1. Protection Devices in Stroke Prevention
CRT 2017
Protection Devices in Stroke Prevention
Does the Data support Routine Use?
Eberhard Grube, MD, FACC, FSCAI
Professor of Medicine, Head, Center of Innovative Interventions in Cardiology (CllC), University Hospital Bonn, Germany
Consulting Professor, Division of Cardiovascular Medicine,
Stanford University School of Medicine, USA

2. Disclosure Eberhard Grube, MD
Physician Name Company/Relationship
Speaker Bureau/Advisory Board: Medtronic: C, SB, AB, OF
LivaNova: C, SB, AB
Highlife: AB, SB
Boston Scientific: C, SB, AB
Equity Interest: Millipede: E, SB, C, InSeal Medical: E, AB, Valtech: E, SB,
Claret: E, AB
Shockwave: E, AB
Valve Medical: E, AB
Mitra/Trilign E, AB, SB
Key
G – Grant and or Research Support E – Equity Interests S – Salary, AB – Advisory Board
C – Consulting fees, Honoraria R – Royalty Income I – Intellectual Property Rights
SB – Speaker’s Bureau O – Ownership OF – Other Financial Benefits‘

3. What is a stroke?
Duration of focal or global neurological deficit* (using specific criteria) 24hrs; OR 24hrs if neuroimaging documents a new hemorrhage or infarct; OR the neurological deficit results in death.
*Neurological deficit = acute episode of a focal or global neurological deficit with at least one of the following: change in the level of consciousness, hemiplegia, hemiparesis, numbness, or sensory loss affecting one side of the body, dysphasia or aphasia, hemianopia, amaurosis fugax, or other neurological signs or symptoms consistent with stroke
Defining stroke has been a contested and evolving discussion in the clinical arena. Definitions vary by clinician, clinical specialties, and studies
Sacco, R et al: AHA/ASA Expert Consensus Document. An Updated Definition of Stroke for the 21st Century. Stroke. 2013;44:

4. In Stroke most damage is unseen
Clinically apparent
Major/disabling stroke
Transient ischemic attack (TIA)
Minor/non-disabling stroke
Subtle and often undetected
“Silent” cerebral infarcts
Clinically unrecognized
….but can have far-reaching effects

5. Cerebral Injury Known consequences of cerebral injury:
obvious - apparent - quiet subtle silent subclinical
2-4% % % % ?%
Known consequences of cerebral injury:
Increased risk of: later CVA, cognitive impairment, vascular dementia

6. TAVR Stroke
Mechanism
Taken together, the evidence suggests that the driving mechanism behind procedural stroke is an impaired washout of embolic debris, which likely results from an interplay of procedural and patient factors
Procedural
Rapid ventricular pacing
Other low cardiac output states
Patient
Carotid artery occlusive disease
Prior CVA
High gradient stenosis
Friction caused by procedural steps
Arch atheroma
Quality and quantity of valve calcification
Sub-therapeutic procedural ACT
Sub-optimal pre and post TAVI antiplatelet protocols
Non-compliance with medications
Hyperviscosity

7. How is stroke impairment severity assessed?
Critical to:
acknowledge which definitions are used when appraising clinical studies
If the patient was impaired at baseline (i.e. floor effect)
The Valve Academic Research Consortium (VARC) has established standard definitions for TAVR practice and research
VARC I uses generally excepted stroke definition, and incorporates the use of the
Modified Rankin Score (mRS) to assign level of impairment.
Minor= mRS of <2 at 30 and 90 days
Major= mRS score >2 at 30 and 90 days
VARC II 2012– as per VARC I, but incorporates patient pre-stroke baseline scores into analysis, and adds level of functioning to classification.
Disabling stroke: An mRS score of ≥2 at 90 days and an increase in at least 1mRS category from an individual’s pre-stroke baseline.
Non-disabling stroke: an mRS score of <2 at 90 days or one that does not result in an increase
in at least one mRS category from an individual’s pre-stroke baseline.

