Author + information
- Received September 17, 2001
- Revision received March 13, 2002
- Accepted April 3, 2002
- Published online June 19, 2002.
- Martin U Braun, MD*,
- Dieter Fassbender, MD‡,
- Steffen P Schoen, MD*,†,
- Markus Haass, MD†,
- Rainer Schraeder, MD§,
- Werner Scholtz, MD‡ and
- Ruth H Strasser, MD, FESC*,* ()
- ↵*Reprint requests and correspondence:
Prof. Dr. med. Ruth H. Strasser, Department of Cardiology, University of Dresden, Fetscherstr. 76, P.O. Box 95, 01307 Dresden, Germany.
Objectives The present study was conducted to determine the safety of the transcatheter closure of a patent foramen ovale (PFO) in patients with cryptogenic cerebral ischemia and the midterm follow-up of recurrent thromboembolic events after interventional PFO closure.
Background Current therapeutic options for stroke prevention in patients with PFO and a history of thromboembolic events include chronic antithrombotics and more invasive treatments such as surgical closure or minor invasive transcatheter permanent closure of the PFO. Promising preliminary and pilot data with the Amplatzer Septal Occluder or the PFO-Star Occluder have been reported. Systematic and long-term data are still missing.
Methods A total of 276 consecutive patients with a PFO and a history of at least one thromboembolic event were recruited in four medical centers and underwent percutaneous PFO closure with the PFO-Star device. Follow-up data were analyzed over an average of 15.1 months, equivalent to 345 patient-years.
Results The implantation was successful in all 276 patients. Peri-interventional reversible complications included transient ST-segment elevations (1.8%) and transient ischemic attack (TIA) (0.8%). Two devices have been removed surgically. During follow-up the annual recurrence rate of thromboembolic events was 1.7% for TIA, 0% for stroke and 0% for peripheral emboli.
Conclusions Interventional PFO closure with the PFO-Star device appears to be a reliable and promising technique resulting in a low recurrence rate of thromboembolic events, especially stroke in patients with a history of cryptogenic ischemia presumably due to paradoxical embolization. To our knowledge, this is the largest coherent and prospective study for interventional PFO closure.
Patients with cryptogenic transient ischemic attack (TIA) or ischemic stroke revealed a high prevalence of a patent foramen ovale (PFO) in up to 40% to 50% of cases (1), suggesting the PFO as a potential risk factor for ischemic stroke due to paradoxical embolism. The incidence of recurrent thromboembolic events in these patients is variable and ranges between 0% and 14% (2,3). Three large and recent studies demonstrated the long-term prognosis of nonselected patients with PFO and stroke, reporting a stroke recurrence rate of 1% to 2%/year, whereas the cumulative annual risk for stroke or TIA was 2% to 4% (4–6). An atrial septal aneurysm (ASA) has been identified as another source for cardiogenic embolism and is frequently associated with a PFO (4,7).
Therapeutic options for secondary stroke prevention in patients with a PFO include two major strategies: 1) the conservative strategy of long-term medical treatment with antithrombotic therapy (platelet antiaggregating drugs) or oral anticoagulation, and 2) the invasive strategy with surgical or interventional closure of the atrial septal defect (8). A variety of studies have shown that the recurrence rate of thromboembolic events in patients with a PFO receiving oral anticoagulation is lower than in patients treated with aspirin (3). However, major bleeding complications of oral anticoagulation correlate with the intensity and duration of anticoagulant therapy (9). Surgical closure of the PFO in patients with paradoxical embolization has been reported by Homma et al. (10)and Ochsenfahrt et al. (11). However, sufficient follow-up data on large patient populations for surgical intervention in preventing stroke recurrence are still lacking.
The percutaneous PFO closure using self-expanding double disc devices such as the Amplatzer Septal Occluder or the PFO-Star Occluder may provide an alternative technique to prevent paradoxical embolism (12,13). The present study analyzes 276 consecutive patients with a PFO and a history of cryptogenic cerebral ischemia for up to 34 months after transcatheter PFO closure.
Patients and methods
From April 1998 until February 2001, 276 consecutive patients with a PFO and at least one documented TIA or stroke underwent transcatheter closure of the defect according to the study protocol approved by the local ethics committee. All patients gave written informed consent.
