Author + information
- Received March 8, 2018
- Revision received April 11, 2018
- Accepted April 17, 2018
- Published online May 22, 2018.
- Won-Keun Kim, MDa,b,c,∗ (, )
- Helge Möllmann, MD, PhDd,
- Christoph Liebetrau, MD, PhDa,c,
- Matthias Renker, MDa,
- Andreas Rolf, MD, PhDa,c,
- Philippe Simona,
- Arnaud Van Linden, MDe,
- Mani Arsalan, MDe,
- Mirko Doss, MD, PhDb,
- Christian W. Hamm, MD, PhDa,c and
- Thomas Walther, MD, PhDe
- aDepartment of Cardiology, Kerckhoff Heart Center, Bad Nauheim, Germany
- bDepartment of Cardiac Surgery, Kerckhoff Heart Center, Bad Nauheim, Germany
- cDepartment of Cardiology, Justus-Liebig University of Giessen, Giessen, Germany
- dDepartment of Cardiology, St. Johannes Hospital, Dortmund, Germany
- eDepartment of Cardiac Surgery, Johann-Wolfgang-Goethe University, Frankfurt, Germany
- ↵∗Address for correspondence:
Dr. Won-Keun Kim, Kerckhoff Heart Center, Department of Cardiology, Benekestrasse 2-8, 61231 Bad Nauheim, Germany.
Objectives The aim of this study was to perform a comprehensive analysis of factors that affect procedural outcomes of transcatheter aortic valve replacement using the ACURATE neo prosthesis (Symetis/Boston, Ecublens, Switzerland).
Background Predictors of procedural outcomes using the ACURATE neo prosthesis are poorly understood.
Methods A total of 500 patients underwent transfemoral aortic valve replacement with the ACURATE neo prosthesis. Device landing zone calcification was stratified as severe, moderate, or mild. Anatomic and procedural predictors of second-degree or greater paravalvular leakage and permanent pacemaker implantation were assessed.
Results Post-procedural second-degree or greater paravalvular leakage was more frequent with increasing device landing zone calcification (mild 0.8% vs. moderate 5.0% vs. severe 13.0%, p < 0.001), whereas permanent pacemaker implantation was independent of device landing zone calcification. More severe periannular calcification (odds ratio [OR]: 1.007; 95% confidence interval [CI]: 1.003 to 1.010; p < 0.001), less oversizing (OR: 0.867; 95% CI: 0.773 to 0.971; p = 0.014), the presence of annular plaque protrusions (OR: 2.756; 95% CI: 1.138 to 6.670; p = 0.025), and aortic movement of the delivery system after full deployment (OR: 5.593; 95% CI: 1.299 to 24.076; p = 0.02), and sinotubular junction height (OR: 1.156; 95% CI: 1.007 to 1.328; p = 0.04) independently predicted second-degree or greater paravalvular leakage. Predictors of permanent pacemaker implantation were pre-existing right bundle branch block (OR: 3.122; 95% CI: 1.261 to 7.731; p = 0.01) and more oversizing (OR: 1.111; 95% CI: 1.009 to 1.222; p = 0.03).
Conclusions Successful transcatheter aortic valve replacement using the ACURATE neo device predominantly depends on careful patient selection with appropriate oversizing and taking into account the individual anatomy and calcium distribution of the aortic root.
Transcatheter aortic valve replacement (TAVR) has become the standard therapy for elderly patients with aortic stenosis who are at high surgical risk or inoperable (1). Moreover, recent data have demonstrated that TAVR is noninferior to conventional surgery in intermediate-risk populations (2,3), which has been recently reflected in clinical guidelines (4). The majority of improvements in the field of TAVR are related to accumulation of knowledge and experience and the development of novel devices and delivery systems. The ACURATE neo (Symetis/Boston, Ecublens, Switzerland) is a new-generation self-expanding device that is characterized by an X-shaped stent design with a unique mechanism of deployment (5). Data from the Conformité Européene mark trial and a large post-market registry demonstrate favorable outcomes with a high rate of procedural success and low 30-day and 1-year mortality (6,7). In this study, we analyzed anatomic and procedural predictors of procedural success in a large cohort of patients implanted with the ACURATE neo valve at our institution.
