Article Text

Download PDFPDF

Patent foramen ovale and ventricular septal defect closure
  1. A Wahl,
  2. B Meier
  1. Cardiology, Swiss Cardiovascular Centre Bern, University Hospital, Bern, Switzerland
  1. Professor B Meier, Cardiology, Swiss Cardiovascular Centre Bern, University Hospital, 3010 Bern, Switzerland; bernhard.meier{at}


Despite the growing recognition of the patent foramen ovale (PFO), particularly when associated with an atrial septal aneurysm, as risk factor for several disease manifestations (above all paradoxical embolism), the optimal treatment strategy for symptomatic patients remains controversial. Percutaneous PFO closure is a minimally invasive procedure which can be performed with high success and low morbidity. For secondary prevention of recurrent embolic events, it appears to be clinically at least as effective as oral anticoagulation. Ventricular septal defects (VSDs) are the most common congenital heart defects. Percutaneous VSD closure is more intricate than PFO closure. It is associated with a significant risk of both peri-interventional and mid-term complications. In suitable patients with congenital VSD, device closure may well be the preferred treatment both for muscular or perimembranous VSDs and for residual defects after surgical VSD closure. The risk of complete atrioventricular conduction block remains a concern in the perimembranous group. The history, technique and clinical role of percutaneous PFO and VSD closure are discussed, with emphasis on current problems and future developments.

View Full Text

Statistics from

A 39-year-old healthy nurse had a stroke and turned permanently aphasic. She had been complaining about debilitating migraine headaches for about 10 years before that. No one had thought to look for her patent foramen ovale (PFO) before it caused havoc. Had the PFO been found in time, its significant risks in light of the associated atrial septal aneurysm (ASA) recognised and a device implanted in a 30-min outpatient procedure, the mother of two teenagers would be talking to her family for the rest of her long life, and, less importantly, most probably her migraine would have improved (fig 1).

Figure 1 Patent foramen ovale (PFO) gaping at 1.22 cm in a 39-year-old mother of two teenage boys, aphasic since 2 years after a stroke almost certainly mediated by the PFO. The patient had had bad migraine headaches for 10 years before the stroke. LA, left atrium; RA, right atrium.

A 34-year-old healthy woman was admitted in the middle of the night with acute lateral ST elevation myocardial infarction. Emergency cardiac catheterisation showed an occluded left circumflex coronary artery with otherwise normal vasculature (fig 2). Paradoxical embolism being the most likely cause, a right heart catheter was inserted during the same procedure. It easily passed the suspected PFO, which was immediately closed with a device. The patient left hospital 2 days later in good shape.

Figure 2 Emergency coronary angiogram in a 34-year-old woman presenting with an acute lateral myocardial infarction. The arrow points to an embolic occlusion of the left circumflex coronary artery.

These two cases characterise what is at stake when talking about the PFO. In the first patient, the chance was flagrantly missed. In the second, the small myocardial infarction may have been a blessing, prompting action before a possible stroke.

The foramen ovale, a valve-like opening of the interatrial septum, is a pivotal feature during intrauterine life. The interatrial septum is made from two overlapping embryological structures—the left-sided partially fibrous septum primum and the right-sided muscular septum secundum. These grow from the periphery to the centre, leaving a central opening, called the foramen ovale, with the septum primum serving as a one-way slit valve for physiological right-to-left shunt during in utero development. The blood flow from the umbilical vein entering the right atrium through the inferior vena cava and bypassing the non-functional lungs keeps the foramen ovale open until after birth. From then on, the postnatal enhancement of the pulmonary circulation, with subsequent increase in left atrial pressure to a level slightly above right atrial pressure, results in functional closure of the foramen ovale by apposition of the septum primum against the septum secundum. In most people, the caudal portion of the septum primum on the left side and the cranial portion of the septum secundum on the right side fuse permanently in the ensuing months. However, autopsy studies have shown that the foramen ovale remains patent—that is, can be opened in about one-quarter of the general population.1 The PFO thus represents the most common cardiac congenital abnormality (fig 3), and permits intracardiac shunting during those periods when right atrial pressure exceeds left atrial pressure (figs 4 and 5). The prevalence of PFO declines from 34% during the first three decades, to 25% during the fourth to eighth decade and to 20% in the ninth decade and beyond. Selective mortality or spontaneous closure even late in life may be reasons for that. An autosomal dominant inheritance has been demonstrated.2 There is no gender preference.

