Comparison of Contemporary Devices Used for Transcatheter Patent Foramen Ovale Closure

ABSTRACT: Background. Transcatheter patent foramen ovale (PFO) closure can be performed with various devices. However, their handling, safety, presence of residual shunt and impact on recurrent thromboembolic events (TEs) are rarely compared with one another. Our goal was to compare the clinical performance of contemporary devices designed for PFO closure. Methods. PFO closure with the Cardia PFO occluder (n = 405), Intrasept (n = 301) and Amplatzer PFO occluder (n = 89) was attempted in 795 patients with presumed paradoxical embolism. Results. The procedure was successful in all patients. The periprocedural complication rate of 1.8% was comparable among the three groups. Residual shunting immediately following the procedure was higher in patients treated with the Cardia PFO occluder (24% vs. 14% [Intrasept] and 16% [Amplatzer]; p = 0.004). After a mean follow-up period of 26 months, no difference in residual shunting was seen (8% [Cardia] vs. 7% [Intrasept] vs. 8% [Amplatzer]; p = 0.736). The annual incidence of recurrent TE was 1.4%, which was not affected by the presence of a residual shunt or the type of device used. New-onset atrial fibrillation (AF) following PFO closure was the only predictor of recurrent TE and was more common in patients treated with the Amplatzer (10% vs. 5% [Cardia]) and 5% [Intrasept]; p = 0.057). Conclusions. The clinical performance of the three PFO occluders evaluated in this study seems comparable. Device design does not seem to impact the success of the PFO closure procedure. AF was the only predictor of recurrent TEs, underscoring the importance of aggressive therapy for atrial arrhythmias early following PFO closure.


J INVASIVE CARDIOL 2008;20:442–447

Transcatheter closure of patent foramen ovale (PFO) has been performed for over 15 years.1 Most frequently, PFO closure is recommended in the context of cryptogenic thromboembolic events (TEs) due to presumed paradoxical embolism via the PFO. Ongoing randomized, controlled studies comparing various devices with medical therapy are slowly enrolling, hence current data on transcatheter PFO closure are entirely based on nonrandomized case series.

Initially, PFOs were closed using devices developed for closure of atrial septal defects (ASDs).2–6 However, given the differences in anatomy between PFOs and ASDs, occluders dedicated for PFO closure were developed. Among the earliest developments are the Amplatzer PFO occluder (APO) (AGA Medical Corp., Golden Valley, Minnesota) and the Cardia PFO occluder (Cardia, Inc., Eagan, Minnesota). The APO emerged from the Amplatzer ASD occluder and consists of two discs of woven nitinol wires and a short connecting waist.7,8 The Cardia PFO occluder is made of two ivalon sails mounted on nitinol struts connected by a short center post. It has been modified over the years, with the most recent development called the Intrasept, which is identical to the third-generation Cardia PFO occluder except for an articulating center post that allows better alignment of the umbrellas.9,10 Comparison of those contemporary devices has rarely been done since most case series combine the results of various devices. One study, however, compared the APO with the earlier generations of the Cardia PFO occluder and found that the APO revealed superior results.11

Whether the more recent generations of the Cardia PFO occluder remain inferior to the APO is unknown. Hence, we evaluated our results of transcatheter PFO closure with the Cardia PFO occluder, the Intrasept and the APO.
 

Methods

Inclusion criteria. All patients undergoing transcatheter PFO closure with either a Cardia PFO occluder, Intrasept or APO at a single institution from April 1998 to November 2007 were included in the present study. A total of 795 closure procedures were performed in patients with PFO following presumed paradoxical TE. All patients were required to meet the following criteria: 1) (a) a history of unequivocal ischemic stroke documented clinically by a neurologist and radiographically by either cranial computed tomography or magnetic resonance imaging, or (b) transitory ischemic attack (TIA), defined as a transient neurological deficit or vision loss with full recovery within 24 hours confirmed clinically by a neurologist, or (c) peripheral thromboembolism verified clinically by an internist or cardiologist, and radiographically by either computed tomography or angiography; and 2) the exclusion of other identifiable causes for the TE, with: (a) ultrasound of cerebral arteries and aorta, (b) 24-hour Holter-electrocardiography, (c) 24-hour blood pressure measurements; and 3) the presence of a PFO with or without atrial septal aneurysm (ASA) with spontaneous or inducible right-to-left shunting during contrast transesophageal echocardiography or transthoracically, if clearly visible. A small shunt volume was defined as 3–20 bubbles, and a large shunt volume as > 20 bubbles passing the PFO into the left atrium. An ASA was defined as an interatrial septum of abnormal mobility with protrusion of the septum into the left or right atrium of at least 10 mm beyond baseline. Informed consent was obtained from all patients or their relatives prior to the procedure. The protocol was approved by the institutional committee on human research.

