Introduction
Atrial arrhythmias are known to be associated with atrial septal defects (ASDs). The surgical closure of ASDs can also eliminate atrial arrhythmias, and on some occasions, can induce new arrhythmias. In many cases, the arrhythmias, which are present before defect closure, continue to persist. Atrial arrhythmias are a source of significant morbidity and mortality, and there is a need to decrease the incidence of postoperative arrhythmias, both new-onset after surgery and the persistence of preexisting arrhythmias.¹ On the other hand, there is some suggestion that early correction of ASDs can prevent the onset of atrial arrhythmias related to such defects.²

Patent foramen ovales (PFOs), on the other hand, have not been directly related to atrial tachyarrhythmias, but there is some evidence that links PFOs to “atrial vulnerability” for arrhythmias.³ Furthermore, operative correction of PFOs seems to lead to atrial tachyarrhythmias.^4–7 Other forms of cardiac surgical intervention have also revealed a predisposition for atrial arrhythmias in patients with PFOs.⁸

It appears that surgical manipulation of the heart can lead to postoperative atrial arrhythmias. In this context, a percutaneous route of closure may provide a less invasive and less arrhythmogenic alternative in the correction of ASDs and PFOs.^9,10

Percutaneous closure of PFOs has gained acceptability as a safe and effective therapy for patients with paradoxical embolism.^7,11–17 However, there are also concerns that the closure of PFOs might induce atrial arrhythmias and ironically predispose a patient to stroke, reversing some of the benefits of PFO closure, or even increasing the risk of stroke compared to preclosure risks.

The present study analyzes the effects of percutaneous closure of PFOs on the resolution of preexisting atrial arrhythmias and the occurrence of new postclosure, long-term atrial arrhythmias.

Methods
This study included patients with a PFO who underwent percutaneous closure between the years 2002 and 2007. The most common indication was one or more cerebrovascular accidents, explained best by an episode of paradoxical embolism. All patients were vigorously investigated to rule out other causes of systemic emboli with reasonable certainty. There were 10 patients with hypoxia not explained by any other etiology who underwent percutaneous closure of their PFO. Eight patients with various symptoms, mainly migraine headaches, who underwent percutaneous closure, were also included. The investigations varied, but included neurologic examination, brain computed tomography (CT), brain magnetic resonance imaging (MRI), extracranial Doppler ultrasonography, 12-lead electrocardiography (ECG), Holter monitoring, 2-dimensional echocardiography, and transesophageal echocardiography with bubble study.

Procedure. Closure of the PFO was performed using the CardioSEAL occluder device (NMT Medical, Inc., Boston, Massachusetts) under mild sedation. PFO closure was performed under intracardiac echocardiography (ICE) guidance in all cases. The implantation technique has been well described previously.¹¹ Patients were routinely placed on clopidogrel for at least 6 months and subacute bacterial endocarditis prophylaxis for 6 months.

Follow-up evaluations. The patients were observed in telemetry units overnight and discharged the following morning if their stay was uneventful. Follow-up echocardiography 3 months after device placement was performed in most cases to confirm placement of the device and ensure closure of the atrial defect. The patients were seen for routine follow-up visits where, in most cases, their pulse was checked and an ECG was performed. Holter monitoring was done in cases where the patient complained of palpitations. The patients in our study were followed for an average of 11.9 months.

Results
Subject characteristics. Data were analyzed using SPSS software, version 11.5 (SPSS, Inc., Chicago, Illinois). The sample included 68 subjects, the majority of whom were male (67.6%). The mean age for the group was 55.75 years (standard deviation [SD] = 14.74), with a range of 19–84 years The subjects’ ages were grouped into categories as a means of more accurately describing the group. Table 1 indicates the age categories of participants, showing that the greatest number of subjects were between 51 and 65 years of age. A small percentage of patients were current smokers (19.1%), 26.5% were former smokers and 54.5% were nonsmokers. Indications for the procedure suggested that most of the subjects had experienced cerebrovascular disorders (cerebrovascular accident [CVA] or transient ischemic attack [TIA]). Before the procedure, pulmonary arterial pressure (PAP) averaged 24.96 (SD = 10.4). Prior to the procedure, all patients underwent baseline ECG, with 4 noted to be in atrial fibrillation.

Procedure results. Most subjects (57.4%) showed good results from the procedure with no complications. Postprocedure complications developed in 29 subjects, with the majority of the complications being cardiac in nature (72.4%) Before the procedure, PAP averaged 24.96 (SD = 10.4), while the postprocedure PAP average was 21.7 (SD = 7.1). Since a number of variables were measured at the categorical level, chi-square (c2) analysis was used to determine whether any predictive factors could be identified. Age category, gender, and smoking status demonstrated significant predictive ability (p < .05) of postprocedure complications (Table 3).

