Patent foramen ovale (PFO) is a frequent finding in young patients with cryptogenic stroke.^1–4 PFO has also been related to migraine,5 platypnea-orthodoxia syndrome (condition in which shortness of breath and hypoxemia occur when upright and resolve when prone),6 and decompression illness in divers.⁷ Long-term anticoagulation⁸ and surgical closure of PFO have been used as therapeutic options,^9–12 however, their results have been mixed with respect to stroke prevention. Recently, percutaneous transcatheter PFO closure techniques have been used more frequently as a therapeutic option in these patients.^13–20
Several studies have also reported varying frequencies of ATs following transcatheter closure of PFO.^15–20 However, the effect of percutaneous transcatheter PFO closure techniques on ATs is not well studied. We report the frequency of ATs in patients who underwent transcatheter closure of PFO at our institution.
Methods
Patient population. The patient population included 71 patients with PFO and ? 1 cryptogenic stroke or orthodoxia between February 2001 and March 2004. All patients underwent complete neurological evaluation prior to the procedure. Patients’ evaluation included 12-lead electrocardiogram (ECG), Holter monitoring, two-dimensional (2-D) echocardiography, extracranial Doppler ultrasound, brain computed tomography scan or magnetic resonance imaging. Patients with other identifiable etiologies for the stroke were excluded. All patients underwent transesophageal echocardiography (TEE) (Hewlett-Packard ultrasound machine, Palo Alto, California) with microbubble injection with or without the Valsalva maneuver prior to the procedure to diagnose PFO. The shunt was defined as small when 3 to 20 bubbles crossed to the left atrium, and as large when there were > 20 bubbles in left atrium. Large shunts defined by this categorization have been shown to be associated with a significant risk of stroke compared to small shunts.²¹
Procedure. Closure of the PFO was performed by using a CardioSEAL device (NMT Medical, Inc., Boston, Massachusetts) or Amplatzer PFO occluder (AGA Medical Corp., Golden Valley, Minnesota). Written consent was obtained from all patients. PFO closure was performed with either TEE guidance, under intracardiac ultrasound guidance, or by using a standard transseptal approach. The PFO closure device was deployed across the transseptal puncture. A bubble study was performed after deployment of the device to evaluate the persistence of shunt by TEE, intracardiac echocardiography or transthoracic echocardiogram. All patients had a 12-lead ECG the next day prior to discharge.
Follow up. Patients were followed clinically at 1 month, 6 months, 12 months and yearly thereafter following the procedure. All patients underwent 2-D echocardiography, color-flow Doppler, and agitated saline solution injection to evaluate the presence of residual shunt. Patients with complaint of palpitation underwent 48-hour Holter monitoring and were evaluated by an electrophysiologist.
Statistical analysis. All continuous variables are expressed as a mean value ± standard deviation. Comparison of continuous values was performed using the Mann-Whitney test. Categorical variables were compared using the Fisher’s Exact test. A p-value < 0.05 was accepted as representing a significant difference.

Results
Baseline characteristics (Table 1). The patient population included 71 patients [31 (44%) men, aged 54 ± 14 years] with PFO and >/= 1 paradoxical emboli (n = 70) or orthodoxia (n = 1) between February 2001 and March 2004. Their cardiovascular risk factors included a history of hypertension in 20 (28%) patients, diabetes mellitus in 6 (9%) patients, hyperlipidemia in 10 (14%), and coronary artery disease in 6 (8%). Three (4%) patients had a history of migraine.
Echocardiographic data (Table 1). Echogardiographic data were reported by different observers. The mean ejection fraction was 60% ± 6%. PFO was associated with an atrial septal aneurysm in 10 (14%) patients. Twenty-two (31%) patients had evidence of at least mild mitral regurgitation by color-flow Doppler, which was severe in only 1 patient. Twenty-nine (41%) patients had at least mild left atrial enlargement (> 5 cm; normal range 2.5–4.5 cm), while 10 (14%) patients had at least mild right atrial enlargement (>/= 5 cm; normal range 2.2–4.4 cm). The mean left atrial dimension was 3.7 ± 0.9 cm on the parasternal view, and was 5.3 ± 1.0 cm on the apical view. The mean right atrial dimension measured on the apical view was 4.7 ± 1.0 cm. Left ventricular hypertrophy was present in 2 (3%) patients.
Procedural results. PFO closure was performed using a CardioSEAL device in 67 (94%) patients, while an Amplatzer occluder was used in 4 (6%) of patients. The procedure was performed under general anesthesia with TEE guidance²² in 23 (33%) patients, with intracardiac ultrasound²³ in 15 (21%) patients, and using a standard transseptal approach24 in 33 (46%) patients. A 33 mm device was used in 18 (25%) patients, while the remaining 53 (75%) patients received a 28 mm device. The average size of device was 29 ± 4 mm.
Follow up. The mean follow up was 248 ± 229 days. One patient committed suicide prior to the first follow-up visit. Another patient died during follow up from metastatic small-cell carcinoma. Six patients (8%) required repeat intervention for significant residual shunt detected by color-flow Doppler on the follow-up echocardiography.
