Author + information
- Received January 16, 2015
- Revision received May 13, 2015
- Accepted May 13, 2015
- Published online February 1, 2016.
- Kentaro Shibayama, MDa,b,
- Hirotsugu Mihara, MDa,
- Hasan Jilaihawi, MDa,
- Javier Berdejo, MDa,
- Kenji Harada, MDa,
- Yuji Itabashi, MDa,
- Robert Siegel, MDa,c,
- Raj R. Makkar, MDa and
- Takahiro Shiota, MDa,c,∗ ()
- aCedars-Sinai Heart Institute, Los Angeles, California
- bHeart Center, Tokyo Bay Urayasu/Ichikawa Medical Center, Urayasu, Japan
- cUniversity of California, Los Angeles, Los Angles, California
- ↵∗Reprint requests and correspondence:
Dr. Takahiro Shiota, Cedars-Sinai Heart Institute, 127 South San Vicente Boulevard, Suite A3411 (TS), Los Angeles, California 90048.
Objectives This study of 3-dimensional (3D) transesophageal echocardiography (TEE) aimed to demonstrate features associated with transvalvular aortic regurgitation (AR) after transcatheter aortic valve replacement (TAVR) and to confirm the fact that a gap between the native aortic annulus and prosthesis is associated with paravalvular AR.
Background The mechanism of AR after TAVR, particularly that of transvalvular AR, has not been evaluated adequately.
Methods All patients with severe aortic stenosis who underwent TAVR with the Sapien device (Edwards Lifesciences, Irvine, California) had 3D TEE of the pre-procedural native aortic annulus and the post-procedural prosthetic valve.
Results In the 201 patients studied, the total AR was mild in 67 patients (33%), moderate in 21 patients (10%), and severe in no patients. There were 20 patients with transvalvular AR and 82 patients with paravalvular AR. Fourteen patients had both transvalvular and paravalvular AR. Patients with transvalvular AR had larger prosthetic expansion (p <0.05), a more elliptical prosthetic shape at the prosthetic commissure level (p <0.01) and more anti-anatomical position (p <0.001), which was defined as malposition of the prosthetic commissures in relation to the native commissures, than the patients without transvalvular AR. Age (odds ratio [OR]: 1.05; 95% confidence interval [CI]: 1.01 to 1.09; p < 0.05) and effective area oversizing (OR: 0.97; 95% CI: 0.93 to 0.99, p <0.05) were associated with mild or greater paravalvular AR.
Conclusions 3D TEE successfully demonstrated the features associated with transvalvular AR, such as large prosthetic expansion, elliptical prosthetic shape, and anti-anatomical position of prosthesis. Additionally, effective area oversizing was associated with paravalvular AR.
Transcatheter aortic valve replacement (TAVR) for severe aortic stenosis (AS) has been shown to be beneficial for a number of high-risk surgical patients (1). However, post-procedural aortic regurgitation (AR) occurs after TAVR in a significant number of patients (1–9). Recent published reports demonstrate that patients with mild or higher levels of AR have a worse outcome compared to patients with no or trivial AR (3,4).
Total AR after TAVR is the combination of paravalvular and transvalvular AR (10). Previous studies have shown that prosthetic undersizing, aortic annular calcification, and prosthetic long-axis malpositioning result in a gap being created between the native aortic annulus and implanted device, leading to paravalvular AR (5–9). However, the mechanism of transvalvular AR has not been evaluated adequately. The primary objectives of this study using 3-dimensional (3D) transesophageal echocardiography (TEE) were to: 1) demonstrate features associated with transvalvular AR; and 2) confirm the fact that a gap between the native aortic annulus and the prosthesis is associated with paravalvular AR.
Between December 2010 and May 2013, 376 consecutive patients underwent TAVR for the treatment of severe AS, using the Sapien (first generation balloon expandable valve) and Sapien XT (second generation balloon expandable valve) valves (Edwards Lifesciences, Irvine, California) and intraprocedural 3D TEE at Cedars-Sinai Heart institute (Los Angeles, California). Of all patients, we excluded patients who had a history of surgical aortic valve replacement or TAVR, those who died during or within 24 h of the procedure, and those with prosthetic high malpositioning, defined as the lowest part of the stent frame above the aortic annulus, by TEE (8,9). We also excluded patients whose 3D TEE data, including pre-procedural native aortic annulus and/or post-procedural prosthetic valve, were not available for analysis. As a result, 201 patients were analyzed in this study. This study was approved by the institutional review board.