Figure 3 Possibilities of diagnosing a patent foramen ovale (PFO). (A) Spontaneous right-to-left shunt in a transoesophageal echocardiogram with colour Doppler. (B) Bubble test at the end of a sustained Valsalva manoeuvre. (C) Contrast injection through a catheter passed through the femoral vein into the right atrium (RA), about to pass the PFO into the left atrium (LA). (D) Closure of the PFO with a device.
Figure 4 Transcranial Doppler recording immediately before (A) and after (B) occluding the patent foramen ovale (PFO) in a 72-year-old patient. An identical right atrial bubble injection after a sustained Valsalva manoeuvre produces marked signals testifying to the passage of bubbles through the PFO before (C) but not after (D) closure.
Figure 5 Eustachian valve (EV) as an additional risk factor for paradoxical stroke in presence of a patent foramen ovale (PFO). (A) Right atrial angiogram showing the EV virtually dividing the right atrium (RA) into two compartments and canalising the inflow from the inferior vena cava (IVC) directly onto the PFO, which is marked by the implanted Amplatzer 25 mm occluder. (B) Transoesophageal echocardiography depicted in an analogous projection demonstrating these features as well. Note that access to the PFO of bubbles injected into an arm vein is hindered by the EV. LA, left atrium.

ASA is a congenital abnormality of the interatrial septum characterised by a redundant, amuscular membrane in the region of the fossa ovalis, corresponding to a sector of the central part of the septum primum (figs 6 and 7). The prevalence of ASA in the general population was about 1% in autopsy series35 and 2.2% in a population-based transoesophageal echocardiographic (TOE) study.6 ASA is associated with a PFO in 50–85% of cases. The instability of an ASA renders postnatal fusion difficult and thus begets PFOs. The criteria for distinction between a floppy interatrial septum and an ASA vary between autopsy, transthoracic and TOE studies. ASA is generally diagnosed if the diameter of the base of the flimsy portion of the interatrial septum exceeds 15 mm, and if the excursion of the aneurysmal membrane exceeds 10 mm in either the left or right atrium, or if the sum of the total excursion is ⩾10 mm.5

Figure 6 Marked atrial septal aneurysm (ASA) associated with a relatively small shunt through the patent foramen ovale (PFO). LA, left atrium; RA, right atrium.
Figure 7 Follow-up angiogram 6 months after closure of a patent foramen ovale (PFO) associated with an atrial septal aneurysm (ASA). The PFO is completely closed and the ASA is still present (motion of the septum below the device demonstrated by the two frames at various periods of the heart cycle). This situation no longer constitutes a risk for systemic embolism.

In most people, a PFO will remain asymptomatic for life. However, since the initial description of a fatal stroke in a young woman linked to a PFO by Cohnheim in 1877, PFO and ASA have been increasingly recognised as potential mediators of several disease manifestations, including paradoxical embolism, orthostatic desaturation in the setting of the rare platypnoea–orthodeoxia syndrome, refractory hypoxaemia due to right-to-left shunt in patients with right ventricular infarction or severe pulmonary disease, neurological decompression illness in divers, migraine with aura, transient global amnesia, obstructive sleep apnoea and, more recently, even high-altitude pulmonary oedema.

Ventricular septal defects (VSDs) are the most common congenital heart defects both in children and adults.7 Their documented incidence has increased dramatically with advances in echocardiographic techniques. They are present in from 0.1 to 5% of live births, with an estimated prevalence of simple VSDs in the adult population of 0.3 per 1000. Their haemodynamic significance is influenced by the size of the defect, the pressure gradient between both ventricles and pulmonary vascular resistance. These openings in the ventricular septum are classified according to their location. Perimembranous VSDs, which include about 70–80% of cases, involve the membranous septum and the adjacent area of the muscular septum. Tricuspid valve aneurysm formation can lead to their functional closure. The far less common muscular VSDs occur in 15% of cases, and can undergo spontaneous closure. The rare infundibular VSDs, which account for around 5% of all VSDs, can functionally close by prolapse of the right aortic cusp. Finally, ventricular septal rupture is a mechanical complication of myocardial infarction which remains associated with very high morbidity and mortality. Before the reperfusion era, it occurred typically after 3–5 days in 2% of patients. It is seen in 0.2% of patients with reperfused myocardial infarcts.

Medium and larger congenital VSDs, as well as virtually all acquired VSDs, mostly after infarction, require closure for prognostic reasons, in order to prevent development of pulmonary arterial hypertension, ventricular dysfunction and arrhythmias. A history of endocarditis constitutes another indication for VSD closure. Only congenital VSDs will be discussed here. Surgical repair of congenital VSDs is performed with low mortality (1–5%), but complications include complete atrioventricular block in up to 3% of patients, as well as other conduction disturbances and arrhythmias, chylothorax, phrenic nerve injury, postpericardiotomy syndrome, wound infection and neurological sequelae of cardiopulmonary bypass. While in the early era 80% closure and 3% complete atrioventricular block rates were reported in catheterised patients at long-term follow-up,8 more recent studies9 report residual shunting in 31% and complete atrioventricular block in up to 3% of patients, with occurrence rates10 of 10 per 10 000 patient-years for pacemaker implantation and 16 per 10 000 for endocarditis. However, most residual shunts after surgical closure are insignificant, and a reintervention is rarely required.10 The rare patients with multiple VSDs, so-called Swiss cheese VSDs, have particularly high residual shunt rates. Adult patients with successful surgical closure need follow-up for aortic valve dysfunction. Those with residual shunt need follow-up for left ventricular volume overload.