No relationship was identified between these variables and the specific occurrence of resolution of postoperative atrial fibrillation. Out of a total of 68 qualifying patients who underwent correction of their defect, 4 (5.88%) developed new-onset atrial fibrillation. Two patients experienced postoperative transient atrial fibrillation lasting < 30 days. More importantly, 3 of the 68 (4.41%) patients with preprocedural atrial fibrillation were noted to have normal sinus rhythm postoperatively. One patient with atrial fibrillation was noted to be in atrial fibrillation postoperatively. The mean age of the entire population was 55.75 years, with a SD of 14.75. The mean age of patients with new-onset permanent atrial fibrillation was 63 years, with a SD of 5.67. The mean age of patients experiencing resolution of their atrial arrhythmias was 57 years, with a SD of 16.46.

Among females (n = 22), new atrial fibrillation was evident in only 1, and among those experiencing resolution of atrial fibrillation, none were female. Other cardiac arrhythmias, mostly nonspecific in nature with questionable relation to the procedure itself, were seen by ECG on a routine postoperative visit and evaluated after complaints of dizziness or palpitations (Table 4). Noncardiac complications, mostly minor in nature, were also seen (Table 5).

Discussion
This study examined the effects of percutaneous closure of PFOs on the incidence of new atrial arrhythmias, the resolution of preexisting ones and the factors associated with these. The literature on percutaneous closure of PFOs is fairly limited, and few studies have explicitly looked at the relationship between the closure of PFOs and atrial arrhythmias. Many studies do report the complication rates of post percutaneous closure, but there is significant variability in the time period for which these patients were followed and the specific events that were included as complications. Consequently, the reported complication rate following a percutaneous procedure varies from as low as 0%¹³ to rates similar to ours, at 52%,¹⁸ if the rates of nonclosure were also included as complications, as was done in our study. The age of the participants in our study was higher than in most studies. One study, which followed the patients for a median period of 1.9 years, reported a total complication rate of 30.8%.⁶ This discrepancy in complication rates is partly due to small sample sizes, which affects the study on percutaneous closure of PFOs. Also, the specific complications examined vary from study to study; the complications that we report above include minor symptoms such as nonspecific dizziness and hiccups, which were questionably related to the procedure itself. Other complications, which may have been present before the procedure, were more likely related to the procedure, such as ventricular and atrial ectopy seen on Holter monitoring performed for other indications. Other complications such as femoral artery pseudoaneurysm and hematoma formation were also relatively minor and expected.

The incidence of new-onset atrial fibrillation was 5.9% in our study, which is comparable to other studies reporting postprocedural atrial fibrillation. The rates in these studies varied from 3.5–8%.^4,5,17 Open-surgical procedures exclusively for PFO closure are infrequently performed. Several studies that were conducted with follow up did not look at atrial arrhythmias. The performance of such procedures with the advent of percutaneous procedures is also of questionable efficacy.^19,20 The incidence of atrial fibrillation after open-surgical closure of ASDs in adults and children has been reported and is within a similar range.^1,21 The incidence of atrial fibrillation following surgical closure of ASDs does appear to depend on age.¹ The physiologic and anatomic differences between ASDs and PFOs are recognized, and we accept that a comparison might not be completely valid.

Our study did not identify any factors such as age or smoking that might predict the appearance of new or previously-present atrial fibrillation. In the study by Kiblawi et al, the development of postprocedural atrial fibrillation was more common in patients > 55 years of age compared to younger patients.¹⁷ A similar conclusion was reached by Gatzoulis et al, who looked at open-surgical repair of ASDs.¹ We did, however, establish a predictive relationship between the overall complication rate with age, smoking history and gender.

The most obvious drawback of our study was its retrospective nature, which limited the ability to gather data on more preprocedural variables. The other major limitation was the small sample size, which also confounds most other studies looking at PFO/ASD closure. Furthermore, paroxysmal atrial fibrillation was not rigorously ruled out preoperatively by Holter monitoring, and postoperative Holter monitoring was symptom-driven. Thus, asymptomatic paroxysmal atrial fibrillation could have been missed. More definitive studies with larger sample sizes specifically studying the effects of percutaneous correction of PFOs on atrial fibrillation are needed.

Acknowledgement. We would like to thank Dr. Faran Bashir for his review of and suggestions regarding this manuscript.