Prosthetic sizing of TAVR
Sizing for TAVR was made at the surgeon’s discretion, using a prospective knowledge of cross-sectional multidetector computed tomography (MDCT) dimensions after May 2011 (8). If good cross-sectional MDCT data were unavailable, cross-sectional 3D-TEE data were used for TAVR sizing (9). Traditional cutoffs for annular size by 2-dimensional (2D) TEE measurement were used for prosthetic selection before May 2011 (8).
Comprehensive 2D and Doppler transthoracic echocardiography (TTE) was performed using the iE33 ultrasonography system (Philips Medical Systems, Andover, Massachusetts) routinely, at baseline and 1 day after TAVR. Intraprocedural TEE was performed using the iE33 ultrasonography system with the patient under general anesthesia. The 2D aortic annular diameter was determined in the long-axis mid-systole and end-diastole frames. The 2D aortic annular area was defined as the area assumed to be a circle. The live 3D dataset including aortic root was acquired before and after prosthetic deployment for later analysis.
Grading of AR
In accordance with Valve Academic Research Consortium criteria (10), the degree of total AR was graded as none or trivial, mild, moderate, and severe by post-procedural TTE, if necessary, including quantitative parameters. Subsequently, the severity levels of transvalvular AR (Figure 1) and paravalvular AR were evaluated separately mainly by using post-procedural TTE assessments (11). If it was difficult to differentiate transvalvular AR from paravalvular AR, we additionally used TEE assessment for the severity of transvalvular and paravalvular AR (12). The degree of each AR was judged independently by 2 physicians who were experienced in reading and assessing TAVR echocardiograms.
3D TEE analyses of native aortic annulus and prosthetic valve
All 3D volumetric images were analyzed offline by using commercial software (QLAB, Philips Ultrasound, Andover, Massachusetts). The cardiovascular mode was used for correct alignment of the aortic root. After the mid-systole and end-diastole frames were chosen, the cross-sectional cutting plane perpendicular to the 2 orthogonal long-axis planes (Figures 2A and 2B) was shifted to find the level of the aortic annulus, which was defined as the plane including each lowest cusp hinge point (Figures 2C-I and 2D-I). The pre-procedural native aortic annulus was then traced to determine the maximum and minimum diameters, areas, and perimeters in the mid-systole and end-diastole frames (Figures 2C-II and 2D-II) (9,13).
For 3D analysis of the post-procedural prosthetic valve, the cardiovascular mode was similarly used for correct alignment of the prosthetic valve. The prosthetic valve stent frame after TAVR was assessed at the 2 true cross-sectional levels, aortic annulus level and prosthetic commissure level, which were perpendicular to the 2 true orthogonal long-axis views of the prosthetic valve (Figures 3A and 3B). The prosthetic valve was traced along the mid edge of the prosthetic frame at both levels (Figures 3C and 3D). The maximum and minimum diameters, areas, and perimeters were similarly measured at each level of the prosthetic valve. Prosthetic expansion by 3D TEE was defined as the mid edge area of prosthetic frame divided by the nominal external valve area (14). Eccentricity was calculated as: [1 − (minimum diameter/maximum diameter)] (7,14). Effective area oversizing by 3D TEE was defined as the prosthetic frame area divided by the native aortic annular area in the mid-systole frame (7). Anti-anatomical position was defined as malposition of the prosthetic valve commissures in an angle of nearly 60° (50° to 70°) compared with the native aortic valve commissures (Figure 4A). ΔArea was defined as the difference between the pre-procedural native aortic annular area in the mid-systole frame (Figure 2C-II) and the post-procedural prosthetic valve frame area at the aortic annulus level (Figure 3C).
Data are mean ± SD for continuous variables or as number with percentages for categorical variables. Differences between groups were analyzed by t tests, Mann-Whitney U test, or Kruskal-Wallis test for continuous variables and by chi-square test or Fisher exact probability test for categorical variables, as appropriate. To identify associated factors with mild paravalvular AR or greater, multiple logistic regressions were used. Variables with probability values <0.10 in individual analyses were included in the multivariate analysis. Correlation between the degree of paravalvular AR and ΔArea was analyzed by Spearman rank correlation coefficient. Intra- and interobserver variability for 3D measurements by TEE were analyzed in 15 randomly selected patients. Two-tailed probability values <0.05 were considered statistically significant. Statistical analysis was performed using SPSS version 21.0 software (SPSS Inc., Chicago, Illinois).
