled diameter size of the THV as per the manufacturer.
Dilatations were first performed using the transfemoral Edwards Commander delivery system (Edwards Lifesciences) alone without a THV to understand the effect of underfilling or overfilling. The delivery system for each valve size were first deployed with nominal volume (17 ml for the 23-mm, 23 ml for the 26-mm, and 33 ml for the 29-mm THVs) and balloon measurements made using scientific digital calipers at the midpoint of the balloon. This step was repeated 10 times to assess for intraobserver variability, and a mean nominal diameter was calculated. Each delivery system was then under and overfilled by ∼10% volume (±1 ml for the 23-mm, ±2 ml for 26-mm, and ±3 ml for 29-mm delivery systems), and balloon measurements were made using calipers. This step was also repeated 10 times and a mean diameter of the delivery balloon diameter recorded for both underfilling and overfilling.
Ex vivo overexpansion
The transfemoral delivery system for each valve size (23, 26, and 29 mm) was used to expand the THV to nominal size (diameter) using the manufacturers recommended nominal filling volume. Overexpansion of the valves was performed in a sequential manner with incremental dilatations of 1 mm up to a maximum of 3 mm above nominal size (i.e., a 23-mm S3 was overexpanded to 24, 25, and 26 mm). True Dilatation balloon valvuloplasty catheters (Bard Vascular, Tempe, Arizona) of various sizes (24, 25, 26, and 28 mm) were used to overexpand the transcatheter heart valves. The True Dilatation balloon is noncompliant allowing high pressure balloon inflation with a consistent balloon diameter. For dilatations where an appropriate sized True Dilatation balloon was not available, a Z-Med II balloon (NuMED, Hopkinton, New York) was used. The Z-Med II is a compliant valvuloplasty balloon, and digital calipers were used to measure the Z-Med balloons to ensure the appropriate balloon diameter was achieved.
At nominal size and with each 1 mm of incremental overexpansion, multimodality imaging was performed. High-resolution photography was performed at the same magnification and same fixed camera height. Micro computed tomography (microCT) was performed both at nominal size and for each 1 mm of valve overexpansion. All images were performed using the Nikon XT H 225 ST microfocus X-ray tomography system (Nikon Metrology, Cambridge, Canada) (Figures 1A and 1B).
Measurements were made using microCT at the inflow, middle, and outflow of the valve (Figure 1B). Inflow of the valve was at the level of the adaptive polyethylene fabric seal. Axial measurements were made using the center of each valve strut as a marker (Figure 1D) to measure diameter (midstrut diameter [MSD]) and area (midstrut valve area [MSVA]). Inner diameter and inner valve area measurements were made within the frame of the valve, and outer diameter measurements were made including the frame of the valve (Figure 1C). To assess for potential blooming artifact the difference in outer and inner diameter was calculated to determine a microCT value of stent strut diameter. This was then compared with a manual measurement of the diameter of the valve stent strut as measured using digital calipers to determine the degree of blooming artifact.
In the event of any potential blooming artifact, a derived inner diameter (DID) and derived outer diameter (DOD) and area were also calculated using the MSD value and manually measured strut diameter.
DID and DOD were calculated using the following equation:where r denotes radius of stent strut as measured by digital calipers.
To calculate the derived inner valve area (DIVA) and derived outer valve area (DOVA) the following equation was used, assuming circularity of the valve:where dr denotes derived radius as calculated from equation 1.
Assessment of leaflet integrity
Valves were carefully examined with each incremental overexpansion step, assessing for leaflet damage and integrity.
Hydrodynamic testing was performed after the final overexpansion step (3 mm above nominal size) using a commercially available pulse duplicator (ViVitro Labs Inc., Victoria, Canada) (Figure 1E). Valves were tested in accordance with International Organization for Standardization (ISO) 5840-3:2013 guidelines for in vitro pulsatile flow testing for heart valve substitutes implanted by transcatheter techniques (9). Valves were placed in a holder fabricated from silicone with a durometer of scale Shore A hardness of 40 ± 5 (Figure 1F). Justification for the selection of sample holder hardness was based on published data on acceptable tissue compliance matched with published data on the silicone material hardness scale (10–12). To prevent paravalvular leakage, testing was conducted with sealing applied around the valve sample perimeter. Test fluid used was 0.9 ± 0.2% sodium chloride test solution maintained at 37 ± 2°C (1 drop of Cosmocil [preservative] [Lonza, Basel, Switzerland] per 1 l).
