Topical Review

Patent Foramen Ovale Closure Versus Medical Therapy

for Cryptogenic Ischemic Stroke
A Topical Review
Jeffrey L. Saver, MD; Heinrich P. Mattle, MD; David Thaler, MD, PhD

M anaging young and middle-aged patients with cryp-

togenic ischemic stroke and a patent foramen ovale
(PFO) is a common clinical dilemma for neurologists, car-
diologists, and primary care physicians. Each year, ≈18 000
patients in the United States and 345 000 worldwide, aged 18
to 60, present with a PFO and an embolic stroke of otherwise
Downloaded from http://stroke.ahajournals.org/ by guest on May 14, 2018
Because venous thrombosis is common in adults, and venous
thrombi generate numerous migrating emboli, right-to-left
shunting can expose patients to substantial ischemic stroke risk
though only a fraction of blood flow actually passes through
the PFO. Silent thrombus formation in the deep and superfi-
cial venous system is a staple of human adult life, occurring

undetermined source (Figure 1).1–5 Leading treatment options
to prevent stroke recurrence include antiplatelet medications,
anticoagulant medications, and percutaneously placed PFO
closure devices. Lacking randomized trial data, choosing
among these options was long a quandary for physicians. But
major new trial results in 2017 to 2018,6–9 added to earlier tri-
als reported in 2012 to 2013,10–12 have transformed the evi-
dential foundation for PFO management. This topical review
appraises and synthesizes the accumulated trial findings and
considers how they may best inform physician and patient
treatment decision-making.
Pathophysiology
Right-to-left shunts enable thrombi that form in, and dislodge
from, the venous system to bypass filtration in the pulmonary
vasculature, paradoxically cross to the arterial tree, and travel
to and occlude recipient cerebral arteries. PFOs are the most
common causes of right-to-left shunts, present in about one
quarter of all adults. During fetal life, the interatrial septum
separating the right and left atria initially forms with a tunnel,
the foramen ovale, that enables maternally oxygenated blood
to bypass the fetal pulmonary circulation and nourish arterial
beds directly. This interatrial passage closes in most individu-
als during the first 3 months after birth. However, in about one
quarter of the population, the passage remains patent lifelong.
The mean diameter of persisting PFOs is 4.9 mm (range, 1–19
mm), more than sufficient to allow passage of emboli large
enough to occlude the middle cerebral artery stem (3 mm) and
major cerebral cortical branches (1 mm).
at a rate of perhaps 4% per year overall, and 10% during each
long-haul plane flight.13 When a deep venous thrombus (DVT)
develops, it can serve as the nidus for a barrage of thromboem-
boli directed at the right atria, 300 to 50 000 emboli per hour.14
Atrial septal aneurysms (ASAs) are an additional atrial
septal abnormality that may interact with, and potentiate risk
of, a PFO. The term ‘aneurysm’ is a misnomer, as the defect
does not consist of a weakened blood vessel wall, but rather a
hypermobile interatrial septum that protrudes alternately into
the right and left atria each cardiac cycle.
Clinical Epidemiology
In the general population, PFOs are detected on transesopha-
geal echocardiography (TEE) in ≈20% to 25% of individu-
als, ASAs are present in 2.2%, and 83% with ASAs also have
a PFO. In contrast, among young and middle-aged patients
with cryptogenic ischemic stroke, PFOs are present more
frequently, in ≈50% to 60%. On the basis of a meta-analy-
sis of case–control studies,4 with additional estimate correc-
tion for potential publication bias (Supplemental Figure I in
the online-only Data Supplement), young and middle-aged
patients with a cryptogenic ischemic stroke have a 2.3-fold
increased relative risk of having a PFO present, compared
with age-matched individuals with a phanerogenic (known
cause) ischemic stroke. From these observations, probability
theory dictates that, when young or middle-aged cryptogenic
stroke patients are found to have a PFO, the PFO is the mecha-
nism of the stroke in 73% of individuals and incidental in 27%
(Supplemental Figure I in the online-only Data Supplement).

Received December 22, 2017; final revision received March 18, 2018; accepted March 20, 2018.
From the Comprehensive Stroke Center and Department of Neurology, David Geffen School of Medicine at UCLA, Los Angeles, CA (J.L.S.); Department
of Neurology, University Hospital Bern, University of Bern, Switzerland (H.P.M.); and Department of Neurology and Comprehensive Stroke Center, Tufts
University School of Medicine, Boston, MA (D.T.).
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.
117.018153/-/DC1.
Correspondence to Jeffrey L. Saver, MD, Geffen School of Medicine at UCLA, 710 Westwood Plaza, Los Angeles, CA 90095. E-mail jsaver@mednet.
ucla.edu
(Stroke. 2018;49:00-00. DOI: 10.1161/STROKEAHA.117.018153.)
© 2018 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org DOI: 10.1161/STROKEAHA.117.018153