A total of 500 consecutive patients with severe aortic stenosis undergoing transfemoral TAVR using the ACURATE neo prosthesis at our center between May 2012 and September 2017 constituted the study population. The design of the ACURATE neo prosthesis (Figure 1) and details of the implantation technique have been described previously (5). In brief, for stable positioning of the device, it is recommended to keep the delivery system in the outer curvature of the arch and to maintain slight forward tension. A radiopaque intersection line in the stent body, commonly referred to as the “marker band,” is frequently used to indicate the correct annular position. The deployment of the device consists of step 1, in which the upper crown and the stabilization arches are released, followed by the full release of the prosthesis in step 2. The prosthesis size was selected in adherence to the recommendation of the manufacturer on the basis of the maximum area-derived effective annular diameter. The final decision was at the discretion of the operator and, particularly regarding borderline sizes, depended on additional factors, including balloon sizing, patient stature, and device landing zone (DLZ) calcification. Baseline data including demographics, comorbidities, risk scores, and echocardiographic results were drawn from a prospective database. Informed consent was obtained from each patient. The study was conducted in accordance with the Declaration of Helsinki and was approved by the local ethics committee.
Factors with potential impact on procedural success
Anatomic parameters were assessed with pre-procedural multidetector computed tomographic (MDCT) imaging, which was performed using a 64-slice or a 192-slice dual-source scanner (Somatom Definition or Force, Siemens Healthcare, Forchheim, Germany) as previously described (8). Datasets were analyzed offline using dedicated U.S. Food and Drug Administration–approved software (3mensio, Pie Medical, Amsterdam, the Netherlands) by a single reader with extensive experience in cardiac imaging who was blinded to clinical data. In addition to standard measurements, we determined the cover index (100 × [prosthesis diameter − MDCT annular size]/prosthesis diameter) for area- and perimeter-derived annulus diameter in systole and diastole, annular eccentricity (maximum/minimum annular diameter), and the aortoannular angle; we also identified annular plaques with intraluminal protrusion <4 or ≥4 mm (Figure 2) and noted the presence of a bicuspid aortic valve. The total calcium load of the DLZ was measured according to the Agatston method using non-contrast-enhanced MDCT imaging (9). The calcium volume of the DLZ (CVDLZ) was measured on contrast-enhanced MDCT images using a scan-specific threshold as described elsewhere (10). Furthermore, we stratified according to CVDLZ into groups of severe (>75th percentile), moderate (25th to 75th percentiles), and mild (<25th percentile) DLZ calcification (11). The aortic valve complex was divided into 3 regions in the craniocaudal axis for separate determination of the calcium volume of the aortic valve, calcium volume of the left ventricular outflow tract, and calcium volume of the annular region (CVAnn) as specified in Figure 2. The calcium distribution across the 3 leaflets (left coronary, right coronary, and noncoronary) was measured for the aortic valve complex (CVDLZ for the left, right, and noncoronary leaflets) and for each region (calcium volume of the aortic valve for the left, right, and noncoronary leaflets; calcium volume of the left ventricular outflow tract for the left, right, and noncoronary leaflets; and CVAnn for the left, right, and noncoronary leaflets) (Figure 2). Asymmetrical distribution was determined by calculating the maximum absolute difference between the lowest and highest value of the calcium volume of each leaflet region (ΔCVDLZ, Δ calcium volume of the aortic valve, Δ calcium volume of the left ventricular outflow tract, and ΔCVAnn) (12).
The implantation depth, relative expansion, and coaxial position of the prosthesis were analyzed on final angiography as previously described (Online Figure 1) (13). In addition, we recorded the position of the delivery system in relation to the ascending aorta and aortic arch (outer curvature vs. midluminal) immediately before full deployment, a horizontal alignment of the prosthesis (<30° against the vertical plane), the annular position of the marker band, any movement of the stent holder right after full deployment (toward the left ventricle, no movement, toward the ascending aorta), the stability of the device position upon full release (stable vs. upward movement), and the position of the upper crown in relation to the native leaflet calcification (adjacent vs. remote) (Figure 3). The use of pre-dilatation, which was omitted in selected patients (13), and operator experience were also taken into consideration. An operator was defined as experienced after a minimum of 50 TAVR procedures with any device as first operator. The learning curve of the center was analyzed by comparison of the first 100 with the last 100 patients treated with the ACURATE neo device.
