Author + information
- Received March 12, 2018
- Revision received May 22, 2018
- Accepted May 29, 2018
- Published online August 15, 2018.
- Andreas Fuchs, MD, PhDa,b,
- Klaus F. Kofoed, MD, PhD, DMSca,b,
- Sung-Han Yoon, MDc,
- Yannick Schaffnerd,
- Gintautas Bieliauskas, MDa,
- Hans Gustav Thyregod, MD, PhDa,
- Raj Makkar, MD, PhDc,
- Lars Søndergaard, MD, DMSca,
- Ole De Backer, MD, PhDa and
- Vinayak Bapat, MDe,∗ ()
- aThe Heart Center, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark
- bDepartment of Radiology, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark
- cHeart Institute for Interventional Technologies, Cedars-Sinai, Los Angeles, California
- dFEops, Ghent, Belgium
- eDepartment of Cardiovascular Surgery, Columbia University Medical Center, New York, New York
- ↵∗Address for correspondence:
Dr. Vinayak Bapat, Department of Cardiovascular Surgery, Columbia University Medical Center, 177 Fort Washington Avenue, MHB 7GN-435, New York, New York 10032.
Objectives The aim of this study was to assess the commissural alignment between bioprosthetic and native aortic valve leaflets following surgical aortic valve replacement (SAVR) and transcatheter aortic valve replacement (TAVR) and to investigate its impact on valvular function and coronary filling.
Background Expansion and geometry have been shown to affect leaflets of implanted transcatheter aortic bioprosthesis, but commissural alignment has not been studied.
Methods Pre- and post-procedural multidetector computed tomography (MDCT) of 28 SAVR patients and 212 TAVR patients were analyzed. Commissural alignment between the bioprosthetic (post) and native (pre) aortic valves was categorized as aligned (0° to 15° angle deviation) or as mild (15° to 30°), moderate (30° to 45°), or severe (45° to 60°) commissural misalignment (CMA).
Results With SAVR, 27 of 28 cases (96%) were aligned and 1 had mild CMA. For all types of transcatheter heart valves (THVs), there was random valve implantation with regard to commissural alignment: 22% of THVs were aligned, 25% had mild CMA, 22% had moderate CMA, and 31% had severe CMA. The degree of commissural alignment was not associated with a difference in transvalvular gradient, paravalvular aortic regurgitation, or simulated coronary filling. However, there was a significantly higher rate of mild central aortic regurgitation in those THVs with moderate or greater CMA compared with those THV with mild or less CMA (7.8% vs. 1.1%; p = 0.03).
Conclusions Commissural alignment is excellent in case of SAVR but random in case of TAVR. There is no association between CMA and transvalvular gradient or coronary filling; however, there is a significantly higher rate of mild central aortic regurgitation in case of moderate or greater CMA.
Transcatheter aortic valve replacement (TAVR) has become an established therapeutic option for patients with symptomatic, severe aortic valve stenosis (AS) who are ineligible or at high risk for surgical aortic valve replacement (SAVR) (1–3). In recent years, TAVR technology is also increasingly used to treat patients with an intermediate- to low-risk profile; this practice is supported by results from the NOTION, PARTNER-II, and SURTAVI trials indicating that TAVR is a viable option for patients with a lower risk profile (1,4,5).
Although the randomized trials have demonstrated equivalence or even better early and short-term results for TAVR, concerns over mid-term and long-term results have been raised. The main concerns cited for transcatheter heart valves (THVs) have been leaflet thrombosis and valve durability, compared with surgical heart valves (SHVs) (6,7). With THVs, various factors can affect complete expansion and geometry of the stent frame, as calcifications in the leaflets, annulus, and left ventricular outflow tract (LVOT) are not removed and the depth of implant can be variable. Geometry of the stent in turn can affect function of the leaflets (8–10). Another factor of relevance may be orientation of the THV within the aortic root and its influence on leaflet function, coronary blood flow, and ease of access to the coronary ostia (11,12).