8. TAVR: Sources of Cerebral Emboli
Shaggy Aorta
1Protruding aortic arch atheromas: risk of stroke during heart surgery with and without aortic arch endarterectomy. Stern et al. American Heart Journal Oct
2Atheromas of the thoracic aorta: clinical and therapeutic update. Tunick et al. J. Am. Coll. Of Card. March 2000
3Atheromatous disease of the thoracic aorta and systemic embolism. Clinical picture and therapeutic challenge. Sheikhzadeh et al. Z Cardiol. Jan. 2004
Manipulation of the degenerative Aortic Valve 9

9. Majority of Strokes Occur Periprocedurally
P. Kahlert et al, Circulation 2012;126:

10. (data from the CoreValve ADVANCE Study)
TAVR vs. SAVR
Randomized Trials
The majority of post-TAVR strokes, regardless of time point, have an ischemic etiology
(data from the CoreValve ADVANCE Study)

11. TAVR vs. SAVR Randomized Trials
As practice evolved, randomized trials between TAVR and SAVR showed no difference in the rate of 30-day stroke using both self-expanding and balloon-expanding valves
p=0.12
p=0.46
p=0.57
p=0.37
p=0.46
p=0.37
p=0.12
p=0.57

12. Weighted average (n=4,795) ~3.5%
TAVR Stroke
Rates with Contemporary Devices
In contemporary practice, the overall stroke rate remains around 3,5%
Weighted average (n=4,795) ~3.5%
1Manoharan, et al., J Am Coll Cardiol Intv 2015; 8: ; 2Moellman, et al., presented at PCR London Valves 2015; 3Linke, et al., presented at PCR London Valves 2015; 4Kodali, et al., Eur Heart J 2016; doi: /eurheartj/ehw112; 5Vahanian, et al., presented at EuroPCR 2015; 6Webb, et. al. J Am Coll Cardiol Intv 2015; 8: ; 7DeMarco, et al, presented at TCT 2015; 8Meredith, et al., presented at PCR London Valves 2015; 10Falk, et al., presented at EuroPCR 2016; 11Kodali, presented at TCT 2016

13. Timing, Risk Factors, Outcomes of Stroke, TIA after TAVR: PARTNER
2,621 participants in the PARTNER trial and continued-access registry: 30-d, 1 & 3 years
Stroke incidence was 3.3% at 30d (of which 85% occurred within 1 week, peaking at day 2)
TF-TAVR stroke rates were 3.8% at 30d, 5.4% at 1y, and 6.9% at 3y
Patients experiencing a stroke or TIA had lower 1-year survival
With stroke: 47% 1-yr survival vs 82% without for TF-TAVR
With TIA: 64% 1-yr survival vs. 83% without for TF-TAVR
Conclusions
Risk of stroke or TIA is highest early after TAVR
Stroke and TIA after TAVR are associated with increased risk of 1-yr mortality
This highlights the need for embolic protection devices, anti-platelet therapy, and procedural modifications
Kapadia, et al. Circ Cardiovasc Interv. 2016;9:e002981

14. Procedural predictors Predictors of late stroke
Neurological Events Following TAVR and Their Predictors: CoreValve Trials 3,687 patients from the two CoreValve US pivotal trials and Continued Access Study followed for a total of 3,581 person-years
8.4% stroke rate 1-year post- TAVR
4.1% in early phase (0-10d)
4.3% in late phase (11-365d)
There were no echo or CTA imaging predictors of stroke
There was no learning curve or site experience effect observed
Patient predictors
of early stroke
Procedural predictors
Predictors of late stroke
NIHSS score >0
Repositioning of prosthesis
Small body surface area
Prior stroke
Longer catheter delivery time
Severe aortic calcification
Prior TIA
Longer procedure time
6mo history of falls
PVD
Rapid pacing used
Angina
Low BMI
Conclusions
The rate of stroke after TAVR remains a significant concern
Readily available imaging data does not contribute to the ability to predict stroke
Further efforts to limit stroke, such as embolic protection devices and new anti-thrombotic regimens, merit investigation
Kleiman, et al. Circ Cardiovasc Interv. 2016;9:e003551

15. TVT Registry shows no significant decline in stroke rate over time
Over 53,000 US TAVR patients
No significant decline in stroke rate over time
Holmes D, et al. ACC 2016

16. TVT Registry shows stroke risk is independent of experience
Over 53,000 US TAVR patients from >350 US centers
No significant decline in stroke rate as centers gain experience
Self-reported rates without prospective neurologist exams pre and post-procedure likely underestimate true rates
Carroll J, et al. ACC 2016

17. TAVI 30d major stroke rates similar across the surgical risk spectrum
n=389 consecutive TAVI patients treated in Bern
Low vs. high
RR (95% CI)
0.73 ( )
Intermediate vs. high
RR (95% CI)
0.97 ( )
Overall p-value p=0.83 no significant differences
STS <3%
STS 3-8%
STS >8%
Wenawesser P, et al. Clinical outcomes of patients with estimated low or intermediate surgical risk undergoing transcatheter aortic valve implantation. Eur Heart Journal :

