1) The CLOSURE-I trial found that percutaneous closure of a patent foramen ovale (PFO) with the STARFlex device plus medical therapy did not provide a significant benefit over medical therapy alone in preventing recurrent stroke or transient ischemic attack in patients under age 60 who had a cryptogenic stroke or TIA and a PFO. 2) The trial observed a higher rate of atrial fibrillation and major vascular complications in the device closure group compared to medical therapy alone. 3) The results were inconclusive as to whether PFO closure benefits patient subgroups based on the degree of shunting or presence of an atrial septal aneurysm.

1. Sydney Adventist Hospital
Hornsby Ku-ring-gai Hospital
Dr Jason Sharp MB BS FRACP FCSANZ
Consultant Cardiologist

2. History
 1877 – Conheim autopsy
 1972 – only 128 cases of unexplained stroke had been
reported in the literature
 1997 – Amplatzer ASD closure device used in animals –
nitinol double umbrella filled with polyester (Dacron)
fabric
 2012 – CLOSURE-I trial published

3. Ms EW case of a persistent neurologist
 43yo female. Sensory stroke symptoms but
clouded with history of possible migraine.
Subtle changes on MRI + page missing but on
further review it was felt there was a right
thalamic stroke fitting with the symptoms.
Mildly abnormal procoagulant screen. OCP
(ceased).
 TOE initially showed negative bubble study
via antecubital vein, no PFO, mobile
interatrial septum.
 Referred to me for second opinion.

4. Ms EW
 RepeatTOE revealed atrial septal aneurysm
and PFO with positive bubble study via right
femoral vein.
 Admitted to hospital for PFO closure. Lesion
unable to be crossed. Multiple bubble studies
negative while patient ventilated.
 Recommendation?

5. Ms EW continued…
 Readmitted to another hospital with
subsequent successful closure.

6. PFO detection
 TOE
 Bubble study
 Femoral vs ante-cubital vein (SVC blood directed
toward tricuspid valve, IVC blood directed toward
PFO).
 Saline vs dedicated echo contrast media
 Valsalva
 Degree of shunting (<5, 5-25, >25 bubbles)
 Transcranial doppler

7. Methods do matter
 Hamann et al: TOE/TCD detection rate was:
 11.4%/4.5% via antecubital injection
 18%/13.6% via antecubital injection plus the Valsalva
manoeuvre
 38.6%/36% via femoral injection alone
 50%/50% via femoral injection plus the Valsalva
manoeuvre
(Neurology 1998, 50: 1423-1428)

8. What is an Atrial Septal
Aneurysm?
 Redundant and hypermobile portion of
interatrial septum with >10mm excursion
from the centreline during the cardiac cycle.
 Some papers define >15mm total excursion.
 2.2% ofTOE patients
 4.3% of PFO patients

9. How does PFO and / or ASA cause
stroke?
1. Embolisation from the venous system (e.g. DVT) to the
arterial system & brain.
• But there is a low rate of DVT found in these patients.
• Look for history of Valsalva manoeuvre at time of
stroke.
2. In situ thrombus formation
3. Atrial dysfunction

10. Is PFO a stroke risk?
Overell, Bone & Lees, 2000 Neurology 55: 1172-1179.
Meta-analysis of case-control studies
Relative risks:
 PFO 1.83 (1.25-2.66; 15 studies)
 ASA 2.35 (1.46-3.77; 9 studies)
 Both 4.96 (2.37-10.39; 4 studies)

11. Is PFO a stroke risk?
Size of defect
Migraine history
More than 1 previous event
Other factors (external)
 Valsalva, cough, OSA
 Mechanical ventilation
 Surgical operations
 (joint replacement, sitting posture)
 Diving, aviation

13. Atrial dysfunction theory
 Rigatelli et al JACC (Cardiovasc Int) July 2009
 98 patients with PFO, previous stroke
 50 AF controls
 70 risk matched controls
 Measured left atrial emptying and several other
atrial function parameters.
 Atrial septal aneurysm was associated with worse
atrial function.
 Atrial function normalised after PFO closure.

14. Age, PFO and stroke
Overell et al 2000
Age range Relative Risk of Stroke
<55 years RR 6
>55 years RR 2.26

15. Randomised Data?
 Thanopoulos et al Catheterization & Cardiovasc
Interventions Nov 2006
 Non-randomised patient preference study of 92
patients with cryptogenic stroke and PFO.
 2 year follow-up of antiplatelet vs closure.
 0% events in closure group, 14.75% in antiplatelet
group.

16. What to do about PFO?
PFO in Cryptogenic Stroke Study PICSS Circulation 2002
 630 strokes; 34% had PFO; half to aspirin, half to
warfarin; 2 year follow-up, endpoints were death or
ischaemic stroke, many older patients
 No significant differences, if on treatment:
 With or without PFO
 Related to size of PFO
 With or without atrial septal aneurysm
 Between treatments
 BUT!! INR target was 1.4-2.8. Only 265 had CS!!
 In crypotogenic stroke with PFO 9.5% risk in warfarin
group, 16.3% in aspirin group but p=0.16

17. PFO and stroke Mas et al NEJM 2001
 Approximately 27% of “normal” people have a PFO.
 581 patients with cryptogenic stroke treated with
aspirin. 4 years follow-up. Prospective data.
Recurrent stroke risk
PFO and ASA 15.2%
PFO alone
or neither PFO nor ASA
2 to 4%
• Therefore aspirin is not providing adequate protection.
• SPARC data also showed ASA at high risk
• Spontaneous passage of bubbles also a risk factor

