0. RESPECT Extended Follow-up Results
John D. Carroll, M.D.
Acknowledgements
Jeffrey L. Saver, M.D. David E. Thaler, M.D., Ph.D. Richard Smalling, M.D., Ph.D. Lee A. MacDonald, M.D. David S. Marks, M.D. David L. Tirschwell, M.D.
for the RESPECT Investigators

1. Disclosure Statement of Financial Interest
Within the past 12 months, I or my spouse/partner have had a financial interest/arrangement or affiliation with the organization(s) listed below.
Affiliation/Financial Relationship
Company
Consulting Fees
Steering Committee Member for RESPECT
St Jude Medical

2. Unmet Need for Secondary Prevention of Recurrent PFO-Related Strokes
Young patients exposed to decades of risk for PFO-related stroke
1-2% event rate per year
RESPECT assessed PFO closure as a treatment option
Hart et al. Lancet Neurol 2014;13(4):

3. Key Aspects of RESPECT Trial
Device trial for secondary prevention
Superiority trial: PFO closure vs. guideline-directed medications
Largest randomized PFO trial: 980 patients
499 AMPLATZER™ PFO Occluder; 481 MM
Assumptions
Paradoxical embolism was cause of initial stroke
Recurrent strokes would be due to recurrent paradoxical embolism

4. RESPECT Trial Population
Included:
Subjects with a PFO who have had a cryptogenic stroke within the last 270 days
Excluded:
Subjects aged <18 years or >60 years
Subjects with identified stroke etiology
Subjects who are unable to discontinue anticoagulants
Carroll et al. NEJM 2012;368:

5. RESPECT Primary Endpoint Results
Enrollment ended when 25 ischemic stroke events occurred - results were reported in NEJM
Analysis Population
Relative Risk Reduction
P-Value
Intention-to-Treat
50%
0.089
Per-Protocol
58%
0.048
As Treated
67%
0.013
Carroll et al. NEJM 2012;368: Note: Per Protocol and As Treated analysis modified from NEJM analysis in response to FDA questions.

6. Extended Follow-up Provides Considerable New Data
AMPLATZER™
PFO Occluder
(N=499)
Medical Management
(N=481)
Mean Follow-up (years)
Initial Analysis
3.0
2.7 
Extended Follow-up
5.5
4.9 
Total Patient-Years of Follow-up
1476
1284
2769
2376

7. Higher Discontinuation Rate in MM Arm 11% of MM Subjects: Off-Label PFO Closure
HR: 0.560
Log-rank p-value: <0.0001 0%
50% 1 2 3 4 5 6 7 8 9 10
Time to Event (Years)
Discontinuation
Rate
10%
20%
30%
40%
AMPLATZER™ PFO Occluder (N=499)
Medical Management (N=481)
499
463
369
212 86 20
481
394
307
168 71
AMPLATZER MM
# at Risk (Discontinuation Rate)
(0%)
(4.9%)
(10.0%)
(15.4%)
(22.8%)
(30.3%)
(14.4%)
(18.8%)
(26.5%)
(33.5%)
(43.3%)

8. 1 out of 5 Patients Were >60 Years in Extended Follow-up Analysis
As patients age, increase in non-cryptogenic strokes expected
PFO closure can only reduce risk for recurrent strokes mediated by paradoxical embolism
Appropriate clinical interpretation of trials requires adjudication for stroke mechanism

9. Blinded Adjudication of Stroke Cause Using ASCOD Phenotyping
ASCOD coding captures presence of possible stroke etiologies, and assigns a probability of relatedness (post-hoc)
Five phenotypes:
A = atherosclerosis
S = small vessel disease
C = cardiac pathology
O = other cause
D = dissection
Recurrent strokes classified as either cryptogenic or of known cause
Amarenco et al. Cerebrovasc Dis 2013;36:1-5

10. Nearly 1/3 of Recurrent Strokes in Extended Follow-up Are of Known Mechanism
Atherosclerosis = 1
Small Vessel Disease = 6
Cardioembolic = 5 (AF = 4, endocarditis = 1)
Other = 1 (radiation arteriopathy)
Dissection = 0
Subjects with Recurrent Ischemic Stroke
Cryptogenic (Possibly Paradoxical Embolism)
Known Mechanisms

11. How Do Recurrent Strokes from Known Mechanisms Confound Interpretation of RESPECT?]

12. 1 out of 3 Recurrent Strokes Had Mechanism That PFO Closure Cannot Prevent Extended Follow-up in ITT Population
0.85
0.90
0.95
1.00 1 2 3 4 5 6 7 8 9 10
AMPLATZER™ PFO Occluder (N=499; # strokes = 18)
Event-free
Probability
Medical Management (N=481, # strokes = 24)
Stroke of Known Mechanism
HR: n/a (non-proportional hazards)
Log-rank p-value: 0.16
499
463
369
212 86 20
481
394
307
168 71 10
AMPLATZER MM
# at Risk (KM Estimates)
(0%)
(1.6%)
(1.9%)
(3.6%)
(6.0%)
(3.2%)
(4.8%)
(5.1%)
(7.0%)
(12.4%)
Time to Event (Years)