Outcomes of interest
The primary endpoints were predictors of second-degree or greater paravalvular leakage (PVL) and permanent pacemaker implantation (PPI), respectively. The secondary endpoints were 30-day mortality and device success according to the Valvular Academic Research Consortium 2 criteria (14). PVL was evaluated post-procedurally after deployment of the prosthesis and post-dilatation when necessary, but before any secondary measures such as implantation of a second valve or surgical valve replacement, and at discharge. For the assessment of PVL, transthoracic echocardiograms were independently reviewed by 2 experienced cardiologists, who were blinded to clinical data, with mutual consent in the case of disagreement, according to established criteria (15). On discharge echocardiography, we also localized the number and position of PVL in a short-axis view and measured the minimum and maximum diameter of the stent prosthesis for calculation of its eccentricity in diastole (Online Figure 2). Follow-up data were obtained at outpatient visits or via telephone interview.
Continuous variables are expressed as median and interquartile range; categorical data are presented as numbers and percentages. Comparison of continuous data was performed using the Mann-Whitney signed rank test for paired data and the Mann-Whitney U test for unpaired data. For categorical data, the 2-sided Fisher exact test was applied. A stepwise logistic regression analysis (forward logistic regression) was carried out to specify independent predictors of second-degree or greater PVL and PPI; after exclusion of multicollinear covariates with variance inflation factors >5, appropriate variables that showed associations in the univariate analysis with p values ≤0.10 were included. A 2-sided p value <0.05 was considered to indicate statistical significance. Statistical analyses were carried out using SPSS version 22.0 (IBM, Armonk, New York).
Patients and procedural results
Baseline characteristics of the study population are summarized in Table 1. The median age was 82.1 years (IQR: 78.8 to 85.3 years), 65.2% were women, and the Society of Thoracic Surgeons score was 4.4% (interquartile range: 3.1 to 6.6). DLZ calcification was severe, moderate, and mild in 124, 252, and 124 patients, respectively. Figure 4 illustrates that second-degree or greater PVL was more common in patients with more severe DLZ calcification, whereas PPI was independent of DLZ calcification. An overview of procedural data is provided in Table 2 and Online Table 1. Device success was achieved in 89.6%. All-cause 30-day mortality was 3.4%.
A comparison of the first 100 and the last 100 cases revealed that over time, patient selection for this device changed, with more distinct oversizing and less calcified DLZs in the last 100 cases with coincident declines in 30-day mortality, device failure, and second-degree or greater PVL (Table 3).
Post-procedure, none or trace, mild, moderate, and severe PVL was noted in 193 (38.8%), 273 (54.8%), 29 (5.8%), and 3 (0.6%) of 499 patients, respectively. Discharge or final echocardiography was available in 499 subjects, with none or trace, mild, and moderate PVL in 199 (39.9%), 276 (55.2%), and 24 (4.8%), respectively. The circumferential distribution of PVL in the short-axis view is illustrated in Online Figure 3. Patients with moderate PVL more frequently had ≥2 leaks, and the leaks were most commonly located at the site of the left coronary cusp.
Predictors of second-degree or greater PVL
Online Table 2 lists the factors that were significantly associated with post-procedural second-degree or greater PVL.
In the multivariate analysis, increased CVAnn, less oversizing (cover index for perimeter-derived annular diameter in diastole), presence of annular plaque protrusions, aortic movement of the delivery system after full deployment, and increased sinotubular junction (STJ) height were identified as independent predictors of post-procedural second-degree or greater PVL (Table 4).