Patients with the following criteria were included: 1) presence of a PFO with spontaneous or provokable right-to-left shunting as confirmed by contrast transesophageal echocardiography (TEE), or 2) history of a cryptogenic cerebral ischemia (TIA or stroke) confirmed clinically by a neurologist. A TIA was characterized as a transient and reversible neurologic deficit lasting <12 h, whereas a stroke lasted ≥24 h and might result in a residual neurologic motor or sensory disability. All patients received cerebral imaging either by cerebral computer tomography (CT) or magnetic resonance imaging (MRI). Patients with any other identifiable cardiovascular thromboembolic risk were excluded, such as the following: 1) arteriosclerotic plaques in the ascending aorta or extracranial arteries as determined by sonography; 2) cardiac arrhythmias such as atrial fibrillation as determined by history, 12-lead electrocardiogram (ECG) or 24-h Holter ECG; 3) prothrombotic coagulation disturbances as determined by extensive coagulation blood tests including protein C and S, antithrombin III, fibrinogen, antiphospholipid antibodies and APC resistance; and 4) all patients with other indications for oral anticoagulation, such as severe decreased left ventricular systolic function or an implanted prosthetic valve.
Degree of right-to-left shunt
The degree of interatrial shunting across the PFO was determined by multiplane TEE (HP Sonos 5500, Hewlett Packard, Palo Alto, California). The maximum number of microbubbles seen in the left atrium in any single frame after intravenous contrast injection during Valsalva maneuver was used to quantify the shunt size. A “small” degree of shunt was defined as 3 to 20 bubbles and a “large” degree as >20 bubbles. According to Mügge et al. (7), 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 the baseline.
PFO-Star device and the implantation procedure
The PFO-Star is a self-locating and self-expanding device consisting of two Ivalon-square discs (Fig. 1, left). Each umbrella has a diameter of 18 mm, 22 mm, 26 mm, 30 mm or 35 mm and is expanded by four or six nitinol arms. Three different generations of PFO-Star devices (generation I to III) were used in this study, representing technical progress in the development of the device system. Generation I devices were constructed with 2-mm center posts, 2-mm-thick foam Ivalon sails and titanium protective end caps. Generation II introduced new 3-mm and 5-mm center posts to accommodate variations in septal thickness. Thickness of the Ivalon sail was reduced to improve endothelization of the implanted device and to allow the use of smaller delivery sheath sizes (10F). Generation III devices were constructed with the left-side Ivalon attached on the outside of the frame.
Interventional procedures were performed under local (96.4% of patients) or general (3.6% of patients) anesthesia. Venous access was gained via the right femoral vein and the PFO was passed with a standard exchange wire using a 6F-multipurpose catheter under fluoroscopic and TEE guidance. All patients were anticoagulated with 10,000 U heparin and received 1.2 g amoxycillin intravenously (IV) and 0.5 to 1.0 mg atropin IV to prevent endocarditis and catheter-induced coronary spasm. The stretched diameter of the PFO was defined by balloon sizing. After the 6F-multipurpose catheter was changed to a 10 to 12 F trans-septal sheath, the selected PFO-Star device was attached to the delivery forceps and loaded into the trans-septal sheath while the system was flushed with saline. The device was advanced in the left atrium. Under fluoroscopic and TEE control the sheath and the partially expanded device were pulled back until contact to the interatrial septum was achieved. In this position the sheath was further pulled back to expand the right atrial disc of the PFO-Star occluder. After correct position of the device was confirmed, it was released from the forceps (Fig. 1, middle). All procedures were performed under multiplane TEE guidance. In some of the later patients magnetic resonance imaging (turbo-spin-echo and gradient-echo-sequence, Magnetom, Siemens, Munich, Germany) was performed during follow-up (after six months) on a 1.5 T whole-body scanner (Fig. 1, right).
Postinterventional treatment included oral aspirin (100 mg once a day) for 12 months, clopidogrel (75 mg once a day) for 6 weeks and low-dose heparin for the first 3 days after intervention. To prevent infectious complications, amoxycillin (0.625 g bid) was given for 3 days. Prophylaxis of infectious endocarditis was performed for six months according to the guidelines of the American Heart Association.