1
 2  Stroke  June 2018
Figure 1. Annual incidence of cryptogenic ischemic stroke in young and middle-aged adults with patent foramen ovale (PFO). Broad esti-
mates of annual incidence in (A) United States; (B) globally.1–5
Diagnosis
Downloaded from http://stroke.ahajournals.org/ by guest on May 14, 2018
PFOs can be detected by transthoracic echocardiography, TEE,
or transcranial Doppler ultrasound (TCD). Of these tests, trans-
thoracic echocardiography is the least sensitive, detecting only
50% to 60% of PFOs found on TEE or TCD. Compared with
gold standard autopsy diagnosis, TEE has a sensitivity of ≈90%
and specificity >95%. TCD detects 90% to 100% of PFOs found
on TEE and up to 10% missed with TEE; TCD may detect
small PFOs missed by TEE in part as stronger Valsalva maneu-
vers can be elicited, and confirmed by waveform monitoring,
during TCD than TEE. TEE and TCD are complementary. Both
provide PFO detection and quantification of shunt size. TEE, if
tolerated by patients, uniquely characterizes PFO anatomy and
presence of an ASA and assesses the presence of competing
proximal sources of embolism, including aortic arch atheroscle-
rosis, atrial appendage thrombi, and signs of atrial cardiopathy.
TCD uniquely quantifies the cerebral burden of paradoxical
embolism, on both bubble study and during 30 minutes of mon-
itoring for spontaneous microembolism, and assesses the pres-
ence of competing distal arterial sources of embolism.
All young and middle-aged patients presenting with ischemic
stroke should undergo investigations to identify large artery
atherosclerosis, cerebral small artery microatherosclerosis,
nonatherosclerotic arteriopathies, structural and dysrhythmic
cardiac sources of embolism, and arterial (and, if right-to-left
shunt present, venous) hypercoagulable states.5 The presence
of a high-grade alternative source of stroke greatly lessens the
likelihood that a medium-grade source like PFO is causative;
and the presence of another medium-grade source reduces the
likelihood to a lesser degree. Even when aortic, cervical, and
cerebral vessel imaging has not identified substantial large
artery atherosclerosis, the mere presence of risk factors for vas-
cular disease, such as hypertension, hyperlipidemia, diabetes
mellitus, and tobacco use, mildly diminishes the likelihood a
detected PFO is causally related to the stroke.15
In young and middle-aged patients with cryptogenic ischemic
stroke, in contrast to older patients, low burden paroxysmal atrial
fibrillation (AF) is infrequently found. In the CRYSTAL-AF
trial (Cryptogenic Stroke and Underlying AF), 3 years of con-
tinuous monitoring with an inserted loop recorder revealed
low burden paroxysmal AF among only 3% of cryptogenic
ischemic stroke patients age <54 years and 4% aged 54 to 61
years.16 In patients in this age range, it is reasonable to screen
for paroxysmal AF with continuous inpatient cardiac telemetry
during the index stroke admission, or, in patients not hospital-
ized, with 24 hours to 7 days of ambulatory monitoring. More
prolonged ambulatory cardiac monitoring is reasonable in the
small subset of patients with structural, electrophysiological, or
blood biomarker evidence of atrial cardiopathy.5
Absolute certainty that a detected PFO is the mechanism
of a particular ischemic stroke is rare—achieved when echo-
cardiography or other imaging catches thrombi in the act,
wriggling through the atrial septum. But the probability that
a discovered PFO is causally related to an otherwise crypto-
genic ischemic stroke is increased by 5 types of additional
clinical features:
• Presence of, or disposition to, venous thrombosis:
Likelihood of PFO complicity in the ischemic stroke
is strongly increased when concurrent deep or superfi-
cial venous thromboembolism is present on noninvasive
lower extremity ultrasound, pelvic computed tomogra-
phy or magnetic resonance venography, or pulmonary
computed tomography arteriography performed within
the first 48 to 72 hours after stroke onset (before time for
venous thrombosis to develop secondarily).17 Negative
noninvasive testing by no means rules out a venous
source—the great preponderance of venous clots are
superficial (eg, calf) or small, deep thrombi discover-
able only on invasive contrast venography18; but invasive
venography causes patient discomfort so is not usually
obtained. Circumstances promoting venous thrombosis
are also suggestive that a PFO is pathogenic, including
recent immobility (such as extended plane or car travel,
surgery, or illness)19; laboratory findings of a venous
hypercoagulable state; imaging showing anatomic
causes of venous congestion, such as May-Thurner syn-
drome20; or history of prior venous thromboembolism.
Pure venous hypercoagulable states, such as protein
C and S deficiencies, factor V Leiden mutation, and pro-
thrombin gene mutation, likely increase the likelihood
the PFO is causally related, as they foster venous throm-
boemboli that may pass through the PFO. In contrast,
mixed arterial–venous hypercoagulable states, such as
 Saver et al   PFO Closure for Cryptogenic Stroke   3
antiphospholipid antibodies or hyperhomocysteinemia,21
have bidirectional effects. They promote both venous
thromboemboli that may pass through the PFO but also
arterial thrombi that may form in situ in the cerebral arte-
rial tree, unrelated to the PFO.
• Increased right-to-left shunt flow, permanently or tran-
siently: The greater the volume of right-to-left flow across
the PFO on imaging, the greater the chance that a venous
thromboembolus will swerve across the interatrial shunt
rather than flow directly from right atrium to right ven-
tricle. Resting greater right-to-left flow may arise from
anatomically large PFO size and from chronic pulmo-
nary hypertension hemodynamically fostering paradoxi-
cal flow. Transient increased right-to-left flow may arise
from Valsalva maneuver, when forceful expiration against
a closed glottis momentarily increases thoracic pressure.
Stroke onset coincident with a Valsalva maneuver supports
PFO culpability. Careful history-taking should determine
whether the ictus occurred during or immediately after
Downloaded from http://stroke.ahajournals.org/ by guest on May 14, 2018
heavy lifting, straining at stool, sexual intercourse, cough-
ing, sneezing, or vomiting.19 Moreover, Muller maneuver,
forceful inspiration against a closed glottis, momentarily
decreasing thoracic pressure, may also be followed by an
increase in right-to-left flow, due to increased intratho-
racic pressure swings. In a patient with sleep apnea, stroke
onset during sleep is suggestive, as a common setting for
Muller maneuver is vigorous inspiration attempts during
apneic periods.19,22 Sleep apnea is also associated with
pulmonary hypertension, which may increase right-to-left
shunt flow across a PFO.
• ASA: ASAs potentiate stroke risk in patients with PFO,
possibly by (1) hemodynamically fostering access to
the PFO of venous thromboemboli arriving in the right
atrium23,24; or (2) causing local thrombosis in the PFO
tunnel or the right or left atria by fostering turbulent flow,
with platelet activation, or flow stasis, in perianeurysm
pockets or via incomplete left atrial emptying, with clot-
ting protein cascade activation.25,26
• Recipient brain artery or territory typical of embolism:
Typical recipient sites in the cerebral circulation for
proximal emboli are large main arterial trunks (caus-
ing large combined superficial-deep infarcts or isolated,
large deep infarcts) and small distal arterial branches
(causing isolated superficial infarcts). Emboli much less
commonly veer into small deep penetrator arteries (caus-
ing isolated, small deep infarcts).27 Large or superficial
infarcts in multiple cerebrovascular territories are par-
ticularly suggestive of a proximal source of embolism,
though this pattern is present in only one sixth of PFO
patients.
• Absence of risk factors for atherosclerosis: As mild ath-
erosclerotic disease is a common competing potential
cause of ischemic stroke, PFO causality is supported by
the absence of demographic and medical risk factors for
plaque development, including younger patient age and
absence of hypertension, hyperlipidemia, diabetes melli-
tus, and tobacco use. The Risk of Paradoxical Embolism
Score uses absence of atherosclerotic risk factors and
infarct topography, though not other features, to quantify
the likelihood that a PFO is causally related to stroke in
individual patients.15
Treatment Options
Treatment options for secondary prevention of recurrent stroke
in cryptogenic ischemic stroke patients with PFO include,
alone or in combination, antiplatelet therapy, anticoagulant
therapy, surgical PFO closure, and percutaneous PFO closure.
Until recently, in the absence of any randomized trials show-
ing that any of these regimens were better than no treatment at
all, the conventional treatment approach was medical therapy,
most often with antiplatelet agents alone, and less often with
vitamin K–dependent oral anticoagulants alone. Among the
medical therapy options, US national practice guidelines weakly
endorsed antiplatelet therapy as preferred.28,29 Although physi-
ological reasoning suggests that anticoagulation might be supe-
rior to antiplatelet therapy, as anticoagulants better avert stasis
thrombi arising in veins, anticoagulation is also associated with
increased bleeding, and comparative studies are only weakly
suggestive of an efficacy advantage. The 2 randomized trials
directly comparing anticoagulant against antiplatelet therapy
in subgroups of patients with PFOs and cryptogenic ischemic
stroke found nonsignificant efficacy differences: PICSS trial
(PFO in Cryptogenic Stroke Study; hazard ratio [HR], 0.52;
95% confidence interval [CI], 0.16–1.67; P=0.28)30 and CLOSE
trial (Patent Foramen Ovale Closure or Anticoagulants versus
Antiplatelet Therapy to Prevent Stroke Recurrence; HR, 0.44;
95% CI, 0.11–1.48; P=0.18).8 In addition, minor bleeding com-
plications were significantly,30 and major bleeding complications
nominally,8 more frequent among anticoagulated patients. These
randomized controlled trial findings align with a larger, indi-
vidual participant data meta-analysis of 12 prospective observa-
tional studies and randomized trials: anticoagulation compared
with antiplatelet therapy (HR, 0.75; 95% CI, 0.44–1.27).31
Compared with older anticoagulants, newer, non–vitamin
K–dependent oral anticoagulants are promising options, given
their more predictable blood levels, reduced bleeding rates, and
comparable efficacy in prevention of DVT,32 but have not yet
been studied in any large cohort of PFO patients.
Surgical PFO closure, requiring thoracotomy and cardiopul-
monary bypass with their attendant risks, is rarely pursued as
a standalone therapy. It may be a useful option when a patient
is undergoing cardiac surgery for another indication. In obser-
vational series, open surgical closure had minimal periopera-
tive mortality, but morbidity included AF, pericardial effusion,
postoperative bleeding, infection, and postpericardiotomy syn-
drome. The annual event rate for recurrent stroke or transient
ischemic attack has ranged from 0% to 9%. Recurrence of
cerebral ischemia may be related to incomplete PFO closure by
open surgical treatment, occurring in up to 73% of patients.33
Percutaneously placed PFO closure devices are a less inva-
sive approach to anatomic management and have advanced
substantially in the quarter century since their first use in
humans.34 Device designs differ in important ways that affect
ease of delivery, efficacy in achieving complete closure, and
adverse effects. Devices tested in randomized trials may be
grouped into 2 broad classes, based on frame shape: the first
generation of devices had umbrella-clamshell contours, with
some frame projection away from the septal surface, whereas
later developed devices had double-disk contours, a more
compact profile potentially less prone to complications.35
 4  Stroke  June 2018