The threshold for prediction of second-degree or greater PVL on the basis of CVAnn was 97.0 mm3 (area under the curve [AUC] = 0.773; 95% confidence interval [CI]: 0.702 to 0.845; p < 0.001), with sensitivity and specificity of 78.1% and 78.1%, respectively. In terms of oversizing, second-degree or greater PVL was discriminated by cover index for perimeter-derived annular diameter in diastole with a threshold of 4.4% (AUC = 0.647; 95% CI: 0.531 to 0.762; p = 0.009; sensitivity 75.1%, specificity 55.6%) or by cover index for perimeter-derived annular diameter in systole with a threshold of 2.5% (AUC = 0.645; 95% CI: 0.535 to 0.755; p = 0.01; sensitivity 79.9%, specificity 46.4%).
In the subgroup with severe DLZ calcification, independent predictors of second-degree or greater PVL were the presence of annular plaque protrusions, increased periannular calcification (ΔCVAnn), and increased STJ height (Online Table 3).
A pre-existing right bundle branch block (RBBB), a higher degree of oversizing (cover index for perimeter-derived annular diameter in systole), and more aggressive pre-dilatation (higher ratio of balloon to annulus size) were significantly associated with the need for PPI (Online Table 4). In the multivariate analysis, independent predictors of PPI were pre-existing RBBB and cover index for perimeter-derived annular diameter in systole. The threshold for prediction of PPI on the basis of cover index for perimeter-derived annular diameter in systole was 3.6% (AUC = 0.594; 95% CI: 0.518 to 0.669; p = 0.03), with sensitivity of 81.6% and specificity of 37.0%.
Given that official sizing recommendations of the manufacturer have not been validated with clinical data, we elaborated a modified sizing chart that takes into account the required minimum oversizing to avoid post-procedural second-degree or greater PVL (Table 5). The new recommendations are based on the thresholds of cover index for perimeter-derived annular diameter in diastole and in systole with minimum oversizing of 4.4% and 2.5%, respectively.
We present a comprehensive analysis of anatomic and procedural factors that affect periprocedural outcomes based on the largest single-center transfemoral TAVR cohort treated with the self-expanding ACURATE neo device. Key findings were as follows: 1) Overall, the best outcomes were achieved in mild to moderate DLZ calcification (Figure 4). Second-degree or greater PVL was more common in patients with more severe DLZ calcification, whereas PPI was independent of DLZ calcification. 2) Predictors of second-degree or greater PVL were more severe periannular calcification (CVAnn), less oversizing, the presence of annular plaque protrusions, increased STJ height, and aortic movement of the stent holder at full release of the prosthesis. 3) Predictors of PPI were pre-existing RBBB and more oversizing.
The device success of 89.6% noted in the present series is slightly lower than in previous reports (7). This may reflect not only real-world data but also the fact that this patient cohort includes many of the earliest patients treated globally with the ACURATE neo prosthesis, as illustrated in Table 3.
Moderate or severe PVL following TAVR has been linked to worse outcomes, which is why there are sustained efforts to minimize this common complication of transcatheter valve therapy. Post-procedure, the frequency of second-degree or greater PVL was somewhat high (6.4%), but after second measures, including the implantation of a second valve or surgical valve replacement, it was reduced to 4.8% at discharge, which is in line with previous results (7,16). As stated earlier, we additionally noted a learning-curve effect, with substantial improvement over time (rate of second-degree or greater PVL of 11.0% in the first 100 cases vs. 3.0% in the last 100 cases, p = 0.03), which may be attributed largely to more oversizing and better patient selection, with less DLZ calcification, but also to a modified implantation technique.
Anatomic factors with impact on second-degree or greater PVL
The association between DLZ calcification and oversizing in relation to second-degree or greater PVL° has been extensively investigated in recent years and was confirmed in the present cohort. Ours is the first demonstration, however, that for this specific device, increased calcification in the periannular region enables the best discrimination of post-procedural second-degree or greater PVL, with a threshold of 97.0 mm3. Periannular calcification is commonly located at the base of the leaflets and may not be shifted aside as well as calcium deposits at the edge of the leaflets; hence, deposits at the base are likely to impair proper expansion of the ACURATE neo prosthesis given its relatively low radial force. The same mechanism accounts for the presence of annular plaque protrusions which were highly predictive of second-degree or greater PVL. The lower threshold of minimally required oversizing to prevent second-degree or greater PVL was 4.4% when using cover index for perimeter-derived annular diameter in diastole and 2.5% when using cover index for perimeter-derived annular diameter in systole. In most previous studies, the cover index is reported to be higher, depending on the type of implanted prosthesis and the applied modality of annular measurement (17,18). The causal association between STJ height and PVL is not clear but may be related to a suboptimal anchoring mechanism of this device in a larger aortic root anatomy.
