The morphology and function of the inter-atrial septum changes dramatically from the period of in utero development until its role in normal adult cardiac physiology is established. Its anatomical construction is complex and involves the eventual formation of a septum secundum, fossa ovalis and septum primum. In utero a mixture of deoxygenated blood from the foetus and oxygenated blood from the placenta are deflected by the Eustacian valve through the fossa ovalis to provide the blood supply to the upper body, brain and coronary arteries of the developing embryo. At birth, blood is rapidly directed to the lungs, resulting in a rise in left atrial volume and pressure that pushes the the two septae together, thereby sealing the fossa ovalis. These tissue planes eventually fuse to form a non-perforate inter-atrial septum that defines the separate right- and left-heart circulations first described by William Harvey in 1628.1 Defects in this multifaceted process, still poorly understood, underlie the development of both atrial septal defects (ASDs) and patent foramen ovale (PFO).
Patent Foramen Ovale
PFO occurs due to the failure of complete fusion of the septum secundum and primum. Right-to-left and left-to-right shunting may occur across the unfused septum depending on transient differences in pressure and compliance between left and right heart chambers. A PFO has been identified at autopsy in 27% of individuals with otherwise normal hearts.2 Alternatively, using transoesophageal echocardiography a PFO was diagnosed in 24.3% of 585 randomly sampled people over 45 years of age in Olmstead County, Minnesota.3 PFO has been implicated in a wide variety of clinical syndromes, with the common pathophysiological theme of transient right-to-left intercardiac shunting. These include cryptogenic stroke, migraine headache (especially with an aura), platypnea-orthodeoxia syndrome, high-altitude pulmonary oedema and brain abscess, among others.4,5
In all likelihood most people with a PFO will never experience an adverse health event related to this remnant of the foetal circulation. Indeed, a recent article examining surgical closure of PFO when found incidentally at heart surgery for unrelated conditions suggested that closure of the PFO under such circumstances may cause more harm than good.6 However, when evaluating a cohort of patients with a specific pathological condition thought to be causally related to PFO, observational studies almost always report a higher prevalence of PFO in patients with the clinical condition in question than in those without.7 Occasionally, individual patients demonstrate the pathophysiological substrate necessary to establish a solid aetiological connection such as thrombus traversing the PFO, but more commonly the inferential link is more obtuse.8 This underscores the importance of results from prospective randomised trials to guide therapy in the vast majority of patients. Many of the syndromes associated with PFO are uncommon and it is extremely unlikely that such scientifically valid trials will ever be conducted, but hopefully the condition of cryptogenic stroke will prove the exception to this rule.
Currently, there are five ongoing randomised trials of percutaneous PFO device closure versus medical therapy (anticoagulation) in patients with first-time cryptogenic stoke, together enrolling almost 3,500 patients9 (see Table 1). Of these trials, only the ClOSURE 1 trial, comparing the STARflex® Septal Closure System versus best anticoagulation as determined by a neurologist, has completed enrolment. Nine hundred patients were randomised (1:1) and results will hopefully be presented in spring 2010. In the meantime practitioners are encouraged to refer eligible patients with first-time cryptogenic stroke to appropriate centres to enable completion of additional trials. All non-trial-related PFO closure in the US is performed off-label using devices approved for other indications, but such a designation has not prevented a 50-fold increase in the weighted national estimate of PFO/ASD closure from 1998 to 2004.10
Individual patients nevertheless demand careful attention and they may be ineligible for randomised trials or refuse to be enrolled. The best treatment for these patients often remains unclear, but some general principles can be outlined. Non-randomised trials comparing device closure of PFO versus medical therapy suggest a lower rate of recurrence after device placement (0–4.9 versus 3.8–12%), but recurrent events persist after device closure and may be related to other aetiologies (e.g. pulmonary AV malformations, aortic arch atherosclerosis) incomplete closure of the PFO or device thrombosis.11