Device implantation. The implantation procedure for the devices used is similar. Details regarding device implantation and postprocedural treatment were published previously.9 Briefly, following conscious sedation, the occluder was deployed under fluoroscopic and transesophageal echocardiographic guidance. The device was delivered through an 8–12 Fr transseptal sheath (Cook Cardiology, Bloomington, Indiana), depending on the size of the occluder chosen. Before and after release, the position of the device and presence of residual shunt was checked by right atrial contrast injection through the transseptal sheath.

Follow up. Acetylsalicylic acid, 100 mg once daily, was prescribed for 6 months. Clopidogrel, 75 mg once daily, given for 4 weeks, became part of the routine postprocedural care in the year 2000. Prophylaxis against infective endocarditis was recommended for 1 year following defect closure. Patients were instructed to return to our institution or their referring cardiologist after 6–12 months, at which time a chest X-ray, as well as a transesophageal contrast echocardiogram for detection of residual shunting were performed. Residual shunting was graded according to the definition used for evaluation of the PFO before closure. Thereafter, patients were followed annually via telephone contact by a trained staff member. Patients were further instructed to follow up with their referring physicians, who were asked to inform us of any change in clinical status. Recurrent TEs (stroke or TIA) were defined according to the inclusion criteria. Events were either recorded during patient follow up, telephone contact or reported through the referring physician. External data control was provided through prospective submission of device implantation data and follow-up information to the Medical Review Board of the Health Insurance Funds.

Statistical analysis. Continuous data are presented as median with range, or mean with standard deviation (SD), as appropriate. The incidence of ischemic events before and after PFO closure was calculated as the total number of events before and after defect closure divided by the cumulative patient years pre- and post-intervention. Categorical variables were compared by the chi-square test or the two-sided Fisher’s exact test. Continuous data were compared using analysis of variance (ANOVA). Survival free from recurrent TEs was calculated with the Kaplan-Meier method, using the log-rank test to identify potential predictors of recurrent TEs. The follow-up time until the first recurrent TE in each individual was included in the analysis. For those patients without recurrent TEs, the available follow-up time was used for analysis purposes. Variables were chosen based on potential confounders described in the literature, including the presence of any cardiovascular risk factors, defined as either hypertension, diabetes mellitus, hyperlipidemia or smoking habit. The alpha level was set at 0.05. Data were analyzed using SPSS software, Version 13.0 (SPSS, Inc., Chicago, Illinois).

Results

Baseline characteristics. The mean age of the 795 patients was 51 ± 13 years, with the majority of patients (59%) under the age of 55 years (Table 1). Comorbidities were common, with hypertension (242 patients, 30%) and hyperlipidemia (172 patients, 22%) being the most frequently observed. ASAs were present in 46% (362) of patients. Prior to PFO closure, patients had a mean of 1.5 TEs, mainly cryptogenic strokes (785 events, 66%), representing an annual incidence of TEs of 3% before transcatheter PFO closure.

Comparing the three groups, patients in the Cardia group were younger (49 years; p = 0.011) and had less common ASAs (155 patients, 38%; p < 0.0001), compared to the Intrasept (52 years; ASA: 166 patients, 55%) and APO groups (52 years, ASA: 41 patients, 46%). Patients closed with an APO more commonly had an underlying thrombophilia (15 patients, 13%; p = 0.016) compared to the other two occluder groups (Cardia: 35 patients, 9%; Intrasept: 21 patients, 7%).

Procedural data. All 795 devices were implanted successfully with a mean fluoroscopy time of 3.7 minutes (Table 2). The periprocedural complication rate was 1.8% (14 patients) consistent of device embolization, pericardial effusions, transient ST-segment elevations and arrhythmias. No emergency surgeries were warranted and no peri-procedural deaths occurred. No differences in complication rates were seen among the 3 different occluder types.