Overexpanded valves were tested on the aortic side of the pulse duplicator with a spring-loaded disc valve (ViVitro Labs) on the mitral side of the pulse duplicator. Measurements were based on average results taken from 10 consecutive cycles. High-speed video was captured at each step condition. Pulsatile forward flow performance was tested at nominal beat rate of 70 ± 1 beats/min, systolic duration of 35 ± 5%, mean aortic pressure of 100 ± 2 mm Hg, and simulated cardiac outputs of 5.0 ± 0.1 l/min.
Results of transvalvular regurgitant performance were compared with minimum performance requirements in accordance with the ISO 5840-3:2013. This was based on final MSD achieved for each valve. The required transcatheter regurgitant fraction (%) for acceptable performance as per ISO guidelines for overexpanded 23-, 26-, and 29-mm valves were ≤15.0%, ≤17.5%, and ≤20.0% respectively.
Assessment of leaflet coaptation defects and restriction
Objective assessment of leaflet coaptation defects and restriction were made using analysis of the high-speed videos for each of the overexpanded valves, after the final overexpansion step was performed. A 23-mm SAPIEN S3 deployed in a 23-mm annulus was used as a reference. Geometric orifice area (pixel) was calculated by plotting the area on both maximal opening and closing of the valve. A custom-made MATLAB Version 7 (The MathWorks, Natick, Massachusetts) program was used to measure geometric orifice area and determine leaflet restriction.
Leaflet coaptation defects were analyzed using the following equation:
Leaflet restriction was quantified by measuring the following ratio: (distance from the midpoint of the free margin of each leaflet to the stent border divided by the radius of the valve orifice)∗100. Using the 23-mm S3 in a 23-mm annulus as a reference this ratio should be >90%, representing leaflet motion in a nominally deployed valve.
Means were calculated for the ten diameter measurements made for each delivery catheter balloon at nominal volume and with underfilling and overfilling. Statistical analyses were performed using SAS software version 9.4 (SAS Institute Inc., Cary, North Carolina).
Underfilling and overfilling the delivery balloon catheter
The mean diameter (midpoint of the balloon) achieved with nominal volume of the delivery catheter alone was consistent with the stated size of the appropriate THV: 22.9 mm for 23-mm THV delivery system, 26.1 mm for 26-mm system, and 29 mm for 29-mm system. For the 23-mm delivery balloon 10% underfilling and overfilling resulted in a diameter reduction of 0.55 mm or an increase of 0.46 mm, for the 26-mm balloon the relative changes in balloon diameter were 0.92 and 0.82 mm, and for the 29-mm balloon 0.73 mm and 1.28 mm, respectively (Table 1).
Assessment of microct measurement artifact
Although the degree of blooming artifact was minimal there was discrepancy between the microCT-derived and manual caliper measurements of strut diameter. The stent frame diameter as measured by microCT and manual calipers was 0.6 mm and 0.55 mm, respectively. Therefore, the midstrut measurements and calculated DID, DOD, DIVA, and DOVA are reported at nominal size and for each incremental step of overexpansion.
Nominal size and overexpansion of THV
Valve diameter and area measurements at the inflow, midvalve, and outflow of the S3 THV, for both nominal size and incremental overexpansion, are detailed in Table 2. The MSD and MSVA measurements at the midvalve of the THV were preferentially used in this study. DID, DOD, DIVA, and DOVA are also reported (Table 2). The DID and DIVA measurements are similar to the 3-dimensional annular measurements in the sizing guide providing by the manufacturer.
When the THV was deployed as specified by the manufacturer, the achieved midvalve MSD for the 23-, 26-, and 29-mm S3 were 22.8, 26.9, and 28.6 mm, respectively. The 23-, 26-, and 29-mm S3 valves were all able to be incrementally overexpanded beyond their labeled nominal diameter and reached maximal midvalve MSD of 26.4, 28.4, and 31.2 mm, respectively (Table 2, Figure 2).
Variation in measurements across the inflow, middle, and outflow of the THV were observed. There was a “dumbbell” appearance to the compliant balloon with greater dilatation achieved at the inflow and outflow component of the valve than the middle (Table 2). This “dumbbell” effect was predominantly observed when overexpansion was performed with a compliant balloon. The 29-mm valve, which had all incremental overexpansion dilatations performed using compliant balloons, had preferential dilatation at the inflow and outflow compared with the middle of the valve. In comparison, the 23-mm valve, which had all incremental overexpansion dilatations performed with a noncompliant balloon, had consistent measurement across the inflow, middle, and outflow of the valve (Table 2).
