Patent Foramen Ovale
A potential causal relationship between patent foramen ovale (PFO) and stroke was first described by Cohnheim in 1877. In the last two decades several studies have investigated the role of PFO in cryptogenic ischemic stroke, migraine headaches, platypnea-orthodeoxia, and decompression sickness. Percutaneous PFO closure has emerged as a treatment option in the last decade, with significant controversy around its indications. Results of randomized trials of transcatheter PFO closure have only been recently reported.
Developmental Anatomy of the Atrial Septum ( Figure 32-1 )
During fetal life, there is a single atrial cavity. A septum primum (SP) develops from the cranial wall of the single atrium and grows toward the endocardial cushions, thereby dividing the single atrium into left- and right-sided chambers. The area between the SP and the endocardial cushions is known as ostium primum (OP). Fenestrations then develop in the middle of the septum primum and coalesce to form ostium secundum (OS). The OS allows right-to-left shunting of oxygenated blood. To the right of septum primum, another septum known as septum secundum (SS) then develops, and covers the OS and in most instances covers the OP as well. A flaplike valve known as PFO is formed between the two septae, which now allows oxygenated placental blood to cross over from right-to-left atrium during the remainder of intrauterine life. Spontaneous fusion of SP with SS occurs in about 75% of the individuals by 2 years of age, leading to closure of PFO. In the remaining individuals there is an oblique crescent-shaped defect resembling a tunnel, which is called PFO. The prevalence of probe-patent PFO is about 27% in necropsy studies with decreasing prevalence for each decade of life.
Paradoxical embolism of a thrombus originating in pelvic or lower extremity veins has been implicated as a mechanism to explain association between PFO and ischemic stroke. The incidence of pelvic deep vein thrombosis (DVT) has been found to be significantly higher in patients with cryptogenic stroke and PFO compared with patients with stroke of determined origin. Some PFOs have a long overlap between the SP and SS referred to as the tunnel. There are several reports of thrombus visualized “in transit” through PFO on transthoracic (TTE) as well as transesophageal echocardiography (TEE). This has also led to speculation that stagnated blood in the tunnel may lead to thrombus formation, which may subsequently embolize in the systemic circulation—an explanation often referred to as the “lurking clot theory.” Studies have shown an association between the presence of atrial septal aneurysm (ASA) and risk of stroke.
Platypnea-orthodeoxia is a rare clinical syndrome characterized by dyspnea and oxygen desaturation in the upright position that is relieved by lying down or recumbence. In the absence of elevated right-sided pressures, or lung disease, this can occur due to an anatomical abnormality that predisposes to right-to-left shunt from a PFO, such as a prominent eustachian valve that directs blood from the inferior vena cava (IVC) toward the interatrial septum, an aortic aneurysm, or an enlarged or elongated and horizontal aortic root that distorts the interatrial septum thereby predisposing to right-to-left shunting during upright position. TTE or TEE in the sitting position may be required to demonstrate flow across the septum by color or contrast if negative in supine position. Cardiac magnetic resonance imaging (MRI) or computed tomography (CT) may help demonstrate aortic abnormalities. Transcatheter PFO closure has been shown to be associated with marked improvement in symptoms.
Decompression sickness occurs when a diver ascends from a dive and nitrogen bubbles entering venous circulation, which usually get diffused in the lungs, enter the systemic circulation via a right-to-left shunting source such as PFO, and embolize to the brain leading to ischemic lesions. A recent prospective study of 104 scuba divers with history of major decompression sickness showed that transcatheter PFO closure appears to prevent symptomatic and asymptomatic (ischemic brain lesions on MRI) decompression sickness.
Migraine is a common disorder affecting about 10% of the adult population and is more common in women. In the last two decades, studies have shown an association between PFO and migraine, especially migraine with aura. Several retrospective observational studies have reported an improvement in migraine after PFO closure. In contrast, the only completed prospective randomized double blind trial (Migraine Intervention with STARFLEX Technology—MIST trial) in patients undergoing PFO closure primarily for migraine control failed to show any significant difference of PFO closure on the primary endpoint of migraine cessation, or secondary endpoints of improvement in migraine compared with a sham procedure. However, the MIST trial was found to have several limitations including unrealistic endpoint of migraine cessation, inadequacy of TTE in screening for PFO as indicated by absence of PFO during closure, and shorter duration of follow-up. This implies that there may be some patients who benefit from device closure who remain to be identified. Recently, a significant reduction in frequency and severity of migraine was demonstrated with PFO closure in patients with large PFO (based on TCD) and subclinical brain MRI lesions. These brain lesions may indicate silent thromboembolism and these patients may be high risk for future embolic events. Similarly, Rigatelli and colleagues recently showed that PFO closure resulted in significant reduction in migraine in patients with high-risk PFO characteristics such as curtain shunt pattern on TCD and TEE (implying larger degrees of shunting), right-to-left shunting during normal respiration, ASA, and presence of eustachian valve. Not all patients with migraine have a PFO and not all patients with PFO suffer from migraine. The onus is on future trials to identify patients who would benefit most from PFO closure based on high-risk PFO morphology, which is best assessed with TEE, and possibly also those with subclinical lesions on brain imaging.