Patients with multiple defects on neurological imaging, PFOs associated with atrial septal aneurysms or particularly large right-to-left shunts may benefit from device closure, although the data from numerous observational studies remain controversial in this regard.12 Recurrent events despite anticoagulation, in the absence of any other compelling reason for stroke, remain the strongest indication for device closure.13 In Europe, some investigators have taken a more liberal attitude towards PFO closure, citing the relatively low risk of the procedure in experienced centres and even comparing the technique to vaccination against future neurological events.14
As can be seen there is a wide variety of opinion concerning indications for PFO closure in adults. The debate might be less intense if the procedure were risk-free. Currently, an ideal PFO closure device does not exist. Such a device would be completely safe and 100% effective and leave behind no residual material, obviating the need for long-term antiplatelet therapy. Double-disc devices, which sandwich the septum between discs made of combinations of polyester fabric over a metal frame deployed in the left and right atrium, account for the vast majority of PFO closure in the US, and include the Amplatzer Cribiform ASD/PFO occluder (AGA Medical Corp., Golden Valley, MN, US), the CardioSeal and STARFlex Septal occluder (NMT Medical, Boston, MA, US) and the Helex Septal Occluder (WL Gore & Associates Inc., Flagstaff, AZ, US). A number of other devices using this same principle are available in other countries.15 A variety of complications have been reported with these device implantations and include stroke, cardiac tamponade, air embolism, device embolisation, transient arrhythmias, device erosion and impingement on other cardiac structures. In general, complication rates are not high with these devices and range from 1.5 to 8% depending on the study.11 Incomplete closure rates also vary depending on detection methods and the device utilised, although it is important to recognise that the relationship of incomplete closure to recurrent events remains unclear. In the FORCAST Registry, patients receiving a CardioSEAL or STARFlex device (n=272) had a 74.2% complete closure rate, with an additional 21% of patients having only trivial shunts.16 A recently published single-centre study using two iterations of the Cardia and Intrasept occluder (Cardia Inc., MN, US) and the Amplatzer PFO occluder (n=795) showed a 100% procedural success with an overall complication rate of 1.8%. Residual shunting was present acutely in 14–24% depending on the device, but this decreased to 8% at mean follow-up of 26 months.17
A number of new approaches to septal closure are being investigated. A common theme is to reduce device profile and leave minimal material behind. Attempts to weld the septum primum and secundum together using radio-frequency energy (Cierra Inc., Redwood City, CA, US and CoAptus Medical Corporation, Redmond, WA, US) appeared to be a relatively safe technique, but low complete closure rates have caused both companies to abandon this approach.18 Similarly, another approach is to attempt percutaneous suturing of the septum. A specially designed system (SuperStitch EL, Sutura, Inc., Fountain Valley, CA, US) using two needles and two suture retaining arms allows the septal sutures to be exteriorised and then synched tight with a knot advanced to the septal surface. An early experience with 20 patients in a single centre showed successful deployment in 65% with complete closure in 45% of cases where the suture was deployed.19 While these closure rates are not high, it represents an early experience and failure can be rectified with placement of a more traditional device. The BioSTAR PFO occulder (NMT Medical, Boston, MA, US), which is available outside the US, employs a bio-absorbable disc made of pure alpha-type collagen that disappears over a 24-month period, at which point it is replaced by native tissue. This methodology eventually leaves behind only a minimal scaffold of thin metal struts and the results of the BEST trial (n=58) demonstrated a 96% complete closure rate at six months with no major complications or adverse tissue reactions.20
Finally, the COHEREX FlatStent (Coherex Medical, Salt Lake City, UT, US) is designed to rest mostly inside the PFO tunnel itself, closing the defect from within, thereby leaving minimal device profile in the left atrium. It is retrievable and repositionable with an initial study showing 100% successful deployment and 94% complete closure rate at 180 days in an early clinical cohort of 49 patients.21
Atrial Septal Defect
The intactness of the atrial septum depends on growth of the septum secundum and primum and proper fusion of the endocardial cushion, which also fashions the superior portion of the ventricular septum and septal leaflets of the mitral and tricuspid valves. ASDs, in general, are the result of abnormalities of this process. Secundum ASDs occur from either too much absorption of the septum primum or insufficient growth of the septum secundum.