Table.  Features of Patent Foramen Ovale Closure Device Trials

Inclusion Criteria Follow-Up Ratio of
Year of Enrollment/ Patient Years (mean)/ Follow-Up
Trial Publication Follow-Up Geography Event Type Timing Age Number Patient-Years Dev/Med*
CLOSURE 2012 E: 2003–2008 United States, Cryptogenic IS ≤ 6 mo 18–60 909 1.7/1555 1.06
Canada or TIA
F: 2003–2010
PC 2013 E: 2000–2009 Europe, Canada, Cryptogenic IS or No <60 414 4.1/1681 1.04
Brazil, Australia periph embolism restriction
F: 2000–2012
RESPECT 2013/2017 E: 2003–2011 United States, Cryptogenic IS ≤ 9 mo 18–60 980 5.8/5688 1.14
Canada (Tissue-Def)
F: 2003–2016
CLOSE 2017 E: 2007–2014 France, Germany Cryptogenic IS ≤ 6 mo 16–60 473 (653)† 5.3/2507 1.04
(Tissue-Def)
F: 2007–2016
REDUCE 2017 E: 2008–2015 Europe, Canada, Cryptogenic IS ≤ 6 mo 18–59 664 3.4/2232 1.10
United States (Tissue-Def)
F: 2008–2016
DEFENSE- 2018 E: 2011–2017 South Korea Cryptogenic IS ≤ 6 mo 18–80 120 1.6‡/ ≈187 1.03
Downloaded from http://stroke.ahajournals.org/ by guest on May 14, 2018

PFO (Tissue-Def)
F: 2011–2017
CLOSE indicates Patent Foramen Ovale Closure or Anticoagulants Versus Antiplatelet Therapy to Prevent Stroke Recurrence; CLOSURE, Evaluation of the STARFlex
Septal Closure System in Patients With a Stroke and/or Transient Ischemic Attack due to Presumed Paradoxical Embolism Through a Patent Foramen Ovale; DEFENSE-
PFO, Device Closure Versus Medical Therapy for Cryptogenic Stroke Patients With High-Risk Patent Foramen Ovale; IS, ischemic stroke; PC Trial, Clinical Trial Comparing
Percutaneous Closure of Patent Foramen Ovale Using the Amplatzer PFO Occluder With Medical Treatment in Patients With Cryptogenic Embolism; REDUCE, Gore
REDUCE Clinical Study; RESPECT, Randomized Evaluation of Recurrent Stroke Comparing PFO Closure to Established Current Standard of Care Treatment; and TIA,
transient ischemic attack.
*Mean duration of follow-up among device patients/mean duration of follow-up among medical patients. Longer follow-up among device patients occurred because
of (1) more end point events in medical patients, ending study participation, and (2) more dropouts in medical patients, in part to pursue device placement outside of
the trials.
†Full results reported for 473 patients randomized to closure and medical antiplatelet therapy groups, pending for 180 randomized to the medical anticoagulation
therapy group.
‡For DEFENSE-PFO, only follow-up years estimated from the Kaplan–Meier curve of the fully-reported time period—the first 2 years after enrollment.