Percutaneous ASD closure was reported by King et al in the 1970s,10a Rashkind in the 1980s,10b and Sideris et al in the 1990s.10c Since Bridges et al proposed in 1992 that percutaneous PFO closure would reduce the incidence of recurrent strokes, percutaneous PFO closure has been shown to be safe and feasible in numerous studies, using a variety of devices (fig 8), according to historical availability.1117 In these studies (table 1), the reported success rates varied between 90 and 100%, with complication rates between 0 and 10%. Complete PFO closure was achieved in 51–100% of patients, and the reported yearly recurrence rates of ischaemic strokes and transient ischaemic attacks varied between 0 and 3.4%.

Figure 8 Clinically used patent foramen ovale occlusion devices (manufacturers indicated inside the figures).
Table 1 Selected studies of percutaneous patent foramen ovale closure

As for ASD and PFO closure, the potential benefits of percutaneous VSD closure, mainly avoidance of open heart surgery and cardiopulmonary bypass, appear intuitive. In 1987, Lock et al attempted percutaneous closure of seven postinfarction, congenital or postoperative congenital VSDs in six patients, using Rashkind double umbrella devices.17a All VSDs were successfully crossed, but one device embolised immediately into the pulmonary artery, and all four patients with postinfarction VSD died from increasing VSD shunting within several days, with visualisation of other VSDs at autopsy. Thereafter, these and other devices originally designed for the closure of other intracardiac defects were used with variable success (77–100%). Besides the very high residual shunt rates (35–100%), their major drawbacks were the large (11 F) delivery sheaths required, complex implantation techniques, inability to reposition and redeploy the device and interference with aortic and tricuspid valves. Specially designed Amplatzer VSD occluders became available in 1998, and initial applications in children were reported by Thanopoulos et al17b and Hijazi et al.17c The main advantages of the Amplatzer membranous and muscular VSD occluders are the possibility of recapturing, redeploying, or removing the device until it has been unscrewed from its delivery cable, the smaller delivery sheaths (7–9 F), the higher closure rates attributable to its mechanism of closure (stenting of the defect with the waist of the device) and its round shape, minimising the potential for valvular or tissue injury. In 50 patients undergoing percutaneous closure of muscular VSDs,18 using the Sideris (n = 2) and Amplatzer muscular VSD occluder (n = 48), device success was 100%, without residual shunt by transthoracic colour Doppler examination at 24 h, and no late complications.

In a phase I multicentre clinical trial of the Amplatzer membranous VSD occluder in 35 patients,19 the only device specifically designed for perimembranous VSD closure, device success was 91%, with a complete closure rate of 96% at 6 months by transthoracic echocardiography. There was no mortality, but 9% of patients had serious complications (one permanent complete atrioventricular block requiring pacemaker implantation after 3 months, one peri-hepatic bleeding requiring transfusions and one failed transcatheter device retrieval with rupture of the tricuspid valve chordae tendinae requiring surgical device removal, tricuspid valve repair and VSD closure), and miscellaneous adverse events occurred in 15 additional patients (43%). One more patient (2%) developed complete atrioventricular block requiring pacemaker implantation 16 months after device implantation. In a European multicentre trial,20 54 paediatric patients underwent transcatheter VSD closure using the Amplatzer membranous VSD occluder, with 91% device success and 94% complete occlusion rate at 1 year, as assessed by transthoracic colour Doppler echocardiography. Main procedural complications precluding device implantation included severe bradycardia with haemodynamic compromise in two patients (4%), and transient Mobitz II (2:1) atrioventricular block in one patient (2%). Furthermore, device embolisation into the descending aorta with percutaneous removal and successful closure with another occluder was seen in two patients (4%), and one patient (2%) developed late complete atrioventricular block requiring pacemaker implantation. In a partially incomplete international registry encompassing 100 patients,21 device success was 93%, and the complete closure rate 84% at 6 months follow-up, with complications in 29% of patients, including complete atrioventricular block requiring pacemaker implantation in 2%.


All patients should undergo contrast TOE before the intervention for initial diagnosis of PFO and detailed delineation of anatomy (ie, associated ASA, Eustachian valve, fig 5) including assessment of right-to-left shunt. Echocardiographic “high-risk” features which should affect decision-making include larger PFO size,2225 a greater degree of right-to-left shunt22 24 26 27 and the presence of an associated ASA.28

The procedure is generally performed using simultaneous fluoroscopic and TOE12 1416 or intracardiac 29 echocardiographic (ICE) guidance. However, we feel that routine echocardiographic guidance provides little additional information to the knowledge gleaned from a hand injection of contrast medium30 in a profile-adjusted view (fig 1). TOE is poorly tolerated by supine patients, and comes therefore at the cost of sedation or general anaesthesia, including intubation to exclude the risk of bronchial aspiration. ICE29 is a costly alternative that is more comfortable for the patient and eliminates the need for general anaesthesia. However, it increases the invasive risk (second access, rigid, unguided intravenous catheter). Moreover, both imaging modalities considerably lengthen the procedure.

