The foramen ovale provides communication between both atria in utero, and although anatomical closure usually occurs in the first two years of life, it may remain patent (patent foramen ovale [PFO]) into adulthood. The autopsy prevalence of PFO was 27% among 965 otherwise normal hearts,1 and transoesophageal echocardiography detected PFO in 24.3% of 585 randomly sampled individuals.2 PFO is associated with other septal anomalies, such as atrial septal aneurysms (ASAs) and prominent Chiari networks.2
PFO has been assumed to be an anatomical variant of no clinical relevance. However, it is an obvious substrate for right-to-left shunting and it has lately been implicated in the genesis of different pathologies, for example:
- cryptogenic stroke, myocardial infarction or peripheral embolisms, as it allows for paradoxical embolism;
- decompression syndrome in divers;3
- platypnea-orthodeoxia syndrome;4,5
- complications from pulmonary embolism6 or right ventricle infarction;7,8
- migraine with aura;9,10 and
- hypoxemia in sleep apnoea syndrome.11,12
The favourable procedural results of the percutaneous techniques for atrial septal defect (ASD) closure led to the expansion of the procedure to PFO closure and to the design of dedicated devices. However, despite percutaneous closure of PFO becoming increasingly popular among clinicians and patients, questions regarding its benefits remain unanswered. There is an absolute lack of randomised trials evaluating the secondary prevention of cryptogenic stroke, and clinical decisions are based exclusively on retrospective or non-randomised studies. There has been only one randomised trial of PFO closure published to date, directed to migraine treatment (Migraine Intervention with STARFlex Technology [MIST]), showing no differences for any of the primary or secondary end-points13 with PFO closure. Therefore, prior to describing the percutaneous options for PFO closure, we consider it essential to expose some of the concerns that arise in this field. We will focus on the relation between PFO and stroke or migraine, being the most studied of these conditions.
Cryptogenic Stroke and Patent Foramen Ovale – A Relationship Yet to Be Clearly Defined
Cryptogenic stroke constitutes 30–40% of all causes of stroke in patients under 55 years of age.14 The role of a PFO in a patient with a stroke is difficult to establish, as it could be just an innocent bystander to the aetiological mechanism of a cryptogenic stroke. To complicate matters more, the term ‘cryptogenic’ is a diagnosis of exclusion and thus dependent on the thoroughness of the clinical evaluation, which sometimes might not be profound enough to unveil uncommon risk factors. Similarly, PFO detection depends on the meticulousness of the echocardiographic evaluation, thus introducing bias in non-blinded prevalence studies, as one could be more prone to dedicate time to find a PFO in those with an unknown cause of stroke than in those in which the aetiology is already known.
Nevertheless, the reported prevalence of PFO is consistently higher in patients with cryptogenic stroke compared with controls and with patients with a known cause of stroke, strongly suggesting a role of PFO in the genesis of ischaemic events. This relationship is more evident among patients under 55 years of age, with Lechat et al. reporting PFO in 40% versus 10% of stroke patients and control patients, respectively (p≤0.001),15 and Overell et al. describing an odds ratio (OR) for PFO of 6.0 (95% confidence interval [CI] 3.72–9.78) in patients with cryptogenic stroke versus patients with a known cause of stroke.16 In patients over 55 years of age, the results have not been so consistent, although recent studies support an increased prevalence of PFO among those with cryptogenic stroke.17
Different studies report that an associated ASA defines a higher-risk subgroup,18,19 and Cabanes et al. found the stroke odds of patients with both PFO and ASA to be 33.3 times higher than the stroke odds of patients without.18
However, despite the results of these prevalence studies, the controversy regarding the real role of PFO in cryptogenic stroke has gained interest after some papers reported that PFO does not increase the risk of first ischaemic events or of recurrences. A population-based study of 585 randomly sampled patients aged 45 years or older showed that PFO was not an independent risk factor for cerebrovascular events during a median follow-up of 5.1 years (hazard ratio 1.46, 95% CI 0.74–2.88).2
When analysing recurrent events, the PICCS study found that in patients undergoing pharmacological treatment after a stroke, PFO was not associated with an increased rate of recurrence at two-year follow-up in the overall sample (14.8 versus 15.4%) or in the cryptogenic subgroup (14.3 versus 12.7%), not even after stratifying for PFO size or when comparing isolated PFO with associated ASA (14.5 versus 15.0%).20 A systematic review of the Quality Standards Subcommittee of the American Academy of Neurology similarly showed no increased risk of subsequent stroke or death in patients with PFO (relative risk 0.95, 95% CI 0.62–1.44).14
Secondary Prevention of Cryptogenic Stroke in Patients with Patent Foramen Ovale – What Is the Best Therapeutic Option?
