Percutaneous closure is a surgical procedure used to treat patients with patent foramen ovale (PFO) and atrial septal defect (ASD). Advancements in device technology and image guidance now permit the safe and effective catheter-based closure of numerous intracardiac defects, including PFO and ASD. [1, 2] With these catheter-based closure procedures becoming more prevalent in adults, it is imperative that clinicians possess a sound understanding of intracardiac shunt lesions and indications for repair or closure.
The anatomy of PFO and ASD are shown in the images below.
Studies on annual recurrences after a cerebral vascular accident or a transient ischemic attack reported an incidence ranging from 3% to 16%. [3, 4] In a large study, the recurrent stroke rate or mortality from an embolic event was 6% to 8% per year.  A pooled analysis suggests that the presence of a PFO alone increased the risk for recurrent events 5-fold, with an even higher risk in the presence of an atrial septal aneurysm. 
Other studies have found no significant influence of an isolated PFO but show a strong influence by a PFO associated with an atrial septal aneurysm. In a trial randomizing patients to acetylsalicylic acid or coumadin, the individual presence of a PFO or an atrial septal aneurysm had no influence on the incidence of recurrent stroke.  There are differing results on the question of whether a PFO with or without hypermobility of the septum leads to recurrent stroke are presumably related to factors (besides patient selection) such as diagnostic accuracy of the tests and definitions used to identify both the PFO and the atrial septal aneurysm.
Homma et al  identified other risk factors for recurrent stroke, including the presence of a Eustachian valve directed toward the PFO, the gaping diameter of the PFO, and the number of microbubbles present in the left atrium during the first seconds after release of a Valsalva maneuver during a bubble test.
The effectiveness of percutaneous closure compared with surgical closure or medical management in preventing recurrent embolic events is currently unknown. Uncontrolled studies evaluating percutaneous device closure of a PFO or, much less often, an ASD in patients with at least one paradoxical embolic event (ie, transient ischemic attack, stroke, or a peripheral embolism) have reported results similar to those reported for surgical closure.  A clinical trial is currently ongoing to determine whether percutaneous closure is more effective than medical management in preventing recurrent embolic events.
Nagpal et al studied 414 patients who underwent percutaneous closure between 2002-2009. They concluded that the long-term rate of recurrent stroke is low and serious long-term complications are rare, particularly in patients with a single neurological event. 
Some studies have suggested that PFO closure may reduce the frequency and severity of migraine headaches in patients with significant right-to-left shunts. This data comes from single-center observational studies.  In a meta-analysis of 11 studies, Butera et al evaluated a total of 1,306 patients who underwent PFO closure, of whom 40% suffered from migraine. Quantitative synthesis showed complete cure of migraine in 46%, with resolution or significant improvement of migraine in 83% of cases. They concluded that a significant group of subjects with migraine headaches, particularly if treated after a neurological event, may benefit from percutaneous closure of their patent foramen ovale. However, no randomized, controlled studies are currently available to support this.
Heparin should be administered to achieve a recommended activated clotting time of greater than 200 seconds throughout the procedure.
Following percutaneous puncture of the femoral vein, a standard right heart catheterization is performed. An angiogram is sometimes performed demonstrate atrial communication.
Catheterization of the left atrium is done using a 45-degree left anterior oblique position and cranial angulation of 35-45 degrees, and then contrast medium is injected into the right upper pulmonary vein.
A 0.035-inch exchange J-tip guidewire is then introduced into the left atrium. Afterwards, a complaint balloon catheter is advanced over the exchange guidewire into the left atrium and used to determine the diameter of the defect.
For ASD closure, transesophageal echocardiography may be preferred. For PFO, intracardiac echocardiography may be preferred. Because the defect in PFO is small, this usually provides adequate images during device implantation.
If balloon sizing is performed in addition to echocardiographic measurements, a stop-flow technique should be used. The stop-flow technique is performed as follows.
Using a balloon specifically designed for sizing atrial communications (an Amplatzer sizing balloon), the catheter is passed over the exchange guidewire directly through the skin. To facilitate this percutaneous entry, an assistant should apply forceful negative pressure with an attached syringe.
Under fluoroscopic and echocardiographic guidance, the balloon catheter is placed across the defect and inflated with diluted contrast medium until the left-to-right shunt ceases as observed by echocardiography.
The balloon is deflated until flow is seen, and then reinflated until the shunting ceases. Measurements can then be made using echocardiographic imaging, fluoroscopy, or by using the sizing plate.
Once the diameter of the defect has been determined, an occlusion device is selected, which should be equal to or, if the identical is not available, 1 size larger than the defect. Then, the balloon catheter is removed, leaving the 0.035-inch exchange guidewire in place.
The delivery cable is passed though the loader and the device is screwed to the tip of the delivery cable. Once securely attached, the device and loader is immersed in cold (less than 5° C) sterile saline solution and the device is pulled into the loader with a jerking motion. Flush the device via the side arm.