Procedural factors with impact on second-degree or greater PVL
Among procedural factors, only the displacement of the stent holder in aortic direction after full release of the prosthesis independently predicted second-degree or greater PVL, which would indicate tensile stress on the delivery system when during positioning of the prosthesis the final movement was in the aortic direction. This in turn might hamper a tight anchoring of the upper crown within the native leaflets, which itself was significantly associated with less second-degree or greater PVL and seems to be required for a better paravalvular sealing.
Interestingly, positioning the delivery system in the outer curvature of the arch and the position of the marker band in relation to the annular plane, both of which are recommended by the manufacturer, had no significant impact on the rate of second-degree or greater PVL rate. This does not necessarily devalue these recommendations, because particularly keeping the delivery system in the outer curvature and maintaining a certain forward tension on the delivery system in the moment of full device release may increase stability during deployment und minimize the risk for aortic movement. However, excessive forward pressure on the delivery system should be avoided because of the risk for device migration into the left ventricle, which occurred in a single case in the present series.
The seemingly paradoxical finding that patients in whom pre-dilatation had not been performed had a lower rate of post-dilatation is attributable to the fact that pre-dilatation was generally omitted in cases with mild or moderate DLZ calcification. Finally, in contrast to other self-expanding transcatheter heart valves (19), a horizontal position of the prosthesis did not affect the PVL rate, which can be attributed to the principle of top-down deployment and better coaxial alignment due to the stabilization arches.
Second-degree or greater PVL in severe DLZ calcification
The high rate of second-degree or greater PVL, at 13% in the group with severe DLZ calcification, would not generally preclude these patients from treatment with the ACURATE neo prosthesis. Among subjects with severe DLZ calcification, the presence of annular plaque protrusions, an asymmetrical distribution of calcification in the periannular region (ΔCVAnn) and increased STJ height predicted second-degree or greater PVL in this specific subgroup, meaning that in the absence of these risk factors, the use of the ACURATE neo may be considered despite of severe DLZ calcification.
The rate of PPI of 10.2% was within the range of previous data (7,16). Our findings that pre-existing RBBB and a higher degree of oversizing were predictors of PPI are consistent with what has been reported for other transcatheter heart valves (20). Theoretically, a cover index for perimeter-derived annular diameter in systole between 2.5%, as a minimum requirement to prevent moderate PVL, and 3.6%, as the upper limit to protect against an increasing frequency of PPI, may represent an ideal range. The only procedural factor that was associated with a higher PPI rate was more aggressive pre-dilatation. Because this approach did not show any benefit with respect to PVL, less aggressive pre-dilatation may reduce PPI rates without affecting the occurrence of PVL, as demonstrated recently (21). In contrast to other THVs, we found no association between implantation depth and the need for PPI, which may be explained by the minimal protrusion into the left ventricular outflow tract and the relatively low radial force of the ACURATE neo.
Apart from limitations inherent to a retrospective study, a key limitation is the fact that PVL was not adjudicated in a core laboratory. Furthermore, results from this single-center analysis may not be broadly generalizable; however, such a standardized and comprehensive in-depth investigation of factors that determine procedural outcome would have been difficult in a retrospective multicenter setting. With regard to the modified sizing recommendation, a residual bias attributable to interobserver and intraobserver variability of annular measurements must be taken into account, although with the use of dedicated software applications divergences are usually minimal (22). Nonetheless, validation in a separate cohort would be necessary to further confirm our findings.