In a randomized trial directly comparing 3 closure devices
against one another, including 1 umbrella-clamshell device
(CardioSEAL STARflex occluder [C]) and 2 double-disk
devices (Amplatzer PFO occluder [A] and HELEX occluder
[H]), better outcomes were observed with the double-disk
devices, including fewer recurrent ischemic events (A, 1.4%;
H, 4.1%; C, 5.9%), less thrombosis on device (A, 0%; H,
0.5%; C, 5.0%), and less AF (A, 3.6%; H, 2.3%; C, 12.3%).36
Six randomized trials have now compared PFO closure
devices with medical management alone and provide the first
firm evidence to guide treatment selection. Collectively the tri-
als enrolled 3560 fully-reported patients followed for 13 850
patient-years (Table). They differed in several important aspects,
including devices tested in the device arm, medications tested in
the medication arm, permitted qualifying events, and extent of
workup mandated to exclude competing causes of stroke (Table;
Figure 2). These differences explain variations among trial
results and provide insights into optimal treatment deployment.
Overall, with regard to clinical efficacy, the trials demon-
strate a beneficial reduction in recurrent ischemic stroke with
PFO device closure plus long-term medical antithrombotic (pri-
marily antiplatelet) therapy compared with long-term medical
antithrombotic therapy (antiplatelet or anticoagulant) alone. In
study-level meta-analysis, device closure, compared with medi-
cal therapy alone, reduced the rate of recurrent ischemic stroke
(HR, 0.30; 95% CI, 0.13–0.68; P=0.004; Figure 3; Supplemental
Table I in the online-only Data Supplement). There was
evidence of heterogeneity of treatment benefit across different
device classes (P [subgroup difference]=0.02), with a substan-
tial, and highly statistically significant, reduction in recurrent
stroke with double-disk devices (HR, 0.20; 95% CI, 0.08–0.54;
P=0.001) contrasting with a less pronounced, and formally non-
significant, effect with an umbrella-clamshell device (HR, 0.90;
95% CI, 0.41–1.98; P=0.79). In the 5 double-disk trials, the event
rate for recurrent ischemic stroke over 5 years after randomiza-
tion was 6.0% on medical therapy versus 1.8% with device clo-
sure plus long-term antiplatelet therapy. Tissue-defined transient
ischemic attacks were not reduced, likely in part because of their
less reliable diagnostic recognition.
The fully-reported medical therapy groups in these trials
consisted of a mix of patients treated with antiplatelet agents
and anticoagulant agents in 4 of the 6 trials and patients
treated with antiplatelets alone in 2 trials. In contrast, the
protocol-permitted early concomitant medical antithrom-
botic therapy among device group patients was confined to
antiplatelet agents alone in 5 of the 6 trials, and early anti-
coagulation was used in <1% of device-treated patients.
Among the 3 trials with separately reported medical therapy
treatment subgroups, there was evidence of heterogeneity of
treatment effect depending on the type of medical antithrom-
botic therapy (P [subgroup difference]=0.02). Compared
with medical antiplatelet therapy alone, device closure plus
long-term medical antiplatelet therapy was superior in avert-
ing recurrent stroke (HR, 0.19; 95% CI, 0.06–0.56; P=0.003;
 Saver et al   PFO Closure for Cryptogenic Stroke   5
Downloaded from http://stroke.ahajournals.org/ by guest on May 14, 2018

Figure 2. Trial populations and event rates. A, Heat map display of stringency of entry criteria in selecting patients likely to have patho-
genic rather than incidental patent foramen ovales (PFOs). Cell coloring: green—most stringent; gray—intermediate; and red—least strin-
gent. B, Heat map display of characteristics of enrolled patients and treatments. Cell coloring conveys likely effect on power of study to
detect PFO-specific treatment effect: green—heightens; gray—intermediate; and red—lessens. ASA indicates atrial septal aneurysm.

Supplemental Figure II in the online-only Data Supplement).
In contrast, in the underpowered comparison of medical
therapy with anticoagulants alone versus device closure plus
long-term antiplatelet therapy, no benefit of device closure
was noted (HR, 1.32; 95% CI, 0.43–4.03).
The trials also provided important information on technical
efficacy in achieving PFO closure and on periprocedural and
long-term safety of the closure devices. After device placement,
there was effective PFO closure (no or only trace, residual
shunting) in 93% to 96% of patients in double-disk trials and in
87% in the umbrella-clamshell trial. Immediate, serious, pro-
cedure-related complications were infrequent among the 1780
of 1889 patients allocated to the device groups who actually
underwent a placement procedures and included access site or
retroperitoneal hemorrhage in 1.01%, pericardial tamponade
in 0.17%, and cardiac perforation in 0.06%. The most com-
mon complication of device placement was AF, but the great
preponderance of AF events were transient episodes occurring
in the first 4 to 6 weeks after device placement, during the first
settling of device elements into atrial tissue. Occurring in 3.2%
of patients undergoing a device procedure, these generally
self-limited, periprocedural AF events have less likelihood of
serving as a new stroke source, compared with AF events that
occur later. Among the 4 double-disk device trials with avail-
able data, device placement showed a nonsignificant increase
in annual risk of AF in the postperiprocedural period, 0.39%

