Author + information
- Received October 19, 2014
- Revision received December 16, 2014
- Accepted January 15, 2015
- Published online April 27, 2015.
- Nicolas M. Van Mieghem, MD, PhD∗∗ (, )
- Nahid El Faquir, BSc∗,
- Zouhair Rahhab, BSc∗,
- Ramón Rodríguez-Olivares, MD∗,
- Jeroen Wilschut, MD∗,
- Mohamed Ouhlous, MD, PhD†,
- Tjebbe W. Galema, MD, PhD∗,
- Marcel L. Geleijnse, MD, PhD∗,
- Arie-Pieter Kappetein, MD, PhD‡,
- Marguerite E.I. Schipper, MD, PhD§ and
- Peter P. de Jaegere, MD, PhD∗
- ∗Department of Cardiology, Thoraxcenter, Erasmus Medical Center, Rotterdam, the Netherlands
- †Department of Radiology, Erasmus Medical Center, Rotterdam, the Netherlands
- ‡Department of Cardio-Thoracic Surgery, Erasmus Medical Center, Rotterdam, the Netherlands
- §Department of Pathology, Erasmus Medical Center, Rotterdam, the Netherlands
- ↵∗Reprint requests and correspondence:
Dr. Nicolas M. Van Mieghem, Department of Interventional Cardiology, Thoraxcenter, Erasmus Medical Center, Room Bd 171, ‘s Gravendijkwal 230, 3015 CE Rotterdam, the Netherlands.
Objectives The aim of this study was to identify variables associated with tissue fragment embolization during transcatheter aortic valve replacement (TAVR).
Background Brain magnetic resonance imaging and transcranial Doppler studies have revealed that cerebrovascular embolization occurs frequently during TAVR. Embolized material may be r thrombotic, tissue derived, or catheter (foreign material) fragments.
Methods A total of 81 patients underwent TAVR with a dual filter–based embolic protection device (Montage Dual Filter System, Claret Medical, Inc., Santa Rosa, California) deployed in the brachiocephalic trunk and left common carotid artery. Both balloon-expandable and self-expanding transcatheter heart valves (THVs) were used. Filters were retrieved after TAVR and sent for histopathological analysis.
Results Overall, debris was captured in 86% of patients. Captured material varied in size from 0.1 to 9.0 mm. Thrombotic material was found in 74% of patients and tissue-derived debris in 63%. Tissue fragments were found more often with balloon-expandable THVs (79% vs. 56%; p = 0.05). The embolized tissue originated from the native aortic valve leaflets, aortic wall, or left ventricular myocardium. On multivariable logistic regression analysis, balloon-expandable THVs (odds ratio: 7.315; 95% confidence interval: 1.398 to 38.289; p = 0.018) and cover index (odds ratio: 1.141; 95% confidence interval: 1.014 to 1.283; p = 0.028) were independent predictors of tissue embolization.
Conclusions Debris is captured with filter-based embolic protection in the vast majority of patients undergoing TAVR. Tissue-derived material is found in 63% of cases and is more frequent with the use of balloon-expandable systems and more oversizing.
Transcatheter aortic valve replacement (TAVR) is associated with procedure-related neurological events (1,2). The 30-day incidence of major or disabling stroke is 3.4% to 8%, and most of these events occur within the first 24 to 48 h after TAVR (3–6). Diffusion-weighted magnetic resonance imaging studies have revealed new brain defects after TAVR in up to 80% of cases irrespective of the access strategy used (7–9). Transcranial Doppler studies have identified balloon valvuloplasty and actual valve positioning and deployment as primary causes of cerebral embolization during TAVR (10). When using cerebral embolic protection filters during TAVR, macroscopic debris is detected in up to 75% of cases (11). Histopathological analysis of this debris revealed tissue originating from major cardiovascular structures and the aortic valve in more than one-half of the patients. Enhanced knowledge of the pathophysiology of tissue embolization during TAVR is relevant a fortiori because the fate of at-first-sight subclinical new brain lesions may be correlated with expedited neurocognitive impairment in the long term (12,13). Predictors of tissue detachment and embolization to the brain during TAVR are currently unknown. This study aimed to analyze the association of patient and procedural variables with embolization of tissue to the brain during TAVR.