18. Study Design
 Prospective, multi-center, randomized, open-label, two-arm
superiority trial designed to test whether PFO closure using
STARFlex® plus medical therapy is superior to medical therapy
alone for preventing recurrent stroke orTIA in patients with
cryptogenic stroke orTIA and a PFO
 Study population: Patients 60 years old or younger with a
cryptogenic stroke orTIA and a PFO documented byTOE, with or
without atrial septal aneurysm, within 6 months of randomization
 DVT, hypercoagulopathy excluded
 Primary endpoint : 2-year incidence of stroke orTIA, all cause
mortality for the first 30 days, and neurological mortality 31 days
to 2 years

19. Baseline Characteristics ITT
STARFlex Medical P value
N randomized 447 462
Mean Age 46.3 (18-61) 45.7(18-61)
Male 52.1% 51.5%
White 89% 90%
Index cryptogenic
stroke
73% 71%
Mod/substantial
shunt*
58%
(231/400)
51%
(228/451)
0.04
ASA > 10 mm* 38%
(151/400)
35%
(160/451)
0.49
* modified ITT

20. 2 Year Primary Endpoint ITT
STARFlex
n = 447
Medical
n = 462
Adjusted
P value*
Composite 5.9%
(n=25)
7.7%
(n=30)
0.30
Stroke 3.1%
(n=12)
3.4%
(n=13)
0.77
TIA 3.3%
(n=13)
4.6%
(n=17)
0.39
*Adjusting performed using Cox Proportional Hazard Regression and adjusting for related patient characteristics including:
age, atrial septal aneurysm, prior TIA/CVA, smoking, hypertension, hypercholesterolemia

21. Adverse Events
STARFlex
N=402
Medical
N=458
P value
Major vascular
complications*
3.2%
(n =13)
0.0% <0.001
Atrial fibrillation 5.7%
(n= 14/23 periprocedural)
0.7%
(n=3)
<0.001
Major bleeding 2.6%
(n=10)
1.1%
(n=4)
0.11
Deaths (all non
endpoint)
0.5%
(n=2)
0.7%
(n=3)
ns
Nervous system
disorders
3.2%
(n=12)
5.3%
(n=20)
0.15
Any SAE 16.9%
(n=68)
16.6%
(n=76)
ns
*Perforation LA (1); hematoma >5cm at access site (4); vascular surgical repair (1); peripheral nerve injury (1);
procedural related transfusion (3);retroperitoneal bleed (3)

22. Composite Primary Endpoint
Baseline Shunt and Atrial Septal Aneurysm
(TEE)
STARFlex
N=400
Medical
N=451
P value
Trace
shunt
7.0%
(n=8/114)
8.0%
(n=10/126)
0.75
Moderate
shunt
5.3%
(n=7/132)
8.4%
(n=12/143)
0.31
Substantial
shunt
3.6%
(n=3/84)
5.3%
(n=3/57)
0.62
No atrial septal
aneurysm
6.4%
(n=15/236)
8.5%
(n=20/236)
0.38
Atrial septal
aneurysm
4.9%
(n=7/142)
6.5%
(n=9/139)
0.58

23. Aspirin versus Warfarin (physician
discretion)
Aspirin alone
(n=243)
Warfarin alone
(n=139)
P value
Composite 6.7%
(n=14)
8.1%
(n=9)
0.63
Stroke 3.9%
(n=8)
2.7%
(n=3)
0.67
TIA 2.9%
(n=6)
6.3%
(n=7)
0.09

24. CONCLUSIONS
 CLOSURE I is the first completed, prospective, randomized,
independently adjudicated PFO device closure study
 Superiority of PFO closure with STARFlex® plus medical therapy over
medical therapy alone was not demonstrated
 no significant benefit related to degree of initial shunt
 no significant benefit with atrial septal aneurysm
 insignificant trend (1.8%) favoring device driven byTIA
 2 year stroke rate essentially identical in both arms (3%)
 Major vascular (procedural) complications in 3% of device arm
 Significantly higher rate of atrial fibrillation in device arm (5.7%)
 60% periprocedural

25. CONCLUSIONS
 Alternative explanation unrelated to paradoxical embolism
present in 80% of patients with recurrent stroke orTIA
 cryptogenic stroke andTIA include multiple etiologies
 in many patients with cryptogenic stroke orTIA a PFO may be
coincidental
 diagnostic criteria for paradoxical embolism are imprecise
 potential efficacy of PFO device closure in better defined patient
subgroups requires further study
 Percutaneous closure with STARFlex® plus medical therapy does not
offer any significant benefit over medical therapy alone for the
prevention of recurrent stroke orTIA in patients < age 60 presenting
with cryptogenic stroke orTIA and a PFO

26. CLOSURE-I trial - Issues
 Procedural success 90%. “Effective closure” 86%.
 So ITT closure only 77%.
 BUT! “Effective closure” included trace shunting
or no shunting. Pre-procedure 114 of 400
Starflex patients had trace shunting.Therefore
real closure rate even lower (possibly as low as
50%).
 Thrombus on device 1%.
 Small absolute numbers of events.
 Slow recruitment. Short follow-up.
 Results incongruent with previous data

27. Incidental PFO?
 Alsheikh-Ali et al, Stroke 2009
 Analysis of 23 case-control studies examining
presence of PFO in pts with CS (total approx
2300 pts).
 In patients with CS
 1/3 of PFOs are likely to be incidental in all age
groups
 1/5 in younger age group
 1/10 if ASA + PFO

28. Starflex vs Amplatzer

29. Where to from here?
 RESPECT trial
 Maybe RCTs are not the answer?
 Good quality registry needed.