Atrial tachyarrhythmias (Tables 2 and 3). Five (7%) patients developed ATs (3 atrial fibrillation (AF) and 2 typical type I atrial flutter) following PFO closure. The onset of ATs ranged from 1 day to 480 days post-procedure. ATs occurred more frequently in those who received a 33 mm device than in those who received a ? 28 mm device [4/18 (22%) versus 1/53 (2%); p < 0.05]. Two patients had a preoperative history of palpitation without any history of documented arrhythmia. One patient with AF had a history of AF post-coronary artery bypass and aortic valve repair. This was controlled on propafenone, with the patient being AF-free for more than 12 months prior to the procedure. Although he was appropriately anticoagulated with warfarin, he experienced multiple episodes of transient ischemic attacks prior to the PFO closure procedure. His TEE was negative for any cardiac source of emboli except or a PFO with a large right-to-left shunt. Another patient with AF developed large clots on both sides of the closure device for which she underwent surgical removal of the device along with a modified Maze procedure. Two patients with typical atrial flutter subsequently underwent radiofrequency ablation of the right subeustachian isthmus. One of these patients later developed frequent, very short episodes of paroxysmal AF detected on 24-hour Holter monitoring.
Patients with ATs showed a trend toward a larger left atrium only on apical view of 2-D echocardiography (6.5 ± 1.4 cm versus 5.1 ± 0.9 cm; p = 0.05). However, the difference in left atrial size on the parasternal view and right atrial size between the 2 groups was not statistically significant (4.5 ± 1.1 cm versus 3.7 ± 0.8 cm and 5.9 ± 2.4 cm versus 4.6 ± 0.7 cm).

Discussion
We report a 7% frequency of ATs in patients who underwent a percutaneous closure of PFO, which is much higher than the frequency of AT in the age-matched general population.²⁵ These arrhythmias occurred more frequently in those patients who received a larger closure device.
The prevalence of AF in the general population has been reported to be < 1% in patients 21 Our study demonstrates that AT occurs with more frequency in patients receiving transcatheter closure devices for PFO compared to the age-matched general population. This may be due to the fact that these devices act as a local irritant for the adjacent myocardium and increase the automaticity in the surrounding area by focal irritation. It is also possible that these devices act as a central anatomical obstacle that provides the foundation for the development of anatomic macroreentry. Another possibility is that these devices cause some local changes in the surrounding tissue, enhancing triggered activity, which may be the mechanism of atrial tachyarrhythmia. The fact that these arrhythmias were more common in patients who received larger devices supports the notion that these devices cause a local effect, increasing the occurrence of arrhythmias.
ATs have also been reported in other studies following transcatheter closure of PFO and atrial septal defect (ASD).^15–20,26 In a study of 276 patients who underwent transcatheter closure of PFO using the Amplatzer Septal Occluder device, the authors reported a 0.8% frequency of brief episodes of AF.¹⁶ Another study reported a 6% frequency of AT following transcatheter closure of PFO.²⁰ Sievert et al¹⁸ reported AF within the first weeks after implantation of the device in 5 (2%) out of 281 patients with PFO and cryptogenic stroke. Beitzke et al¹⁹ reported AT in 6 (4%) patients (2 with supraventricular tachycardia, 1 with atrial flutter and 3 with AF) which occurred between 1 and 4 weeks after implantation of the device. All patients with AF required medical or electrical cardioversion to sinus rhythm, while in the other patients, the arrhythmias disappeared spontaneously after several weeks. Although the CardioSEAL and Amplatzer occluder devices were used with similar frequencies in their study (73 and 77 cases, respectively), 5 of these arrhythmias occurred with the CardioSEAL, and only 1 with the Amplatzer PFO occluder. Thus it was suggested by the authors that the angular design of the CardioSEAL device might predispose patients to arrhythmias by focal irritation. However, other studies that compared these devices did not show a difference between devices.^14,18 Thus we do not believe that the high frequency of ATs in our group was due to the fact that most patients in our study received a CardioSEAL device. Atrial tachyarrhythmias have also been reported following transcatheter closure of atrial-septal defect (ASD). Hill et al²⁶ used ambulatory ECG monitoring to prospectively assess the electrocardiographic effects of transcatheter closure of ASD using the Amplatzer septal occluder device. Ambulatory Holter monitoring was performed pre- and immediately post-transcatheter closure in 41 patients with secundum ASD. Changes in atrioventricular (AV) conduction, including intermittent second-degree AV block type II, and complete AV dissociation post-closure occurred in 3 patients (7%). Supraventricular ectopy was noted in 26 patients (63%) post-closure, including 9 patients (23%) with nonsustained supraventricular tachycardia, of which only 3 had short runs of supraventricular tachycardia prior to the closure. There was also a significant increase in the post-closure number of supraventricular premature beats per hour (p = 0.047).
Study limitation. Forty-eight hour Holter monitoring was obtained only in patients with symptoms of palpitation. However, some patients may have had significant atrial dysrhythmias without symptoms which may have resulted in underreporting the true incidence of such ATs.
In conclusion, ATs are relatively common after transcatheter closure of PFO. They are more common in patients who receive larger devices. Possible mechanisms may be focal irritation caused by the larger device, or the role of these devices as a central anatomic obstacle resulting in a circuit susceptible to macroreentry.