All patients who underwent TAVR with cerebral protection using the Montage Dual Filter embolic protection device (EPD) (Claret Medical, Inc., Santa Rosa, California) between December 2011 and December 2013 were included in this study (Figure 1). Baseline characteristics and procedural and outcome data were prospectively collected in a dedicated database in accordance with local institutional review board guidelines. Patients were considered at high operative risk by heart team consensus. Pre-procedural planning included a contrast multislice computed tomography examination. Aortic root calcification was quantified by the Agatston score and further graded semiquantitatively as follows: grade 1, no calcification; grade 2, mildly calcified (small isolated spots); grade 3, moderately calcified (multiple larger spots); and grade 4, heavily calcified (extensive calcifications of all cusps) (14). Patient eligibility for the EPD required an appropriately sized left common carotid artery (≥5 mm) and brachiocephalic artery (≤ 9 mm) without significant stenosis (≥70%). The Claret Montage EPD consists of 2 polyurethane mesh filters with 140-μm pores mounted on a nitinol frame. Its concept has been discussed in detail elsewhere (15). Briefly, via right radial artery access, 1 filter is deployed in the brachiocephalic trunk and 1 filter in the left common carotid artery. Of note, the left vertebral artery is not protected. One trained operator performed all EPD implantations.
TAVR procedure and filter handling
All TAVR procedures were performed with patients under general anesthesia. Patients were pre-loaded with dual-antiplatelet therapy (aspirin and clopidogrel). A standardized anticoagulation regimen with heparin was initiated, aiming for an activated clotting time (ACT) between 250 and 300 s. The Montage Dual Filter EPD was deployed before the introduction of the large-bore TAVR access sheath in the groin. The remainder of the TAVR evolved according to standard practice. Both self-expanding and balloon-expandable transcatheter heart valves (THVs) were used in this study. After valve implantation, the Claret Montage EPD filters were recaptured and removed. Filters were visually inspected on site. Filters were cut, and the debris was stored in a buffered formalin (4%) solution and sent for analysis to the Department of Pathology.
Debris was dehydrated and embedded in paraffin and cut into 3- to 4-μm-thick sections. Staining was done with hematoxylin and eosin and Movat pentachrome. Material of very small size (<0.25 mm) was processed following the Cellient procedure and stained with both Giemsa and hematoxylin and eosin. Additional staining techniques were performed whenever applicable to identify specific tissue origin, as previously described (11).
Categorical variables are presented as frequencies and percentages as appropriate and compared using the Pearson chi-square test or the Fisher exact test. Continuous variables are presented as mean ± SD or median (IQR) in case of a normal or skewed distribution, respectively, and compared using analysis of variance. Normality of the distributions was assessed using the Shapiro-Wilks test. To study the independent predictors of tissue embolization, logistic regression was performed. The following variables were deemed pathophysiologically relevant and included in the multivariable logistic regression model: valve type, cover index as a surrogate for oversizing, valve area, need for post-dilation, and peripheral arterial disease. A maximum of 5 variables were allowed to enter the multivariable analysis given the absolute event rate of 51 (63%) in keeping with the frequency of the dependent variable y (n/10).
Statistical analyses were performed using SPSS software version 21.0 (SPSS Inc., Chicago, Illinois). Statistical significance was defined as p < 0.05.
A total of 81 patients underwent TAVR with the Montage Dual Filter EPD. Tables 1 and 2 ⇓⇓summarize baseline patient and procedural characteristics. The mean age was 79 years (interquartile range [IQR], 73 to 84), and 57% were male. One-third and 26% of patients were considered to be, respectively, frail or technically inoperable. Atrial fibrillation was present in 26%. There was significant aortic root calcification with a median Agatston score of 3,000 (IQR: 1,800 to 4,100). The transfemoral route was the predominant access strategy. The CoreValve (Medtronic, Minneapolis, Minnesota) was used in 46 cases (69%), the SAPIEN XT (Edwards Lifesciences, Irvine, California) in 24 (30%), and Portico (St. Jude Medical, St. Paul, Minnesota) in 1 case. The mean ACT 30 min after a heparin bolus was 230 ± 68 s. Balloon pre-dilation was routine (90%), and post-dilation was required in more than one-fourth of cases. One patient experienced a transient ischemic attack 1 week after successful CoreValve implantation during an episode of new-onset atrial fibrillation. Two patients experienced a disabling stroke, 1 patient immediately after TAVR with a CoreValve complicated by pericardial effusion and 1 patient after TAVR with a SAPIEN XT valve complicated by valve embolization and rescue CoreValve implantation, respectively. Both patients required prolonged cardiopulmonary resuscitation. These 2 patients died of multiorgan failure, respectively, 2 and 6 weeks after the index TAVR procedure. The overall 30-day mortality rate was 3% (n = 2).