Figure 3. Forest plot of trials comparing PFO closure devices plus long-term medical antithrombotic (primarily antiplatelet) therapy
against medical antithrombotic (antiplatelets or anticoagulants) therapy alone, for deterrence of recurrent ischemic stroke. Wherever avail-
able, data reflect ischemic stroke defined using the modern tissue-based definition. Forest plot analyzes hazard ratio, rather than odds or
risk ratio, because of the imbalance in duration of study follow-up between treatment groups in all trials. CI indicates confidence interval;
IV, inverse variance; and MT, medical therapy.
 6  Stroke  June 2018
versus 0.26% per year (HR, 1.50; 95% CI, 0.77–2.93; P=0.24;
Supplementary Figure III in the online-only Data Supplement).
DVT and pulmonary embolism tended to occur more fre-
quently in follow-up in the device than medical group in 1 trial,
but not the other 4. The risk was particularly increased among
the small proportion of patients, 3.6%, who had a history of
previous, clinically manifest, unprovoked DVT before trial
entry.6 For these uncommon patients, lifelong anticoagulation
to avert recurrent DVT and pulmonary embolism may be ben-
eficial,37 with or without concomitant PFO closure to further
prevent paradoxical embolization.
The trials provide evidence that certain patient subgroups
may have greater or lesser reduction in recurrent ischemic
stroke from device placement). In trials of double-disk
devices, 3 patient features were associated or likely associated
with magnified reduction in recurrent ischemic stroke with
device therapy: (1) ASA (HR, 0.13 versus 0.43; P [subgroup
difference]=0.005; Supplemental Figure IV in the online-only
Data Supplement); (2) larger shunt sizes, across 4 size lev-
Downloaded from http://stroke.ahajournals.org/ by guest on May 14, 2018
els (P [subgroup difference]=0.07; Supplemental Figure V
in the online-only Data Supplement); and (3) index ischemic
stroke topography not confined to a single, deep, penetrating
artery (HR, 0.34 versus 2.25; P [subgroup difference]=0.04;
Supplemental Figure VI in the online-only Data Supplement).
The magnitude of the benefit conferred by PFO closure
device placement is moderate overall, higher in select patients,
and in a range likely to be deemed clinically worthwhile by
an important proportion of patients and families. Among all
patients enrolled in the randomized trials and randomized
among device closure plus antiplatelet therapy compared with
antiplatelet therapy alone, the number needed to treat to pre-
vent 1 recurrent ischemic stroke over 5 years was 24. Among
patients with ASA, comparing device therapy plus antiplate-
lets to medical therapy, the number needed to treat to avert 1
stroke over 5 years was 13, and among patients with moder-
ate-substantial shunts, the number needed to treat was 18. The
5-year time frame for this analysis is likely conservative with
regard to benefit estimate. Young and middle-aged ischemic
stroke patients with PFO and first cryptogenic ischemic stroke
have decades of future at-risk years; modest annual reductions
in stroke rates, if sustained, have many years to additively
accrue value. These estimates may also be conservative as
some patients with perceived high recurrence risk on medical
therapy were treated with device closure outside randomized
trials, rather than enrolled and randomized.
An important reason that the annual recurrent ischemic
stroke reduction is modest is that event rates in patients treated
with medical therapy alone are low to start with, 1.2% per
year across the 6 RCTs, placing a ceiling on how much further
benefit device placement can confer. This low recurrence rate
on medical therapy is less than with other causes of stroke and
accords with the status of PFOs as medium-, not high-, grade
risk sources for embolism.38 Consequently, patients have a
welcome choice between 2 good options—medical therapy
alone, with a modest recurrence rate, or closure device com-
bined with antiplatelet therapy, with a dramatic relative, and
moderate absolute, further reduction in risk.
Device placement dramatically reduces recurrent ischemic
strokes of otherwise undetermined origin, but not recurrent
ischemic strokes of determined, non-PFO origin, which con-
tinue to accumulate at low frequencies in both treatment groups.6
Likely, some of the patients with a recurrent stroke of determined
cause harbored other sources of stroke that were discounted or
inapparent at the time of the initial investigation. In addition,
as patients age, new risk factors for recurrent stroke emerge,
including atherosclerosis and AF. Closing the PFO will prevent
future PFO-related strokes, but not strokes due to other causes.
The mechanisms of recurrent ischemic strokes of determined
cause in patients treated with double-disk devices were diverse;
intrinsic small vessel disease was most common, but also large
artery atherosclerosis, dilated cardiomyopathy, and other mecha-
nisms.6 Antithrombotic therapy can reduce recurrent events from
many of these. Accordingly, it is reasonable to treat patients who
undergo device closure with lifelong gentle antiplatelet (or pos-
sibly anticoagulant) therapy as well, rather than discontinuing all
antithrombotic agents after device endothelialization, to provide
ongoing protection against recognized and unrecognized pos-
sible competing causes of the initial stroke.
Management Recommendations
Selection of a treatment strategy should be patient-centric,
informed by the clinical trial evidence, and tailored to the pre-
sentation, findings, and preferences of each individual. Young
and middle-aged cryptogenic stroke patients with PFO should
know that both of the well-studied treatment options are good
choices: recurrence rates are low with both antiplatelet ther-
apy alone and device therapy added to antiplatelet therapy;
and that recurrence rates are lower with device placement plus
antiplatelet therapy. In addition, long-term anticoagulation
alone and device placement plus long-term anticoagulation
(after initial antiplatelet therapy until device endothelializa-