After groin anaesthesia, the PFO is crossed via the right femoral vein under fluoroscopic guidance in the anteroposterior view either by a standard length normal 0.035 inch (0.9 mm) guidewire, or with the help of a catheter, typically a 6 French multipurpose catheter. Balloon sizing is not required. Indeed, the maximal opening of the flap-like PFO is not instrumental for the success of closure. Moreover, there is a finite risk of laceration of a thin septum primum. Usually a TOE precedes percutaneous closure and provides sufficient information about the presence and length of the tunnel and the solidity of the septum primum (ASA). This obviates the need for balloon measurements for all devices. The device-specific delivery system is then inserted over this wire. To keep the indwelling time of the sheath short, the device should have been prepared previously. Keeping the proximal sheath exit below heart level and the distal sheath exit away from the atrial wall, oozing through the sheath should be ascertained before device insertion to avoid air embolism. Using Amplatzer PFO occluders, a 25 mm device can be selected for all cases except those with a particularly large ASA. In 5% of cases, this device will have to be exchanged for a larger one during the procedure because of suboptimal anchoring (negative Pacman sign30). The left-sided disc is then deployed and gently pulled back against the interatrial septum under fluoroscopic guidance in a left anterior oblique projection. To deploy the right atrial disc, tension is maintained on the delivery cable while the delivery sheath is further withdrawn. Right atrial contrast angiography by a hand injection through the sidearm of the delivery sheath serves to delineate the atrial septum. The so-called “Pacman sign” refers to the aspect of the device on fluoroscopy that should be achieved before release. Seen in profile, the cranial halves of the left and right atrial discs should appear like open jaws biting into the thick septum secundum, reminiscent of the arcade game figure Pacman about to gobble up a dot30 (fig 9). Of course, TOE or ICE can also be used to ascertain correct device position, but all devices so far available permit injection of dye through the introducer sheath, which is all that is needed to ascertain correct placement before release, provided that the correct view is selected. Finally, the device is unscrewed from the delivery cable, the sheath is removed and haemostasis is achieved by manual compression, which can be done by the patient himself. The procedure can be performed on an outpatient basis and may take less than 30 min, with a fluoroscopy time of 5 min. Patients can be released to unrestricted physical activity as soon as a few hours after the intervention.

Figure 9 Pacman sign used to assess angiographically proper seating of the occluder device before release. (A) Hand injection of contrast medium into the right atrium through the introducer delineating the salient features of a PFO—that is, the device being spread cranially (to the left) by the thick, muscular septum secundum (SS) while the discs are hardly separated by the paper-thin septum primum (SP, dashed line) caudally (to the right). (B) The aspect is reminiscent of the arcade figure Pacman, which is superimposed on the very picture of the left hand panel.

Antibiotics during the intervention are commonplace in virtually all laboratories, and prevention against endocarditis is usually recommended for a few months until the device is completely covered by tissue. However, the use of antibiotics is not supported by evidence. Failed implantation, due to inability to canulate the PFO, is extremely rare (<1%). A trans-septal puncture for device implantation has been recommended by some authors in these situations, as well as with funnel-shaped PFOs. However, a PFO that cannot be canulated must represent a negligible risk for spontaneous passage of particulate matter, and the Amplatzer PFO occluder is designed to adjust to funnel-type PFOs, in contrast to fixed-connector-distance devices. When the newest generation of devices is used, the periprocedural complication rate is <1% in experienced centres, and complete closure rates >90% can be expected. Follow-up treatment for antithrombotic protection until full device endothelialisation empirically includes acetylsalicylic acid (80–300 mg daily) for 1–6 months, with the addition of clopidogrel (75 mg once a day) for 1–6 months at some centres. At 3–6 months after percutaneous PFO closure, a contrast TOE should be repeated, to assess for a residual shunt after endothelial overgrowth, and exclude thrombosis of the device, the thrombus rate varying significantly between the different devices.31 If the PFO proves to be completely closed, all drugs can be discontinued, unless required for another indication, such as associated coronary artery disease. When a small residual shunt persists, we recommend the continuation of acetylsalicylic acid, and a further contrast TOE at 1 year, since small residual shunts may close late. When a moderate or large residual shunt persists, implantation of a second device is recommended, which results in complete closure in about 90% of cases.


Haemodynamically relevant muscular VSDs, though in theory ideally suited for percutaneous closure, are rare beyond infancy. The more common perimembranous VSDs, being close to the aortic and atrioventricular valves, with the inherent risk of injuring valvular and trabecular structures or conductive tissues, are less suited for percutaneous closure.