Compared with nominal size, each incremental 1-mm overexpansion lead to progressive restriction of the valve leaflets on macroscopic inspection (Figure 2). This affected all 3 S3 sizes but was particularly pronounced in the 23-mm S3. Even at 1-mm overexpansion there was visible leaflet restriction, which increased in severity at 3-mm overexpansion (Figures 3A and 3B). Valve leaflets were less restricted after overexpansion of the 26- and 29-mm S3 compared with the 23-mm S3 (Figure 2).
After maximal overexpansion of the 26 mm S3 there was evidence of leaflet tear on macroscopic inspection. The site of the tear was at point of leaflet insertion into the THV frame (Figures 3C and 3D). No other valves had evidence of leaflet tear or damage on macroscopic examination.
Hydrodynamics varied by valve size. High-speed video of the overexpanded 23-mm valve showed that there was a central coaptation defect on closing, with marked leaflet restriction on opening (Figure 4, Online Video 1). The overexpanded 26 mm valve had a small central coaptation gap. The perforated leaflet was “splayed” on opening, with delayed closure, and also did not close in alignment with the other leaflets (Figure 4, Online Video 2). The overexpanded 29-mm valve had no visible coaptation gap and leaflet motion was symmetrical (Figure 4, Online Video 3).
The mean pressure gradient across the overexpanded 23-, 26-, and 29-mm valves were 8.6 ± 0.1, 6.9 ± 0.1, and 8.5 ± 0.1 mm Hg, respectively. The effective orifice area for the overexpanded 23-, 26-, and 29-mm were 2.0, 2.2, and 2.2 cm2, respectively.
Transvalvular regurgitant fraction for the overexpanded 23-, 26-, and 29-mm valves was 15.9%, 18.9%, and 7.1%, respectively. The overexpanded 23- and 26-mm regurgitant fractions were above minimum acceptable ISO guidelines of ≤15% and ≤17.5%, respectively. The overexpanded 29-mm valve still had excellent hydrodynamic function, with a regurgitant fraction significantly below minimum acceptable ISO guidelines of ≤20% (Figure 5). The regurgitant volumes for the overexpanded 23-, 26-, and 29-mm valves were 8.4 ± 0.7, 9.0 ± 0.3, and 1.2 ± 0.4 ml, respectively.
Assessment of leaflet coaptation defects and restriction
The coaptation defects for the overexpanded 23-, 26-, and 29-mm valves were 1.354%, 0.054%, and 0%, respectively. Leaflet restriction (%) for the overexpanded 23-, 26-, and 29-mm valves were below 90%, indicating restriction in all 3 valves (Figure 6). The nominally deployed 23-mm valve used as a reference had leaflet restriction values >90%.
This bench study demonstrates that the S3 THV can be overexpanded beyond the manufacturer’s recommended nominal diameters. Although minor degrees of overexpansion may be safe, excessive overexpansion risks compromising leaflet integrity, hydrodynamic function, and durability. A better understanding of these issues may have value to implanters where clinical circumstances require informed judgement regarding the role of overexpansion.
Our bench study demonstrated that each size of S3 THV tested could potentially be overexpanded by a specific diameter (0.5 mm for a 23-mm THV, 0.8 mm for a 26-mm THV, and 1.3 mm for a 29-mm THV) by overfilling the delivery catheter balloon by ∼10%. However, with even larger commercially available balloons larger increases in diameter (by as much as 3 mm) can be achieved. Consistency of overexpansion across the valve structure also varied depending on balloon choice. When a noncompliant balloon was used overexpansion was similar at the inflow, middle and outflow of the valve. However, more compliant balloons may result in greater expansion at the outflow, with the inflow constrained by the annulus and fabric seal (3). It is unknown if moderate asymmetric expansion impacts hydrodynamics and durability. The maximal amount a S3 THV can be overexpanded may ultimately be limited by the constraints of the THV frame and external sealing skirt. Although the THV frame may determine the mechanical limit of overexpansion, the more pertinent practical limit of overexpansion is what happens to the tissue leaflets, and the point at which leaflet function is compromised.
Importantly, nominal THV size appeared to be a determinant of leaflet function. Smaller THVs may be at greater risk of compromised leaflet function with overexpansion. We found progressive visible evidence of leaflet restriction with progressive overexpansion across the range of available THV sizes. However, this was most noticeable for the smallest available 23-mm S3, for which just 1 mm of overexpansion resulted in visible leaflet restriction. We were not able to test the even smaller 20-mm S3, however, we did find that larger THVs (26- and 29-mm sizes) had lesser degrees of leaflet restriction, even with greater degrees of overexpansion. Similarly, nominal THV size also appeared to be a determinant of hydrodynamic function with overexpansion. Although the overexpanded 23- and 26-mm valves had unfavorable hydrodynamics, the 3 mm overexpanded 29-mm valve in comparison still had excellent hydrodynamic function, minimal visible leaflet restriction, and a regurgitation fraction still well below the ISO guidelines. This may partly explain the favorable procedural outcomes with modest overexpansion of 29-mm S3 valves in patients with very large annuli exceeding recommended guidelines (4,13).