The aim of this study was to assess the commissural alignment of the bioprosthetic and native aortic valves following SAVR and TAVR and investigate its impact on the short-term valvular function and coronary filling.
Patients with severe tricuspid AS who underwent SAVR with a Perimount (Edwards Lifesciences, Irvine, California), Magna (Edwards Lifesciences), Epic (Abbott Laboratories, Abbott Park, Illinois), Trifecta (Abbott Laboratories), or MitroFlow (Sorin Group, Milan, Italy) or TAVR with a SAPIEN 3 (Edwards Lifesciences), Centera (Edwards Lifesciences), Lotus (Boston Scientific, Natick, Massachusetts), Acurate (Boston Scientific), CoreValve (Medtronic, Minneapolis, Minnesota), Evolut R (Medtronic), or Portico (Abbott Laboratories) were included in the study. All TAVR and SAVR procedures were performed at Rigshospitalet in Copenhagen, Denmark, or Cedars-Sinai Hospital in Los Angeles between April 2013 and December 2016.
The patients included in this study were randomly selected; the only exclusion criteria were a bicuspid AS and a glomerular filtration rate <45 ml/min. All patients underwent multidetector computed tomographic (MDCT) scanning twice: the first time was 1 to 3 months prior to the aortic valve replacement (AVR) procedure, and the second time was 1 to 3 months after the AVR procedure.
All patients are part of the SAVORY (Subclinical Aortic Valve Bioprosthesis Thrombosis Assessed With 4D CT) registry (NCT02426307) or the RESOLVE (Assessment of Transcatheter and Surgical Aortic Bioprosthetic Valve Thrombosis and Its Treatment With Anticoagulation) registry (NCT02318342). In accordance with the institutions’ policies, all patients gave written informed consent for AVR and the use of anonymous data for research in accordance with the ethics committee review board approval (RH-2016-111, Suite #04625).
CT examinations were performed according to site-specific institutional CT protocols, comprising contrast-enhanced, multiphasic, electrocardiographic gated cardiac CT imaging using latest generation MDCT scanners. A cardioselective beta-blocker was administered orally approximately 1 h before scanning in patients with heart rates >60 beats/min and no contraindications to this drug. Contrast phases were reconstructed with 0.5-mm slice thickness and a 0.5-mm increment. Anonymized datasets were provided for image analysis at Rigshospitalet. For this analysis, all datasets were transferred to a dedicated post-processing platform (Vitrea 6.3, Vital Images, Minnetonka, Minnesota).
CT image analysis
The commissural orientation was assessed for each patient both in the native aortic valve (on the pre-AVR CT scan) and in the prosthetic aortic valve (on the post-AVR CT scan) in the end-diastolic phase of the heart cycle. Analyses of pre- and post-AVR CT scans were made separately, unaware of findings in the other scan.
To evaluate commissural alignment, the origin of the right coronary artery (RCA) was identified in a cross-section perpendicular to the axis of the aorta. The RCA was rotated around the center of the aorta until it was on top at 12 o’clock in the image and a vertical line could be drawn through RCA and the center of the aorta. At the cross-sectional level of the aortic leaflet coaptation, where the leaflets meet in the middle in a triangular formation, 3 angles through the center of the aorta were measured ad modum Bailey et al. (11): from RCA to the right coronary cusp (RCC)/left coronary cusp (LCC) commissure, from RCA to the LCC/noncoronary cusp (NCC) commissure, and from RCA to the NCC/RCC commissure. For each of these 3 angles, a Δ angle deviation between the pre-AVR and post-AVR CT scan was calculated. From these 3 Δ angle deviations, 1 mean angle deviation could be calculated per patient to indicate the commissural alignment (Figure 1).
Definition of commissural alignment
Commissural alignment between native and prosthetic aortic valves was defined in 4 categories: aligned (angle deviation 0° to 15°), mild commissural misalignment (CMA) (angle deviation 15° to 30°), moderate CMA (angle deviation 30° to 45°), and severe CMA (angle deviation 45° to 60°) (Figure 1).