The best therapeutic strategy for these patients is still to be defined. They have traditionally been treated with antiplatelet regimens (especially warfarin and aspirin), but after the recognition that PFO could play a role in the genesis of cryptogenic strokes, percutaneous (and sometimes surgical) closure of the defect is increasingly being used.
- Pharmacological strategies: both antiaggregant and anticoagulant regimens have been used. The PICCS study compared the outcomes of patients randomly assigned to aspirin or warfarin and found no significant differences in stroke recurrence or death among patients with PFO (two-year event rates 13.2 versus 16.5%; p=0.49).20
- PFO closure: a case series of patients who received either pharmacological treatment (aspirin and clopidogrel: 44 patients) or PFO closure (Amplatzer PFO occluder: 48 patients) after a cerebrovascular ischaemic event showed a one-year recurrence rate of 15.9% with pharmacological treatment and 0% after PFO closure (p<0.005).21 Similarly, a systematic review of articles describing the outcomes of stroke patients undergoing pharmacological treatment or transcatheter PFO closure reported a one-year recurrence rate of 3.8–12% with the former strategy and of 0–4.9% with the latter.22
These studies strongly suggest a benefit of percutaneous PFO closure over pharmacological regimens. Nevertheless, the significant baseline differences in these non-randomised studies, with a higher prevalence of risk factors for atherosclerosis in the pharmacologically treated group, limit the possible conclusions about the superiority of one particular approach. Well-controlled, randomised trials are thus imperative to determine the safety and effectiveness of PFO closure and to compare it with standard medical treatment after a stroke in order to really establish its benefit over pharmacological regimens.
In fact, PFO closure does not definitely prevent ischaemic recurrence, neither surgically23,24 nor percutaneously,22,25 suggesting that other factors may also play a role in the genesis of recurrent ischaemic events, i.e. non-diagnosed hypercoagulability states. Incomplete defect closure22,26 and device thrombosis27,28 have been described and might explain some of the recurrences.
Migraine and Patent Foramen Ovale
A higher prevalence of PFO has also been reported among patients with migraine with aura compared with patients with migraine without aura (48 versus 23%, OR 3.13, 95% CI 1.41–7.04) and with controls (48 versus 20%, OR 3.66, 95% CI 1.21–13.25), whereas the prevalence in the last two groups was similar (OR 1.17, 95% CI 0.32–4.45).29
Retrospective studies reported a reduction in the frequency of attacks of migraine with aura after the percutaneous closure of PFO in patients with cryptogenic stroke.9,10 To better define the value of PFO closure in this setting, the MIST trial (the first randomised trial ever completed for percutaneous PFO closure) compared interventional treatment with sham control in patients with migraine with aura. Despite preliminary reports at the ACC 2006 Scientific Sessions suggesting that headache burden and headache days were significantly reduced in the device-treated group, the final published results showed no differences for any of the primary or secondary end-points.13
Initial Strategies for Percutaneous Closure of Patent Foramen Ovale
The main aim of PFO closure is first to eliminate the substrate for right-to- left shunting, and second to avoid lifelong antiplatelet treatment. The procedures for PFO and ASD closure are similar, as are the devices used (and, in fact, devices for ASD closure are sometimes used for PFO closure), and they imply a femoral venous access to reach the right atrium, the interatrial septum and the left atrium through the defect.
