Clinical impact of cerebral protection during transcatheter aortic valve implantation

Embolization of debris can complicate transcatheter aortic valve implantation (TAVI) causing stroke. Cerebral embolism protection (CEP) devices can divert or trap debris.


| INTRODUCTION
Stroke from cerebral embolism is a dreaded complication during transcatheter aortic valve implantation (TAVI) as it is associated with serious mortality and morbidity.The epidemiological dimension and clinical impact of cerebral embolism during TAVI procedures can vary, depending on various factors, including patient age, comorbidities and the procedural technique.][6][7][8][9][10] The rationale for using cerebral embolism protection (CEP) devices during TAVI is to reduce the risk of stroke and/or the volume of ischaemic embolic lesions by trapping and removing embolic debris and thrombotic material that might otherwise find their way through supra-aortic vessels.There is significant scientific evidence from the published literature supporting the use of cerebral embolism protection devices in TAVI procedures.Many studies have suggested that the use of these devices can efficiently retrieve embolizing debris or thrombotic material and potentially improve outcomes.However, there are still some uncertainties regarding their overall clinical efficacy.The lack of an ultimate outlook for their clinical impact is holding back CEP adoption.In fact, their use is not all plain sailing, as they can add further complexity to the TAVI procedure and increase the cost.In addition, accessibility to their use can largely vary among centres.
In this context, our aim was to provide a comprehensive quantitative synthesis of all clinical evidence available to date around the clinical impact of CEP use during TAVI procedures.

| METHODS
We developed a protocol for the review on 4 December 2022 and registered it in the AsPredicted study registry of the University of Pennsylvania on 25 January 2023 (https:// aspre dicted.org/ VLD_ 5D7), with the registration number 119812.

| Data sources and searches
We searched PubMed, Scopus and Doaj electronic databases from 1 January 2014 through 4 December 2023 using the following keywords and corresponding MeSH (Medical Subject Headings) terms: ((tavr OR tavi) AND (embolic AND protection AND device OR cerebral AND protection AND device)).We also checked the reference lists of eligible studies and screened scientific abstracts and relevant websites.

| Study selection
Two investigators (F.B. and S.D.R.) independently screened search records to identify eligible trials.No disagreements occurred.Inclusion criteria were randomized and observational controlled trials of patients with severe aortic stenosis undergoing TAVI that compared an intervention with a cerebral embolic protection device versus standard treatment and reported the incidence of stroke as primary endpoint.Exclusion criteria were duplicate publications, trials published in a language other than English, and studies in which the endpoint measure was not specified.

| Data extraction and quality assessment
Two reviewers (F.B. and N.C.) independently extracted data about study characteristics and event rates from full articles.Two investigators (F.B. and S.D.R.) independently assessed study quality for each trial by using the Cochrane Risk of Bias Tool.The following domains were evaluated: randomization method; allocation concealment; blinding of patient, investigator and outcome adjudication committee; reporting bias; attrition bias; and any other potential sources of bias, such as those related to trial designs, or the risk for contamination or crossover between the groups.The assessment was done at the study level and focussed on the main study outcome (stroke).Disagreements were resolved by consensus.

| Data synthesis and analysis
We focussed our primary summary of data on trials evaluating use of CEP devices during TAVI that are currently available commercially and used data extracted from the original primary publications.We based our analyses on the primary endpoint of stroke.We analysed death, major bleeding, vascular complications, acute renal insufficiency and total volume of new ischemic lesions as secondary endpoints.Baseline characteristics evaluated were age, sex, EUROSCORE and STS score, type of device used, pre-existing hypertension, dyslipidaemia, diabetes, prior myocardial infarction or angina, atrial fibrillation, peripheral artery disease, previous TIA or stroke, porcelain aorta and ejection fraction during echocardiography (Table 1).We used the risk ratio (RR) with 95% CI as the summary measure.Heterogeneity was assessed by using the Cochran Q test by means of a chi-squared function.p values below 0.10 were considered indicative of heterogeneity.I 2 values were calculated to estimate variation among        studies attributable to heterogeneity.Meta-analysis results were displayed with forest plots in which the measure of effect for each study is represented by a square, and the area of each square is proportional to study weight.A meta-regression analysis was performed to assess the impact of moderator variables on meta-analysis results.
To determine the actual power of this meta-analysis, trial sequential analysis (TSA) was conducted, as previously described. 11,12Analyses were performed by using Comprehensive Meta-Analyst, Open Meta-Analyst and Trial Sequential Analysis 0.9.5.10 Beta (Copenhagen Trial Unit, Centre for Clinical Intervention Research).