Follow up and residual shunting. Clinical follow-up information is available for 91% of the patients (721 patients). The mean follow-up duration was 26 months, with longer follow up for the Cardia device due to earlier availability (Table 3). Transesophageal echocardiography 6 or 12 months following the procedure was performed in 583 patients (73.3%). Residual shunting immediately following device placement was seen in 19% of the cases (150 patients) and decreased to 8% (58 patients) during follow up. Patients in the Cardia group more commonly had residual shunting immediately following device placement (95 patients, 24% vs. 41 patients, 14% [Intrasept] vs. 14 patients, 16% [APO]; p = 0.004). However, during follow up, the residual shunt with all three devices was comparable, ranging between 7 and 9% (Residual shunt during follow up: 33 patients, 9% [Cardia] vs. 19 patients, 7% [Intrasept] vs. 6 patients, 7.5% [APO]; p = 0.736).

Recurrent thromboembolic events. During follow up, 24 recurrent TEs in 22 patients were noted, of which 15 were TIAs and 9 were documented strokes (Table 3). This translates into an annual incidence of recurrent TEs of 1.4%. Events occurred between 2 weeks to 42 months following the PFO closure procedure. The annual incidence of recurrent TEs ranges between 1.1–2.5% depending on the occluder (Table 3). In a Kaplan-Meier analysis, device type had no impact on the occurrence of recurrent events (p = 0.419) (Table 4). Among the possible cofactors for recurrent TEs, only new-onset atrial fibrillation (AF) during follow up was a predictor of TEs (Table 4 and Figure 2).

Complications during follow up. Thrombus formation and wire fractures were seen almost exclusively in the first two generations of the Cardia occluder (10/11 patients and 14/14 patients, respectively). Pericardial effusions were noted in 4 patients. Nine patients died during follow up, 7 of whom underwent PFO closure with a Cardia and 2 with an Intrasept. Seven of the deaths were not related to the procedure or occluder placement. Two patients died of cancer, 1 of pneumonia, 1 of an acute coronary syndrome, 1 of an acute aortic dissection, 1 of a pulmonary embolism and 1 secondary to a motor vehicle accident. Occluder placement as a possible cause of death could not be excluded in 2 cases. One patient died from a stroke and 1 of unknown causes.

Two devices were explanted. The first patient had bacteremia following a motor vehicle accident with multiple traumas and the occluder was thought to be the culprit for persistent bacteremia. The second patient experienced 2 recurrent TEs and ultimately underwent device removal and surgical PFO closure during coronary artery bypass surgery for later-diagnosed 3-vessel disease. New-onset AF was noted in 6% of patients (45 patients) and was more common in the group of patients treated with the APO (10.1%, 9 patients; p = 0.057) compared to patients who underwent closure with a Cardia (5%, 20 patients) or Intrasept device (5.3%, 16 patients).

Discussion

This study demonstrates the equivalence of two different designs of PFO occlusion devices with regard to safety, handling and efficacy to prevent residual shunting and recurrent TEs. The differences in design of these devices, a double- umbrella made of ivalon in the case of the Cardia PFO occluder, and Intrasept versus the APO’s double disc of woven nitinol strands, have been thought to be responsible for differences observed in safety and procedural success in a prior smaller study, suggesting inferiority of the Cardia PFO occluder.11

Why are our results different than the ones in an earlier report using apparently identical devices? Most importantly, Schwerzmann et al used early-generation devices of the Cardia PFO occluders. Earlier Cardia PFO devices (generations 1 and 2) had softer nitinol arms mounted on the outside of thicker ivalon sails compared to the third-generation Cardia PFO occluder or the Intrasept. The design of these early versions of the Cardia PFO occluder are responsible for several adverse events such as wire fractures and presumably increased residual shunting rates and, more importantly, device-related thrombus formation.9 However, with further development of the device, these complications have been eradicated. This has also been documented by Braun et al, who compared the Cardia PFO occluder including third-generation devices with the APO, and found no differences in safety or recurrent TEs.12 Comparing procedural characteristics of the Cardia PFO occluder, Intrasept and APO with older devices developed for ASD closure, they are clearly superior; for example, the Sideris buttoned device was plagued by incomplete PFO closure problems, the CardioSeal device has an unacceptably high rate of wire fractures, and the ASDOS occluder is cumbersome to deploy and is associated with a high rate of periprocedural complications.2,13,14