tion is completed) are promising, not yet sufficiently stud-
ied, options. In patients with features which may potentiate
the benefit of device placement, including ASA, substantial
shunt size, and embolic topography of the index stroke, device
placement plus antiplatelet therapy may be especially attrac-
tive. In patients with a history of overt deep venous thrombo-
sis, and possibly patients with venous hypercoagulable states,
long-term anticoagulation with or without device placement
may be preferred, to avert DVT and PE, as well as recurrent
ischemic stroke. In patients in whom anticoagulation is con-
traindicated due to past bleeding complications or bleeding-
prone conditions, device placement is preferred. Similarly, for
patients in whom frequent or prolonged interruption of oral
anticoagulation may be anticipated, because of multiple surgi-
cal or invasive procedures, pregnancy, or medication nonad-
herence, device placement is an attractive option.39
At the systems level, the management of patients with PFO
and cryptogenic stroke requires close coordination between neu-
rologists and cardiologists expert in the evaluation and treatment
of neurocardiovascular disease.40 An emerging model is a jointly
run specialized outpatient clinic at which cryptogenic stroke
patients, with or without PFO, can be efficiently evaluated by
both specialties. Neurologists’ insights on distinctive features
of the stroke presentation and topography, and cardiologists’
insights on distinctive features of the PFO anatomy and other
cardiac structural and rhythm findings, can be synthesized into
an integrated perspective on likely stroke mechanism and likely
 Saver et al   PFO Closure for Cryptogenic Stroke   7
benefits and risks of the different therapeutic options for consid-
eration by patients and families. Such joint specialty collabo-
ration is facilitated by the US Food and Drug Administration
labeling for PFO closure devices, which specifically states
that the population for whom closure devices are indicated are
patients who have been “…deemed by a neurologist and a cardi-
ologist to have had a cryptogenic stroke following an evaluation
to exclude known causes of ischemic stroke.”
Future Directions
Among the important areas that would benefit from further
study are:
• Predictive models to identify patients who benefit a lot, a
little, or not at all: A pooled, individual participant-level
data meta-analysis of the completed trials is desirable
to develop predictive models identifying combinations
of patient characteristics associated with magnified,
reduced, and absent benefit of device closure over medi-
cal therapy alone.
Downloaded from http://stroke.ahajournals.org/ by guest on May 14, 2018
• Role of anticoagulation, including newer oral antico-
agulants: Additional randomized trials are needed to
determine the role of (1) anticoagulation alone, and (2)
anticoagulation combined with PFO closure; both are
promising, but currently not fully proven, treatment
options.
• Role of new PFO closure devices: Three newer classes
of PFO closure devices have entered development: (1)
tunnel insertion devices; (2) bioabsorbable devices; and
(3) suture devices.41 All reduce device surface expo-
sure in the left atrium that may provoke thrombosis.
Tunnel insertion devices are especially well suited for
closure of long PFO tunnels. The new device classes
may improve closure rates above those attained by dou-
ble-disk devices which yield excellent, ≈95%, rates of
effective closure (no or only trace shunting), but only
moderate, ≈75%, rates of complete closure (no shunt-
ing at all).
• Patients over age 60: Older patients are more likely to
have competing low-medium risk potential causes of the
index stroke, reducing the likelihood that a discovered
PFO is causally related to the cryptogenic stroke.4 But
older individuals are also more prone to venous throm-
boembolism,42 and those with a PFO are prone to more
right-to-left shunting, in part because of a higher prev-
alence of sleep apnea, increasing the likelihood that a
discovered PFO is causally related to the cryptogenic
stroke. Also, older individuals may be more prone to
the complication of closure device–induced AF. Given
these age-related differences, RCTs in individuals over
age 60 are needed. Pending those studies, PFO device
closure might be deployed highly selectively, especially
in younger old patients, age 61 to 65, and after more
extensive workup for paroxysmal AF, with at least 2 to 4
weeks of ambulatory monitoring.
• Refractory migraine with aura: The validation of PFO
closure as a treatment to prevent recurrent ischemic
stroke provides support for additional studies of PFO
closure as a treatment for refractory migraine with aura
preceding most attacks, which may be due to minor par-
adoxical emboli provoking episodes of cortical spread-
ing depression.
Conclusions
The development of PFO closure device therapy has followed
the 5-stage life cycle common to technological advances: tech-
nology trigger, peak of inflated expectations, trough of disillu-
sionment, slope of enlightenment, and plateau of productivity.43
Initially (2000–2011), based on nonrandomized case series,
enthusiasm for device closure was rampant, slowing enroll-
ment in RCTs as some physicians were unwilling to refer
patients to trials, so certain were they of dramatic benefit. Next
(2011–2016), when initial randomized trials showed only sug-
gestive, not definitive, evidence of modest treatment advantage,
disappointment was sharp, and several authorities prematurely
pronounced that PFO closure devices were not beneficial.28,44,45
Finally, in the past year (2017–2018), additional results from
randomized trials have ushered in a degree of enlightenment,
confirming that PFO closure devices are of genuine benefit, but
that the magnitude of benefit is moderate and, to date, dem-
onstrated only for carefully selected young and middle-aged
patients with cryptogenic ischemic stroke. PFO closure devices
combined with antiplatelet therapy are a useful addition to the
therapeutic armanterium. Additional reduction in recurrent