| Role of the funding source
The funding bodies had no role in the study's design, conduct, review or reporting or the decision to submit the manuscript for publication.

| RESULTS
We retrieved 236 records in total, as a result from a combined search with Scopus, PubMed and DOAJ.Using a PRISMA-based screening protocol (Figure 1), we identified 27 studies, both randomized and observational, comparing the use of CEP during TAVI versus standard TAVI that met our prespecified selection criteria.We excluded one of these studies from our analysis, because it evaluated a device that is not yet approved on the market. 13inally, 26 studies including 540.247 patients were available for meta-analysis.Of these, eight were randomized trials and 18 observational studies (Table 2), published from 2014 to 2023.  Resu of study quality assessment are reported in Tables S1 and S2, in the supplement file.Regarding randomized trials (Table S1), the methods adopted for randomization and allocation concealment were generally adequate.Blinding of patients and operators was not warranted for technical reasons; however, the endpoint adjudication committees were blinded to patients' allocation to the treatment arm.Risk of selection, detection, attrition and reporting bias were generally judged as low.Similarly, no blinding of patients and TAVI operators was warranted in nonrandomized studies (Table S2).Although these studies were generally judged of good quality, the study design did not exclude the risk of selection and attrition bias.
Different devices were used in these studies, all designed to minimize the risk of cerebral damage by capturing/deflecting embolic debris away from cerebral circulation during Transcatheter Aortic Valve Implantation (TAVI) and other transcatheter structural interventions.Specifically, the Triguard™ (Keystone Heart Ltd.) device is the only CE marked device designed to cover and protect all three major cerebral aortic arch vessels: It is placed through a femoral 9-Fr sheath with a parallel use of a pigtail (Table S3).It consists of a Nitinol frame with a dome-shaped mesh deflector, designed to 'self-position' and conform to adapt to aortic arch anatomies while providing stability to the filter that is designed to allow adequate blood flow to the brain while diverting emboli downstream.Similarly, the Embol-X™ (EXIS, Edwards Lifesciences) system features a self-expandable and self-fitting nitinol-based frame, covered by a semipermeable polyester mesh that is placed inside the aorta where it captures particles in the bloodstream (Table S3).The Sentinel™ device (Claret Medical) is a dual system filter.It is released in the brachiocephalic trunk and the left common carotid advanced in a 6-Fr sheath from the right upper extremity; the curve of the device can be adjusted to accommodate anatomic variations of the aortic arch, but one limitation is the incomplete cerebral coverage (Table S3).The Embrella™ (Edwards Lifesciences) was the earliest dedicated device for TAVI.It is used via the right radial (or brachial) artery with a 6-Fr sheath and advanced into the aortic arch where the device covered brachiocephalic trunk and left common carotid, leaving left vertebral artery unprotected in most of the cases (Table S3).