13. Significant Reduction in Recurrent Cryptogenic Stroke 54% Relative Risk Reduction in ITT Population
0.85
0.90
0.95
1.00 1 2 3 4 5 6 7 8 9 10
Time to Event (Years)
Event-free
Probability
HR: 0.460
Log-rank p-value: 0.042
AMPLATZER™ PFO Occluder (N=499; # cryptogenic strokes = 10)
Medical Management (N=481, # cryptogenic strokes = 19)
Device not in place
499
463
369
212 86 20
481
394
307
168 71 10
AMPLATZER MM
# at Risk (KM Estimates)
(0%)
(1.2%)
(1.5%)
(2.5%)
(2.7%)
(4.1%)
(5.2%)
(10.8%)

14. 70% Relative Risk Reduction in Recurrent Cryptogenic Stroke With Device In Place
0.85
0.90
0.95
1.00 1 2 3 4 5 6 7 8 9 10
Event-free
Probability
HR: 0.302
Log-rank p-value: 0.004
AMPLATZER™ PFO Occluder Implanted (N=464; # cryptogenic strokes = 7)
Not Implanted
(N=516, # cryptogenic strokes = 22)
464
445
357
206 82 20
516
412
319
174 75 10
AMPLATZER
Not Implanted
# at Risk (KM Estimates)
(0%)
(0.9%)
(1.9%)
(3.0%)
(4.6%)
(5.7%)
(11.2%)
Time to Event (Years)

15. Additional Sensitivity Analysis
Is the superiority of PFO Closure more clearly seen in younger patients?
Analysis not dependent on stroke etiology phenotyping

16. Freedom from Recurrent Stroke of Any Mechanism: <60 Yrs 52% Relative Risk Reduction in ITT Sensitivity Analysis
0.85
0.90
0.95
1.00 1 2 3 4 5 6 7 8 9 10
Time to Event (Years)
Event-free
Probability
HR: 0.476
Log-rank p-value: 0.035
AMPLATZER™ PFO Occluder (N=475, # strokes = 12)
Medical Management (N=463, # strokes = 22)
475
417
308
166 69 15
463
353
254
124 51 9
AMPLATZER MM
# at Risk (KM Estimates)
(0%)
(1.8%)
(2.1%)
(3.3%)
(3.4%)
(4.9%)
(5.4%)
(6.9%)
(14.7%)

17. Does anatomy and physiology of PFO matter in terms of treatment effect?
Atrial septal aneurysm (ASA) and substantial right-to-left shunts are used by clinicians to identify PFOs that may not be incidental

18. Greater Benefit in Substantial Shunt or ASA Subgroup 75% Relative Risk Reduction in Recurrent Cryptogenic Stroke in ITT Population
0.85
0.90
0.95
1.00 1 2 3 4 5 6 7 8 9 10
HR: 0.245
Log-rank p-value: 0.007
Event-free
Probability
AMPLATZER™ PFO Occluder (N=319, # cryptogenic strokes = 4)
Medical Management (N=301, # cryptogenic strokes = 13)
319
299
229
134 52 11
301
243
186
105 45 7
AMPLATZER MM
# at Risk (KM Estimates)
(0%)
(0.6%)
(1.0%)
(1.5%)
(3.6%)
(4.8%)
(6.6%)
Time to Event (Years)

19. Summary of Efficacy Findings in Extended Follow-up
Analysis Population
(Endpoint)
Relative Risk Reduction
P-Value
Analysis Conclusion
ITT (All-Cause Stroke)
n/a*
0.16
Confounded by strokes of known mechanism
ITT (Cryptogenic Stroke)
54%
0.042
Efficacy for cryptogenic stroke prevention
Device In Place (Cryptogenic Stroke)
70%
0.004
Accounting for device placement increases efficacy
ITT: <60 years old
(All-Cause Stroke)
52%
0.035
Supportive sensitivity analysis
ITT: ASA/SS Subgroup
(Cryptogenic Stroke)
75%
0.007
Additional benefit in patients with ASA or SS
* non-proportional hazards (not appropriate to estimate)

20. Procedure or Device Related SAEs SAEs Adjudicated by DSMB
No intra-procedure strokes
No device embolization
No device thrombosis
No device erosion
Very low rate of major vascular complications (0.9%) and device explants (0.4%)

21. AMPLATZER™ PFO Occluder
Adjudicated SAEs of Interest Favorable SAE Profile for AMPLATZER™ PFO Occluder
Event Type
AMPLATZER™ PFO Occluder
(N=499)
[2769 Pt-Yrs]
Medical Management
(N=481)
[2376 Pt-Yrs]
Events
Rate*
Atrial fibrillation 7
0.25 4
0.17
Major bleeding 17
0.61 14
0.59
Death from any cause 6
0.22 10
0.42
DVT/PE 3
0.12
* Rate expressed as number of events per 100 patient-years
DVT/PE rate of unclear significance
Not associated with procedure/access site, thrombophilia evaluation not done in trial, and warfarin was allowed in MM group

22. Strengths and Limitations of RESPECT
High procedural success and effective closure rates
Longest follow-up of PFO closure RCTs
Adjudication of stroke mechanism
Limitations
Powered to detect overly optimistic treatment effect
Differential dropout rate could lead to bias
Significant rate of off-label PFO closure (11%)

23. Conclusions
AMPLATZER™ PFO Occluder is superior to medical management in reducing recurrent cryptogenic ischemic stroke
Treatment effect is fully manifest in types of strokes for which closure is intended
Superiority is substantial and sustained
Procedure and device are safe
RESPECT reinforces need for comprehensive risk factor modification