stroke rates can be achieved when neurologists and cardiolo-
gists, working collaboratively, perform detailed etiologic evalu-
ations, exclude other causes of stroke, identify high-risk patient
features, and provide nuanced treatment recommendations to
young and middle-aged ischemic stroke patients with PFO and
otherwise cryptogenic ischemic stroke.
Disclosures
Dr Saver and Dr Thaler report contracted payments for service on
clinical trial steering committee from St. Jude Medical/Abbott. The
other author reports no conflicts.
References
1. Benjamin EJ, Blaha MJ, Chiuve SE, Cushman M, Das SR, Deo R, et al;
American Heart Association Statistics Committee and Stroke Statistics
Subcommittee. Heart disease and stroke statistics-2017 update: a report
from the American Heart Association. Circulation. 2017;135:e146–
e603. doi: 10.1161/CIR.0000000000000485.
2. Krishnamurthi RV, Feigin VL, Forouzanfar MH, Mensah GA, Connor
M, Bennett DA, et al. Global and regional burden of first-ever ischaemic
and haemorrhagic stroke during 1990–2010: findings from the global
burden of disease study 2010. Lancet Glob Health. 2013;1:e259–e281.
3. Hart RG, Diener HC, Coutts SB, Easton JD, Granger CB, O’Donnell MJ,
et al. Embolic strokes of undetermined source: the case for a new clinical
construct. Lancet Neurol. 2014;13:429–438.
4. Alsheikh-Ali AA, Thaler DE, Kent DM. Patent foramen ovale in crypto-
genic stroke: incidental or pathogenic? Stroke. 2009;40:2349–2355.
5. Saver JL. Cryptogenic stroke. N Engl J Med. 2016;374:2065–2074.
6. Saver JL, Carroll JD, Thaler DE, Smalling RW, MacDonald LA, Marks
DS, et al; RESPECT Investigators. Long-term outcomes of patent
foramen ovale closure or medical therapy after stroke. N Engl J Med.
2017;377:1022–1032. doi: 10.1056/NEJMoa1610057.
7. Søndergaard L, Kasner SE, Rhodes JF, Andersen G, Iversen HK, Nielsen-
Kudsk JE, et al; Gore REDUCE Clinical Study Investigators. Patent fora-
men ovale closure or antiplatelet therapy for cryptogenic stroke. N Engl J
Med. 2017;377:1033–1042. doi: 10.1056/NEJMoa1707404.
8. Mas JL, Derumeaux G, Guillon B, Massardier E, Hosseini H, Mechtouff
L, et al; CLOSE Investigators. Patent foramen ovale closure or antico-
agulation vs. antiplatelets after stroke. N Engl J Med. 2017;377:1011–
1021. doi: 10.1056/NEJMoa1705915.
9. Lee PH, Song JK, Kim JS, Heo R, Lee S, Kim DH, et al. Cryptogenic
stroke and high-risk patent foramen ovale: the defense-pfo trial [pub-
lished online ahead of print February 28, 2018]. J Am Coll Cardiol. doi:
10.1016/j.jacc.2018.02.046.
 8  Stroke  June 2018
10. Furlan AJ, Reisman M, Massaro J, Mauri L, Adams H, Albers GW, et al;
CLOSURE I Investigators. Closure or medical therapy for cryptogenic
stroke with patent foramen ovale. N Engl J Med. 2012;366:991–999. doi:
10.1056/NEJMoa1009639.
11. Carroll JD, Saver JL, Thaler DE, Smalling RW, Berry S, MacDonald LA,
et al; RESPECT Investigators. Closure of patent foramen ovale versus
medical therapy after cryptogenic stroke. N Engl J Med. 2013;368:1092–
1100. doi: 10.1056/NEJMoa1301440.
12. Meier B, Kalesan B, Mattle HP, Khattab AA, Hildick-Smith D, Dudek
D, et al; PC Trial Investigators. Percutaneous closure of patent foramen
ovale in cryptogenic embolism. N Engl J Med. 2013;368:1083–1091.
doi: 10.1056/NEJMoa1211716.
13. Scurr JH, Machin SJ, Bailey-King S, Mackie IJ, McDonald S, Smith
PD. Frequency and prevention of symptomless deep-vein thrombosis in
long-haul flights: a randomised trial. Lancet. 2001;357:1485–1489. doi:
10.1016/S0140-6736(00)04645-6.
14. Nicholls SC, O’Brian JK, Sutton MG. Venous thromboembolism: detec-
tion by duplex scanning. J Vasc Surg. 1996;23:511–516.
15. Kent DM, Ruthazer R, Weimar C, Mas JL, Serena J, Homma S,
et al. An index to identify stroke-related vs incidental patent fora-
men ovale in cryptogenic stroke. Neurology. 2013;81:619–625. doi:
10.1212/WNL.0b013e3182a08d59.
16. Thijs VN, Brachmann J, Morillo CA, Passman RS, Sanna T, Bernstein
Downloaded from http://stroke.ahajournals.org/ by guest on May 14, 2018
RA, et al. Predictors for atrial fibrillation detection after cryptogenic
stroke: results from CRYSTAL AF. Neurology. 2016;86:261–269. doi:
10.1212/WNL.0000000000002282.
17. Lapergue B, Decroix JP, Evrard S, Wang A, Bendetowicz D, Offroy MA,
et al. Diagnostic yield of venous thrombosis and pulmonary embolism by
combined CT venography and pulmonary angiography in patients with
cryptogenic stroke and patent foramen ovale. Eur Neurol. 2015;74:69–
72. doi: 10.1159/000437261.
18. Lethen H, Flachskampf FA, Schneider R, Sliwka U, Köhn G, Noth J, et
al. Frequency of deep vein thrombosis in patients with patent foramen
ovale and ischemic stroke or transient ischemic attack. Am J Cardiol.
1997;80:1066–1069.
19. Ozdemir AO, Tamayo A, Munoz C, Dias B, Spence JD. Cryptogenic
stroke and patent foramen ovale: clinical clues to paradoxical embolism.
J Neurol Sci. 2008;275:121–127. doi: 10.1016/j.jns.2008.08.018.
20. Kiernan TJ, Yan BP, Cubeddu RJ, Rengifo-Moreno P, Gupta V, Inglessis
I, et al. May-Thurner syndrome in patients with cryptogenic stroke
and patent foramen ovale: an important clinical association. Stroke.
2009;40:1502–1504. doi: 10.1161/STROKEAHA.108.527366.
21. Caplan LR, Saver J. Treatment. In: Caplan LR, ed. Caplan’s Stroke: A
Clinical Approach. Cambridge University Press; 2016:145–211.
22. Mojadidi MK, Bokhoor PI, Gevorgyan R, Noureddin N, MacLellan WC,
Wen E, et al. Sleep apnea in patients with and without a right-to-left
shunt. J Clin Sleep Med. 2015;11:1299–1304. doi: 10.5664/jcsm.5190.
23. Bushi D, Grad Y, Einav S, Yodfat O, Nishri B, Tanne D. Hemodynamic
evaluation of embolic trajectory in an arterial bifurcation: an
in-vitro experimental model. Stroke. 2005;36:2696–2700. doi:
10.1161/01.STR.0000190097.08862.9a.
24. Kobayashi N, Karino T. Flow behavior and distribution of embolus-model
particles at the terminal bifurcation of the human internal carotid artery.
World Neurosurg. 2016;90:469–477. doi: 10.1016/j.wneu.2016.03.012.
25. Silver MD, Dorsey JS. Aneurysms of the septum primum in adults. Arch
Pathol Lab Med. 1978;102:62–65.
26. Rigatelli G, Aggio S, Cardaioli P, Braggion G, Giordan M, Dell’avvocata