Although short-term outcomes have been favorable, one must be cognizant of the potential early and unknown late implications of overexpansion. While compromised THV hydrodynamics and durability may be acceptable in patients with a limited life expectancy, reduced durability may be a major concern in patients with the potential for longevity. In such patients, the balance of risks may shift in favor of surgical aortic valve replacement. We did observe leaflet damage in the 26-mm valve, and the mechanism is potentially multifactorial due to both overexpansion, and also leaflet damage due to use of a noncompliant balloon. Given the risks of both early and late leaflet failure, excessive overexpansion should be discouraged. However, an understanding of the issues described in this study may be of value to clinicians when faced with clinical challenges; such as valve selection in a patient with intermediate annulus dimensions and concerns about annular injury or paravalvular regurgitation due to selection of a valve that is too small or too large, or an already implanted valve that is clearly too small.
Ex vivo bench testing may not entirely reflect how a THV will expand in a patient’s native annulus or within a surgical bioprosthesis. The lack of a noncompliant balloons larger than 28 mm in diameter necessitated the use of larger partially compliant balloons which may have introduced some variability, although this reflects clinical practice. The unfavorable hydrodynamics of the overexpanded 26-mm valve is likely skewed due to the leaflet tear. Repetition of bench testing with more THVs would be desirable.
Overexpansion of balloon-expandable valves is possible and reproducible. However excessive overexpansion may be associated with impaired hydrodynamic function and the possibility of early or late leaflet failure. Smaller valves may be at greater risk with overexpansion than larger valves. In general, overexpansion is best avoided unless clinical circumstances are compelling.
WHAT IS KNOWN? There are some clinical situations in which clinicians may wish to overexpand a THV beyond nominal diameter. However, objective information to guide practice is limited.
WHAT IS NEW? Overexpansion of balloon-expandable valves is possible. However, excessive overexpansion may be associated with impaired hydrodynamic function, acute leaflet failure, and reduced durability. Smaller valves may be at greater risk with overexpansion than larger valves.
WHAT IS NEXT? Future studies assessing the impact of overexpansion on long-term durability would be important. Long-term clinical follow-up of overexpansion cases is also desirable.
Dr. Cheung has served as a consultant to Abbot Vascular, Medtronic, and Neovasc. Dr. Ye has served as a consultant for Edwards Lifesciences. Dr. Dvir has served as a consultant for Edwards Lifesciences, Medtronic, and St Jude Medical. Dr. Blanke has served as a consultant for Edwards Lifesciences. Dr. Wood is a consultant to Edwards Lifesciences. Dr Pibarot has received grant support from Edwards Lifesciences and Medtronic. Dr. Leipsic has received institutional research support from Edwards Lifesciences and Medtronic. Dr. Webb has served as a consultant for Edwards Lifesciences and ViVitro Labs. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- micro computed tomography
- derived inner diameter
- derived inner valve area
- derived outer diameter
- derived outer valve area
- International Organization for Standardization
- midstrut diameter
- midstrut valve area
- SAPIEN 3
- transcatheter heart valve
- Received April 16, 2018.
- Revision received June 4, 2018.
- Accepted June 13, 2018.
- 2018 American College of Cardiology Foundation
- Vahl T.P.,
- Kodali S.K.,
- Leon M.B.
- Barbanti M.,
- Leipsic J.,
- Binder R.,
- et al.
- Shivaraju A.,
- Kodali S.,
- Thilo C.,
- et al.
- Barr P.,
- Ormiston J.,
- Stewart J.,
- Nand P.,
- Ramanathan T.,
- Webster M.
- Chhatriwalla A.K.,
- Allen K.B.,
- Saxon J.T.,
- et al.
- Johansen P.,
- Engholt H.,
- Tang M.,
- Nybo R.F.,
- Rasmussen P.D.,
- Nielsen-Kudsk J.E.
- Webb J.,
- Gerosa G.,
- Lefèvre T.,
- et al.
- ↵(2013) Cardiovascular Implants – Cardiac Valve Prosthesis. Part 3: Heart Valve Substitutes Implanted by Transcatheter Techniques (International Organization for Standardization, Geneva).
- Mathur M.,
- McCabe J.M.,
- Aldea G.,
- Pal J.,
- Don C.W.