By standard transthoracic echocardiography, left ventricular ejection fraction, aortic transvalvular gradient (both mean and peak), aortic valve area, aortic regurgitation (AR; both central and paravalvular), and mitral regurgitation were assessed before discharge and at 3-month follow-up. Assessment of AS and AR was according to Valve Academic Research Consortium (VARC)-2 definitions.
Computational fluid dynamics simulations
Computational fluid dynamics simulations were used to investigate the impact of valve alignment on coronary filling. For this purpose, patient-specific models containing LVOT, calcified native leaflets, ascending aorta, and coronary arteries were generated on the basis of pre-operative MDCT images (Mimics version 18.0, Materialise, Leuven, Belgium). MDCT examinations with suboptimal image quality of the coronary arteries were not considered for this analysis. Subsequently, finite-element computer modeling (Abaqus/Explicit version 6.12, Dassault Systèmes, Paris, France) was used to virtually implant an Evolut R THV (Medtronic) in the same position as observed at the post-operative MDCT images, leading to a prediction of device and aortic root deformation (13,14). Skirt and prosthetic leaflets were added to the implanted frame, aligning the prosthetic to the native leaflets (0° angle deviation). Nonalignment was then obtained by rotating the prosthetic skirt and leaflets by 60° corresponding to severe CMA. Finally, a drop in pressure from the ascending aorta to the coronary arteries was quantified for both the aligned and severe CMA configuration using computational fluid dynamics (OpenFOAM version 5, OpenCFD, Bracknell, United Kingdom).
Descriptive statistics are expressed as mean ± SD for continuous variables and as frequency and percentage for discrete variables. The differences in means between groups were determined using the Wilcoxon rank sum test, whereas the Fisher exact test was used to test for associations between discrete variables. To identify independent predictors of central AR, all stent frame variables with p values ≤0.10 on univariate analysis were included in a stepwise multivariate logistic regression model. A 2-tailed p value <0.05 was considered to indicate statistical significance. Statistical analyses were performed using commercially available software (SPSS version 23.0, SPSS, Chicago, Illinois).
In this study, 240 patients with symptomatic, severe AS who underwent AVR were included; of these, 28 patients were treated with SAVR and 212 with TAVR. The mean age was 80 ± 7 years, and the mean Society of Thoracic Surgeons surgical risk score was 3.8% for the entire study population (2.8% in the TAVR group vs. 3.9% in the SAVR group, respectively, p = 0.045). All baseline variables are reported in Table 1.
An overview of the different types of SHVs and THVs implanted can be found in Table 2. The majority of TAVR cases were performed by transfemoral approach (97%), with implantation of 82 balloon-expandable THVs (39%), 88 self-expanding THVs (41%), and 42 mechanically expanding THVs (20%).
On the basis of MDCT analysis, it was shown that 27 of 28 SHVs (96%) had been implanted aligned with the native aortic valve commissures. Only 1 Magna valve had been implanted with mild CMA (Table 2).
The THVs were shown to be randomly oriented within the aortic root when considering commissural alignment with the native aortic valve. In total, 47 THVs (22%) were shown to be implanted aligned with the native valve commissures, whereas 53 THVs (25%) were implanted with mild CMA, 46 THVs (22%) with moderate CMA, and 66 THVs (31%) with severe CMA. This random pattern was observed for all types of THVs (Table 2). Examples of different degrees of CMA following TAVR are shown in Figure 2.
Valvular function of bioprosthetic aortic valves
Echocardiographic control before discharge showed 1 of the 28 SAVR patients (3.6%) with a mean transvalvular gradient ≥20 mm Hg, and no single SHV had a paravalvular leak. It also showed well-functioning THVs, with a mean transvalvular gradient of 9 ± 5 mm Hg (n = 212) and only 3 THVs (1.4%) with mean gradients ≥20 mm Hg. Concerning AR assessment, there was mild central AR in 8 THVs (4%), mild paravalvular AR in 59 THVs (28%), and moderate paravalvular AR in 8 THVs (4%). At 3-month follow-up, these transthoracic echocardiographic findings were nearly unchanged. The moderate paravalvular AR rate in the TAVR group decreased from 4% (n = 8 of 211) to 2% (n = 3 of 190); this decrease did not reach significance (p = 0.295) (Table 3).