A total of 201 patients were analyzed. After TAVR, total AR was none or trivial in 113 patients (57%), mild in 67 patients (33%), and moderate in 21 patients (10%). Greater than moderate total AR was not observed in this study. Mild and moderate transvalvular AR were present in 19 patients (9%) and 1 patient (0.5%) of all patients, respectively. Mild and moderate paravalvular AR were present in 64 patients (32%) and 18 patients (9%) of all patients, respectively. Both transvalvular and paravalvular AR were present in 14 patients (7%).
Patient characteristics and procedures
As shown in Table 1, there were significant differences in left ventricular (LV) ejection fraction, LV diameter, mean transaortic pressure gradient, and aortic valve area between the patients with and those without transvalvular AR. TAVR procedures were similar regardless of the presence of transvalvular AR (Table 2). After post-dilation, moderate or greater paravalvular AR was reduced to less than moderate AR in 11 of 14 patients (79%). There was only 1 patient who had mild transvalvular AR after dilation, but he had already had mild transvalvular AR before post-dilation.
3D features of transvalvular AR
Table 3 shows 2D and 3D measurements of the native aortic annulus and the prosthetic valve. There were no differences in size and shape of the native aortic annulus and prosthetic valve at the aortic annulus level, although the prosthetic valve shape at the commissure level in patients with transvalvular AR was more elliptical than that in the patients without transvalvular AR (p < 0.001). Prosthetic expansion at each level was significantly larger in patients with transvalvular AR than in the patients without transvalvular AR (p < 0.05). There was more anti-anatomical position in the patients with transvalvular AR than in patients without transvalvular AR (p < 0.001). Patients with a small prosthesis had significantly larger prosthetic expansion than those with a large prosthesis (p <0.01). There were not significant differences in the prosthetic shape between the patients with and without anti-anatomical position (annulus level: p = 0.3, commissure level: p = 0.6).
Analysis for the associated factors with paravalvular AR
By multivariate analysis with 8 parameters (age, male, ejection fraction, LV end-diastolic diameter, maximum and minimum diameters and eccentricity at native aortic annulus, effective area oversizing), age (odds ratio [OR]: 1.05; 95% confidence interval [CI]: 1.01 to 1.09; p < 0.05) and the effective area oversizing (OR: 0.96; 95% CI: 0.93 to 0.99; p < 0.05) were identified as significant factors associated with mild or greater paravalvular AR. ΔArea in the patients with none, trivial, mild, and moderate paravalvular AR were 0.38 ± 0.44 cm2, 0.53 ± 0.55 cm2, 0.69 ± 0.54 cm2, and 1.50 ± 0.85 cm2, respectively. Figure 5 shows mild correlation between the degree of paravalvular AR and ΔArea (r = 0.42; p <0.0001).
Intraobserver and interobserver variability
For 3D measurements, both intra- and interobserver variables for area and perimeter of native aortic annulus and prosthetic valve and ΔArea are shown in Table 4.
This is the first study to demonstrate that large prosthetic expansion, elliptical prosthetic shape, and anti-anatomical position are 3D features associated with transvalvular AR. The effective area oversizing (OR: 0.97; 95% CI: 0.93 to 0.99; p < 0.05) was identified as an apparent cause of paravalvular AR.
Prevalence of AR after TAVR
The prevalence of AR has been reported after implantation of the first generation balloon expandable valve in previous papers (1–9). Smith et al. (1) demonstrated that trivial, mild and moderate, and severe paravalvular AR were present in 65.2% and 12.2% of the 287 patients in the TAVR group at 30 days, respectively. They also demonstrated that trivial, mild and moderate, and severe transvalvular AR were present in 63.6% and 1.0% of the 291 patients, respectively (1). Toggweiler et al. (2). showed that 63% of all 88 patients had mild paravalvular AR and that 11% had at least moderate paravalvular AR after TAVR. They also showed that mild or greater transvalvular AR was present in 10% of all patients after TAVR (2). In our study, we found mild and moderate paravalvular AR in 32% and 9% of all 201 patients after TAVR, and mild and moderate transvalvular AR in 9% and 0.5%, respectively, which is consistent with findings in previous reports.