18. In-hospital stroke rate
TAVI and SAVR in-hospital stroke rates are similar across the surgical risk spectrum Complete dataset of TAVI and SAVR patients treated in 2013 in Germany (n=20,340 patients)
In-hospital stroke rate
Low risk
Intermediate risk
High risk
Overall in-hospital stroke rate of 2.3%
Stroke occurred more frequently in low-risk patients treated with trans-apical TAVI (TA-AVI)
There were no statistically significant differences in stroke rates for all other comparisons
Möllmann H, et al. Clin Res Cardiol 2016

19. What are predictors of stroke and cerebral damage in TAVI?
Logistic EuroSCORE is NOT a predictor of stroke in TAVI1
Predictors of neurological events in TAVI studies have included:
Trans-arterial vs Trans-apical access2
Post-dilatation and valve dislodgement3
Smaller indexed valve area4
Cerebrovascular disease4
Predictors of new cerebral damage (DW-MRI) in TAVI studies have included:
Age, hyperlipidemia, post-dil5
Severity of atheroma (arch and descending), catheterization time, age6
Peak trans-aortic gradient5
Zeinah, et al EU TAVR Registry Review and Meta Analysis. ACTA 2015
Eggebrecht, et al. Eurointervention 2012
Nombela-Franco, et al. Circulation 2012
Miller, et al. PARTNER study. JTCVS 2012
Samim, et al. Clin Res Cardiol 2015
Fairbairn, et al. Heart 2012

20. Stroke in TAVR is likely more frequent that thought
The stroke rate in the ‘Control’ arm of SENTINEL independently adjudicated by Stroke neurologists
The true incidence of stroke is most likely UNDER-reported in many trials!!
SENTINEL used comprehensive prospective serial neurological assessment and CEC adjudication by two stroke neurologists 30

21. Stroke Rates in Context
The true prevalence of stroke is most likely under estimated because:
Patients are not routinely neurologically examined before and after TAVR
Neurologic examination is predominantly performed by interventionalists and surgeons and not stroke neurologists
Registries have generally used voluntary reporting of outcomes and retrospective chart abstraction
Standardized definitions for stroke (VARC II) have not been universally adopted
Often only MAJOR or Disabling strokes are reported
Messe S et al: Improving Outcomes From Transcatheter Aortic Valve Implantation Protecting the Brain From the Heart JAMA August 9, 2016 Volume 316, Number 6

22. Stroke Prevention - Current State of Play -

23. Procedural Stroke Prevention
Optimized Anticoagulation
Embolic Protection

24. Current Cerebral Protection Devices
Main Attributes
TriGuard Embolic Deflection Device (Keystone Heart)1
Sentinel Cerebral Protection System (Claret Medical)2
Embrella Embolic Deflector System (Edwards Lifesciences)3
Pore Size: 130 µm
Delivery Sheath: 9F
Access: Transfemoral
Mechanism: Debris deflection
Pore Size: 140 µm
Delivery Sheath: 6F
Access: Brachial or radial
Mechanism: Debris capture and retrieval
Pore Size: 100 µm
Access: Brachial
1Lansky, et. al. , presented at TCT 2015; 2Van Mieghem, et al., presented at TCT 2015; 3Rodes-Cabau, et al., J Am Coll Cardiol Intv 2014;7:
1Lansky, et. al. , presented at TCT 2015; 2Van Mieghem, et al., presented at TCT 2015; 3Rodes-Cabau, et al., J Am Coll Cardiol Intv 2014;7:

25. Debris and Fragments of aortic valve leaflet
Neurologic Injury
How Does it Happen?
Van Mieghem, et al., placed Claret Montage filters into the brachiocephalic and left common carotid arteries during TAVR, and examined the contents after the procedure.
The key findings:
Macroscopic debris was released into the circulation in ~90% of procedures
The debris was composed of thrombotic material, fragments of valve leaflet, calcified particles, myocardial tissue, and plastic fragments from interventional tools
Debris and Fragments of aortic valve leaflet
1Van Mieghem, et al., J Am Coll Cardiol Intv 2015; 8:

26. Embolic Protection Device
Neurologic Injury
Embolic Protection Devices
Every study performed to date has shown that embolic protection devices reduce the amount of ischemic damage measured in the brain by MRI
All studies were small, which affected statistical outcomes
Embolic protection is a viable treatment strategy that should be studied further and considered for patients at increased risk for neurologic events
Acute (Day ≤ 7)
Study
Embolic Protection Device
+EPD
Control
P value
PROTAVI-C Pilot
Edwards Embrella
43.0 mm3
47.5 mm3
0.583
DEFLECT III
Keystone TriGuard
111 mm3
202 mm3
Not reported
MISTRAL-C
Claret Sentinel
95 mm3
197 mm3
0.171
CLEAN-TAVI
Claret Montage
471 mm3
800 mm3
0.024
SENTINEL
294 mm3
309.8 mm3
0.8076
1Rodes-Cabau, et al., J Am Coll Cardiol Intv 2014;7: ; 2Lansky, et al., Eur Heart J 2015;36:2070-8; 3Van Mieghem, et al., EuroIntervention 2016;12: ; 4Haussig, et al., JAMA 2016;316: ;5Kapadia, et al., J Am Coll Cardiol 2016; epub ahead of print