| Stroke
All studies selected reported the incidence of stroke after TAVI.Altogether, a significantly lower stroke rate was found in the CEP arm (1.6%) compared with the standard procedure without CEP (1.8%) (RR = 0.763 95% CI: 0.638-0.912;p = 0.003) (Figure 2A).Sensitivity analysis using the leave-one-out approach confirmed the solidity of the results.In fact, the general overlook of the results did not change significantly when any of the study was removed (Figure S1C).Cumulative meta-analysis revealed that the outlook of the results changed to a significantly lower stroke rate in the CEP arm from year 2017, remaining substantially unaltered afterwards (Figure S1D).Subgroup analysis showed that these results were mainly driven by the larger number of studies using the Sentinel device (RR = 0.734 95% CI: 0.599-0.900;p = 0.003), while no significant difference was found within the smaller subgroup of studies using other devices (Figure 3A).The overall degree of heterogeneity (I 2 = 53%; Q = 51; p = 0.001) was moderated and mostly driven by the subgroup of studies using the Sentinel device (I 2 = 64%), accounting for the largest proportion of all studies selected.The difference was maintained even after stratification of the studies included by their design and by device (Figure S1A; Figure S1B).Assessment of funnel plots did not show a substantial risk for small study effect (Figure S7A).The trial sequential analysis (TSA) showed that the results obtained are adequately powered.According to the trial sequential analysis, the minimum required sample size for this meta-analysis to be adequately powered would have been reached with 173,425 patients.Hence, the actual size of the meta-analysis with 540.247 patients included largely outweighs the minimum required sample size (Figure S8).

| Infarct area
Five of the studies selected, involving 914 patients, reported data on stroke lesion volume measured at   magnetic resonance imaging (MRI).The summary effect showed a trend of smaller lesion volume in the CEP arm, although no statistically significant difference was found (SMD = −0.25195%CI: −0.598-0.096;p = 0.156) (Figure 3B).Similarly, the overall degree of heterogeneity (I 2 = 81%; Q = 22; p = 0.001) was moderated.No relevant changes were found at sensitivity analysis, using the leave-one-out approach (Figure S2A) or at cumulative meta-analysis (Figure S2B).Assessment of funnel plots did not suggest the presence of small study effect (Figure S7B).

| Death
Altogether, 21 studies including 530.000 patients, reported death rates.The reported death rate was significantly lower in the CEP arm compared with the control arm (RR = 0.610 95% CI: 0.482-0.771;p < 0.001) (Figure 2B), even in the subgroup analysis for device type (RR = 0.577 95% CI: 0.448-0.744;p < 0.001) (Figure 3C).Sensitivity analysis showed consistency of results as the alternative removal of single studies had no relevant impact on meta-analysis results (Figure S3B).Cumulative meta-analysis showed a positive outlook for CEP with statistically significance starting from the studies published in 2021 (Figure S3C).Heterogeneity was moderate (I 2 = 51%; Q = 39; p = 0.005).Subgroup analysis showed that results were mostly driven by the studies using the Sentinel device (RR = 0.565 95% CI: 0.436-0.734;p < 0.001) (Figure 3C).Assessment of funnel plots did not show a substantial risk of small study effect (Figure S7C).

| Bleeding
Altogether, 14 studies including 379,610 patients reported data on bleeding complications.No difference was found between the CEP and the standard procedure arms (RR = 1.053 95% CI: 0.793-1.398;p = 0.721).Despite the pronounced heterogeneity (I 2 = 80%; Q = 65; p < 0.001), sensitivity analysis performed by multiple approaches, such as leave-one-out (Figure S4B), cumulative metaanalysis (Figure S4C) or subgroups analysis (Figure 4A) showed consistency of the results.Assessment of funnel plots suggested the presence of small study effect, particularly for studies with larger RRs (Figure S7D).

| Vascular access complications
Among the 13 studies (149,863 patients) reporting on vascular access complications, the use of CEP was not associated with an increased complication rate (RR = 0.937 95% CI: 0.820-1.070;p = 0.334).The overall degree of heterogeneity (I 2 = 0%; Q = 11; p = 0.004) was very low.Sensitivity analysis by means of leave-one-out (Figure S5B), cumulative meta-analysis (Figure S5C) and subgroup analyses (Figure 4B) confirmed the consistency of the results.Assessment of funnel plots did not show a substantial risk of small study effect (Figure S7E).