F, et al. Left atrial dysfunction in patients with patent foramen ovale
and atrial septal aneurysm: an alternative concurrent mechanism for
arterial embolism? JACC Cardiovasc Interv. 2009;2:655–662. doi:
10.1016/j.jcin.2009.05.010.
27. Thaler DE, Ruthazer R, Di Angelantonio E, Di Tullio MR, Donovan JS,
Elkind MS, et al. Neuroimaging findings in cryptogenic stroke patients
with and without patent foramen ovale. Stroke. 2013;44:675–680. doi:
10.1161/STROKEAHA.112.677039.
28. Kernan WN, Ovbiagele B, Black HR, Bravata DM, Chimowitz MI,
Ezekowitz MD, et al; American Heart Association Stroke Council,
Council on Cardiovascular and Stroke Nursing, Council on Clinical
Cardiology, and Council on Peripheral Vascular Disease. Guidelines for
the prevention of stroke in patients with stroke and transient ischemic
attack: a guideline for healthcare professionals from the American Heart
Association/American Stroke Association. Stroke. 2014;45:2160–2236.
doi: 10.1161/STR.0000000000000024.
29. Messé SR, Gronseth G, Kent DM, Kizer JR, Homma S, Rosterman
L, et al. Practice advisory: recurrent stroke with patent foramen ovale
(update of practice parameter): Report of the Guideline Development,
Dissemination, and Implementation Subcommittee of the American
Academy of Neurology. Neurology. 2016;87:815–821. doi:
10.1212/WNL.0000000000002961.
30. Homma S, Sacco RL, Di Tullio MR, Sciacca RR, Mohr JP; PFO in
Cryptogenic Stroke Study (PICSS) Investigators. Effect of medical treat-
ment in stroke patients with patent foramen ovale: patent foramen ovale
in Cryptogenic Stroke Study. Circulation. 2002;105:2625–2631. doi:
10.1161/01.CIR.0000017498.88393.44.
31. Kent DM, Dahabreh IJ, Ruthazer R, Furlan AJ, Weimar C, Serena
J, et al. Anticoagulant vs. antiplatelet therapy in patients with
cryptogenic stroke and patent foramen ovale: an individual par-
ticipant data meta-analysis. Eur Heart J. 2015;36:2381–2389. doi:
10.1093/eurheartj/ehv252.
32. Almutairi AR, Zhou L, Gellad WF, Lee JK, Slack MK, Martin JR, et
al. Effectiveness and safety of non-vitamin K antagonist oral anticoagu-
lants for atrial fibrillation and venous thromboembolism: a systematic
review and meta-analyses. Clin Ther. 2017;39:1456.e36–1478.e36. doi:
10.1016/j.clinthera.2017.05.358.
33. Schneider B, Bauer R. Is surgical closure of patent foramen ovale the
gold standard for treating interatrial shunts? An echocardiographic
follow-up study. J Am Soc Echocardiogr. 2005;18:1385–1391. doi:
10.1016/j.echo.2005.03.037.
34. Bridges ND, Hellenbrand W, Latson L, Filiano J, Newburger JW, Lock
JE. Transcatheter closure of patent foramen ovale after presumed para-
doxical embolism. Circulation. 1992;86:1902–1908.
35. Nassif M, Abdelghani M, Bouma BJ, Straver B, Blom NA, Koch KT, et
al. Historical developments of atrial septal defect closure devices: what
we learn from the past. Expert Rev Med Devices. 2016;13:555–568. doi:
10.1080/17434440.2016.1182860.
36. Hornung M, Bertog SC, Franke J, Id D, Taaffe M, Wunderlich N, et
al. Long-term results of a randomized trial comparing three different
devices for percutaneous closure of a patent foramen ovale. Eur Heart J.
2013;34:3362–3369. doi: 10.1093/eurheartj/eht283.
37. Kearon C, Akl EA, Ornelas J, Blaivas A, Jimenez D, Bounameaux
H, et al. Antithrombotic therapy for VTE disease: CHEST guide-
line and expert panel report. Chest. 2016;149:315–352. doi:
10.1016/j.chest.2015.11.026.
38. Thaler DE, Ruthazer R, Weimar C, Mas JL, Serena J, Di Angelantonio
E, et al. Recurrent stroke predictors differ in medically treated patients
with pathogenic vs. other PFOs. Neurology. 2014;83:221–226. doi:
10.1212/WNL.0000000000000589.
39. Doherty JU, Gluckman TJ, Hucker WJ, Januzzi JL Jr, Ortel TL,
Saxonhouse SJ, et al. 2017 ACC expert consensus decision pathway for
periprocedural management of anticoagulation in patients with nonval-
vular atrial fibrillation: a report of the American College of Cardiology
Clinical Expert Consensus Document Task Force. J Am Coll Cardiol.
2017;69:871–898. doi: 10.1016/j.jacc.2016.11.024.
40. Ning M, Lo EH, Ning PC, Xu SY, McMullin D, Demirjian Z, et al. The
brain’s heart - therapeutic opportunities for patent foramen ovale (PFO)
and neurovascular disease. Pharmacol Ther. 2013;139:111–123. doi:
10.1016/j.pharmthera.2013.03.007.
41. Melikian N, MacCarthy PA. The spectrum of pfo closure devices.
Cardiology Today. 2017;11:55–57.
42. Heit JA. Epidemiology of venous thromboembolism. Nat Rev Cardiol.
2015;12:464–474. doi: 10.1038/nrcardio.2015.83.
43. Fenn J, Raskino M. Mastering the Hype Cycle: How to Choose the Right
Innovation at the Right Time. Harvard Business School Press; 2008.
44. Saver JL, Carroll JD, Smalling R, Thaler D. Letter by Saver et al regard-
ing article, “Guidelines for the prevention of stroke in patients with stroke
and transient ischemic attack: a guideline for healthcare professionals
from the American Heart Association/American Stroke Association”.
Stroke. 2015;46:e85–e86. doi: 10.1161/STROKEAHA.115.007311.
45. Powers WJ, Rabinstein AA, Ackerson T, Adeoye OM, Bambakidis
NC, Becker K, et al; American Heart Association Stroke Council. 2018
guidelines for the early management of patients with acute ischemic
stroke: a guideline for healthcare professionals from the American Heart
Association/American Stroke Association. Stroke. 2018;49:e46–e110.
doi: 10.1161/STR.0000000000000158.
KEY WORDS: aneurysms ◼ cardiologists ◼ embolism ◼ risk ◼ thrombosis
 Patent Foramen Ovale Closure Versus Medical Therapy for Cryptogenic Ischemic Stroke:
A Topical Review
Jeffrey L. Saver, Heinrich P. Mattle and David Thaler
Downloaded from http://stroke.ahajournals.org/ by guest on May 14, 2018

Stroke. published online May 14, 2018;

Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2018 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/early/2018/05/11/STROKEAHA.117.018153.citation

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.

Reprints: Information about reprints can be found online at:

http://www.lww.com/reprints

Subscriptions: Information about subscribing to Stroke is online at:

http://stroke.ahajournals.org//subscriptions/