The Montage Dual Filter EPD was successfully deployed and retrieved in all patients. Figure 2 displays the frequency and distribution of the captured debris. Debris of any type was identified in 86% of patients. The median size of debris was 1 mm (IQR: 0.6 to 1.5 mm) and varied between 0.1 mm and 9 mm. Fibrin and thrombotic material (size varied from 0.2 to 6.2 mm) was found in 74% of patients. Tissue-derived debris was present in 63%. Typical features of the degenerative aortic valve leaflets, notably amorphous calcified material and collagenous and proteoglycan matrix with elastic tissue surrounded by endothelial cells (size, 0.2 to 5.5 mm), was identified in 27 patients (33%) (Figure 3). Arterial vessel wall characteristics are depicted in Figure 4. Collagenous tissue of undetermined origin (either vessel wall or aortic valve leaflet) (size, 0.2 to 1.5 mm) was recognized in 14 patients (17%). Endothelium strands (size, 0.2 to 9 mm) were noted in 39 patients (48%). In 13 patients (16%), the retrieved specimen contained cardiomyocytes representing myocardial tissue (size, 0.1 to 1.7 mm) (Figure 5). Small foreign-body polymer material was present in 8 patients. Of note, compared with TAVR with self-expanding systems, tissue-derived debris was found more often with the balloon-expandable THVs (79% vs. 56%; p = 0.05). Conversely, there was no difference in the presence of thrombotic material in both THV designs.
Predictors of tissue embolization
Balloon-expandable THVs (odds ratio: 7.315; 95% confidence interval: 1.398 to 38.289; p = 0.018) and cover index (odds ratio: 1.141; 95% confidence interval: 1.014 to 1.283; p = 0.028) were independent predictors of tissue embolization, and a trend toward more tissue embolization was seen with balloon post-dilation (odds ratio: 2.607; 95% confidence interval: 0.675 to 10.073; p = 0.17) (Table 3). Further analysis demonstrated no difference in debris size between the self-expanding and balloon-expandable platforms.
The main findings of our study using a filter-based EPD device during TAVR are the following: 1) cerebral embolization is almost ubiquitous; 2) thrombotic material and tissue-derived debris en route to the brain was captured in 74% and 63%, respectively; 3) the captured tissue fragments predominantly originated from large arterial structures and the aortic valve; and 4) use of balloon-expandable THVs and more oversizing were independent predictors of tissue embolization.
The incidence of clinically apparent neurological events after TAVR varies in the literature. Variability in clinical endpoint definitions, reporting bias, and relative underdiagnosis has hampered relevant data comparison between various reports. The Valve Academic Research Consortium endorsed a consensus document on uniform endpoint definitions (16,17). A meta-analysis of 13 studies using these Valve Academic Research Consortium definitions reported 30-day neurological event and major stroke rates of 5.7% and 3.2%, respectively (18). Randomized, controlled trials comparing TAVR with surgical aortic valve replacement (SAVR) have shown similar 30-day major/disabling stroke rates with both treatment modalities, 3.8% for balloon-expandable TAVR versus 2.1% with SAVR (p = 0.20) and 3.9% for self-expanding TAVR versus 3.1% with SAVR (p = 0.55) (1,2). Importantly, neurological events may be subtle and remain undiagnosed. Indeed, in a recent study in which trained neurologists evaluated all patients before and after SAVR, the stroke rate was as high as 17% (19). The high cerebral embolization rate as detected by using filters during TAVR is consistent with brain magnetic resonance imaging and transcranial Doppler studies (7–11). Although most of these defects remain at first sight clinically unnoticed, the long-term impact of transient brain ischemia and subclinical infarcts is unclear, although studies suggest associations with neurocognitive decline and premature dementia (12,13,20).
The histopathology results presented here extend our previously reported data (11). In fact, captured debris seemed even more frequent in the current study (86% vs. 75%). The presence of thrombotic material in 74% of patients can be partly explained by the suboptimal ACT levels as measured 15 to 30 min after the initial heparin bolus. The mean ACT of 230 s was lower than the targeted ACT level of 250 to 300 s. Optimized heparin protocols and potentially more reliable alternative anticoagulants may reduce thrombus formation. We found tissue-derived debris in 63% of patients. In one-third of the patients, the tissue fragments were consistent with aortic valve tissue. In another 17%, the identified collagenous tissue stemmed either from the aortic valve or arterial wall. Endothelium strands were noted in one-half of the study patients. Transcranial Doppler studies confirmed high-intensity transient signals as a surrogate for microembolization when crossing a degenerated aortic valve and subsequent instrumentation within the aortic root including valve positioning and placement. It is conceivable that these essential manipulations are equally responsible for dislodgment of material from the aortic valve and aorta. The identification of myocardial tissue fragments in 4 patients is a novel finding and most probably the result of friction against the myocardium by the stiff guidewire that typically is introduced into the left ventricle to serve as a rail to advance the THV delivery across the aortic valve. Our findings underscore that detachment and subsequent embolization of solid tissue fragments may be inherent in contemporary TAVR practice. Accordingly, a report from the PRAGMATIC Plus Initiative demonstrated significant reductions in vascular and bleeding complications with growing TAVR experience yet no impact on stroke rate, suggesting that experience alone may not affect cerebrovascular embolization during TAVR (21).