3D features of transvalvular AR
Large prosthetic expansion
The previous study by Yared et al. (4) found that a larger aortic annulus might be associated with central AR. However, the present study showed that there were no significant differences between the native aortic annulus in patients with transvalvular AR and that of patients without transvalvular AR. Although controversy remains over an increase in central regurgitation caused by balloon post-dilation (15), previous papers demonstrated that balloon post-dilation was not associated with any significant increase in central AR (3,16), which is consistent with findings of the present study. The present study demonstrated large prosthetic expansion, which may relate to undersized prostheses, was associated with transvalvular AR. In other words, prosthetic overexpansion may potentially cause mild central leaflet separation, which may lead to transvalvular AR.
Elliptical prosthetic shape
The study by Binder et al. (7) demonstrated that prosthetic inflow and outflow eccentricity rates were 2.4% to 3.5% and 2.4% to 4.3%, respectively. The present study showed that the eccentricity at prosthetic annulus and commissure levels in all patients were 3.3 ± 2.7% and 4.7 ± 3.7%, respectively. Therefore, the present study findings are consistent with those in the paper by Binder et al. (7) in regard to first generation balloon expandable valve prosthetic circularity. On the other hand, the fluoroscopic study by Zegdi et al. (17) revealed that the distal extremity of the first generation balloon expandable valve prosthesis (supporting the free edge of the leaflets) was not strictly circular. The present 3D TEE study was consistent with this previous study by Zegdi et al. (17) in terms of prosthetic eccentricity, particularly in the patients with transvalvular AR. It is natural that elliptical prosthetic shape at commissure level results in inequality of the intercommissural distance; consequently that may lead to valve distortion and transvalvular AR.
Anti-anatomical position did not significantly affect the prosthetic shape in the present study. On the other hand, the paper by Zegdi et al. (18) demonstrated that the severity of valve distortion markedly depended on positioning of the prosthetic commissures inside the orifice at the prosthetic commissure level. Although a self-expandable stent was used in that paper (18), anti-anatomical position of the prosthetic valve in relation to the native aortic annulus may not fit the original structure well, which is consistent with the present study. This abnormal positioning of the prosthetic valve may cause valve distortion and transvalvular AR.
Impact of transvalvular AR on TAVR
Because of the low rate of moderate and severe transvalvular AR (only 1% in this study) and the unknown long-term outcome of transvalvular AR (1), it has been presumed that transvalvular AR has a more benign impact on TAVR than paravalvular AR. However, there were more patients with transvalvular AR after TAVR (65%) than after surgical aortic valve replacement (44%; p < 0.001) according to the paper by Smith et al. (1). Cereijo et al. (19) demonstrated that there was a strong relationship between mechanical stressors, which results from valve distortion and early calcification in bioprostheses, in a human study. Consequently, even mild transvalvular AR due to valve distortion may lead to the development of structural valve deterioration.
Associated factor with paravalvular AR
The previous paper by Delgado et al. (20) demonstrated that differences between the aortic valve annulus areas and internal areas of the deployed prosthesis at the ventricular level were larger in patients with moderate AR than in patients with no or mild AR (p = 0.028). Furthermore, the paper by Binder et al. (7) recently showed that the effective area oversizing by MDCT in patients with mild or more paravalvular AR (n = 67) was significantly smaller than that in patients with no paravalvular AR in individual analysis (p = 0.025). The present study also showed that the effective area oversizing by 3D TEE was a predictor of paravalvular AR by multivariate analysis (p < 0.05), which is consistent with previous papers (7,20). Therefore, a small effective area oversizing, which was caused by a significant mismatch between a larger native aortic valve annulus area and a smaller prosthesis area, may increase the risk of developing mild or more paravalvular AR.
Novelty and clinical implication for TAVR
To the best of our knowledge, this is the first study to evaluate the features of transvalvular AR by using intraprocedural 3D TEE. In addition, 3D anatomy of the prosthetic valve was also successfully analyzed by 3D TEE, although previous papers demonstrated that MDCT provided valuable information on prosthesis deployment (7,20). The post-implanted prosthetic geometry, such as expansion and eccentricity, may be related to the mechanism of transvalvular AR. To prevent abnormal prosthetic geometry, pre-procedural 3D imaging by MDCT or 3D TEE has a critical role in planning for TAVR. Furthermore, anti-anatomical position of prosthetic valve for native aortic valve might be a novel concept to clarify mechanism of transvalvular AR after TAVR. An anatomical deployment of the prosthetic valve may be possible if the markers of the prosthetic commissures are visible by intraprocedural TEE.