The technique of percutaneous VSD closure is similar to PFO closure, but technically far more challenging (fig 10). Echocardiographic guidance, although not necessarily required, is used in most cases, in order to facilitate device deployment and immediate identification of adverse events, such as aortic regurgitation caused by impingement of the device on the valve leaflets, or mitral or tricuspid regurgitation caused by interference with the chordae tendinae. VSD sizing is problematic, since oversizing of the device bears a higher risk of complete atrioventricular block and aortic valve dysfunction. In practice, the average of the echocardiographic and the angiographic measurements seems most accurate. The major difficulty is passage of the defect, which is typically crossed from the left ventricular side using a guidewire which is then snared with a venous catheter, and used to direct a long venous sheath and, ultimately, the device across the VSD from the right ventricular side. The distal disc is then opened and positioned on the left ventricular side of the VSD, followed by deployment of the proximal disc on the right ventricular side. Device delivery from the arterial side is an exception in cases where the sheath cannot be passed over the wire from the venous side. Placement of the device is more challenging than for PFO closure, with frequent inaccurate deployment of both discs in the left ventricle. Reasons for abandoned procedures include size of the defect, residual shunting, aortic valve dysfunction and inability to advance the sheath. As for PFO closure, follow-up treatment empirically includes acetylsalicylic acid and prophylaxis of infectious endocarditis for around 6 months.

Figure 10 (A) Muscular ventricular septal defect (VSD; arrow) in a 31-year-old woman. (B) Closure with a 10 mm Amplatzer muscular VSD occluder. LV, left ventricle; RV, right ventricle.


Paradoxical embolism

Interestingly, despite the high prevalence of a PFO as an embryonic remnant in the general population,1 the absolute lifetime risk of stroke is small. Our lifetime risk of stroke is much smaller even than the risk of recurrent embolic events in symptomatic patients, thus underlining our limited understanding of the pathophysiology of cryptogenic stroke in general, and of the risk associated with a PFO, in particular. In younger patients, the aetiology of ischaemic stroke remains unknown in up to 40% despite an extensive diagnostic evaluation, and is referred to as cryptogenic.23 32 Several case–control studies using contrast echocardiography showed a strong relation between the presence of PFO and cryptogenic stroke in adults aged <55 years,3338 whereas this relationship is still controversial in older age groups.3537 39 According to a meta-analysis,28 in patients younger than 55 years, a PFO confers a relative risk of 3 (95% CI 2 to 4) comparing ischaemic stroke with non-stroke control subjects, and a relative risk of 6 (95% CI 4 to 10) comparing cryptogenic stroke with control subjects with a known cause of stroke.

Traditionally, most patients with presumed paradoxical embolism were treated medically with antithrombotic drugs. However, data are scarce about the efficacy of oral anticoagulation as opposed to antiplatelet therapy, and the duration of treatment required. Furthermore, patients with cryptogenic stroke related to PFO are at increased risk for recurrence despite medical treatment,4044 and the bleeding risk (around 1% of patients have severe bleeding a year) has to be considered. Both, larger PFO size2225 and a greater degree of right-to-left shunt22 24 26 27 signify a higher risk for paradoxical embolism. There are major differences in the baseline characteristics of the patient populations studied, which may account for the disparate recurrence rates reported, ranging from extremely low,42 barely justifying chronic low dose acetylsalicylic acid intake, and certainly prohibiting anything associated with a higher risk, to extremely worrisome, and thus requiring more aggressive treatments, even if accepting potentially serious complications. The optimal treatment strategy of symptomatic patients remains undefined, with non-randomised data suggesting an advantage of percutaneous PFO closure over medical treatment (fig 11).45 46 In particular, medical treatment alone may offer insufficient protection, at least for some subgroups—that is, patients with an associated ASA, and for those having had a recurrent embolic event despite medical treatment.

Figure 11 Cumulative risk of death, recurrent ischaemic stroke and transient ischaemic attack during 4 years in a cohort of patients with different treatment modalities after an initial presumed paradoxical cerebrovascular embolism in the presence of a patent foramen ovale (PFO).46

Younger patients with an isolated PFO having had their first ever embolic event are likely to have a very low risk of recurrence,42 regardless of treatment. In these cases, PFO closure might be an alternative to lifelong antithrombotic treatment. However, patients have to be told that there is a small risk of recurrent events despite successful PFO closure—in our experience <1% per year. Stroke pathophysiology is multifactorial, and a diagnosis of PFO-mediated paradoxical embolism is presumptive. Both a PFO and a cryptogenic stroke may coexist without causal relation in a given patient. In this case, PFO closure will not reduce the risk of recurrence. Thrombus formation31 on the left-sided disc may also beget stroke. For patients with multiple embolic events, percutaneous PFO closure is probably warranted,46 especially if medical treatment has already failed.