The frequency of tissue embolization during TAVR was higher with balloon-expandable THVs compared with self-expanding THVs (79% vs. 56%; p = 0.05). Use of a balloon-expandable THV and more oversizing were particularly associated with tissue embolization. Valve oversizing and inflation of a balloon larger than the actual annular size may hypothetically coincide with a more vigilant impact on the aortic root during the actual valve implantation, which may predispose to tissue dislodgment. Conversely, a transcranial Doppler study found no difference in the overall number of high-intensity transient signals between the self-expanding and the balloon expandable THVs; however, more high-intensity transient signals occurred during the valve positioning with the balloon-expandable THV and conversely during the valve deployment with the self-expanding THV (10).
During TAVR, the THV will push aside the degenerated native aortic valve. With the first-generation THV designs, some degree of oversizing is recommended to reduce the incidence of paravalvular regurgitation. As a consequence, more oversizing and a higher cover index will impose more displacing forces on the degenerated aortic leaflets, which could explain tissue detachment and embolization. The trend of more tissue embolization with balloon post-dilation is in concordance with the findings of a large multicenter study on cerebrovascular events after TAVR that identified balloon post-dilation as a significant predictor for acute neurological events (odds ratio: 2.46; 95% confidence interval: 1.07 to 5.67) (5). Second-generation THV designs require less oversizing and need for post-dilation. Whether these devices will generate less tissue detachment and cerebral embolization needs to be determined. Balloon aortic valvuloplasty (BAV) before THV implantation was standard practice in this case series. BAV by itself may dislodge debris. The need for pre-BAV may be debatable. Whether direct THV implantation without previous BAV would result in less tissue embolization requires further research.
As the indications for TAVR are shifting to lower risk and younger patients with severe aortic stenosis, efforts to reduce cerebral embolization seem appropriate and valuable. Filter-based embolic protection during TAVR may be an effective barrier to prevent tissue debris from reaching the brain. Randomized studies are underway to investigate the impact of filter-based EPDs on TAVR-related brain lesions as assessed by magnetic resonance imaging.
This was a single-center descriptive study with a relatively small number of patients. Predictors of tissue embolization by multivariable analysis in this series should be considered hypothesis generating, and our findings require confirmation in larger patient cohorts. However, the high frequency of debris embolization to the brain in general and tissue fragment embolization in particular offer interesting perspectives. Our results may underestimate the true incidence of cerebral embolization because the filter-based EPD used in this study leaves the left vertebral artery unprotected, and sampling error in the histopathology analysis could occur.
Debris is captured with filter-based embolic protection in the vast majority of patients undergoing TAVR. Tissue-derived material is found in 63% of cases and is more frequent with the use of balloon-expandable systems and more oversizing.
WHAT IS KNOWN: Cerebral embolization of debris is common with transcatheter aortic valve replacement and correlates with ischemic brain lesions by brain magnetic resonance imaging studies.
WHAT IS NEW: This study establishes the common finding of cerebral embolization with transcatheter aortic valve replacement procedures, which seems more frequent after the use of balloon-expandable transcatheter heart valves and with more oversizing.
WHAT IS NEXT: Larger studies are required to confirm our findings and determine whether use of embolic protection devices could reduce cerebral embolization during transcatheter aortic valve replacement.
Dr. Van Mieghem has received research grants from Claret Medical Inc. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- activated clotting time
- balloon aortic valvuloplasty
- embolic protection device
- interquartile range
- surgical aortic valve replacement
- transcatheter aortic valve replacement
- transcatheter heart valve
- Received October 19, 2014.
- Revision received December 16, 2014.
- Accepted January 15, 2015.
- American College of Cardiology Foundation
- Daneault B.,
- Kirtane A.J.,
- Kodali S.K.,
- et al.
- Nombela-Franco L.,
- Webb J.G.,
- de Jaegere P.P.,
- et al.
- Athappan G.,
- Gajulapalli R.D.,
- Sengodan P.,
- et al.
- Ghanem A.,
- Muller A.,
- Nahle C.P.,
- et al.
- Kahlert P.,
- Knipp S.C.,
- Schlamann M.,
- et al.
- Rodes-Cabau J.,
- Dumont E.,
- Boone R.H.,
- et al.
- Kahlert P.,
- Al-Rashid F.,
- Dottger P.,
- et al.
- Van Mieghem N.M.,
- Schipper M.E.,
- Ladich E.,
- et al.
- Gress D.R.
- Tops L.F.,
- Wood D.A.,
- Delgado V.,
- et al.
- Leon M.B.,
- Piazza N.,
- Nikolsky E.,
- et al.
- Kappetein A.P.,
- Head S.J.,
- Genereux P.,
- et al.
- Genereux P.,
- Head S.J.,
- Van Mieghem N.M.,
- et al.
- Messe S.R.,
- Acker M.A.,
- Kasner S.E.,
- et al.
- Gaita F.,
- Corsinovi L.,
- Anselmino M.,
- et al.