Commissural alignment and THV function
When comparing those TAVR patients with mild or less CMA and those with moderate or greater CMA, there was no difference in baseline characteristics, LVEF, transvalvular gradient, aortic valve area, or paravalvular AR (p > 0.05) (Online Table 1, Table 3). However, there was a significantly higher rate of mild central AR on pre-discharge transthoracic echocardiography in THVs with moderate or greater CMA (n = 7 of 112 [6.3%]) compared with THVs with mild or less CMA (n = 1 of 100 [1.0%]) (p = 0.046). This was confirmed on 3-month transthoracic echocardiography (central AR 1.1% vs. 7.8%; p = 0.030) (Table 3, Figure 3A).
In a multivariate regression analysis, stent frame deformation and moderate or greater CMA were shown to be independent predictors of mild central AR with odds ratios of 6.30 (95% confidence interval: 1.05 to 18.85; p = 0.044) and 8.66 (95% confidence interval: 1.05 to 35.72; p = 0.045), respectively (Figure 3B).
Subclinical leaflet thickening was detected in 30 of 212 THVs (14.2%) on post-TAVR MDCT scans. The incidence of leaflet thickening was not different in THVs with mild or less CMA (14 of 100 [14.0%]) compared with those with moderate or greater CMA (16 of 112 [14.3%]) (p > 0.05).
Commissural alignment and coronary filling
Computational fluid dynamics simulations were used to investigate the impact of commissural alignment on coronary filling in 24 patients treated with an Evolut R THV. The drop in pressure from the ascending aorta to the coronary arteries was quantified for both the aligned (0° angle deviation) and severe CMA (60° angle deviation) configuration. Severe CMA did not result in a significantly larger pressure drop (0.9 ± 1.2% for the RCA and 1.7 ± 4.4% for the left main coronary artery, p > 0.05) compared with the perfectly aligned configuration (Figure 4).
In this study, including 240 patients who underwent SAVR or TAVR, we found that: 1) surgical bioprosthetic aortic valves are implanted with commissure-to-commissure alignment to the native aortic valve; 2) the orientation of implanted THVs is random with respect to the commissures of the native aortic valve; 3) this random THV orientation does not translate into a difference in transvalvular gradient, aortic valve area, or paravalvular AR; but 4) THVs implanted with moderate or greater CMA are associated with a significantly higher rate of mild central AR; and 5) in a computational fluid dynamics simulation, coronary filling is not affected by difference in commissural alignment.
TAVR is now a well-established therapeutic option for patients with severe AS. Multiple randomized trials and large studies continue to define its role across the spectrum of lower risk and younger patients. Earlier challenges such as large-caliber delivery systems contributing to a higher incidence of vascular complications, significant incidence of malposition, and paravalvular leak have been overcome to a large extent. As TAVR also increasingly starts to replace SAVR in lower risk and younger patients, one of the questions raised is the durability of THVs (7). All bioprosthetic valves degenerate, and the question being posed is whether it will happen faster with THVs and which factors could be modified to increase their durability.
In SAVR, the calcified leaflets are excised, and the annulus and LVOT are debrided and smoothed out before implantation of the valve. As the surgical bioprosthetic aortic valves are sutured to the annulus under vision, the geometry of the stent frame is maintained, which ensures optimal leaflet coaptation. Stent frame deformation has, however, been reported after SAVR and is known to lead to early structural valve deterioration as changes in stent frame geometry alter the leaflet dynamics (10,15). In TAVR, the device is implanted at variable depth and without excision of the leaflets and debridement of the annulus or LVOT. It has been well documented that asymmetric, noncircular deployment is not uncommon in TAVR, and this can potentially limit THV durability (16). Similarly, depth of implantation has been shown to adversely affect durability (17).