ASA has been associated with cerebral ischaemic events in numerous case–control studies.5 6 38 41 42 47 48 The combination of a PFO with an ASA constitutes a situation of particularly high risk with a relative risk of 16 (95% CI 3 to 86) comparing ischaemic stroke with non-stroke control subjects, and a relative risk of 17 (95% CI 2 to 134) comparing cryptogenic stroke with control subjects (age <55 years) with a known cause of stroke.28 ASA may act as a facilitator for paradoxical embolism. The presence of ASA may increase the PFO diameter owing to the highly mobile atrial septal tissue, leading to a more frequent and wider opening of the channel.22 ASA may also promote a right-to-left shunt by redirecting flow from the inferior vena cava towards the PFO.27 ASA has been indicted as being a nidus for local thrombus formation with subsequent embolisation.47 49 This suggestion can, however, be discarded (except perhaps for patients in atrial fibrillation) as a misdiagnosis of the pre-TOE era.50 Patients with both PFO and ASA constitute a high-risk population with a three- to fivefold increased risk for recurrent embolic events compared with patients with PFO alone.28 Secondary prevention with acetylsalicylic acid alone has been found to be insufficient for protection against recurrent cerebrovascular events in patients with both PFO and ASA,41 42 with a risk of recurrent stroke or transient ischaemic attack of 19% at 4 years. With transcatheter treatment, an ASA associated with the PFO had no influence on device success, nor on the risk of periprocedural complications, the residual shunt rate, or the risk of recurrent events (5% at 4 years).17

Concern has been raised that the current focus on cryptogenic stroke as an indication for PFO closure may deprive the elderly, who have the highest risk51 52 of paradoxical embolism, of a simple preventive treatment.53 Of note, percutaneous PFO closure proved equally safe, feasible and effective for all ages.

At present, the most restrictive indications are applied in the USA, where failed medical treatment for secondary stroke prevention constitutes the sole FDA accepted indication for PFO closure. Two devices were approved by the FDA under the Humanitarian Use Device regulations in patients with presumed paradoxical embolism for whom medical treatment had failed, but these exemptions were later withdrawn owing to the potential to exceed treating 4000 patients a year. However, considering that a PFO can be closed in 30 min, it is difficult to understand why patients should have to have had two strokes before being allowed a simple and plausible treatment. Nevertheless, it must be emphasised that the true therapeutic efficacy of percutaneous PFO closure as an adjunct or alternative to medical treatment needs to be ascertained by randomised studies. The three randomised trials currently underway (PC-trial, RESPECT and CLOSURE) will take several more years before completion. Furthermore, these studies are likely to be underpowered, since the risk of recurrent events has probably been overestimated.

Miscellaneous indications

The risk of a PFO in the perioperative period, with its increased presence of potential emboli (air, venous blood clots or fat) in association with non-physiological intrathoracic pressures, has not been systematically assessed. It has been suggested that high-risk patients should be screened for the presence of a PFO before elective susceptible surgical procedures.54 The platypnoea–orthodeoxia syndrome55 is a rare clinical entity characterised by dyspnoea related to arterial oxygen desaturation induced by the upright position and relieved by recumbency. Both surgical and percutaneous PFO closure have been shown to be curative by elimination of the right-to-left shunt with resolution of systemic oxygen desaturation and thereby relief of symptoms. In the presence of a PFO, both right ventricular infarction56 and severe pulmonary disease can be complicated by refractory hypoxaemia (ie, unresponsive to oxygen administration) owing to PFO-mediated right-to-left shunt. In light of the minimally invasive nature of percutaneous PFO closure, and the severity of the frequently associated comorbid conditions, we believe that percutaneous PFO closure should be the preferred treatment in these patients. The PFO has also been identified as an independent risk factor for mortality (odds ratio 11) and a complicated in-hospital course (odds ratio 5) in patients with major pulmonary embolism.57 This has been related to the PFO-mediated right-to-left shunt in the presence of raised pulmonary artery pressures predisposing to paradoxical embolism with cerebral and vascular embolic complications, including death. The PFO gains similar importance in patients with raised pulmonary artery pressures due to chronic pulmonary parenchymal or vascular disease, and in patients with raised right atrial pressures due to congenital heart disease. Some reports have highlighted a connection between obstructive sleep apnoea and the PFO. This might be explained by a sort of paradoxical platypnoea–orthodeoxia (desaturation by right-to-left shunt in the supine position).58 59

Decompression illness in divers

Neurological decompression illness in divers is caused by regional gas nucleation in predominantly fat-containing tissue as well as by arterial gas embolism, with both conditions being associated with ischaemic brain lesions. It has been hypothesised that a right-to-left shunt in divers with PFO allows for systemic embolisation of venous gas bubbles, and therefore increases the risk of decompression illness.60 Although divers with PFO are more likely to have decompression illness and appear to have more ischaemic brain lesions than divers without PFO, diving itself is associated with a higher incidence of ischaemic brain lesions regardless of the presence of a PFO.61 At present, there is no evidence to ban people with a PFO from scuba diving, and no data on the efficacy of PFO closure for primary or secondary prevention of decompression illness.