Postprocedural control and follow-up
A transthoracic echocardiography was performed within the first few days after transcatheter PFO closure. All patients were followed up prospectively for up to 34 months. After percutaneous PFO closure, clinical examinations were carried out at 1 month, 6 months, 12 months and every 12 months thereafter, including neurologic and medical examination, a 12-lead-ECG and a contrast TEE (Echovist, Schering AG, Berlin, Germany) at rest and during Valsalva maneuver. The neurologic examinations were performed by neurologists. Patients with suspected recurrent thromboembolic events were reevaluated by their neurologist and additionally by cerebral CT and/or MRI.
Analyses of the data included the description of peri-interventional event rates and complications as well as the complications during the follow-up period. No further statistical analyses were included.
A total of 276 patients with a PFO and a cryptogenic cerebral or peripheral thromboembolic event underwent transcatheter closure with a PFO-Star device. The patients’ characteristics are shown in Table 1. An ASA was present in 62 patients (22%) and a hypermobile septum (as a minor form of an ASA) was present in 31 patients (11%). Cardiovascular arteriosclerotic promoters such as smoking (36%), systemic hypertension (15%), hyperlipidemia (11%) and diabetes mellitus (7%) have been identified.
Implantation procedure of the PFO-Star system
The implantation procedure was performed successfully in all 276 patients (100%). General anesthesia was required in 10 patients (3.6%). Three different generations of the PFO-Star device were used: generation I in 14% (n = 38), generation II in 27% (n = 74) and generation III in 59% (n = 164) of the patients. The sizes of the devices were 18 mm (n = 2), 22 mm (n = 25), 26 mm (n = 94), 30 mm (n = 139) and 35 mm (n = 16). Trans-septal puncture was necessary in one patient because of the short distance of the PFO channel to the anterior mitral valve leaflet. Procedural time ranged from 9 to 48 min (mean: 24 ± 8 min), with fluoroscopy times of 1.4 to 10.2 min (mean: 4.1 ± 2.3 min).
Peri-interventional complications (within 24 h)
Device implantation was performed successfully in all patients. The overall peri-interventional complication rate was 4% (11 of 276 patients); most complications were minor and all were reversible (Table 2, upper part). Four patients developed transient ST-segment elevations in the inferior leads, presumably due to air embolism through the trans-septal sheath during device delivery or to mechanically induced spasm of the coronary arteries. The ECG changes resolved within <3 min. In one patient an additional reversible atrioventricular block III developed, with a slow ventricular escape rhythm requiring temporary pacing. In two patients a reversible peri-interventional TIA was observed. One of these patients suffered a right arm paresis, the other a mild left hemiparesis. These neurologic symptoms completely resolved within 4 h. No irreversible cerebral or peripheral ischemic events related to the implantation procedure occurred. Most complications were observed in the early phase of the study, likely because of the original cumbersome underwater introduction system, which may have produced embolization of small air bubbles. After the introduction system was changed to a more practicable and safer high-pressure flushing-loading system, and additionally through the learning curve of the physicians, peri-interventional complications were significantly reduced. Major complications occurred in 2 of 276 patients (0.8%). In one patient, a 30-mm generation II device embolized in the systemic circulation and was successfully removed with a snare from the bifurcation of the aorta. Dislodgement of the device into the proximal orifice of the PFO channel after the device was released from guidewire forceps has been observed in another patient, who had a relatively long PFO channel. According to repetitive TEEs the position of the device was stable without further significant dislocation. Because of low warfarin compliance, the patient preferred surgical removal of the device 12 days after intervention and recovered without any problems.
The total follow-up time of 276 consecutive patients was 345 patient-years; the mean follow-up interval was 15.1 months (Table 2, bottom part).
Clinical examination of the patients revealed mild unspecific thoracic sensations in a small percentage (9.4%) of all patients within the first four weeks after device implantation. The reason for these symptoms is unknown. However, it has been observed that antiphlogistic drugs such as ibuprofen diminish these symptoms, supporting the notion that postinterventional inflammation might be involved. One patient developed an allergic-like skin reaction of unknown etiology one week after implantation. Although it was not certain that the device was causative, the system was surgically removed and the PFO was closed at the same time. However, the histologic examination of the device-associated atrial tissue revealed no inflammation or infiltration, suggesting that not the device itself but another antigen was responsible for the allergic skin reaction.