During the early days of SAVR with mechanical valves, it was well known that valve orientation of either the tilting disk valve or bileaflet valve affects blood flow patterns, efficiency of blood flow, and coronary blood flow (18,19). With the introduction of surgical bioprosthetic aortic valves, this aspect has become of less significance, as most of implantations are carried out in an anatomic orientation (i.e., commissures of the bioprosthetic valve are matched to the native aortic valve commissures). This was confirmed in our study, in which all SHVs, except one valve, were implanted with commissure-to-commissure alignment to the native valve. CMA can be observed when SAVR is performed for a true or functionally bicuspid valve, but bicuspid valves were not included in this study. During TAVR, the valve orientation in the aortic root is random, as shown in this study. This was despite the fact that 5 of the 6 TAVR devices have a pre-defined orientation within the delivery system, and the delivery system is introduced in the peripheral vessels in a particular, standardized manner. Only with the SAPIEN XT or SAPIEN 3, the valve orientation within the delivery system could be variable, as the valve is not crimped in a particular orientation. This observed variation in commissural alignment in TAVR can most likely be explained by inter-patient difference in the tortuosity of the access vessels and orientation of the aorta, as well as the different use of access sites (e.g., left vs. right femoral artery).
While studying THV durability in a simulated model, Bailey et al. (11) found that valve orientation within the aortic root is important in this respect. The investigators placed a SAPIEN-like device in 8 possible orientations with reference to original commissure and found that frames were distorted in all orientations to some degree, and stress within the leaflets varied with rotation and increased as the valve orientation changed from 0° to 60°, with the maximum stress measured at a 60° angle, corresponding to severe CMA. In an idealized setting in which the orientation was 0° and fully expanded, the stress was least. On the basis of these results, the investigators concluded that the most preferred orientation was commissure-to-commissure alignment (0° angle), whereas the least preferred orientation was between 30° and 60° and that this could potentially represent a sufficient increase in stress to alter the life span of the device.
An important observation in our study is the significant increase of mild central regurgitation in patients with moderate or greater CMA. The cause of this association can only be hypothesized. In a multivariate regression analysis, 2 THV-related variables, stent frame deformation and moderate or greater CMA, were shown to be independent predictors of mild central AR. Whether there is an “interaction” between both factors cannot be confirmed or ruled out in this study because of the limited sample size. Possibly, both observations could be attributed to stent frame geometry, which in turn affects leaflet coaptation. In addition, this higher rate of mild central AR could also be linked to a less efficient bioprosthetic leaflet closure during the early diastolic phase in case of moderate or greater CMA.
Preliminary work on the effect of THV orientation on coronary blood flow using a computational fluid dynamics simulation model did not demonstrate any variation comparing the best- and the worst-case scenarios in this study. However, it cannot be excluded that THV orientation can still have an impact on coronary flow and filling in the longer term, as observed with different orientations in mechanical surgical valves.
Another important aspect is access to the coronary ostia. As we continue to implant THVs in lower risk and younger patients, it is not inconceivable that significant numbers of patients may need coronary intervention at a later date. The construction of most THVs may allow easier access to coronary ostia if there is commissure-to-commissure alignment. This is well known after SAVR, as a stent post in front of a coronary ostium can be a challenge for coronary catheterization.
Finally, commissural alignment may also be of importance in case a THV-in-THV implantation is considered. In some patients, there is increased risk for coronary artery obstruction when planning for a valve-in-valve procedure. In such cases, the BASILICA procedure, in which an aortic bioprosthetic leaflet is intentionally lacerated, could prevent this coronary artery obstruction (20). However, in case of moderate or severe CMA, this procedure may also have its limitations.