Aside from its safety characteristics, the most important outcome parameter for a PFO occlusion device is its ability to prevent recurrent TEs. Recognizing the various shortcomings of retrospective case series and the lack of prospective, randomized, controlled trials, the annual incidence of recurrent TEs of 1.4%, as seen in our study, appears to compare favorably with patients receiving medical therapy following cryptogenic TEs and underlying PFOs who have a quoted annual recurrence rate of 3.8–12%.15 Comparing the Cardia PFO occluder with the APO, Schwerzmann et al found a strong trend towards more frequent recurrent TEs with the former device.11 Again, our data do not support their finding, since device type had no impact on recurrent TEs according to the Kaplan-Meier analysis. Given the equivalent rates of post-procedural residual shunting and safety features among the tested devices, a difference in rates of recurrent TEs would be an unexpected finding. In fact, a more recent evaluation of the same patient population initially studied by Schwerzmann and coworkers confirms our finding that the type of device has no impact on the occurrence of recurrent TEs.16

Interestingly, of all parameters we evaluated for their potential impact on recurrent TEs, only new-onset AF following the procedure was a predictor of these events. AF is a common arrhythmia following PFO closure and its occurrence is generally associated with TEs.17 The unknown question is whether episodes of AF were truly new in onset and potentially caused by the occluder device, or if the patients had undiagnosed preexisting AF causing the initial TE, assumed to be cryptogenic in nature. Either way, our finding underscores the importance to effectively diagnose and treat AF in the postprocedural period, as this may prevent a large proportion of recurrent TEs. The seemingly higher incidence of AF in the group of patients treated with the APO, although not statistically significant, is worrisome in this context. The interpretation of this finding is difficult. One explanation is that the APO indeed is more commonly associated with AF and one can speculate that the difference in design and stiffness of the APO compared to the Cardia and Intrasept devices may be responsible for the higher incidence of AF. Another possible interpretation of the seemingly higher rate of AF in the APO group is that it occurred by chance due to the small sample size in this group. Support for the latter interpretation comes from the paradoxical finding that the APO group has the lowest recurrence rate of TEs, yet the highest rate of new-onset AF, which was found to be the only independent predictor of recurrent TEs.

Our data support the accumulating evidence that elderly patients, defined as above the age of 55 years, may benefit from PFO closure in the setting of cryptogenic stroke, as age does not seem to have an impact on the recurrence rate of TEs. Although earlier case-control studies showed no statistically significant association with PFOs and cryptogenic TEs in this age group, a recent study clearly demonstrated that there is an association between those two entities.18,19 Hence, age should not be a factor in deciding whether a patient is a potential candidate for PFO closure following a cryptogenic TE.

Although the PFO occluders used in current practice seem safe and effective, several devices are currently in development with early human experiences. Among the next-generation devices are partially or completely bioabsorbable occluders, devices sealing the PFO by delivering a stitch to the interatrial septum or applying radiofrequency energy, and devices that are placed exclusively inside the tunnel of the PFO.20 Whether these devices and technologies are superior to those used in current clinical practice remains to be seen.

Study limitations. Our study has several shortcomings. First, our data are based on a single-center registry and consequently, patients are not randomized to the type of occluder. Second, our follow up is not complete and we may have missed recurrent TEs or other complications during follow up due to the only intermittent telephone contact. Third, residual shunt data are based on TEE evaluations in only three-quarters of the patients, with the remaining data being based on transthoracic echocardiography reports in patients who refused to undergo follow up TEE examination. Finally, our study cannot establish a causal relationship between PFOs and cryptogenic TEs, nor does it answer the remaining question of whether PFO closure is superior to medical therapy in the patient population studied. Nevertheless, the present study shows that PFO closure with contemporary devices is comparably simple and safe and that the annual incidence of recurrent TEs following defect closure is low. Finally, AF is the only identified predictor of recurrent TE, underscoring the importance of aggressive therapy of atrial arrhythmias occurring early after PFO closure.

Acknowledgements. We would like to thank the staff of the catheterization laboratory, M. Möller, S. Lüdke, L. ârnjak, D. Feldeisen, and S. Hergert, for their expertise in performing the intraprocedural TEEs.