Recurrent thromboembolic events
During follow-up, six recurrent TIAs (1.7%/year) were encountered after the PFO-Star device was implanted (Table 2, bottom part; Table 3). Strokes or peripheral arterial embolism have not been observed. All thromboembolic events occurred within the first six months after transcatheter closure of the PFO and were completely resolved within ≤30 min after onset. No thromboembolic event was observed ≥6 months after PFO closure. One patient, who had a minimal residual shunt, suffered a transient aphasia during a hypertensive crisis. Cranial CT scan did not show any new defect and the symptoms resolved completely. In the other five patients, no residual shunt of the closed PFO was found by contrast-TEE at rest and during Valsalva maneuver, suggesting other reasons than the PFO were responsible for these TIAs. In all these cases repeated TEEs and/or fluoroscopic evaluation demonstrated the PFO-Star device in the correct position, with no evidence of dislocation, arm fracture or thrombus formation on the device.
One of these patients still had a positive bubble test, suggesting a right-to-left shunt at a different site than the PFO or the atrial septum. Another patient had a newly discovered arterosclerotic lesion of the right internal carotid artery as possible cause for the TIA. A third patient underwent cerebral angiography after the recurrent ischemic event. An intracerebral arteriovenous fistula was detected and the patient underwent neurosurgical ligation with uneventful outcome. Other potential risk factors for thromboembolic events such as an ASA or a hypermobile septum were only present in one patient with a recurrent TIA. During the follow-up period after interventional PFO closure, the annual risk of recurrent TIA in patients with a PFO alone was 2.1% and 0.9% in those with both PFO and ASA. Cardiovascular risk factors could be identified in all cases with a recurrent TIA (Table 3).
In conclusion, the average annual recurrence rate after transcatheter closure of the PFO with a PFO-Star device was 1.7% for TIA, 0% for stroke, 0% for peripheral embolism and 1.7% for the combined end point of these thromboembolic events in our population.
Complete PFO closure as assessed by TEE with contrast injection during Valsalva maneuver was achieved in 83% after 1 month, in 96% after 6 months and in 99% after 1 year and 2 years. All patients with a residual shunt had a significant reduction of the shunt degree after device implantation, as determined by the number of microbubbles counted in the left atrium (data not shown). One of these patients with a minimal residual shunt suffered a TIA. However, according to our data the presence of a minimal residual shunt after transcatheter closure of the PFO cannot be considered as a risk factor for recurrent thromboembolic events (Table 4).
Device-adherent thrombus formations were detected in eight patients (2.9%), either at the center of the left atrial disc (three patients) or attached to the expanded nitinol arms (five patients) in generation I and II devices. The patients were put on oral anticoagulation initially with intravenous heparin and later with warfarin. Repetitive TEE examinations showed that the thrombotic material on the device was completely resolved in all cases. During the whole observation period none of these patients were symptomatic or suffered from any thromboembolic event.
Arm fractures of generation I devices were noted in 7 out of 38 patients. However, no embolization of broken arms/material was associated with the broken struts. After the introduction of generation II and III devices, only three arm fractures of generation II devices were detected since. None were seen in generation III devices. During the whole follow-up period all patients were clinically asymptomatic and did not suffer a thromboembolic event or sequalae from the fracture.
Detailed analysis of the 12-lead-ECG at rest and the 24-h Holter-ECGs, which were performed in 75% of the patients 6 months after device implantation, revealed no episodes of cardiac arrhythmias such as atrial fibrillation. A supraventricular tachycardia with 10 to 18 beats was observed in three patients. A sinusatrial or an atrioventricular block was not documented in the follow-up.
The present study is based on 345 patient-years and demonstrates in 276 patients with cryptogenic cerebral ischemia the peri-interventional complication rate and midterm follow-up data after transcatheter PFO closure using the PFO-Star device.
Reliability of device implantation
Percutaneous PFO closure with the PFO-Star device was performed with a success rate of 100% and an overall periprocedural complication rate of 4%, with most complications being minor and all being reversible. These included ST-segment elevations in the inferior leads and a TIA in two patients, most likely due to air embolism through the trans-septal sheath during device delivery. The risk of air embolism was significantly reduced after changing the early cumbersome underwater loading to a high-pressure flushing delivery system. Air embolism is a well known peri-interventional complication during transcatheter PFO closure, which in our study appears comparable to or less frequent than other studies (13,14). Irreversible neurologic deficits, retroperitoneal hematoma or catheter-induced pericardial effusion were not observed.