First, we did not analyze the comparison between the first generation balloon expandable valve and CoreValve (Medtronic, Edgewater, Maryland) by 3D TEE. Second, this report is not a technical validation of the accuracy of 3D TEE for analysis of the aortic annulus, which would require a comparison between 3D TEE and MDCT. However, it has been established that commercial 3D TEE software can analyze precise morphological anatomy of the aortic annulus in previous reports, which have demonstrated strong intermodal and intersubject correlation between 3D TEE and MDCT (9,13). Third, 3D analysis for accurate tracing of the prosthetic outer frame can be difficult. However, it was possible to trace the mid edge of the prosthetic frame as shown in Figure 3. Fourth, 3D assessment of the aortic annulus was performed immediately before and after TAVR; however, AR grade assessment after TAVR was done by TTE 1 day after TAVR. Finally, the number of patients with transvalvular AR (n = 20) was too small to perform a detailed analysis. Therefore, further evaluation for the differences in the baseline characteristics and the outcome of transvalvular AR is required.
3D assessment of the features associated with transvalvular AR, such as large prosthetic expansion, elliptical prosthetic shape, and anti-anatomical position of prosthesis, was successfully performed by 3D TEE. Additionally, effective area oversizing was identified as a predictor of paravalvular AR.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: Evaluation of 3D anatomy of the prosthetic valve using intraprocedural TEE showed that AR after TAVR was associated with large prosthetic expansion, elliptical prosthetic shape, and anti-anatomical position of the prosthesis. A mismatch between a larger native aortic valve annulus area and a smaller prosthesis found by intraprocedural 3D TEE may increase the risk of developing mild or greater paravalvular AR.
TRANSLATIONAL OUTLOOK: Abnormalities related to transvalvular AR after TAVR, as found in this study, may contribute to further deterioration of the prosthesis, warranting careful prospective studies to assess the long-term prognosis of these patients.
The authors thank Dr. Minoru Tabata for statistical review, Dr. Maiko Shiota for careful professional assistance with the manuscript, and Dr. and Mrs. Paul I. Terasaki for their kind support and encouragement.
Dr. Jilaihawi is a consultant for Edwards Lifesciences, St. Jude Medical, and Venus Medtech. Dr. Makkar has received research grants from Edwards Lifesciences, Medtronic, Abbott, Capricor, and St. Jude Medical; and is a proctor for Edwards Lifesciences and consultant for Medtronic. Dr. Shiota is a speaker for Philips. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Ami E. Iskandrian, MD, served as Guest Editor for this paper.
- Abbreviations and Acronyms
- aortic regurgitation
- aortic stenosis
- left ventricular
- multidetector computed tomography
- transcatheter aortic valve replacement
- transesophageal echocardiography
- transthoracic echocardiography
- Received January 16, 2015.
- Revision received May 13, 2015.
- Accepted May 13, 2015.
- American College of Cardiology Foundation
- Toggweiler S.,
- Humphries K.H.,
- Lee M.,
- et al.
- Hayashida K.,
- Lefèvre T.,
- Chevalier B.,
- et al.
- Yared K.,
- Garcia-Camarero T.,
- Fernandez-Friera L.,
- et al.
- Athappan G.,
- Patvardhan E.,
- Tuzcu E.M.,
- et al.
- Binder R.K.,
- Webb J.G.,
- Toggweiler S.,
- et al.
- Jilaihawi H.,
- Kashif M.,
- Fontana G.,
- et al.
- Jilaihawi H.,
- Doctor N.,
- Kashif M.,
- et al.
- Kappetein A.P.,
- Head S.J.,
- Généreux P.,
- et al.
- Zamorano J.L.,
- Badano L.P.,
- Bruce C.,
- et al.
- Sinning J.M.,
- Vasa-Nicotera M.,
- Chin D.,
- et al.
- Ng A.C.,
- Delgado V.,
- van der Kley F.,
- et al.
- Unbehaun A.,
- Pasic M.,
- Dreysse S.,
- et al.
- Nombela-Franco L.,
- Rodés-Cabau J.,
- DeLarochellière R.,
- et al.
- Zegdi R.,
- Ciobotaru V.,
- Noghin M.,
- et al.
- Delgado V.,
- Ng A.C.,
- van de Veire N.R.,
- et al.