Migraine affects about 10% of population—6% of men and 16% of women. It is a complex disorder, in which both genetic and environmental factors play an important role, and its pathophysiology is not fully understood.62 In several studies of patients with migraine, especially of migraine with aura, the prevalence of either right-to-left shunt detected by transcranial Doppler,63 64 presumably through a PFO, or of a directly visualised PFO65 was similar to the prevalence of PFO in patients with cryptogenic stroke. Furthermore, in patients with symptomatic PFO, such as paradoxical embolism66or decompression illness,67 a higher prevalence of migraine has been reported.

At least seven retrospective studies (table 2) have reported an improvement in migraine prevalence or migraine frequency after PFO closure for other reasons, mostly for secondary prevention of paradoxical embolism or after diving accidents. A recent prospective study described a decrease in migraine prevalence of 36% 1 year after percutaneous PFO closure, while it was unchanged in the control group.68 However, the only prospective randomised controlled study with a sham procedure, the MIST trial (Migraine Intervention with STARFlex Technology), including 147 patients with treatment-resistant migraine with aura, failed to reach its primary end point (3% of patients “cured” in each group at 6 months follow-up).68a However, a 50% reduction in migraine days was achieved in 42% of patients of the interventional group as compared with 23% of patients in the placebo group. Since the device used in this study is known to have a relatively high residual shunt rate, it would be interesting to know the relationship between residual shunt and migraine.

Table 2 Prevalence of migraine before and after percutaneous patent foramen ovale closure

Overall, while percutaneous PFO closure seems to improve mid-term migraine symptoms, the total number of patients included in these studies is quite small, and follow-up limited. Furthermore, the completeness of PFO closure does not seem to be associated with migraine relief, which is disturbing from a pathophysiological perspective. Nevertheless, since even migraine refractory to medical treatment is a prevalent condition and since migraine attacks are much more frequent than recurrent cerebrovascular events, the first evidence for PFO closure might come from migraine studies (table 3). A study reporting migraine improvement after pulmonary fistulae closure in Osler patients corroborates the link between right-to-left shunt and migraine.69

Table 3 Current patent foramen ovale (PFO)/migraine randomised studies


With PFO closure, initial device-related complications inflicted by large delivery systems, device dislodgement and embolisation, structural failure, thrombus formation and inability to reposition or remove the device have been subsequently reduced by improvements in device design. Complication rates have fallen to <1% with new devices in experienced centres,17 and consist mostly of venous access problems. There are important differences in complication rates, residual shunt and thrombosis,31 between the currently available devices,17 70 which are probably clinically relevant. Since, in our experience, a residual shunt is a risk factor for recurrent events,11 13 17 complete PFO occlusion is desirable, a goal which can be achieved in >90% of patients with the most recent devices.15 17 Erosion of the free atrial walls, device endocarditis and a need for surgical explantation are exceedingly rare. These are important findings for conditions that have low recurrence rates in natural history. One particular concern is the later development of supraventricular arrhythmias, possibly induced by the device itself, which may require left atrial ablation. In this respect, trans-septal puncture is, if anything, facilitated after device implantation (fig 12). Schoen et al recently reported a new or worsened mild or moderate aortic regurgitation on follow-up TOE in 10% of patients after PFO closure with the CardiaSTAR occluder.71 However, none of them was clinically relevant and we have not seen this phenomenon in our patients.

Figure 12 Demonstration of the large safe zone in which to perform a trans-septal puncture after device closure of a patent foramen ovale. The thin septum primum caudal (to the right) of the device provides a safe area of several square centimetres, reaching from the lower end of the device (ideal place to puncture) to the bottom of the picture. The diameter of the device is 25 mm for reference.

As compared with PFO and ASD closure, percutaneous VSD closure is far more complex and less standardised. Although closure of congenital VSDs appears safe and effective in experienced hands, its indications and limitations are less well defined, and the procedure is associated with a relatively high risk of both peri-interventional and potential long-term complications. These include device embolisation and migration, residual shunting, haemolysis, myocardial perforation, arrhythmias, possibly systemic and pulmonary embolism and, especially with perimembranous VSDs, also aortic and tricuspid valve injury and late complete atrioventricular block. For percutaneous closure of muscular VSDs, closure rates >95% and no long-term complications such as arrhythmias requiring pacemaker implantation or need for surgical explantation have been reported so far. For perimembranous VSDs, even if the rates of residual shunting appear comparable to those of contemporary surgery, the rates of complete atrioventricular block, estimated at 2–6%, are probably higher. Indeed, for isolated surgical perimembranous VSD closure, a risk of complete atrioventricular block as low as 0.7% has been reported.72 Possible mechanisms for early block include trauma from catheter or device manipulation and inflammatory reaction induced by the device, and for late block, both trauma from continuing expansion of an oversized device and device-induced fibrosis. Thus, it appears of paramount importance to avoid oversizing of the device. Overall, these figures are of concern, especially in paediatric patients with a long life expectancy.