PFO can be detected using various echocardiographic techniques, including TTE, TEE, and transcranial Doppler (TCD). More recently, three-dimensional echocardiography (3DE), CT, and MRI have been used, although none as the primary diagnostic tool in routine practice. Agitated saline is commonly used for diagnosing right-to-left shunts. Although the definition of positive contrast study on TTE or TEE remains controversial, it is generally accepted that a right-to-left shunt is diagnosed if at least 3 micro-bubbles appear in the left atrium, either spontaneously or after provocative maneuvers such as cough or Valsalva, within 3 cycles of complete opacification of the right atrium ( Figure 32-2 ). A provocative maneuver increases right atrial filling, thus increasing the RA pressure and opening the foramen ovale. Valsalva maneuver can be calibrated (40 mm Hg strain measured by spirometry and sustained for 10 seconds). A good Valsalva maneuver is sometimes more difficult to obtain during TEE, especially if the patient is heavily sedated, as compared with TTE or TCD. Some studies have shown that sensitivity of detection of PFO was increased when a femoral vein was used for contrast injection instead of the antecubital vein. This is likely due to different inflow pattern into the RA after injection through the femoral vein. Contrast through the inferior vena cava is directed toward the interatrial septum, often potentiated by a eustachian valve ( Figure 32-2 ), whereas contrast through the superior vena cava is directed toward the tricuspid valve. Different morphological characteristics of PFO such as size, degree of shunting, and tunnel length should be taken into account when evaluating a patient with PFO and cryptogenic stroke ( Figure 32-2 ). TCD of the middle cerebral artery after injection of contrast can similarly be used to diagnose right-to-left shunting.
TEE has been shown to correlate very well with autopsy findings, with a sensitivity and specificity approaching 100% in the diagnosis of PFO. Due to its high sensitivity and greater image resolution of the interatrial septal area allowing detailed characterization of PFO morphology, TEE is the current gold standard to diagnose and characterize PFO ( Figure 32-2 ). The drawbacks of TEE are its semi-invasiveness and occasional inability to obtain a good Valsalva maneuver in sedated patients.
The 3DE using reconstruction techniques as well as real-time analysis have been used to evaluate a wide range of pathologies including patent foramen ovale. In a recent comparison, diagnostic accuracy of real-time 3D TTE was significantly higher than that of contrast TTE: sensitivity 83% versus 44% (p <0.001) and close to that of contrast TEE.
Small studies using contrast-enhanced MRI and cardiac CT ( Figure 32-2 ) showed good concordance with TEE in the diagnosis of PFO. However, larger studies have shown both modalities to be inferior compared with TEE in detecting PFO.
Currently the best therapeutic modality for primary or secondary prevention of stroke in patients with PFO is debatable. There are data to suggest that PFO is more common in patients with cryptogenic stroke compared with those with a known cause of ischemic stroke. However, a PFO is fairly common in the “control” population without stroke (prevalence of about 25%), and there are many unknown causes of “cryptogenic stroke.” The available options include antiplatelet therapy, anticoagulant therapy with warfarin, transcatheter closure, and surgical closure. It is very likely that not all PFOs are “culprits” responsible for PTE, especially due to the high prevalence of PFO in the general population. Supporting this, a recent meta-analysis showed that one third of detected PFOs in patients with cryptogenic stroke are likely to be incidental and not benefit from closure, suggesting the importance of patient selection in therapeutic decision making.
At present, there is no consensus regarding antiplatelet versus anticoagulant therapy in patients with cryptogenic stroke and PFO, as reflected by the heterogeneity in the medical arms of the published randomized PFO closure trials. Data from the Warfarin-Aspirin Recurrent Stroke Study (WARSS) show that there was no difference between treatment with warfarin or aspirin in the prevention of recurrent ischemic stroke or death in a large cohort of patients with cryptogenic stroke. The Patent Foramen Ovale in Cryptogenic Stroke Study (PICSS), which is a substudy of WARSS with patients undergoing TEE examination, showed that there was a nonsignificant trend toward lower 2-year risk of stroke or death among warfarin-treated cryptogenic stroke patients with PFO compared with those receiving antiplatelet treatment (9.5% vs. 17.9%; hazard ratio [HR] 0.52; confidence interval [CI], 0.16-1.67). In the PFO-ASA study, consisting of over 580 patients with ischemic stroke of unknown origin, recurrent stroke occurred more commonly despite aspirin therapy in patients with PFO and ASA compared with those with PFO, or ASA alone (HR for combination of PFO and ASA, 4.17; 95% CI, 1.47-11.84), suggesting that preventive strategies in addition to aspirin may be needed for such patients with high-risk PFO anatomy. A recent meta-analysis of retrospective studies also suggests benefit of anticoagulation over antiplatelet therapy for prevention of recurrent neurologic events in patients with PFO and cryptogenic stroke.
Transcatheter Patent Foramen Ovale Closure
Retrospective studies and meta-analyses have shown potential benefit of PFO closure in patients with cryptogenic stroke. However, the completed prospective randomized trials failed to show such benefit as detailed below. This suggests that the real-world selection of high-risk patients where PFO-related PTE is the cause of stroke may potentially be a beneficial approach. Of course, results from retrospective studies, meta-analyses and randomized trials, and their limitations must be discussed with patients. Pending further studies, informed individualized therapeutic decisions should be made depending on patient preferences and perceived risk of PFO and recurrent stroke.
Randomized Trials of Patent Foramen Ovale Closure for Cryptogenic Stroke
In the first randomized report of transcatheter PFO closure, the CLOSURE 1 (Evaluation of the STARFlex septal closure system in patients with stroke and/or transient ischemic attack due to presumed paradoxical embolism through a PFO) trial, 909 patients with cryptogenic stroke or TIA were randomized to medical therapy or transcatheter PFO closure using the STARFlex device. With a success rate of 89% for closure, there was no difference in the outcomes of recurrent stroke (2.9% vs. 3.1%; p = 0.79) or TIA (3.1% vs. 4.1%; p = 0.44) with closure compared with medical therapy. There were many limitations of the CLOSURE 1 trial. The majority of recurrent events in this study (20 of 23 patients in closure group and 22 of 29 patients in medical therapy group) were not related to PTE, and alternative explanations for recurrent neurologic events were observed, including atrial fibrillation, subcortical lacunar infarcts, aortic arch atheroma, complex migraine, vasculitis, etc. This increases the likelihood that the initial neurologic event may not have been related to PFO and PTE. Thus, patients in the CLOSURE 1 trial may not have been the ideal population to study PFO closure. Only a third of patients in this trial had high-risk features such as ASA, and only about a half had significant shunting. Closure of insignificant or incidental PFOs may have diluted the beneficial effects of PFO closure. Patients with hypercoagulable testing or DVT were excluded from this study, thus excluding patients in whom the mechanism of stroke was perhaps most likely to be related to PTE. Moreover, even though the CLOSURE trial is being viewed as a “negative” trial, it shows that PFO closure is an effective alternative to medical therapy in reducing stroke.