The dilator is inserted into the delivery sheath and secured to the sheath with the locking mechanism. The dilator/delivery sheath assembly is then introduced through the groin.
Once the delivery sheath has reached the inferior vena cava, the dilator is removed to allow back bleeding to purge all air from the system. Then, the hemostasis valve is connected and flushed with a syringe before the atrium is entered. The sheath is advanced over the guidewire through the communication into the left upper pulmonary vein.
The correct position of the delivery sheath is verified by a test hand injection of contrast medium or by echocardiography. The guidewire is then removed and the sheath flushed with sterile saline.
The loading device is attached to the delivery sheath. Then, the device is advanced into the sheath by pushing the delivery cable. Under fluoroscopy and echocardiographic guidance, the left atrial disc is deployed and the device is pulled gently against the atrial septum.
With tension on the delivery cable, the sheath is pulled back and deployed into the right atrial disc. The sheath is pulled back by approximately 5-10 cm. A gentle “to-and-fro” motion with the delivery cable assures a secure position across the atrial septal defect, which can also be observed by echocardiography.
Correct placement should then be confirmed. See the image below.
If the device is unsatisfactory or if the device does not reconfigure to its original shape, the sheath can be advanced while retracting the delivery cable to recapture the device into the sheath and redeploy or replace with a new device.
The videos below demonstrate percutaneous closure.
In late 2001, the U.S. Food and Drug Administration (FDA) approved two PFO occluder devices—the Amplatzer Septal Occluder and the CardioSEAL Septal Occlusion System—for percutaneous atrial septal defect closure under a humanitarian device exemption for the treatment of patients with recurrent cryptogenic stroke due to presumed paradoxical embolism through a PFO who had failed conventional drug therapy. [12, 13]
However, the FDA withdrew approval for these devices effective October 31, 2006, after finding that the eligible population in the United States described by the approved indication was significantly greater than the 4000 patients per year limit permitted by an humanitarian device exemption. 
Another device, the CardioSEAL Septal Occlusion System has been used off-label for ASD and PFO closure but is FDA approved for use only for closure of certain complex ventricular septal defects. The humanitarian device exemptions for two patent foramen occluder devices, the CardioSEAL STARFlex Septal Occlusion system and the Amplatzer PFO occluder, were withdrawn in October 2006. These devices are now available in the United States for investigational use only.
Favorable outcomes have been reported in the great majority of patients treated with the Amplatzer device (see the image below).
In a series of 100 patients, for example, the Amplatzer device was successfully implanted in 93 with a procedure time ranging from 30 to 180 minutes.  The total ASD occlusion rate at 3 months was 99%.
The safety and efficacy of the HELEX device were evaluated in a nonrandomized, noninferiority study involving 247 patients (median age of 5.5 years) who were treated with either the device or open surgical repair.  In addition to usual criteria for closure of a secundum ASD, patients in the device arm were required to have a defect diameter less than or equal to 22 mm and adequate septal rims to secure the device, both of which were determined by transesophageal or intracardiac echocardiography. [17, 18]
At 12 months, there was no statistically significant difference between the two groups for the combined primary endpoint of device safety, the need for repeat procedures to the target ASD, and documentation of complete occlusion or clinically insignificant leak. The noninferiority of the device compared with surgery persisted even after controlling for baseline differences. The most common adverse outcome in the device group was device embolization requiring catheter retrieval (1.7%).
Monitoring and Follow-up
All patients should be kept overnight for observation. A limited transthoracic echocardiography should be performed prior to discharge. Patients with any observed small pericardial effusion following device implantation should be closely monitored with serial echocardiograms performed until resolution of the pericardial effusion. Higher risk patients should be followed more closely, including clinical follow-up with echocardiogram 1 week after device implantation.
All patients should be on low-dose aspirin for life. Clopidogrel should be continued for a total of 3 months. Transthoracic echocardiography should be repeated at 1 month and 3 months after the procedure. A final transesophageal echocardiogram is recommended at 6 months after the procedure.
Device implantation (and anticoagulation) can be associated with complications. In addition to the 6% procedural complication rate reported by Windecker et al,  including 4 cases of device embolization, there is the issue of device-related thrombus formation. In one report, device-related thrombi were reported in 2.5% of 593 implants for PFO. 
Technical failures have become extremely rare (eg, inability to cannulate the PFO is less than 1%). Complications may include cardiac tamponade, symptomatic air embolism, loss of device, or puncture site problems; however, these are rare. Complete closure at follow-up can be expected in 90-95% cases with the two devices currently in use. Some trivial residual shunt may be acceptable, albeit undesirable, as the device will act as a filter for particulate matter.
Events have recurred in cases where the PFO was not responsible for the index event, in cases where small emboli formed on the left side of the device, or in cases where closure is incomplete.  Recurrent events may come close to the natural course for the first year (about 3%), after which they are extremely rare. In contrast, the natural course under platelet inhibitors or warfarin tends to have a steady or even increasing rate of events over the years.  Hence, the follow-up curves do seem to diverge in favor of device closure in nonrandomized comparisons.