| DISCUSSION
Our results confirm the effectiveness of CEP, as it is associated with a significant lower rate of periprocedural stroke.These results summon all clinical evidence available, 40,41 including randomized control trials and observational studies.These results represent a relevant piece of information, in the context of heterogeneous results from recent studies.The PROTECTED TAVR trial (Stroke protection With Sentinel During Transcatheter Aortic Valve Replacement) enrolled 3000 patients, but failed to meet its primary endpoint, namely reducing the stroke rate.Despite the rate of stroke was numerically lower in patients treated with the CEP device, this reduction was not statistically significant.The study did however document F I G U R E 3 Meta-analysis results for device type subgroups.The figure reports the results of the meta-analysis comparing TAVI performed with CEP against standard TAVI without CEP.Every panel reports the results of a specific outcome, including Stroke (Figure 3, panel A), the volume of cerebral lesions measured with MRI (Figure 3, panel B), all-cause death (Figure 3, panel C), bleeding (Figure 4, panel A), vascular access-related complications (Figure 4, panel B), acute kidney injury (Figure 4, panel C).Studies names are listed on the left with the year of publication for each panel.Individual risk ratios with 95% confidence intervals are provided in the middle part of the figure, together with the calculations for the meta-analysis.The Forest plot on the right is the graphical representation of meta-analysis results.The position of the square indicates the mean effect size in each single study, whereas the area of the square reflects study weight.The black horizontal line depicts the 95% Confidence Interval of the effect size in each study.The red diamond at the very bottom of the Forest plot shows the overall summary effect, while summary effects are also provided for device type subgroups (blue diamonds).A summary effect positioned on the left of the vertical line implies a lower rate of the single endpoint in the CEP arm.CEP, cerebral embolic protection; CI, confidence interval; RR, risk ratio; Std, standard deviation.
a significantly lower incidence of invalidating stroke (a pre-defined secondary endpoint) with CEP.However, the study was underpowered for this outcome.For this reason, the ongoing PROTECT-TAVI-BHF (ISRCTN16665769) was designed, in order to be adequately powered to detect an eventual meaningful prevention of disabling stroke.
The trial is open to all adult patients scheduled for TAVI at participating specialist cardiac centres across the UK who can receive the CEP device.The trial aims to recruit 7730 participants and is anticipated to provide insights that would validate our results.
Measuring the impact of CEP on periprocedural TAVI complications is not a simple exercise.It should be highlighted that the risk of stroke in patients undergoing TAVI can be divided into at least two components.The periprocedural risk is related to valve dilatation, handling of the delivery system in the aorta, and the release of the prosthesis itself, but also to indirect periprocedural phenomena linked to the activation inflammation and thrombosis due to both pharmacological and mechanical stress but also by the presence of bulky catheters and the TAVI prosthesis itself. 42The rationale for using CEP devices is to positively impact some of the above-mentioned mechanisms.In this regard, it is important to underline the different concepts adopted by the multiple devices used in the studies included in this meta-analysis.As we know, CEPD are designed to block embolic debris, either by means of filtering mechanism (similar to a net that removes emboli from circulation), or deflection mechanism (redirecting emboli away from cerebral circulation towards the systemic circulation).The effectiveness of the device depends upon the ability to protect the three openings of the major branches of the aortic arch. 43The Embrella Embolic Deflector (EED) (Edwards Lifesciences; Irvine, California, United States) works by deflecting the emboli, with a dual membrane system with two opposing petals positioned along the greater curvature of the aorta covering the ostia both the brachiocephalic and the left common carotid arteries.Using a dual membrane system with two opposing petals positioned along the greater curvature of the aorta, the Embrella Embolic Deflector (EED) (Edwards Lifesciences) deflects emboli from the openings of both the brachiocephalic and left common carotid arteries. 20,44imilarly, the Sentinel Cerebral Protection Device (Claret Medical) is also placed in the brachiocephalic trunk and the left common carotid artery but removes the emboli debris through a filtering mechanism. 18One limitation of the above-discussed CEPD devices is that they do not offer full protection, as the left subclavian artery is not covered.TriGuard (Keystone Heart Ltd.) also works by deflection mechanism but is unique in the aspect of covering the left subclavian artery providing protection against all territories of the brain. 45ewer CEPD devices are under development to improve cerebral protection and capture both cerebral and noncerebral debris.Among them, the Emboliner has not yet been made available commercially in many countries and clinical studies are still ongoing.The company behind this development has initiated the ProtectH2H trial (NCT05684146), which is a prospective, randomized, open-label, multicentre, two-arm investigation.Its primary objective is to demonstrate the safety and effectiveness of the Emboliner device when compared to a control device, specifically the Sentinel cerebral protection system from Boston Scientific Corporation.This trial focuses on patients undergoing TAVI. 46,47The large heterogeneity brought about by the clinical question under evaluation here makes it difficult to speculate about the potential explanation for the different results among studies.Most probably, the general outlook of the approach represents the result of a complex interaction of multiple factors, including procedural ease and reproducibility along with effectiveness in filtering debris.Finding the sweet spot, perfectly balancing effectiveness, safety and practical easiness of the procedure might be therefore the aim to improve patients' outcomes in the real world.In this context, it is very reassuring that we found no impact on procedural complications and bleeding, suggesting that CEP is generally a safe procedure in experienced centers.Similar meta-analyses were performed before on this topic.However, to the best of our knowledge, this is the