In RESPECT (Randomized evaluation of recurrent stroke comparing PFO closure to established current standard of care treatment) trial, 980 patients with cryptogenic stroke were randomized to medical therapy or transcatheter PFO closure. In the intention-to-treat cohort, recurrent stroke occurred in 9 patients in the closure group and 16 in the medical therapy group (HR with closure 0.49; 95% CI, 0.22-1.11; p = 0.08). In contrast, there was statistically significant reduction in the risk of recurrent stroke with PFO closure when analyses were performed in prespecified per-protocol cohort (HR 0.37; 95% CI, 0.14-0.96, p = 0.03), and as-treated cohort (HR 0.27; 95% CI, 0.10-0.75, p = 0.007). In addition, closure was found to provide greater benefit in patients with severe right-to-left shunt and in those with an atrial septal aneurysm. Strengths of the RESPECT trial over CLOSURE 1 trial include longer follow-up, more stringent inclusion criteria with exclusion of patients with TIA and lacunar infarcts, and use of Amplatzer PFO occlude device, which provides more effective closure rates with much less device-related complications such as thrombosis and atrial fibrillation. Limitations of the RESPECT trial include high drop-out rate (17% in medical therapy group and 9% in closure group) and nonadherence to protocol in some patients with important implications on outcomes (3 out of 9 patients with recurrent ischemic stroke in the closure group of the intention-to-treat population did not have a device at the time of recurrent stroke).
In the PC (Comparing Percutaneous closure of PFO using the Amplatzer PFO Occluder with medical treatment in patients with cryptogenic embolism) trial, 414 patients were randomized to transcatheter PFO closure or medical therapy. Recurrent stroke occurred less frequently in the closure group compared with the medical therapy group; however, this was not statistically significant (0.5% vs. 2.4%; HR, 0.20; 95% CI, 0.02-1.72; p = 0.14). Closure also did not reduce recurrent TIAs compared with medical therapy alone (2.5% vs. 3.3%; HR, 0.71; 95% CI, 0.23-2.24; p = 0.56). Limitations of the PC trial include inclusion of TIA in the primary endpoint and difficulty recruiting patients with a long recruitment period.
Despite lack of benefit in reduction of recurrent neurologic events with PFO closure compared with medical therapy in randomized trials, there are signals pointing toward benefit with closure, particularly in a select group of patients at high risk, such as those with ASA and large shunting. Whenever possible, patients with cryptogenic neurologic events and PFO must be enrolled in ongoing randomized trials such as Patent Foramen Ovale Closure or Anticoagulation versus Antiplatelet Therapy to Prevent Stroke Recurrence (CLOSE, ClinicalTrials.gov number, NCT00562289), Device Closure versus Medical Therapy for Cryptogenic Stroke Patients with High-Risk Patent Foramen Ovale (DEFENSE-PFO, NCT01550588), and Gore Helex Septal Occluder/Gore Septal Occluder for Patent Foramen Ovale (PFO) Closure in Stroke Patients (REDUCE, NCT00738894).
Indications for Transcatheter Patent Foramen Ovale Closure
In the United States, transcatheter PFO closure is not Food and Drug Administration (FDA) approved. As discussed, this procedure is still controversial, given discordance in retrospective and randomized trial data. Off-label PFO closure has been performed in patients with cryptogenic stroke, other paradoxical embolic events presumed to be related to PFO, platypnea-orthodeoxia syndrome, decompression sickness, and migraine headaches. The author found no difference in the risk of stroke in patients with PFO and an implantable intracardiac device such as pacemaker or defibrillator compared with those without an intracardiac device. Another study found an increased risk of stroke/TIA in patients with implantable devices with PFO compared with those without PFO. The authors study was different from this study in that the patient population consisted of PFO patients only with the goal of studying whether device implantation had any impact on the outcome of stroke in patients with PFO. Additionally, the latter study included patients with prior stroke/TIA while the authors did not, as prior stroke itself is an important predictor of future stroke. The authors have also found no difference in stroke risk between patients with atrial fibrillation with and without PFO. As such, PFO closure cannot be recommended in patients with pacemakers, implantable defibrillators, or atrial fibrillation.
Transcatheter PFO closure has been performed off-label with devices that are used for transcatheter atrial septal defect (ASD) closure ( Figure 32-3 ). These include the Helex septal occlude (WL Gore, Flagstaff, Arizona), Amplatzer atrial septal occluder (ASO) (St. Jude Medical), Amplatzer multifenestrated, or cribriform ASO. In addition, Amplatzer PFO occluder, CardioSEAL, STARFlex, and Premere PFO closure system have also been used. Currently only the Amplatzer and Helex systems are used in the United States.
The Amplatzer PFO occluder, used in the RESPECT and PC trials, is a self-expanding, double-disk device made of 0.005-inch nitinol wire and polyester patches sewn within each disk to occlude blood flow ( Figure 32-3 ). The waist is thin and mobile and the right atrial disk is larger than the left atrial disk as opposed to the Amplatzer ASO device. There are 3 device sizes available, based on the right atrial disk diameter—18, 25, and 35 mm. Device sizing depends on distance from the PFO to the SVC or aorta. The 25-mm device is used in the vast majority of cases. The Amplatzer cribriform ASO device, used for closure of fenestrated secundum ASDs, also has been used for PFO closure. It consists of a thin waist and equal-sized left and right atrial disks and is available in 4 sizes—18, 25, 30, and 35 mm. The Amplatzer ASO device is discussed in the ASD closure section.