Initial designs consist of two discs that are permanently implanted on both sides of the atrial septum covering the defect, which eventually become encapsulated by fibrous healing tissue. Some of them are made of a nitinol frame and a polyester fabric. Antiaggregation is recommended temporarily after the procedure while waiting for the endotelisation to be completed, as is prophylaxis for endocarditis.
Several series have shown the technique for septal closure to be feasible and safe using different devices, with reported procedural success rates of over 95%.9,21,30,31 Published series have sometimes combined different occlusors and both ASD and PFO procedures. Among these devices are the Amplatzer ASD/PFO occluder (AGA Medical Corp., Golden Valley, MN, US), the STARFlex and the CardioSEAL Septal occluder (NMT Medical, Boston, MA, US), PFO STAR device and Intrasept device (Cardia, Eagan, MN, US), Helex Septal Occluder (WL Gore & Associates Inc., Flagstaff, Arizona, US), Premer (St Jude Medical Inc., St Paul, Minn, US), Solysafe septal occluder (Carag, Baar, Switzerland) and SeptRx (Cordis, Miami, Flo, US), etc.
Major procedural complications are currently rare, with a rate of 1.5% reported in a systematic review by Khairy et al.,22 and include cardiac tamponade, device embolisation, air embolisation, cerebrovascular events and haemorrhage, etc. Minor complications are more frequent, i.e. arrhythmias (especially atrial fibrillation) and vascular complications, with an incidence of 7.9% reported in this same study. Mid- and long-term complications include erosion of adjacent structures,32 late dislodgement and device thrombosis.27 Another important factor is the rate of incomplete defect closure. Residual shunt rates after PFO procedures are very variable and have been reported to depend on device design, ranging from 6 to 33% at six months.33
One of the longest and more recent series (using the Intrasept device) on 430 patients after a suspected paradoxical embolism showed no procedural deaths, 0.2% major procedural complications and 11.5% overall complications. The mid-term complication rate (median follow-up of 0.8 years) was 2.6%, basically due to atrial arrhythmias and one thrombus detected by echo in the left side of the device. Residual shunt was detected in 10.3% immediately after the procedure and in 13.2% at follow-up.25
Newer Strategies for Patent Foramen Ovale Closure
Despite the good procedural results obtained with these first-generation devices, there are some concerns regarding device-related complications. Erosion of adjacent structures and late cardiac tamponade,32 as well as device thrombosis and stroke,34 have been described. In addition, the existence of a permanent implant on the septum may preclude future transeptal access to the left atrium being necessary. Recent designs try to overcome these issues, either using bioabsorbable devices or avoiding permanent implants in the septum.
The BioSTAR device (NT Medical, Boston, Mass, US) has already obtained European and Canadian regulatory approval. Its design is similar to previous devices, with a double-umbrella structure mounted on a nitinol frame, but the occluders are made of a bioabsorbable material (a purified acellular porcine collagen layer) that is absorbed and replaced with healthy native tissue over time. This way, 90–95% of the implant is bioabsorbed and disappears at follow-up (24 months). In the BEST study, successful implantation was achieved in 98% of the patients, with no procedural complications related to the device or the delivery system.35
Radiofrequency applied to the atrial septum from the right atrium is able to weld the septum primum and secundum and close the PFO without a device being left in place. The initial clinical experience with the PFx closure system (Cierra Inc., Redwood City, CA, US) nevertheless showed a procedural success rate clearly less than that of current devices, with 43% of PFO closed after the first procedure and 63% after a second attempt.36 No serious procedure-related events occurred. Furthermore, PFO closure with this strategy implies scarring of the septum, and future consequences (especially regarding arrhythmias) need to be evaluated. The initial two systems under evaluation (Cierra Inc., Redwood City, CA, US, and CoAptus Medical Corporation, DCA, JEE, Redmond, WA, US) are no longer pursuing this avenue actively.