first one demonstrating an adequately powered conclusive finding, as supported by the results of trial sequential analysis and cumulative meta-analysis.This is relevant evidence in support of our findings, especially in a contest of studies with multiple sources of heterogeneity, including study F I G U R E 4 Meta-analysis results for device type subgroups.The figure reports the results of the meta-analysis comparing TAVI performed with CEP against standard TAVI without CEP.Every panel reports the results of a specific outcome, including Stroke (Figure 3, panel A), the volume of cerebral lesions measured with MRI (Figure 3, panel B), all-cause death (Figure 3, panel C), bleeding (Figure 4, panel A), vascular access-related complications (Figure 4, panel B), acute kidney injury (Figure 4, panel C).Studies names are listed on the left with the year of publication for each panel.Individual risk ratios with 95% confidence intervals are provided in the middle part of the figure, together with the calculations for the meta-analysis.The Forest plot on the right is the graphical representation of meta-analysis results.The position of the square indicates the mean effect size in each single study, whereas the area of the square reflects study weight.The black horizontal line depicts the 95% Confidence Interval of the effect size in each study.The red diamond at the very bottom of the Forest plot shows the overall summary effect, while summary effects are also provided for device type subgroups (blue diamonds).A summary effect positioned on the left of the vertical line implies a lower rate of the single endpoint in the CEP arm.CEP, cerebral embolic protection; CI, confidence interval; RR, risk ratio; Std, standard deviation.
design, the architecture of the CEP device, and the time the studies were run.In fact, the rapid evolution of TAVI devices, delivery systems and preprocedural planning produces an evolving landscape.The lower-than-expected stroke rates after TAVI in the studies included can be attributed to several factors.First, advancements in TAVI technology and procedural techniques have improved over time, leading to enhanced precision and reduced complications.Moreover, the experience and expertise with TAVI procedures have increased, contributing to better patient outcomes.Additionally, improved patient selection criteria and preprocedural assessments help identify individuals who are more likely to have successful outcomes with lower risks, including stroke.Similarly, the progressive expansion of the indications to TAVI to younger patients and lower clinical risk groups has certainly contributed.Finally, with the definition of the study endpoint itself might play a role.
While the primary focus of CEP is understandably on preventing overt strokes, it is crucial to acknowledge the significance of 'silent' brain lesions.Even in cases where overt stroke rates may not be notably high, the long-term implications of these subtle lesions cannot be overlooked.These inconspicuous abnormalities likely play a role in contributing to cognitive impairment and early dementia.These outcomes gain increased relevance as TAVI extends its reach to younger patients with extended life expectancies.This represents a crucial element also for the design of future studies.In fact, adopting stroke as the main study outcome becomes problematic with the relatively low stroke rates observed in the meta-analysis after TAVI (1,8% overall).In fact, about 38,500 patients would be required for a superiority trial with an 80% power to detect the ability of CEP to reduce stroke prevalence by 20% in the experimental group.Alternatively, approximately 24,000 patients would be required for a superiority trial with an 80% power to detect the ability of CEP to reduce stroke prevalence by 25%.Hence, efforts are needed to adopt composed yet clinically relevant endpoints, and consider hierarchic classification and possibly a win ratio approach in this setting.Furthermore, results obtained so far suggest that the identification of more stringent selection criteria to identify the higher risk patients that will benefit most from CEP during TAVI might simplify the design of a trial to finally answer this research question, at the same time increasing the net benefit to patients and the cost-effectiveness of CEP.
Despite advancements, the adoption of embolic protection devices during TAVI remains constrained.This limitation may stem not only from the notable clinical discordance between the rates of clinically evident strokes and the discovery of asymptomatic lesions on MRI scans but also from challenges in accurately assessing clinical strokes.In fact, discrepancies in stroke rates are influenced by factors such as the presence or absence of routine neurological assessments and reliance on self-reporting.Additionally, a significant deterrent to the widespread utilization of CEP devices during TAVI is their inability to provide comprehensive protection to all cerebral vessels.For instance, the sentinel device falls short in safeguarding the left vertebral artery and the descending aorta.Overcoming the limited adoption of these devices requires addressing the multifaceted factors involved, including operator preferences, accurate risk assessment, cost perception by payers and a comprehensive approach to cerebral vessel protection.Hence, enhancing cost-effectiveness considerations might play a pivotal role in encouraging broader acceptance and implementation. 48lthough TAVR is a continually growing procedure, the femoral approach often cannot be performed.Recently, new alternative access such as trans-axillary 49 access or trans-caval access are emerging.The median Society of Thoracic Surgeons predicted risk of mortality was similar among patients treated by trans-carotid, trans-apical and trans-aortic access, but there is little information on the use of embolic protection devices when TAVR is performed with accesses other than the femoral one.Despite this, many CEP devices have been developed with varying access routes (i.e., brachial).However, the use of thromboembolic protection devices does not influence the access route from which TAVR is performed.Theoretically, for anatomical reasons, the trans-axillary access route may be burdened by a lower embolization rate.On the other hand, a trans-caval access could be burdened by a higher embolization rate and the routine use of embolic protection devices could be more rational.
Consequently, conclusions from this analysis have to be drawn with caution.Future studies are warranted to better define the potential role of routine use of CEP devices during TAVI.