The Helex device is a nonself-centering double disk device composed of single nitinol wire covered with polytetrafluoroethylene (PTFE) with a left atrial eyelet, center eyelet, and right atrial eyelet ( Figure 32-3 ). This device is FDA approved for closure of secundum ASDs >18 mm in diameter. The device is available in 5-mm increments, from 15 to 35 mm. The gray catheter attaches to the right atrial eyelet and is used to retract or extrude the device. The mandrel attaches to the left atrial eyelet and contains the locking loop; pulling the mandrel releases the device and locks it in place.
Transcatheter PFO closure is performed in the cardiac catheterization laboratory under conscious sedation (we use midazolam and fentanyl) with fluoroscopic and ultrasound guidance (TEE, or now usually intracardiac echocardiography [ICE]) ( Figures 32-4 and 32-5 ) ( and ). Aspirin 325 mg is usually administered before the procedure and Clopidogrel 600 mg loading dose at the end of the procedure. Femoral venous access is obtained in bilateral groins with an 8 Fr or 9 Fr sheath each (or both sheaths in the same vein), one of which is for ICE. We prefer a long 30-cm sheath for the ICE catheter to easily traverse the iliac vein into the inferior vena cava, particularly if inserted into the left vein. The ICE catheter is advanced into the right atrium and the interatrial septum adequately interrogated, and bubble study is performed through the contralateral femoral venous sheath. A Goodale-Lubin (GL) catheter is advanced with a 0.035-inch J-tipped guidewire into the SVC. The guidewire is removed and the GL catheter connected to the manifold. Right atrial angiography can then be performed if needed. The GL catheter is then directed toward the interatrial septum and the PFO crossed with or without the 0.035-inch J-tipped guidewire using ICE and fluoroscopic guidance. Once across the PFO, intravenous heparin is administered in order to achieve ACT >250 seconds. The catheter and guidewire are placed in the left superior pulmonary vein, taking care to ensure that the wire tip is not in the left atrial appendage to avoid perforation. The 0.035-inch J-tipped guidewire is exchanged for a 0.035-inch J-tipped Amplatz extra-stiff wire, again taking care to ensure that the tip is not in the appendage. PFO diameter is then measured with a sizing balloon, taking care to inflate the balloon gently to avoid tearing the interatrial septum ( Figures 32-4 and 32-5 ). The next steps depend on the device used.
For Helex device, the system is prepped and flushed as recommended by the manufacturer. A device to balloon-stretched diameter of at least 2 : 1 is recommended. A 9 Fr sheath is used without guidewire, or an 11 Fr sheath with guidewire. After initial prepping, the green delivery catheter is placed in the left atrium over the 0.035-inch extra-stiff guidewire. The guidewire is removed and the left atrial disk is deployed using the “push-pinch-pull” technique under fluoroscopic and ICE guidance to ensure positioning in the left atrium away from the roof and appendage. The entire system is then pulled against the left side of the interatrial septum. The right atrial disk is then deployed. Placement and positioning is confirmed with ICE, and left anterior oblique (LAO) projection on fluoroscopy, with the right and left atrial disks straddling the septum ( Figures 32-4 and 32-5 ). Once acceptable placement is confirmed, the mandrel is pulled, which moves the locking loop off the left atrial eyelet to around the right atrial eyelet. The device can be retrieved and redeployed at any point prior to lock release. It can also be retrieved from the body after lock release if the position is not favorable. Once correct positioning is confirmed, the device is released.
For Amplatzer devices, the initial steps are similar to the Helex device. The device is loaded on the delivery cable and prepped as per the manufacturer’s instructions to ensure no air in the system. The device is then introduced from the loader into the delivery sheath, which is placed in the mid-left atrium, and carefully pushed under fluoroscopy to ensure absence of air bubbles. Once the device reaches the tip of the delivery sheath, the sheath is withdrawn gently, exposing the left atrial disk, under fluoroscopic and ICE guidance. After making sure there is adequate opening of the left atrial disk in the left atrium, the system is pulled against the left atrial side of the interatrial septum such that the left atrial disk abuts the septum. The sheath is then withdrawn, exposing the right atrial disk on the right atrial side under fluoroscopic and ICE guidance. Once the device position is confirmed with fluoroscopy and ICE, and felt to be stable and fully expanded without obstruction or impingement of nearby structures, the device is released. Bubble study or right atrial angiogram may be performed at the end of the procedure. Femoral venous sheaths are removed and hemostasis achieved by manual compression.
It is our practice to administer two doses of antibiotics 12 hours apart. Patients are monitored with telemetry overnight, and chest x-ray and TTE with bubble study are performed the following morning to confirm accurate positioning. Aspirin 81 mg daily and Clopidogrel 75 mg daily for 6 months are prescribed. TTE with bubble study is repeated at 6 months. Endocarditis prophylaxis is advised for 6 months.
Transcatheter PFO closure is a safe procedure; however, complications can occur in 1% to 4% of patients with most complications being mild. The most frequent reported complication after PFO closure is the occurrence of atrial arrhythmias including atrial fibrillation and atrial flutter. In retrospective studies, new atrial fibrillation (AF) was observed in 3.9% of patients, while the rate of AF was very low in the RESPECT trial (0.2%). Device thrombosis occurs in 0.6% patients and device embolization can occur in 0.07% of patients. Device fracture was observed in older generation devices, but extremely rare in the current devices. Serious bleeding from vascular complications occurred in ≤0.5% in the RESPECT and PC trials. Pericardial effusion or tamponade has been reported in 0.3% of patients. Air embolism is a potentially disastrous complication that can occur due to inadequate flushing of the device systems or while introducing the device systems into the delivery sheath. This complication can be easily avoided by careful flushing and paying meticulous attention to fluoroscopy while advancing the device through the delivery sheath.