The CIERRA PFx closure system is a new percutaneous system that employs radiofrequency energy to close the PFO by welding the tissues of the septum primum and septum secundum together, thus avoiding the need for an implantable device. In brief, after crossing the PFO using a guidewire, the PFx catheter is advanced into the right atrium under fluoroscopic and TOE or ICE guidance. Vacuum pressure is then applied to hold the septum primum and secundum in place, and radiofrequency energy is applied to close the PFO. At present, this device is plagued by low feasibility, and high residual shunt rates, requiring subsequent device closure.

Other developments include transcatheter suture systems, and absorbable devices. Indeed, once the PFO is endothelialised, there is no need for a device any more. The bioabsorbable BioSTAR device, a development from the STARFlex device, consists of an acellular porcine intestinal collagen matrix coated with a heparin benzalkonium chloride complex mounted on the nitinol frame of its parent double umbrella device. The collagen matrix is gradually absorbed over a period of about 24 months and replaced by host tissue, leaving almost only the nitinol frame behind. In a phase I multicentre trial, successful device implantation was achieved in 57 of 58 patients (98%) with a PFO or small ASD.73 Five patients required treatment for transient supraventricular arrhythmias, one patient developed urticaria after device implant and one patient had a right atrial thrombus at 1 month TOE. Closure rates at 1 month were 48/52 (92%) and 54/56 (96%) at 6 months, with one patient lost to follow-up. However, residual shunt assessment was conducted using transthoracic contrast echocardiography only, and only moderate or large shunts were reported. Therefore, and also considering the relatively high residual shunt rates of the parent STARFlex device, much higher residual shunt rates have to be assumed demonstrable by contrast TOE. Of interest, follow-up TOE residual shunt data at 1 month were not reported.

For VSD closure, new developments include surgical device placement techniques for patients with multiple VSDs, thus reducing the invasiveness of surgical closure and avoiding cardiopulmonary bypass.


Despite growing recognition of the PFO, particularly when associated with an ASA, as a risk factor for several disease manifestations, paradoxical embolism in particular, the optimal treatment strategy for symptomatic patients remains controversial. Percutaneous PFO closure is a minimally invasive procedure, which can be performed with high success and low morbidity in patients. It has thus supplanted surgical PFO closure, which is associated with the well-known complications of open-heart surgery.74 With respect to secondary prevention of recurrent embolic events, percutaneous PFO closure appears to be clinically at least as effective as medical treatment.45 46 However, it has to be emphasised that the true therapeutic efficacy of percutaneous PFO closure as an adjunct or alternative to medical treatment can only be ascertained from randomised studies, which are far from being completed.

Percutaneous VSD closure is more experimental, and associated with a relatively high risk of both peri-interventional and potential long-term complications. At present, in suitable patients with congenital VSDs, if the expertise is available, device closure appears to be the preferred treatment both for muscular VSDs and for residual defects following unsuccessful surgical VSD closure, and probably also for perimembranous VSDs beyond infancy (>4 year-old?), the risk of complete atrioventricular block being a major concern in this last group.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 10a.
  12. 10b.
  13. 10c.
  14. 11.
  15. 12.
  16. 13.
  17. 14.
  18. 15.
  19. 16.
  20. 17.
  21. 17a.
  22. 17b.
  23. 17c.
  24. 18.
  25. 19.
  26. 20.
  27. 21.
  28. 22.
  29. 23.
  30. 24.
  31. 25.
  32. 26.
  33. 27.
  34. 28.
  35. 29.
  36. 30.
  37. 31.
  38. 32.
  39. 33.
  40. 34.
  41. 35.
  42. 36.
  43. 37.
  44. 38.
  45. 39.
  46. 40.
  47. 41.
  48. 42.
  49. 43.
  50. 44.
  51. 45.
  52. 46.
  53. 47.
  54. 48.
  55. 49.
  56. 50.
  57. 51.
  58. 52.
  59. 53.
  60. 54.
  61. 55.
  62. 56.
  63. 57.
  64. 58.
  65. 59.
  66. 60.
  67. 61.
  68. 62.
  69. 63.
  70. 64.
  71. 65.
  72. 66.
  73. 67.
  74. 68.
  75. 68a.
  76. 69.
  77. 70.
  78. 71.
  79. 72.
  80. 73.
  81. 74.
  82. 75.
  83. 76.
  84. 77.
  85. 78.
  86. 79.
  87. 80.
  88. 81.
  89. 82.
  90. 83.
  91. 84.
  92. 85.
  93. 86.
  94. 87.
  95. 88.
  96. 89.
  97. 90.
  98. 91.
  99. 92.
  100. 93.
  101. 94.
View Abstract


  • Funding: BM: research grant and speaker bureau for AGA Medical.

  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.