Author + information
- Received March 15, 2018
- Revision received May 17, 2018
- Accepted June 5, 2018
- Published online August 25, 2018.
- Julia Seeger, MDa,
- Renu Virmani, MDb,
- Maria Romero, MDb,
- Birgid Gonska, MDa,
- Wolfgang Rottbauer, MDa and
- Jochen Wöhrle, MDa,∗ ()
- aDepartment of Internal Medicine II, Cardiology, University of Ulm, Ulm, Germany
- bCV Path Institute, Gaithersburg, Maryland
- ↵∗Address for correspondence:
Prof. Dr. Jochen Wöhrle, University Hospital of Ulm, Albert-Einstein-Allee 23, 89081 Ulm, Germany.
Objectives The aim of this study was to evaluate the debris captured by the Claret Sentinel cerebral embolic dual-filter protection device during transfemoral transcatheter aortic valve replacement (TAVR) with different valve types.
Background Risk for embolization of debris during TAVR may vary by TAVR device.
Methods The filters of 100 consecutive patients were collected and captured debris was analyzed by histopathology and histomorphometry. Three valve types were implanted: the balloon-expandable Edwards SAPIEN 3 (n = 42), the self-expandable Medtronic Evolut R (n = 35), and the mechanically implantable Boston Scientific Lotus (n = 23).
Results Among the 3 groups there was no difference in baseline data, including Society of Thoracic Surgeons score for mortality, calcification, or pre-dilation. The type of captured debris did not differ among the 3 valve types in the proximal or distal filter. With the balloon-expandable valve, there were significantly more patients with large debris measuring ≥1,000 μm. The number of particles in the proximal filter was significantly lower with the Lotus (89.8 ± 106.3) compared with the Evolut R (187.3 ± 176.9) and Edwards SAPIEN 3 (172.3 ± 133.5) valves (p = 0.035). Total tissue area in the proximal filter was significantly smaller for the Lotus compared with the other 2 valve types (7.1 ± 6.3, 20.1 ± 19.0, and 21.3 ± 15.1 mm2; p = 0.0014). In contrast, for the distal filter, there were no differences with respect to valve type for total tissue area, particle size, and number of particles.
Conclusions A significant difference was observed in the size and number of captured tissue particles with the double-filter embolic protection device among different valve types in patients undergoing TAVR. The largest particles were observed in patients treated with a balloon-expandable valve.
Periprocedural stroke during transfemoral transcatheter aortic valve replacement (TAVR) is a serious complication substantially increasing acute and long-term morbidity and mortality (1–4). Stroke rates have been reported in up to 10.0% of patients after TAVR (5,6), with stroke risk being highest during the TAVR procedure. On cerebral magnetic resonance imaging (MRI), new cerebral lesions after TAVR have been reported in up to 98% of patients (5). The U.S. Food and Drug Administration recently approved the Sentinel Cerebral Protection System (Claret Medical, Santa Rosa, California) (7). Use of the double-filter protection device was associated with a 44% decrease in lesion volume on cerebral MRI in the randomized SENTINEL trial (6). Clinically overt strokes were reduced from 8.2% in unprotected TAVR patients to 3.0% (p = 0.05) in patients undergoing protected TAVR within the first 72 h (8). Histopathologic analysis revealed debris captured in 99% of patients. Nonrandomized data recently demonstrated possible differences in stroke rates dependent on the TAVR device used (9–11).
The aim of this study was to evaluate differences in debris captured by the Claret Sentinel cerebral embolic dual-filter-based protection device among different TAVR devices.
Since 2016 the double-filter Sentinel cerebral embolic protection device has been the standard of care for patients undergoing TAVR at the University of Ulm (12). The Sentinel device is a dual-filter-based intraluminal embolic protection device inserted through a 6-F sheath introduced via the right radial, ulnar, or brachial artery prior to passage of any other device across the aortic arch. The proximal filter consists of a radiopaque nitinol frame with a 140-μm pore polyurethane filter and is positioned in the brachiocephalic trunk. The distal filter is similarly constructed and is inserted in the left common carotid artery. The 2 filters cover all brain areas supplied by the right vertebral and right and left carotid arteries, constituting more than 90% of the cerebral blood flow. Only the left vertebral artery remains unprotected.
Among a series of more than 600 patients undergoing protected TAVR for symptomatic severe native aortic stenosis, the filters of 100 consecutive patients were collected, and captured debris was analyzed for histopathology and histomorphometry. During this period 3 different TAVR devices were used: the balloon-expandable Edwards SAPIEN 3 valve (Edwards Lifesciences, Irvine, California), the self-expandable Medtronic Evolut R (Medtronic, Minneapolis, Minnesota), and the mechanically implantable Boston Scientific Lotus (Boston Scientific, Marlborough, Massachusetts).