ASD is the most common congenital heart defect presenting in adults after bicuspid aortic valve and accounts for 6% to 10% of all defects at birth. It affects twice as many females as males. Left-to-right shunting at the atrial level with right-sided volume overload and eventually pulmonary vascular disease and pulmonary hypertension are responsible for the clinical presentation. Since FDA approval of a device for transcatheter ASD closure in December 2001, there has been a shift from surgical closure to transcatheter closure with excellent results and good prognosis in treated patients.
The development of the interatrial septum has been discussed in the section on PFO ( Figure 32-1 ). There are 4 types of ASDs depending on location. The most common is secundum ASD (75% of all ASDs), which is a defect in the region of the fossa ovalis. The primum ASD (15% to 20%) is located in the inferior portion of the atrial septum near the crux of the heart and occurs due to deficiency of endocardial cushion tissue. It is often associated with a cleft in the anterior mitral valve leaflet or ventricular septal defect (common atrioventricular canal defects). The sinus venosus type of ASD (5% to 10%) is located in the superior or inferior part of the septum, near the entrance of the superior or inferior vena cava into the right atrium or SVC. The superior sinus venosus ASD is often associated with anomalous pulmonary venous drainage into the right atrium. Coronary sinus septal defect (<1%) is located in the wall separating the ostium of the coronary sinus from the left atrium. Only the secundum ASDs can be repaired by transcatheter closure; the other types require surgical closure. ASDs are associated with Down syndrome (particularly ostium primum ASD), Holt-Oram syndrome, and DiGeorge syndrome. In addition to the above, other associated lesions with ASD can include mitral valve prolapse and valvular pulmonic stenosis.
ASD leads to shunting at the atrial level. The magnitude and direction of shunting depends on defect size and the relative compliance of the ventricles. Usually the shunt is from left-to-right atrium due to higher compliance of the right ventricle. With increasing age, the left ventricular compliance decreases and left atrial pressure rises, and the magnitude of left-to-right shunt increases. This leads to volume overload and enlargement of the right atrium, right ventricle, and pulmonary artery. Over time, high pulmonary blood flow occurring for several years leads to pulmonary vascular bed remodeling, increase in pulmonary vascular resistance, and pulmonary hypertension. Left untreated, pulmonary vascular changes become irreversible, leading to severe pulmonary hypertension, right-sided pressure overload, and reversal of shunting leading to right-to-left shunting.
During childhood, patients with ASD are usually asymptomatic and may have a pulmonary outflow murmur or fixed splitting of the second heart sound detected incidentally during routine examination. Some children may present with recurrent respiratory infections or even heart failure. Typically, most young adults have a prolonged asymptomatic course. With increasing age, symptoms of reduced exercise tolerance, progressive exertional dyspnea, and heart failure occur with progressive left-to-right shunting as a result of decreased left ventricular compliance and increased left atrial pressure. Arrhythmias including supraventricular arrhythmias, atrial fibrillation, or atrial flutter may be the presenting sign. Paradoxical embolism resulting in stroke or ischemia of other organ systems may also occur. Untreated ASDs can lead to pulmonary vascular disease and pulmonary hypertension in the absence of other causes, but typically not until adulthood.
Physical exam findings that lead to evaluation for ASD include right ventricular heave, wide and fixed split of the second heart sound (due to delayed pulmonary valve closure), ejection systolic murmur best heard at the left sternal border (reflecting increased blood flow through the pulmonary valve), and loud pulmonic component of second heart sound in patients with pulmonary hypertension. Ostium primum ASDs may have associated mitral and tricuspid regurgitation murmurs. Electrocardiographic findings include right atrial enlargement (P-pulmonale), right axis deviation, right ventricular hypertrophy (tall R wave in V1), and incomplete right bundle branch block (rSR’ or rsR’ in leads V1-V3) in secundum ASD. In primum ASD, left axis deviation may be seen. First-degree atrioventricular block can be seen in any kind of ASD. Chest x-ray findings include right atrial and right ventricular enlargement, dilated pulmonary artery, and increased pulmonary plethora.
Echocardiography is the diagnostic modality of choice for ASD ( Figure 32-6 ). In children TTE provides most of the information; however, in adults TEE is important for complete evaluation. The defect is usually seen on TTE from subcostal view of the interatrial septum or apical four-chamber view. Septal dropout is an important limitation of TTE, which can lead to false-positive diagnosis of ASD. Saline contrast echocardiography leads to accurate diagnosis in most cases. In addition to making a diagnosis, TTE demonstrates presence of right atrial and right ventricular enlargement and enables assessment of pulmonary artery pressure using the tricuspid regurgitation jet velocity. The magnitude of left-to-right shunt using noninvasive calculation of pulmonary to systemic blood flow ratio (Qp/Qs) can also be assessed, but is rarely used due to inaccuracies. Additionally, TTE allows evaluation for other associated congenital anomalies such as pulmonary valve disease, mitral valve prolapse, and pulmonary venous drainage. Prior to transcatheter ASD closure, complete assessment using TEE or ICE (usually performed just prior to closure) is critical ( Figure 32-6 ). This allows assessment of various rims/margins for suitability for device closure, drainage of all four pulmonary veins, exclusion of sinus venosus-type ASDs, which can be missed by TTE, and detailed evaluation of mitral valve disease if present. Recently three-dimensional echocardiography has also been used for evaluating ASDs ( Figure 32-6 ). MRI is another noninvasive imaging modality that can be used if echocardiography does not provide all required information. MRI enables direct visualization of the defect, pulmonary venous drainage, calculation of shunt size, and quantification of RV volume and function. Contrast-enhanced cardiac CT can also provide similar anatomic information.