Procedural success of transfemoral implantation of the ACURATE neo device requires careful patient selection with appropriate oversizing and recognition that best outcomes may be achieved in cases with only mild to moderate DLZ calcification. Less aggressive pre-dilatation may help reduce PPI rates, and maintaining forward pressure on the delivery system during deployment may decrease the likelihood of post-procedural PVL.
WHAT IS KNOWN? The ACURATE neo is a self-expanding device with good clinical outcomes but with room for improvement.
WHAT IS NEW? The findings of the present study may help identify patients with aortic root anatomies that are suitable for the ACURATE neo device and help practitioners in refining implantation techniques to improve results.
WHAT IS NEXT? The selection of the appropriate type and size of a prosthesis tailored to patients’ individual anatomic characteristics will become more important in transcatheter valve therapies. For future investigations, the focus should be shifted from mere prevention of severe complications to pursuing optimal outcomes, hence, to identify patients with an anatomy that is particularly suitable for a specific device.
The authors thank Elizabeth Martinson, PhD, from the KHFI Editorial Office for her editorial assistance.
Dr. Kim is a proctor for Symetis and St. Jude Medical; and has received lecture honoraria from Symetis, St. Jude Medical, and Edwards Lifesciences. Dr. Möllmann has received proctoring fees and/or speaking honoraria from Abbott, Biotronic, Edwards Lifesciences, St. Jude Medical, and Symetis. Dr. Liebetrau has received lecture honoraria from Abbott, AstraZeneca, Bayer, Berlin Chemie, Boehringer Ingelheim, Daiichi-Sankyo, and Pfizer/Bristol-Myers Squibb; and meeting expenses from Bayer and Daiichi-Sankyo. Dr. Renker has received lecture honoraria from St. Jude Medical; and meeting expenses from Daiichi-Sankyo. Dr. Rolf has received lecture honoraria from AstraZeneca, Boehringer, Pfizer, Merck Sharpe & Dohme, and Bristol-Myers Squibb. Dr. Doss is a proctor for St. Jude Medical. Dr. Hamm is a member of the advisory board for Medtronic. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Kim and Möllmann contributed equally to this work.
- Abbreviations and Acronyms
- area under the curve
- confidence interval
- calcium volume of the annular region
- calcium volume of the device landing zone
- device landing zone
- Left ventricular outflow tract
- multidetector computed tomographic
- permanent pacemaker implantation
- paravalvular leakage
- right bundle branch block
- sinotubular junction
- transcatheter aortic valve replacement
- Received March 8, 2018.
- Revision received April 11, 2018.
- Accepted April 17, 2018.
- 2018 American College of Cardiology Foundation
- Gaede L.,
- Blumenstein J.,
- Liebetrau C.,
- et al.
- Baumgartner H.,
- Falk V.,
- Bax J.J.,
- et al.
- Mollmann H.,
- Hengstenberg C.,
- Hilker M.,
- et al.
- Mollmann H.,
- Walther T.,
- Siqueira D.,
- et al.
- Agatston A.S.,
- Janowitz W.R.,
- Hildner F.J.,
- Zusmer N.R.,
- Viamonte M. Jr..,
- Detrano R.
- Kim W.-K.,
- Renker M.,
- Rolf A.,
- et al.
- Kim W.K.,
- Blumenstein J.,
- Liebetrau C.,
- et al.
- Khalique O.K.,
- Hahn R.T.,
- Gada H.,
- et al.
- Kim W.K.,
- Liebetrau C.,
- Renker M.,
- et al.
- Kappetein A.P.,
- Head S.J.,
- Genereux P.,
- et al.
- Husser O.,
- Kim W.K.,
- Pellegrini C.,
- et al.
- Kong W.K.,
- van Rosendael P.J.,
- van der Kley F.,
- et al.
- Mylotte D.,
- Dorfmeister M.,
- Elhmidi Y.,
- et al.
- Nazif T.M.,
- Dizon J.M.,
- Hahn R.T.,
- et al.
- Toggweiler S.,
- Nissen H.,
- Mogensen B.,
- et al.
- Korosoglou G.,
- Gitsioudis G.,
- Waechter-Stehle I.,
- et al.