At present only secundum ASDs are amenable to percutaneous closure and the following remarks are confined to this type of defect. An ASD causes left-to-right shunting at the atrial level, which results in increased volume flow in the right ventricle and pulmonary artery compared with the systemic circulation. This can cause reactive pulmonary arteriolar hyperplasia and fibrosis which may result in Eisenmenger’s syndrome, or much more commonly in adults, who have avoided this relatively rare complication, progressive right heart dilatation and complaints of dyspnoea on exertion or arrhythmias. Increased pulmonary blood flow pre-disposes patients to infection and the only randomised trial of ASD closure (surgical) versus conservative medical therapy in adults showed a benefit of ASD closure mostly driven by a reduction in hospitalisation for pneumonia.22 After successful defect closure, however, the majority of patients experience symptomatic improvement with improved exercise tolerance and a reduction in arrhythmias. A decrease in right-heart chamber dimensions may take place (negative remodelling), and some studies have also suggested improved long-term survival.23 Most experts believe ASD closure should be undertaken in symptomatic individuals, those with evidence of right-heart enlargement, left-to-right shunts greater than 1.5/1.0 or episodes of intercardiac embolisation. Although surgical closure is usually well-tolerated and safe, studies comparing percutaneous closure to surgery generally show shorter lengths of stay, fewer complications, lower costs (US$11,000 versus US$21,000) and equal effectiveness to device closure.24,26
The first report of percutaneous closure of an ASD was by King et al., more than 30 years ago.25 Smaller ASDs can be closed with a variety of devices also used to close PFOs, but ASDs greater than 15mm are almost universally sealed with the Amplatzer ASD occluder (AGA Medical, Golden Valley, MN, US). The device consists of two nitinol discs embedded with Dacron cloth joined together by a central waist that is sized to the ASD diameter. Secundum ASDs are not usually located in the exact middle of the atrial septum, but rather displaced away from the centre, which sometimes complicates percutaneous closure. A deficiency of septal tissue rim (<5mm), especially when located inferiorly towards the mitral valve, often precludes successful device closure. Near the aortic valve, a degree of superior rim deficiency can often be overcome by anchoring the device against the aortic wall, although care must be taken not to oversize the occluder in these circumstances since several cases of erosion through the aorta or roof of the right atrium have been reported, often with catastrophic results.27 Large defects (>38mm), those with a higher percentage of circumferential rim deficiency or those associated with other congential defects (anomalous pulmonary venous return) are best referred for surgical repair.