In the current era, cardiac catheterization is not required to establish a diagnosis in the presence of adequate noninvasive imaging. Right heart catheterization with measurement of oxygen saturations (shunt run) and measurement of pulmonary artery pressure and coronary angiography in patients >40 years of age are usually performed at the time of planned transcatheter closure. The author usually performs pulmonary angiogram to confirm absence of anomalous pulmonary venous drainage at the time of closure. Invasive hemodynamic assessment to determine shunt size may be needed when the hemodynamic significance is not clear by echocardiography and also when there is need to determine PVR and pulmonary vascular reactivity in the presence of pulmonary hypertension.
Management and Indications for Atrial Septal Defect Closure
Small ASDs with diameter <5 mm and no evidence of RV volume overload may not require closure as these do not usually impact the natural history. Unrepaired ASDs with significant shunting can result in right-sided volume overload, with progressive heart failure, arrhythmias, hemodynamically significant tricuspid regurgitation, pulmonary hypertension, and reduced survival. Current ACC/AHA guidelines recommend (Class I) ASD closure in the presence of right-sided volume overload, that is, right ventricular or right atrial dilatation in a symptomatic or asymptomatic patient. Closure in presence of symptoms or right-sided heart enlargement prevents further deterioration and helps normalize the right-sided dilatation. Natural history studies of ASD closure show reduced survival after closure in patients older than 24 years of age or with pulmonary hypertension (systolic PAP ≥ 40 mm Hg). Additionally, closure in patients over 40 years of age, while improving symptoms and mortality compared with a medically managed group, did not reduce the risk of atrial arrhythmias. Therefore ASD closure should be performed in a timely fashion in appropriate patients to prevent long-term complications. An ASD other than secundum ASD should be repaired surgically. Closure of ASD may be considered in some patients regardless of evidence of right-sided enlargement, for example, in professional divers and patients undergoing pacemaker implantation due to risk of paradoxical embolism. Similarly, ASD closure may be considered prior to pregnancy. In patients with PAH, pulmonary vasodilator testing to assess for reversibility and test occlusion of ASD should be performed. Inhaled nitric oxide is used commonly as a pulmonary vasodilator. A positive vasoreactivity response is defined as a reduction of mean PAP of >10 mm Hg with resultant mean PAP of 40 mm Hg or less, without fall in cardiac output. Closure in such patients may be performed if there is net left-to-right shunting, PA pressure <2/3 systemic levels, PVR <2/3 SVR, or when responsive to either pulmonary vasodilator testing or test occlusion. A favorable response is indicated by a fall in mean pulmonary artery pressure with test occlusion with no decrease in cardiac output and no rise in right atrial pressure. In presence of unfavorable response, pulmonary vasodilator therapy should be initiated and hemodynamics reassessed a few months later. ASD closure is also indicated in presence of paradoxical embolism and documented platypnea-orthodeoxia. An absolute contraindication (Class III) for closure is irreversible PAH and no evidence of left-to-right shunt.
Devices for Atrial Septal Defect Closure
Percutaneous transcatheter closure has largely replaced surgical repair for a vast majority of secundum ASDs with appropriate morphology in the absence of any other associated defects due to good outcomes and low rates of complications. The two devices approved for transcatheter ASD closure are the Amplatzer ASO (AGA Medical Corporation, Golden Valley, Minnesota) device and the Helex septal occlude (discussed in the PFO section). The Amplatzer ASO is a self-expandable, double-disk device made of nitinol wire mesh that is tightly woven into 2 disks with a 3- to 4-mm connecting waist between the 2 disks ( Figure 32-3 ). The super-elastic properties of nitinol allow the device to be stretched and delivered via sheath size of 6 Fr to 8 Fr. The device size is determined by the waist diameter and ranges from 4 to 40 mm (4 to 20 mm at 1-mm increments, 22 to 40 mm at 2-mm increments; the 40-mm device is not available in the United States). The disk diameters increase with increasing size and the left atrial disk is 6 to 8 mm larger than the right atrial disk depending on device size, as shunting is from left-to-right. The Amplatzer delivery system supplied separately from the device consists of a loader, hemostasis valve with extension tube and stopcock, delivery sheath of varying size and length (depending on device size to be used), a dilator, and a delivery cable. All delivery sheaths have a 45-degree tip (45-degree TorqVue Delivery Sheath). There are sheaths with 180-degree turn available but they are typically not used for ASD closure. A 60-degree angulated Hausdorf Sheath (Cook Medical, Bloomington, Indiana) can be used for ASD with poor posterior inferior rim.
ASD closure is usually performed in the cardiac catheterization laboratory with conscious sedation and ICE and fluoroscopic guidance ( Figures 32-7 and 32-8 ) ( and ). For complex septal anatomy, such as multiple ASDs, TEE may be preferred. Advantages of ICE over TEE include no need for general anesthesia or additional cardiologists to perform the procedure, better views of the posteroinferior part of the interatrial septum, and shorter procedure times. Most operators use the AcuNav ICE catheter (Siemens Medical Solutions distributed by Biosense Webster, Diamond Bar, California). The initial steps are similar to that described for PFO closure. Aspirin 325 mg is usually administered before the procedure and Clopidogrel 600 mg loading dose at the end of the procedure. Femoral venous access is obtained in bilateral groins with an 8 Fr or 9 Fr sheath each (or 2 sheaths in the same vein), one of which is for ICE. We prefer a 9 Fr 35-cm sheath for the ICE catheter to easily traverse the iliac vein into the inferior vena cava, particularly for left femoral vein insertion. Heparin is administered to maintain ACT >250 seconds and a dose of intravenous antibiotic is administered prior to device deployment.