Recurrence of thromboembolic events
During a mean follow-up of 15.1 months 6 patients reported a recurrent TIA after transcatheter PFO closure, resulting in an average annual recurrence rate of 1.7% for TIA, 0% for stroke and 0% for peripheral arterial ischemia. The risk of TIA recurrence was highest during the first six months after interventional PFO closure, whereas no patient suffered a neurologic ischemic event beyond this time point. In all six cases TEE demonstrated a correct position of the device with no (five patients) or a minimal (one patient) residual shunt through the PFO after contrast injection during Valsalva maneuver. Further evaluation with transcranial Doppler ultrasound showed a still detectable right-to-left shunt after injection of saline/air suspension (bubble test) in one case, suggesting a site of shunt between the venous to the systemic circulation other than the PFO, which might be the basis of a recurrent paradoxical embolism. In the others no residual shunt was present. Similar results were found by Windecker et al. (13)and others (14,15), who reported recurrent neurologic ischemic events after successful percutaneous PFO closure (no residual shunt).
Several studies in smaller populations with PFO and cryptogenic cerebral ischemia were performed, reporting the risk of recurrent stroke between 0% and 14% (2,3). However, these data have to be considered with caution, because stroke prophylactic therapy varied among these groups and most data resulted from the retrospective analysis of a relative low number of clinical events and cases. Two large series of nonselected patients with PFO and stroke were performed in the mid-1990s. In the multicenter retrospective French study, 132 patients with a PFO and a history of cryptogenic cerebral ischemia had an annual recurrence rate of 1.2% for stroke and 3.4% for TIA (average follow-up 22.6 months) (5). Similar recurrence rates on medical treatment have been demonstrated by Bogousslavsky et al. (6)in the prospective Lausanne study. In this study, 140 consecutive patients with PFO and TIA (16%) or stroke (84%) were treated with aspirin or warfarin for secondary prevention of recurrent thromboembolic events. During a mean follow-up of three years, the annual risk of having a recurrent stroke was 1.9% versus 3.8% for the combined endpoint of stroke + TIA. A very recent study of Mas et al. (4)showed slightly lower but comparable recurrence rates of thromboembolic cerebral events (TIA + stroke) by ∼2% per year in a larger population. It should be pointed out that the data on the stroke recurrence rate in the present study are not directly comparable to the follow-up results in these three studies, because of: 1) the differences in the patient population, 2) variations in the peri-interventional treatment (such as clopidogrel for six weeks after PFO closure), and 3) the various follow-up periods. However, the average annual recurrence rate of 1.7% for the combined end point of stroke or TIA after PFO closure with the PFO-Star device suggests this technique is a promising alternative in the prevention of thromboembolic events in these patients. Further studies are necessary to directly compare the medial and interventional treatment modalities.
Association of the PFO with an ASA
An ASA has been identified as another source of cardiogenic embolism with a potential risk for ischemic stroke (4,16). Cabanes et al. (17)and Mas et al. (4,5)reported that the coexistence of an ASA + PFO may be an indicator of a higher risk of recurrent thromboembolic events. Another important finding of our study is that only 1 patient out of 62 with an ASA + PFO had a recurrent thromboembolic event after successful percutaneous PFO closure, which is in contrast to the spontaneous course. None of the patients with a hypermobile septum had a cerebral ischemic event after PFO closure. Therefore, these data assume that patients with the coexistence of an ASA + PFO or hypermobile septum + PFO have a special benefit from percutaneous PFO closure, which might be explained by the stabilization of the aneurysmatic atrial septum between the both discs of the device, and the closure of the larger PFO opening associated with an ASA or hypermobile septum.
postinterventional device-associated thrombus formation.