The most recent technique for PFO closure uses standard surgical polypropylene stitches to suture the septum; these stitches are deployed via a femoral access and a 12F sheath using the SuperStitch EL system (Sutura Inc., Fountain Valley, CA, US).37 It emulates surgical closure but without the handicaps of an open surgery, and avoids permanent implants other than the 4-0 sutures. This is a modified design from the Sutura SuperStitch, a CE-marked vascular suturing device (also available for use in the US in general surgery), which has been used extensively for vascular stitching in open surgery, minimally invasive surgery and fluoroscopy-guided procedures. It consists of a catheter with a distal metallic case that contains two needles and two suture-retaining arms. The needles pierce through the tissue, engaging the polypropylene sutures in the retaining arms. The sutures are then exteriorised and a suture knot (KwiKnot, Sutura) is advanced over them and released once on the sutured tissue.
Ex vivo and in vivo animal trials showed the feasibility of the technique, and the promising ‘first-in-man’ experience has already been published.37 Nevertheless, it is a new strategy that is clearly in its infancy, with fewer than 10 patients treated to date. A prospective trial is currently ongoing in our centre, in which two patients have already been included.
Percutaneous closure of PFO using the traditional principle of permanently implanting a double-umbrella structure device in the atrial septum to cover the PFO is a feasible and safe procedure, and several of these devices are currently in use for either PFO or ASD closure. Newer designs try to overcome the limitations of initial devices and are focused on avoiding permanent foreign implants in the atrial septum, and thus have the following theoretical important advantages over traditional left-in-place devices:
- avoiding device-related complications, such as device embolisation, thrombosis or erosion of adjacent structures;
- allowing for future transeptal access to the left atrium if eventually needed; and
- avoiding the need for antiplatelet treatments.
However, despite the favourable procedural results, evidence derived from randomised trials to support the benefits of PFO closure over pharmacological regimens is still lacking. The completion of ongoing studies is essential to defining the real potential of PFO closure strategies in treating these patients, to the point at which no device for PFO closure is commercially available in the US, as the US Food and Drug Administration (FDA) considers that the ‘proof of principle’ (that PFO closure reduces recurrence of events) cannot be demonstrated in this setting unless supported by randomised studies.38 Inclusion into ongoing randomised trials is slow, and the biggest impediment to patient enrolment is the extended off-label use of devices for PFO closure, as reported by the FDA Circulatory System Devices Panel (CSDP).39
The principle of percutaneously suturing the septum with an extendedly used surgical suture is especially appealing. When we can close the defect without leaving a foreign body in place (other than a suture with well-known long-term safety and biocompatibility in humans) and avoid septal scarring, we will have overcome the long-term complications of the procedure. When we can achieve this without significant procedural complications it will become a nearly harmless intervention that will clearly introduce a paradigm shift in the treatment of patients with PFO.
Would not PFO closure be an acceptable option to prevent major events (such as stroke), even if its benefit over long-term pharmacological regimens is not clearly established by randomised trials, if it can be achieved with an intervention that is practically free from acute and long-term complications? Only time will tell.
Uncertainty remains regarding the real role of PFO in the genesis of cryptogenic stroke and the best therapeutic option for these patients, mainly due to the lack of randomised trials comparing pharmacological regimens with PFO closure strategies. Nevertheless, even though the available data (derived from retrospective studies) have to be interpreted cautiously due to significant baseline differences among treatment groups, they consistently suggest a benefit of percutaneous PFO closure over medical treatment in relation to stroke recurrence in these patients. Initial percutaneous approaches for PFO closure have the handicap of leaving a permanent implant in the atrial septum, which may be the substrate for long-term complications and limit future transeptal access to the left atrium. Second-generation percutaneous designs avoid permanent implants and include bioabsorbable devices, radiofrequency PFO closure through the induction of septal scarring and closure of the PFO with stitches, emulating open surgery interventions. Despite these new developments, it is important to emphasise that, to date, no randomised trial has shown a benefit of PFO closure over traditional pharmacological strategies.