A complete right heart catheterization is first performed to measure shunt fraction, pulmonary artery pressures, and pulmonary capillary wedge pressure. In patients older than 40 years, a coronary angiogram is also performed. We also perform pulmonary angiogram with levo phase imaging to assess drainage of all four pulmonary veins into the left atrium. Some operators perform right upper pulmonary vein angiogram in 35-degree LAO cranial projection, which provides an angiographic roadmap of the interatrial septum to facilitate closure.
The ICE catheter is advanced into the right atrium and the interatrial septum adequately interrogated for assessment of various rims, measuring defect size and confirming pulmonary venous drainage. A rim is considered to be deficient if its length is <5 mm, and absent if it is ≤1 mm. The rims should not be deficient (except anterior rim, as many patients lack the anterior rim and it is not a contraindication). The directions include a “warning” that a deficient aortic rim may incur increased risk of erosion, but data are insufficient as discussed later. After completion of hemodynamic assessment, angiography, and ICE assessment, a Goodale-Loubin (GL) catheter is advanced with a 0.035-inch J-tipped guidewire into the SVC. The GL catheter is moved in a caudal direction and then directed toward the interatrial septum and the ASD crossed with or without the 0.035-inch J-tipped guidewire using ICE and fluoroscopic guidance. The catheter and guidewire are placed in the left superior pulmonary vein, taking care to ensure that the wire tip is not in the left atrial appendage to avoid perforation. The 0.035-inch J-tipped guidewire is exchanged for a 0.035-inch 1-cm Amplatz super-stiff wire, again taking care to ensure that the tip is not in the appendage. Balloon sizing is the next step and is usually performed with an AGA sizing balloon or NuMed sizing balloon. Under fluoroscopic and ICE guidance, the balloon catheter is placed in the defect over the extra-stiff guidewire and the balloon is gently inflated until no flow is visualized by color Doppler on ICE imaging ( Figures 32-7 and 32-8 ). It is very important to stop inflating when flow ceases (stop-flow diameter) to avoid oversizing the defect. This diameter is measured on ICE as well as fluoroscopy. For ASO, device size should be equal to but no larger than 1 to 2 mm above the stop-flow diameter. Helex septal occluder size should be at least twice the stop-flow diameter. For defects >18 mm, ASO is preferable over Helex device.
The next steps depend on the device used. For the ASO device, delivery sheath size ranges from 6 Fr to 12 Fr depending on device size chosen. The balloon-sizing catheter is removed, leaving the 0.035-inch wire in place. The delivery cable is passed through the loader and the device is screwed to the tip of the delivery cable. The device and loader are immersed in sterile saline solution and the device is pulled into the loader while flushing through the side arm. The delivery sheath is prepped and the dilator is inserted into the sheath. The short sheath in the femoral vein is removed and the delivery sheath/dilator is then advanced over the 0.035-inch wire, which has been placed in the left upper pulmonary vein. The dilator is removed once it reaches the right atrium and the sheath is de-aired. The sheath is then advanced over the wire into the left atrium, taking care to avoid suction of air in the system. The guidewire is removed and the sheath is flushed carefully. The loading device is then attached to the delivery sheath. Under fluoroscopic guidance, the device is advanced, carefully watching for any sign of air in the system. Once the device is at the tip of the delivery sheath in the left atrium, under fluoroscopic and echocardiography guidance the left atrial disk is deployed by retracting the sheath over the delivery cable. The device is gently pulled against the interatrial septum and with tension on the delivery cable; the sheath is retracted further to deploy the right atrial disk. After deployment, the position is checked by ICE and, if needed, a gentle “to and fro” motion (Minnesota wiggle) can be performed with the delivery cable to assure stable positioning. ICE assessment should include Doppler flow, which may still demonstrate flow through the waist (but should not be present around the disk), and evaluation for obstruction of adjacent structures including atrioventricular valves. If the positioning is unsatisfactory or there is impingement of adjacent structures, the device is retracted back into the delivery sheath and redeployed or replaced with a new device as appropriate. Device positioning can also be confirmed by angiography in LAO cranial projection, which allows separation of the left and right atrial disks. In cases where the device impinges or indents on the aortic root, there may be higher risk of erosion. Once satisfactory positioning is confirmed, the device is released by attaching the plastic vise to the delivery cable and rotating it counterclockwise. Deployment of the Helex septal occlude device is discussed in the section on PFO closure.
After device deployment, sheaths are removed and hemostasis is achieved. We administer two doses of antibiotics 12 hours apart. Patients are monitored with telemetry overnight, and chest x-ray and TTE with bubble study are performed the following morning to confirm accurate positioning. The RV size and device are re-assessed at 6 months with TTE. Endocarditis prophylaxis is advised for 6 months.