Device-adherent thrombus formation was observed by TEE in eight patients (2.9%), either at the center of the left atrial disc or the left atrial nitinol arms itself. Interestingly, no thrombotic clots were detected on the right atrial disc. Because of sufficient anticoagulation, initially with IV heparin and later oral warfarin, the size of the thrombi decreased markedly and they were not detectable by repetitive TEE examination after a maximum of six weeks. All patients were completely asymptomatic during the whole follow-up period. Possible explanations for the thrombus aggregation on the device surface were: 1) insufficient antiplatelet therapy, 2) a thrombotic potential of the left atrial nitinol material, and 3) prothrombotic factors of unknown origin. Therefore, the initial postinterventional therapeutic regime was empirically changed from aspirin (100 mg/day) alone to a combination of aspirin and clopidogrel (75 mg/day). Additional progress in the development of the devices was made from generation II to generation III, where the expanded nitinol arms were located on the inner side with the atrial foam facing the left atrial cavum, thus preventing exposure of the nitinol material to the left atrium. After these changes, no further thrombotic material was detected by TEE. Thrombotic clots on the device surface after transcatheter closure of an atrial septal defect (PFO and/or ASD) is not a specific complication of the early generations of the PFO-Star device, but have been reported in other devices, including the Sideris button (18), the ASDOS (14)and the DAS-Angel-Wings-device (19). Therefore, an effective prophylaxis or inhibition of thrombus formation on the device surface seems to be very important.
Endothelization of the PFO-Star device
Endothelization of the implanted device by atrial endothelial cells is important to gain complete percutaneous PFO closure. Animal experimental data demonstrated a sufficient cell proliferation, covering the device four weeks after implantation (20). Such endothelization in human heart was observed one to two weeks after implantation of the device in one patient, who underwent surgical explantation of the device, which was slightly dislocated in the PFO channel after release (see peri-interventional complications). However, in human heart the detailed time course of endothelization is not known.
Clinical predictors of stroke recurrence after percutaneous PFO closure
Windecker et al. reported in 80 patients using different occlusion systems that the presence of a postprocedural residual shunt might be associated with an increased risk of stroke recurrence after PFO closure (13). In contrast, the data of the present study demonstrate in a larger patient population that the existence of a residual shunt did not correlate with an increased recurrence rate of thromboembolic events after transcatheter PFO closure. Thus, five patients with a recurrent TIA did not have a residual shunt, whereas only one patient had both a TIA and a minimal residual shunt. As specified, all of the patients with a thromboembolic recurrent event were found to have at least one cardiovascular risk factor.
The present study is based on 276 consecutive patients with a PFO and a history of cryptogenic cerebral or peripheral ischemia who underwent percutaneous PFO closure. The presented data report the transcatheter PFO closure with the PFO-Star device as a reliable and feasible technique with a relatively low thromboembolic recurrence rate during the follow-up period of up to 34 months. However, prospective and randomized clinical trials are necessary to directly compare the interventional PFO closure with other treatment modalities such as antiplatelet therapy, oral anticoagulation and surgical repair of the PFO.
- atrial septal aneurysm
- computer tomography
- magnetic resonance imaging
- patent foramen ovale
- transesophageal echocardiography
- transient ischemic attack
- Received September 17, 2001.
- Revision received March 13, 2002.
- Accepted April 3, 2002.
- American College of Cardiology Foundation
- Hanna J.P.,
- Sun J.P.,
- Furlan A.J.,
- Stewart W.J.,
- Sila C.A.,
- Tan M.
- French Study Group on Patent Foramen Ovale and Atrial Septal Aneurysm,
- Mas J.L.,
- Zuber M.
- Mügge A.,
- Daniel W.G.,
- Angermann C.,
- et al.
- Homma S.,
- Di T.M.,
- Sacco R.L.,
- Sciacca R.R.,
- Smith C.,
- Mohr J.P.
- Ochsenfahrt C.,
- Hemmer W.,
- Oertel F.,
- Hannekum A.
- Chan K.C.,
- Godman M.J.,
- Walsh K.,
- Wilson N.,
- Redington A.,
- Gibbs J.L.
- Windecker S.,
- Wahl A.,
- Chatterjee T.,
- et al.
- Bridges N.D.,
- Perry S.B.,
- Parness I.,
- Keane J.F.,
- Lock J.E.
- Gallet B.,
- Malergue M.C.,
- Adams C.,
- et al.
- Cabanes L.,
- Mas J.L.,
- Cohen A.,
- et al.
- International Buttoned Device Trial Group,
- Rao P.S.,
- Berger F.,
- Rey C.,
- et al.
- Rickers C.,
- Hamm C.,
- Stern H.,
- et al.