27. TriGuard (Keystone) DEFLECT III | Day 4 Imaging
Complete freedom from neurologic injury was 57% higher in TriGuard patients
Lesions that formed were 44% smaller in TriGuard patients
p=0.07
1Lansky, et al., Eur Heart J 2015;36:2070-8

28. TriGuard (Keystone) DEFLECT III | Neuro-function
Protected patients experienced less neurologic impairment at the time of hospital discharge
p=0.011
1Lansky, et al., Eur Heart J 2015;36:2070-8

29. Montage (Claret) CLEAN-TAVI | Day 7 Imaging
98% of patients (protected and unprotected) showed some form of neurologic injury on MRI. This high rate results from the sensitivity of the 3-T scanner
Montage significantly reduced total lesion volume by 57%
Lesion Number per Patient
Total Lesion Volume per Patient
1Linke, et al., presented at TCT 2014

30. Claret Medical SENTINEL Study Design
The first FDA approved. randomized pivotal IDE study with prospective neurologist exams confirming the therapeutic importance of embolic debris capture and removal during TAVR
Objective: Assess the safety and efficacy of the Claret Medical Sentinel Cerebral Protection System in reducing the volume and number of new ischemic lesions in the brain and their potential impact on neurocognitive function
Primary Investigators:
Samir Kapadia, MD
Cleveland Clinic
Susheel Kodali, MD
Columbia University Medical Ctr
Axel Linke, MD
Leipzig University, Germany
Population: Subjects with severe symptomatic calcified native aortic valve stenosis who meet the commercially-approved indications for TAVR with the
Edwards Sapien THV/XT/S3 or Medtronic CoreValve/EvolutR
N=363 subjects randomized 1:1:1
at sites in the U.S and Germany.
SAFETY ARM
TAVR with Sentinel
TEST ARM
TAVR with Sentinel
CONTROL ARM
TAVR only
Histopathology
Safety Follow-up
Safety Follow-up
MRI Assessments
Neurological and Neurocognitive Tests
Primary Efficacy Endpoint: Reduction in median total new lesion volume as assessed by 3-Tesla DW-MRI with baseline subtraction
Primary Safety Endpoint: Occurrence of all MACCE at 30 days.
CAUTION: Investigational device. Limited to investigational use by United States law.

31. 30-Day Clinical Outcomes
SENTINEL
30-Day Clinical Outcomes
Device Arm
Control Arm
p-value
30-day Clinical Outcomes
Any MACCE†
7.3% (17/234)
9.9% (11/111)
0.40
Death (all-cause)
1.3% (3/234)
1.8% (2/111)
0.65
Stroke
5.6% (13/231)
9.1% (10/110)
0.25
Disabling
0.9% (2/231)
0.9% (1/109)
1.00
Non-disabling
4.8% (11/231)
8.2% (9/110)
0.22
AKI (Stage 3)
0.4% (1/231) 0%
TIA
Sentinel Access Site Complications
N/A
0.53
†MACCE defined as Death (any cause), Stroke (any), Acute Kidney Injury (Stage 3)

32. SENTINEL / Safety
The Sentinel device is SAFE with successful delivery in almost all patients! … and is compatible with standard cath lab workflow
99% one filter and 94% both filters successfully deployed
0.4% access site complications
1% interaction with other devices (during TAVR)
91% Sentinel devices deployed in < 10 mins

33. SENTINEL Stroke within 30 days Test and Control Arms
NO peri-procedural strokes in the Test (Sentinel) Arm
NO disabling strokes in the Test (Sentinel) Arm
NO early strokes in Sentinel protected areas