Large Atrial Septal Defect with Deficient Rims
ASDs larger than 25 mm are most often associated with rim deficiency. The Helex device cannot be used to close large ASDs; hence most data exist for such defects with the ASO device. In cases of large ASDs with deficient superior, anterosuperior, or posteroinferior rims, the left atrial disk can prolapse through the defect during regular deployment. Several techniques have been described for these situations to increase chances of success. In pulmonary vein approach for large ASD with deficient anterior or posterior rim, the delivery sheath is placed in the left upper or right upper pulmonary vein and the left atrial disk is partially deployed in the pulmonary vein. The sheath is then withdrawn and the remainder of the device is rapidly deployed, keeping the delivery cable fixed and stable. This technique allows the disks to be parallel to the septum. Similarly, in the left atrial roof approach, the delivery sheath is placed near the orifice of the right upper pulmonary vein (not inside the vein), and the left atrial disk deployed in the roof of the left atrium (disk is perpendicular to the spine). The right disk is then deployed by withdrawing the sheath. This allows the posterior edge of the left atrial disk to stay in the left atrium as the remainder of the device is deployed. Other approaches for closing large ASDs with deficient rims include sheath modifications or using special sheaths. The Hausdorf sheath (Cook Medical, Bloomington, Indiana) is a double curve sheath with an angled tip that keeps the disk parallel to the septum and away from the aortic rim. A modification of the Mullins sheath has been described for large ASDs with deficient anterior or posteroinferior rims, where the distal curved portion of the sheath is cut off, resulting in a straight side-hole (SSH) delivery sheath. The sharp end of the cut sheath is then trimmed to reduce the risk of perforation. The SSH technique allows the device to exit the tip of the delivery sheath at an angle parallel to the septum. Other techniques for large ASDs with deficient posterior rims include using a right Judkins catheter technique or a steerable curved guiding catheter such as the Agilis catheter (St Jude Medical Inc., Minneapolis, Minnesota). The balloon-assisted technique consists of using a balloon as a buttress to prevent prolapse of the left atrial disk through the ASD during deployment.
Multiple or Fenestrated Defects
Multiple defects are present in 10% of cases. These can often be treated with a single device. A second device is usually required if the distance between the primary and secondary defect is 7 mm or more. Other methods for closure with a single device include using a nonself-centering device such as Helex device or the Amplatzer cribriform device.
The vast majority of complications with transcatheter ASD closure are minor. The multicenter pivotal studies for both the ASO (St. Jude Medical Inc., St. Paul, Minnesota) and the HSO (W. L. Gore and Associates, Flagstaff, Arizona) showed differences in the efficacy between surgical and device closure of secundum ASD; however, there were differences in safety outcomes. The ASO pivotal study, which enrolled 442 patients in the device group and 152 patients in the surgical group, had a major adverse cardiac event (MACE) rate of 1.6% in the device group compared with 5.2% in the surgical group. Similarly, in the Helex pivotal study, enrolling 119 patients in the device group and 128 patients in the surgical group, the MACE rate was lower in the device group compared with surgical group (5.9% vs. 10.9%). Cardiac erosions are the most feared and life-threatening complications of ASD closure using ASO device. There have not been any cases of erosions with the Helex device. The erosion rate has been reported to be 0.1% to 0.3% (1 to 3 per 1000 implants) with the ASO device. Erosion most commonly occurs in the roof of the atria or the aorta. Large device size and lack of anterior or superior rims have been proposed as risk factors associated with erosion. Lack of wiggle room in case of a large device with deficient rim can lead to constant impact on the atrial or aortic tissue from the edge of the device. A high-risk ASD is a large defect in the superior portion of the septum in close proximity to the aorta with absent or negligible aortic rim. Based on their findings, the AGA expert panel made recommendations regarding erosions, which include avoiding overstretching while balloon sizing, using stop-flow technique for sizing, gentle to and fro motion while assessing stability, and closer follow-up of those with large ASO (greater than 1.5 times native ASD size) and those with deformation of the ASO device at the aortic root with significant splaying of the device edges by the aorta. These are based on an expert opinion, and not verified conclusively in a prospective manner. There are still differences in opinion regarding risk of erosion. A survey of members of the Congenital Cardiovascular International Study Consortium (CCISC) showed that 71.7% felt that a device in which the disks approximated each other and touched/protruded into the aorta without splaying were at the highest risk of erosion. The Circulatory System Devices Panel of the FDA met on May 24, 2012, to discuss current knowledge about the safety and effectiveness of the Amplatzer ASO device and Gore Helex ASD occlude as transcatheter ASD occluder devices used for the closure of secundum ASD. Since CE mark approval in 1998, there have been 97 cases of confirmed or presumed erosion worldwide in patients with on-label use of the ASO device. Within the 97 erosions, 8 deaths have occurred. Almost 90% of erosions occur within 1 year of being implanted, although one case of erosion was reported 8.5 years after the implant. All reported deaths occurred within 16 months following implant, and no deaths occurred in patients <16 years of age. The panel made some recommendations regarding more frequent follow-up in the first year after closure since events frequently occurred within 12 months, collection of ongoing device data for identification of risk factors for erosion, and thorough discussion of risks and benefits with patients.
Device malposition/embolization can potentially occur with any ASD closure. Various causes include large, eccentric defect, inadequate rims, improper sizing, and operator-related technical issues. Embolization is the most frequent complication reported at a rate of 0.5% to 3%. An analysis of the device embolizations reported to the FDA’s MAUDE database (Manufacturer and User Facility Device Experience) showed that in 77% of cases the device was retrieved using transcatheter approach and in 17% of cases surgical retrieval was needed. There were two deaths related to device embolization. Most embolizations occur at the time of deployment or within 24 hours of the procedure. The operator must be familiar with transcatheter retrieval techniques using a gooseneck snare or bioptome. The most common site of embolization is left atrium, followed by right atrium, pulmonary artery, right ventricle, left ventricle, and aorta. If the device is stuck in the ventricle and entangled within the atrioventricular valve apparatus, the patient should be referred for surgical removal. Device embolization appears to be more frequent with the Helex device compared with ASO device.
Other complications include wire frame fracture in the case of Helex device, thrombus formation on the device, new onset atrial arrhythmias, and impingement of adjacent structures including the atrioventricular valves.
Clinical Trial Data
Results are excellent with both Amplatzer and Helex devices. Closure rates with the Amplatzer ASO device are >95% at 1-year follow-up, and >91% with Helex device. Safety outcomes and complications have been discussed previously. Overall, the risk of complications is lower with the transcatheter closure approach compared with surgical approach.