| Study limitation
The included studies are heterogeneous, including the study design (not all of them are randomized trials), patient characteristics, the implanted prosthesis, type of CEP devices and the investigated primary endpoints.Furthermore, additional hurdles make it even more difficult to decipher the actual clinical impact of CEP as a strategy to prevent periprocedural stroke in TAVI.Among these is the impact of atrial fibrillation (AF) and emergent periprocedural AF, as well as the presence of thrombotic stratifications on native valve tissues or the prosthesis itself. 50,51In addition, despite periprocedural cerebral events are largely prevalent, some post-procedural events and especially silent strokes-on which CEP cannot exert any impact-might have been counted together with periprocedural events, as several hours are often necessary for their diagnosis using imaging techniques. 52

| CONCLUSIONS
Results of the present meta-analysis of clinical studies investigating the impact of CEP during TAVI suggest a significant reduction in the risk of stroke, without any signal for increased rates of complications, such as AKI, bleeding, or vascular access complications.However, this field is still under development and ongoing larger scale RCTs will soon provide additional evidence to guide the clinical use of CEP devices in specific population subgroups.

F I G U R E 1
PRISMA flow chart.The figure depicts the PRISMA-based screening protocol.T A B L E 2 Randomized and observational trials.

2
Meta-analysis results.The figure reports the results of the meta-analysis comparing TAVI performed with CEP against standard TAVI without CEP.The first panel (panel A) reports the results of our primary endpoint (stroke).The second panel reports the results of all-cause death (panel B).Studies names are listed on the left with the year of publication for each panel.Individual risk ratios with 95% confidence intervals are provided in the middle part of the figure, together with the calculations for the meta-analysis.The Forest plot on the right is the graphical representation of meta-analysis results.The position of the square indicates the mean effect size in each single study, whereas the area of the square reflects study weight.The black horizontal line depicts the 95% Confidence Interval of the effect size in each study.The red diamond at the very bottom of the Forest plot shows the overall summary effect. 19 23 33