34. SENTINEL Debris Removal
The Sentinel device successfully captured and removed debris in 99% of patients.

35. Primary Efficacy Endpoint
SENTINEL
Primary Efficacy Endpoint
p = 0.33
42.2% reduction [95% CI: -3.2,67.6)

36. Latib and Pagnesi MRI Meta-Analysis (JACC 2017) shows Claret filters significantly reduce cerebral new lesion volume on MRI
Total patients
Studies included
Patient level
Statistical method used
Notes
314
165 w/EP
149 no EP
CLEAN-TAVI
MISTRAL-C
SENTINEL No
Study-level pooled analysis
MRI data only
Measure
Value
95% CI
p-value I²
Weighted Mean Difference in Total New Lesion Volume (All Territories, mm3)
mm3
(Weighted Mean Diff.)
(-218.2, -10.5)
0.031
80%
Forest plot shows results for new total lesion volume in patients undergoing transcatheter aortic valve replacement with versus without
cerebral embolic protection (CEP) filters. The weighted mean difference (WMD) among groups equals mm3 (95% confidence interval
[CI]: mm3 to 10.5 mm3), thus confirming a significant reduction in the analyzed endpoint (p value of 0.031). CLEAN-TAVI Claret
Embolic Protection and TAVI (transcatheter aortic valve implantation); MISTRAL-C MRI (magnetic resonance imaging) Investigation in TAVI
with Claret; SENTINEL Cerebral Protection in Transcatheter Aortic Valve Replacement.
Latib and Pagnesi JACC 2017; 69(4):378-80
CAUTION: Investigational device. Limited to investigational use by United States law.

37. Giustino, et al. Clinical Event Meta-Analysis (JACC 2017) shows significant >40% reduction in risk of stroke or death with protection
Total patients
Studies included
Patient level
Statistical method used
625
376 w/EP
249 no EP
CLEAN-TAVI
MISTRAL-C
SENTINEL
DEFLECT III
EMBOL-X No
Aggregate data meta-analysis according to PRISMA guidelines
Random effects model
Measure
Value
(Relative Risk Ratio M-H, Fixed)
95% CI
p-value I²
Death or stroke
0.57
(0.33 to 0.98)
0.04 0%
Death
0.42
(0.14 to 1.26)
0.12
Stroke
0.66
(0.36 to 1.23)
0.20
Clinical Outcomes in Patients Undergoing TAVR With Versus Without Embolic Protection Devices Pooled effect estimates for the risk of death or stroke according to the use of cerebral embolic protection versus not during TAVR. CI = confidence interval; CLEAN-TAVI = Claret Embolic Protection and TAVI; DEFLECT-III = A Prospective, Randomized Evaluation of the TriGuard HDH Embolic Deflection Device During TAVI; EP = embolic protection; M-H = Mantel-Haenszel; MISTRAL-C = MRI Investigation With Claret; SENTINEL = Cerebral Protection in Transcatheter Aortic Valve Replacement; TAVR = transcatheter aortic valve replacement.
Giustino, et al. JACC 2017; 69(4):465-6
CAUTION: Investigational device. Limited to investigational use by United States law.

38. Clinical Event Meta-Analysis of Cerebral Embolic Protection in TAVR shows significant >40% reduction in risk of stroke or death with protection
Clinical Outcomes in Patients Undergoing TAVR With Versus Without Embolic Protection Devices Pooled effect estimates for the risk of death or stroke according to the use of cerebral embolic protection versus not during TAVR. CI = confidence interval; CLEAN-TAVI = Claret Embolic Protection and TAVI; DEFLECT-III = A Prospective, Randomized Evaluation of the TriGuard HDH Embolic Deflection Device During TAVI; EP = embolic protection; M-H = Mantel-Haenszel; MISTRAL-C = MRI Investigation With Claret; SENTINEL = Cerebral Protection in Transcatheter Aortic Valve Replacement; TAVR = transcatheter aortic valve replacement.
Meta-analysis of 5 randomized trials of cerebral protection in TAVR
Including 625 patients (376 with, and 249 without protection)
>40% reduction in risk of stroke or death (6.4% vs 10.8%; RR: 0.57; 95% CI: ; p=0.04; I2 = 0%)
Number-needed-to-treat (NNT) = 22 to reduce one stroke or death
“In conclusion, the totality of the data suggests that use of embolic protection during TAVR appears to be associated with a significant reduction in death or stroke.”
Giustino, et al. JACC 2017; 69(4):465-6
CAUTION: Investigational device. Limited to investigational use by United States law.

39. Conclusion
TAVI is an embologenic procedure and stroke remains a consistent issue despite device and procedure iteration
SENTINEL has confirmed, definitively, the link between new DW-MRI lesions and neurocognition
Patients expectations of TAVI vs SAVR are high as populations move to younger ‘lower-risk’ profiles
The concept of ‘silent ischemia’ is a myth – cerebral lesions do not just ‘disappear over time’: Ischemia is Ischemia.
Further studies are ongoing, examining the role of different valve types and of different pharmacological options to impact stroke that occurs beyond the procedural timepoint

40. Shall we wait for conclusive Data before we use CEP routinely?
Let’s use them until Data prove otherwise!

41. Thank you for your Attention !