TAVR procedures were performed transfemorally with local anesthesia under conscious sedation as described elsewhere (12). Pre-procedural multislice computed tomographic images were analyzed using dedicated software (3mensio, Pie Medical Imaging, Maastricht, the Netherlands). Left ventricular outflow tract calcification was assessed according to Barbanti et al. (13). Valve calcification was graded according to Rosenhek et al. (14). There were no patients with bicuspid valves in the series. Periprocedural outcome was assessed within 48 h post-TAVR.
The protocol complied with the Declaration of Helsinki and was approved by the local ethics committee (NCT02162069). Written informed consent was obtained from all patients. The interdisciplinary heart team of the participating centers made the decision for transcatheter approach.
Grossing and processing
A total of 200 filters (100 proximal and 100 distal) from the Claret Medical Sentinel Cerebral Protection System containing embolic debris captured during TAVR by the Ulm University Hospital team were shipped to the CVPath Institute fixed in 10% neutral buffered formalin. Each filter was digitally photographed (EOS Rebel Xsi, Canon, Tokyo, Japan) prior to any physical alteration. The samples were carefully opened using scissor blades, and all contents were removed and then filtered through a Falcon 40-μm Nylon Cell Strainer (Thermo Fisher Scientific, Waltham, Massachusetts). The material collected by the cell strainer was photographed again, then carefully folded and placed in a Shandon Nylon Biopsy Bag (6774009, Thermo Fisher Scientific). The biopsy bag containing the sample was then transferred to an appropriately bar code–labeled biopsy cassette and submitted for processing. Samples were processed in a graded series of ethanol and xylene (Tissue-Tek VIP 6, Sakura, Torrance, California) and embedded in paraffin. Each paraffin block was serially cut at 4 to 5 μm, with 2 consecutive sections affixed per charged slide. Slides from each sample were stained with hematoxylin and eosin and Movat’s pentachrome stain. The first, sixth, and twelfth slides cut were stained with hematoxylin and eosin, and the second and seventh slides cut were stained with Movat’s pentachrome.
Each sample was evaluated for the presence of thrombus (acute vs. organizing) and its composition, determined by including presence of platelets, red blood cells, inflammatory cells, necrotic core, foamy macrophages, calcification, valve tissue, arterial wall, collagenous tissue, foreign material, and myocardium. These values and all combinations were recorded and reported for both filters separately.
In histomorphometry, particle size (all sizes, ≥150 μm, ≥500 μm, and ≥1,000 μm) and total particle area were automatically measured.
Categorical parameters are presented as counts and percentages and were compared using the Pearson chi-square test and Fisher exact test as appropriate. Continuous variables are presented as mean ± SD and were analyzed using analysis of variance. Baseline data, procedural data, and captured debris were compared among the 3 different valve types (balloon-expandable, self-expandable, and mechanically implantable aortic valves). A p value <0.05 was considered to indicate statistical significance, and tests were 2 sided. Statistical analysis was performed using Statistica version 10 (StatSoft, Tulsa, Oklahoma).
In the present study, the Claret Medical Sentinel dual-filter embolic protection device was analyzed to distinguish the type and quantity of embolic material captured during TAVR in patients with severe symptomatic calcified native aortic valve stenosis. Of the 200 filters received from 100 patients, all 100 proximal and 100 distal filters were deployed properly in the intended position. In each patient, the predominant captured embolic material was composed of acute platelet-rich thrombus in the presence of other tissue (99%). Typical samples are depicted in Figure 1. This was followed by arterial wall (84%), fibroelastic tissue consistent with valve tissue (84%), and calcifications (58%). Foreign material (33%), myocardial fibers (14%), and necrotic core (12%) were found less frequently. Organizing thrombus (7%) and acute thrombus without any associated debris (1%) were rarely identified.
In each proximal filter, the pattern was similar. Acute platelet-rich thrombus was the predominant captured material (89%), followed by arterial wall (69%), valve tissue (66%), calcification (38%), foreign material (21%), myocardial fibers (9%), necrotic core (6%), organizing thrombus (3%), and acute thrombus without any other debris (1%). In each distal filter, again, the pattern was similar. Acute platelet-rich thrombus was the predominant captured material (96%), followed by valve tissue (77%), arterial wall (76%), calcification (43%), foreign material (19%), necrotic core (10%), myocardial fibers (8%), and organizing thrombus (4%).