Goals for future design of THVs may include the ability to achieve commissure-to-commissure alignment with the native aortic valve (e.g., with an anatomically based locator) for several reasons: 1) to prevent possible detrimental effects on valvular function such as a higher rate of central AR or increased shear stress on the leaflets; 2) to prevent a negative impact on stent geometry and/or coronary blood flow in the longer term; and 3) in order not to compromise coronary access in case of later need for coronary or redo-intervention.
This study has a limited sample size and is more explorative in nature. In order to study the complex interplay between multiple aspects of the host and implanted THV – and this for different THV types – larger sample sizes will be needed in future studies. Also, a larger patient cohort with longer-term follow-up will be needed to study a possible (late) impact of THV orientation on coronary blood flow.
We report a large variation in commissural alignment after TAVR, independent of the valve type, with increased incidence of central AR in case of moderate or greater CMA. Further studies are needed to assess the effect of CMA on the geometry of the stent frame and long-term valvular function. We also believe that efforts should be made to eliminate CMA, as lower risk and younger patients will increasingly be treated with TAVR in the near future.
WHAT IS KNOWN? Different factors, including expansion and geometry, have been shown to affect leaflets of implanted transcatheter aortic bioprosthesis.
WHAT IS NEW? Commissure-to-commissure alignment of transcatheter aortic bioprosthesis with the native aortic valve has not been studied. This study shows that transcatheter aortic bioprosthesis are randomly implanted within the aortic root and that mild central AR is more present in case of moderate or severe CMA.
WHAT IS NEXT? Whether this may affect stress distribution on valve leaflets and valve durability needs to be studied in longer term follow-up studies. Goals for future design of THVs may include the ability to achieve commissure-to-commissure alignment with the native aortic valve.
Computational fluid dynamics simulations were performed by FEops (Ghent, Belgium).
Dr. De Backer has been a consultant for Abbott. Prof. Dr. Søndergaard has been a consultant for and received institutional research grants from Abbott, Boston Scientific, Edwards Lifesciences, and Medtronic. Dr. Bieliauskas has been a consultant for Abbott, Edwards Lifesciences, and Medtronic; and received an institutional research grant from Boston Scientific. Dr. Bapat has been a consultant for Abbott, Edwards Lifesciences, and Medtronic. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- aortic regurgitation
- aortic valve stenosis
- aortic valve replacement
- commissural misalignment
- computed tomographic
- left ventricular outflow tract
- multidetector computed tomographic
- right coronary artery
- surgical aortic valve replacement
- surgical heart valve
- transcatheter aortic valve replacement
- transcatheter heart valve
- Received March 12, 2018.
- Revision received May 22, 2018.
- Accepted May 29, 2018.
- Thyregod H.G.,
- Steinbruchel D.A.,
- Ihlemann N.,
- et al.
- Beohar N.,
- Kirtane A.J.,
- Blackstone E.,
- et al.
- Reardon M.J.,
- Van Mieghem N.M.,
- Popma J.J.,
- et al.
- Chakravarty T.,
- Sondergaard L.,
- Friedman J.,
- et al.
- Capodanno D.,
- Petronio A.S.,
- Prendergast B.,
- et al.
- Sturla F.,
- Ronzoni M.,
- Vitali M.,
- et al.
- Fuchs A.,
- De Backer O.,
- Brooks M.,
- et al.
- Bailey J.,
- Curzen N.,
- Bressloff N.W.
- Yudi M.B.,
- Sharma S.K.,
- Tang G.H.L.,
- Kini A.
- De Jaegere P.,
- De Santis G.,
- Rodriguez-Olivares R.,
- et al.
- Rocatello G.,
- El Faquir N.,
- De Santis G.,
- et al.
- Morganti S.,
- Brambilla N.,
- Petronio A.S.,
- Reali A.,
- Bedogni F.,
- Auricchio F.
- Travis B.R.,
- Heinrich R.S.,
- Ensley A.E.,
- Gibson D.E.,
- Hashim S.,
- Yoganathan A.P.
- Khan J.M.,
- Dvir D.,
- Greenbaum A.B.,
- et al.