Closure of the ASD results in an increased volume of blood delivered to the left ventricle since left-to-right shunting at the atrial level is suddenly eliminated. In older adults this occasionally unmasks restrictive left ventricular diastolic dysfunction and it is important that interventional operators be aware of this possible haemodynamic response. A rapid increase in volume delivered to a poorly compliant left-ventricular chamber produces a swift rise in left-atrial pressure, pulmonary venous engorgement and occasionally pulmonary oedema. Ewert et al. evaluated a cohort of consecutive patients over 60 years of age referred for ASD closure, finding seven of 18 patients who responded to balloon -occlusion of the ASD with a rapid rise in left atrial pressure.28 In two of these patients ASD closure was performed with one patient developing worse dyspnoea on exertion afterwards, while the other went into pulmonary oedema requiring intubation and a large dose of diuretics at discharge two weeks later. There were no obvious differences in clinical characteristics between patients who exhibited this response and those who did not, except for a higher baseline mean left-atrial pressure before (7 versus 18mmHg) and after (7 versus 26mmHg) balloon occlusion, which are important to measure prior to attempted closure in any adult patient over 60 years of age or with suspected diastolic dysfunction. Some experts have argued it is possible to aggressively treat such patients with anti-hypertensives and diuretics to prepare them for closure, but there are no reported series to date testing the outcome of this strategy. Another approach to this dilemma is to fenestrate the ASD occluder with a small hole that allows some blood to ‘pop off’ into the right atrium while reducing the overall magnitude of the left-to-right shunt although, again, only isolated cases have been reported.29
Another controversial group of adult patients who present for closure are patients with ASD and pulmonary hypertension. If the pulmonary hypertension is largely reactive and driven by high-volume flow, closing the ASD will ameliorate the pulmonary hypertension over the long-term. If, however, as some experts believe, pulmonary hypertension is driven by progressive, irreversible fibrotic obliteration of the pulmonary arteriolar bed, then closing the ASD will hasten the pace of right ventricular failure by removing the right-to-left ‘pop off’ function of the ASD. There is no clear resolution of this conundrum. In a recent analysis of 54 patients with moderate or severe pulmonary hypertension, only 44% normalised their pulmonary pressures after ASD closure and in the absence of long-term follow-up it is not possible to know whether the pulmonary hypertension will steadily progress.3 These patients may benefit from an assessment of the response of the pulmonary vasculature to vasodilators prior to attempted closure since the more reversible the pulmonary hypertension, the less likely the vascular pathology is fixed.
Percutaneous ASD closure is, overall, a safe procedure with major complications, including device embolisation, thrombosis, cardiac perforation, stroke or vascular injury reported to be around 1–2% in recent series’.31 Device embolisation is the most common major complication, which given the complexity of the anatomy of many ASDs, is not surprising and happens to even very skilled practitioners. It is imperative, therefore, that interventionalists who perform ASD closure be proficient in the techniques of device retrieval using a combination of snares and large-bore catheters. Imaging the defect as accurately as possible to determine proper device sizing thus becomes critical to the success of the procedure, since undersizing predisposes to embolisation while oversizing has been associated with erosion of the device, which can cause tamponade and death. A variety of techniques exist to accomplish accurate defect sizing and include measuring the waist of a balloon as it is indented by the walls of the defect or the balloon diameter at which Doppler flow, determined by echocardiography, ceases. Many operators prefer intracardiac echocardiography using an 8 or 11Fr probe inserted into the right atrium for visualisation, but we prefer transoesophageal echocardiography as it enables a wider field of view in adults and allows 3D interrogation of the defect. We have found 3D imaging to be extremely useful since it can display the entire septum en face, allowing the diameter and shape of the defect to be clearly seen (see Figure 1). The ability to visualise the entire outline of the defect also facilitates interrogation of the surrounding rim and relationship of the defect to other cardiac structures. In addition, fenestrated ASDs can be readily identified, along with unusual anatomic variations, as was the case recently when we performed a closure where the defect was significantly undersized on 2D-echo because a flimsy tissue remnant was mistaken for the border of the ASD.32
Conclusions
Defects of the inter-atrial septum represent a varied and fascinating anatomical landscape for the interventionalist. PFOs, which were once felt to be a curiosity for pathologists, are now appreciated to be contributing to a number of clinical conditions whose aetiology had, before now, seemed a mystery. Indications for PFO closure in many cases are still unclear and we await the results of randomised controlled trials to provide further clarification, especially for a first-time presentation of cryptogenic stroke. Newer devices are in development, which we hope will leave a smaller residual device profile. Secundum ASD closure has, for the most part, left the realm of the surgeon for that of the interventionalist, but nonetheless certain secundum defects remain risky to close, exposing patients to risks of embolisation or device erosion. Newer imaging techniques will hopefully enable the identification with greater certainty those patients better served by referral for traditional surgical repair, while further improvements in closure-device design will continue to enhance the efficacy and safety of the procedure. Since its initial description 400 years ago by William Harvey as one of the heart’s solid borders separating the right and left circulations, the inter-atrial septum has indeed come a long way.