Comparison among aortic valve types: baseline and procedural data
Among 100 patients, 23 patients received the Lotus valve, 35 patients the Evolut R, and 42 patients the SAPIEN 3 valve. Among groups there were no differences in baseline data (Table 1), including Society of Thoracic Surgeons score for mortality, diabetes mellitus, chronic renal failure, history of cardiac surgery, history of stroke or intracerebral bleeding, atrial fibrillation, carotid artery stenosis, peripheral vascular disease, reduced left ventricular ejection fraction (<45%), platelet count, calcification of aortic cusp, and left ventricular outflow tract. Patients receiving the SAPIEN 3 valve were significantly younger compared with those treated with the Lotus or the CoreValve Evolut R valve. Activated clotting time during TAVR was similar among groups (Table 2). Procedural data including pre-dilation and post-dilatation did not differ among groups. Rate of valve repositioning was significantly different between groups (Table 2), because the mechanically implantable Lotus can be repositioned and completely retrieved even after full deployment, and the self-expandable CoreValve Evolut R can be repositioned and fully retrieved during the implantation process.
Comparison among aortic valve types: histopathology and histomorphometry
Histopathologic and histomorphometric data with respect to the different valve types are shown for the proximal filter (Table 3, Online Table 1, Figure 2) and distal filter (Table 4, Online Table 2, Figure 3). Regarding the observed rate of acute thrombus, organizing thrombus, valve tissue, arterial wall, calcification, foreign material, myocardium, necrotic core, or any debris, there was no difference among the 3 valve types in the proximal filter (Table 3, Figure 2) or in the distal filter (Table 4, Figure 3).
Captured debris was also analyzed by histomorphometry. In the proximal filter there were significant differences regarding the size of captured debris (Figure 4A). In patients treated with the SAPIEN 3 valve there were significantly more patients with debris measuring ≥2,000 μm as well as in the range between 1,000 and 2,000 μm, whereas regarding smaller particles, there was no difference among valve types. In addition, the number of particles in the proximal filter was significantly smaller with the Lotus valve compared with the CoreValve Evolut R and SAPIEN 3 valve (Table 3, Figure 5A). In addition, total tissue area was significantly smaller for the Lotus valve compared with the other 2 valve types in the proximal filter (Table 3, Figure 6A). In contrast, for the distal filter, there was no difference with respect to valve types for total tissue area (Table 4, Figure 6B), number of patients with different particle sizes (Table 4, Figure 4B), and number of particles (Table 4, Figure 5B). In the presented patient population, the rate of all strokes was 2% (2 of 100). There was 1 nondisabling stroke on day 1 in a patient treated with the Lotus valve and 1 disabling stroke on day 2 in a patient treated with the Evolut R.
We were able to demonstrate debris captured in 96% in proximal filters and in 97% in distal filters during TAVR with the Sentinel double-filter embolic protection device. Although baseline data did not differ (except age) among the balloon-expandable SAPIEN 3, mechanically implantable Lotus, and self-expandable CoreValve Evolut R valve, we observed significant differences in captured debris of the proximal filter. Captured debris from patients treated with the Lotus valve was significantly smaller in total tissue area. In contrast, the number of patients with large tissue particles was highest with the SAPIEN 3 valve.
In large reported series, clinical reported stroke rates differ among the used valve types. Clinical observed stroke was reported in 3.0% for the Lotus (9), 3.8% for the CoreValve Evolut R (10), and 1.4% for the SAPIEN 3 valve (11) within 30 days. However, those series did not include a neurologist for routine assessment of stroke. The analysis of captured debris with the use of a double-filter embolic protection device gives new insights in the pathogenesis of periprocedural stroke risk during TAVR with different valve types.
Analyzed data so far showed that embolic debris was captured in 75% of patients (n = 40) when an early version of the intraluminal filter device was used and varied in size from 0.15 to 4.0 mm (15). The debris in this series consisted of calcified material, valve tissue, fibrin, collagenous tissue representing vessel wall, and thrombotic material, similar to our studied population. In another series of 81 patients, captured debris measured up to 9.0 mm (14). On multivariate logistic regression analysis, the use of a balloon-expandable SAPIEN XT valve was an independent predictor of tissue embolization (compared with a self-expandable CoreValve), which supports our findings (16). The largest series, with 322 filters, showed thrombus in 91%, arterial wall tissue in 68%, valve tissue in 53%, calcification in 46%, and foreign material in 30% (17), similar to our population. However, in this large series of 322 filters, results were not given with respect to different valve types.
In our consecutive 100-patient series, baseline data did not differ among the 3 valve types (except age). Although calcification of the annulus, calcification of the left ventricular outflow tract, and cardiovascular risk factors as well as pre-dilatation did not differ among groups, the total tissue area and the number of large particles were significantly greater in patients treated with the balloon-expandable valve. This is of special interest because those patients were significantly younger compared with patients treated with the other 2 valve types. With routine use of pre-dilatation in almost all of our patients, the technique of valve implantation may be the most important factor triggering size and amount of embolized debris. The forces at the native calcified aortic valve during balloon inflation may be highest compared with a self-expandable or mechanically implantable valve type, possibly triggering our observed results.
The use of double-filter cerebral embolic protection has been linked to numerically fewer new lesions and smaller total lesions on cerebral MRI (18), lower new lesion volume on cerebral MRI (8), and a lower risk for stroke (8,12). On the basis of our data, there seem to be substantial differences regarding size and total number of embolized debris captured with the double-filter Sentinel embolic protection device favoring the mechanically implantable Lotus device. Results may differ without routine use of pre-dilation. In a recent analysis, significantly more debris per patient was found for valve tissue in patients with pre-dilatation compared with patients without pre-dilation (17). However, in this 100-patient series, 8 different valve types were used, with only 5 cases being treated with the mechanically implantable Lotus valve. The relation of pre-dilation to clinical relevant stroke is controversial. In a large meta-analysis, TAVR with or without pre-dilation was associated with a similar risk for stroke (18).
With TAVR advancing to lower risk and younger patients, the risk for stroke and the size of embolic debris must be taken into account. Large series have demonstrated a significant reduction in periprocedural stroke during TAVR with cerebral double-filter embolic protection compared with unprotected procedures (12). The type of used valve for TAVR should not trigger the use of a cerebral protection device. The difference in captured debris in our series was more apparent in the larger proximal filter, which might be explained by the position in the brachiocephalic trunk, which originates first from the aortic arch and protects the majority of the cerebral blood flow, whereas the distal filter is positioned in the smaller common carotid artery. Because embolic material was captured in the proximal and distal filters in almost all patients undergoing TAVR, a double-filter embolic protection device could be considered for patients undergoing a TAVR procedure.
This was a single-center experience. Only 200 filters were analyzed, so our results must be considered hypothesis generating. Larger randomized controlled trials are needed to confirm these findings. Implantation technique, including pre-dilation, post-dilation, size of valve, and type of valve for a specific patient, may differ among various centers. Results may differ without routine use of pre-dilatation.
We observed a significant difference in the size and number of captured tissue particles with the double-filter embolic protection device among different valve types in patients undergoing TAVR. The largest particles were observed in patients treated with a balloon-expandable valve.
WHAT IS KNOWN? During TAVR there is an almost 100% rate of captured debris with dual-filter cerebral embolic protection.
WHAT IS NEW? We observed that the size and number of captured debris during TAVR substantially differed among 3 different valve types (balloon-expandable, mechanically expandable, and self-expandable valves).
WHAT IS NEXT? Future studies comparing different valve types should focus on periprocedural stroke risk for patients undergoing TAVR with the use of dual-filter cerebral embolic protection.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- magnetic resonance imaging
- New York Heart Association
- transfemoral transcatheter aortic valve replacement
- Received March 15, 2018.
- Revision received May 17, 2018.
- Accepted June 5, 2018.
- 2018 American College of Cardiology Foundation
- Fairbairn T.A.,
- Mather A.N.,
- Bijsterveld P.,
- et al.
- Haussig S.,
- Mangner N.,
- Dwyer M.G.,
- et al.
- Kapadia S.R.,
- Kodali S.,
- Makkar R.,
- et al.,
- for the SENTINEL Trial Investigators
- Rogers T.,
- Alraies C.,
- Torguson R.,
- Waksman R.
- Kapadia S.R.
- Falk V.,
- Wöhrle J.,
- Hildick-Smith D.,
- et al.
- Kalra S.S.,
- Firoozi S.,
- Yeh J.,
- et al.
- Wendler O.,
- Schymik G.,
- Treede H.,
- et al.
- Seeger J.,
- Gonska B.,
- Otto M.,
- Rottbauer W.,
- Wöhrle J.
- Barbanti M.,
- Yang T.H.,
- Rodès Cabau J.,
- et al.
- Van Mieghem N.M.,
- El Faquir N.,
- Rahhab Z.,
- et al.
- Schmidt T.,
- Akdag O.,
- Wohlmuth P.,
- et al.
- Van Mieghem N.M.,
- van Gils L.,
- Ahmad H.,
- et al.
- Bagur R.,
- Kwok C.S.,
- Nombela-Franco L.,
- et al.