Author + information
- Received October 23, 2004
- Revision received December 18, 2004
- Accepted December 20, 2004
- Published online April 19, 2005.
- Abhay Divekar, MBBS⁎,⁎ (, )
- Tidimogo Gaamangwe, MSc, PEng‡,
- Nasir Shaikh, MBBS†,
- Michael Raabe, MD§ and
- John Ducas, MD†
- ↵⁎Reprint requests and correspondence:
Dr. Abhay Divekar, FE-241, 685 William Avenue, Winnipeg, Manitoba R3E 0Z2, Canada.
Objectives Amplatzer septal occluder (ASO)-associated cardiac perforation (CP) at our institution prompted this retrospective review.
Background Cardiac perforation is a rare complication after transcatheter atrial septal defect (ASD) closure.
Methods To identify CP after transcatheter ASD closure with ASO, cardiac events (CE) describing definite CP, hemopericardium, pericardial effusion, cardiovascular collapse, or sudden death were analyzed. Cardiac events were identified from published literature (MEDLINE), medical device regulating agencies in North America and the European Commission, and AGA Medical Corporation (Golden Valley, Minnesota). Institutional cases were reviewed. Cardiac events were defined as early (pre-discharge) or late (post-discharge).
Results Twenty-nine CEs were identified. Five were excluded because findings were inconclusive for device-related CP. Ten patients were <18 years of age. Late CEs occurred in 66.6%; 25% presented weeks later (longest, three years). Three deaths were reported. Cardiac perforation occurred predominantly in the anterosuperior atrial walls and/or adjacent aorta.
Conclusions Amplatzer septal occluder-associated CP uniquely involves the anterosuperior atrial walls and adjacent aorta. Pathophysiology remains poorly understood.
Transcatheter closure of atrial septal defect (ASD) is now offered as an acceptable alternative to surgery (1,2). Cardiac perforation (CP) is a unique complication of transcatheter ASD and patent foramen ovale (PFO) closure (1,3–8). Incidence varies from 0.1% to 4% for various devices (1,3). The Amplatzer septal occluder (ASO) (AGA Medical Corp., Golden Valley, Minnesota) and implantation technique have been described. The ASO-associated CP at our institution prompted retrospective review.
To identify CP after transcatheter ASD closure with ASO, cardiac events (CE) describing definite CP, hemopericardium, pericardial effusion, cardiovascular collapse, or sudden death were analyzed. Cardiac events were defined as early (pre-discharge) or late (post-discharge).
Retrospective review was conducted of institutional cases including procedural transesophageal echocardiography (TEE) and catheterization data.
A literature search was conducted via MEDLINE using key words: Amplatzer, Amplatzer septal occluder, complications, perforation, erosions, cardiovascular collapse, sudden death, and hemopericardium.
During review, ASO-associated CPs were identified from the U.S. Food and Drug Administration (FDA) website prompting a search of CEs reported to other medical device regulating agencies in North America and the European Commission (websites and/or direct communication).
A 24-year-old woman (case #29, 160 cm, 40.8 kg; Table 1)was the 13th patient undergoing device closure in March 2002. An ASD measuring 15 × 17 mm by TEE with a stretched balloon diameter (SBD) of 23 mm and deficient anterosuperior rim was closed with a 26-mm ASO using fluoroscopy and TEE. The device splayed over the aorta (Fig. 1);there was no residual shunt or impingement on cardiac structures. She was discharged the same evening on aspirin and clopidogrel. Chest radiograph was unremarkable. She was asymptomatic before collapse three days later. After cardiopulmonary resuscitation, emergent surgical exploration found a hemopericardium, and perforations were identified in the anterosuperior right atrial wall and the contiguous right posterior aortic wall. The device was removed, the ASD was closed, and the perforations were repaired. Life support was withdrawn three days later secondary to irreversible neurologic insult. Postmortem examination confirmed intraoperative findings (Fig. 2).Guidewire/delivery catheter-related injury was suggested, citing atraumatic rounded edge of the device and lack of published complications.
A 42-year-old woman (case #8, 160 cm, 56 kg; Table 1) was the 23rd patient undergoing closure. An ASD, measuring 10 mm × 13 mm by TEE with a 17-mm SBD and adequate rims was closed with a 20-mm ASO using fluoroscopy and TEE. There was no residual shunt, impingement on cardiac structures, or splaying over the aorta (Fig. 1). Aspirin and clopidogrel were prescribed. Hospital policy after this incident resulted in hospitalization for three days. Echocardiogram 24 h later showed good device position and no effusion. Three days later chest pain and shortness of breath were associated with an expanding effusion. Emergent surgical exploration found a hemopericardium, and the rim of the device was eroding through the atrial perforation. Anterosuperior right and left atrial perforations were noted at the edge of the respective retention disks (Fig. 3).The device was removed, the ASD was closed, and the perforations were repaired. Recovery was uneventful.
In both patients device integrity was maintained.
Analysis of reported cases
All cases reported to the manufacturer and regulating agencies and published in the literature have been accounted for and reported only once (Table 1). The U.S. FDA and Health Canada were the only agencies who had received reports of CP (Table 2).Only the U.S. FDA had free and easily accessible online information via the Manufacturer And User facility Device Experience (MAUDE) database.
Twenty-nine CEs were identified. Five were excluded because findings were inconclusive for device-related CP. All patients presented with chest pain, shortness of breath, hemodynamic collapse, or sudden death. Ten patients were <18 years of age, and 76% were women.
Among the remaining 24 CEs, 14 had defined CP and hemopericardium, 3 had defined CP and fistula formation, and 5 had hemopericardium only. The ASO size ranged from 12 to 38 mm. Relationship between SBD and ASO size is shown in Table 3.Device malposition or breach in structural integrity was not reported. Information regarding ASD morphology and technical details was not consistently available.
The CEs occurred early in 20.8% and late in 66.6% (unknown in three patients). The CEs presented 1.5 h to 3 years after intervention: 5 within 1 day, 10 within 3 days, and 6 after 3 days (3 weeks to 3 years). Good outcome was reported in 14 patients; 3 patients had neurological deficits, and 3 patients died (unknown in 4 patients).
Sites of CP are shown in Table 4.All CP (except case #4) occurred in the anterosuperior atrial walls and/or adjacent aorta. All patients with CP and hemopericardium had cardiac tamponade; pericardiocentesis was performed in 10 patients. Surgical exploration was performed in 19 of 24 patients. The ASO was explanted in 15 patients and remains implanted in 7 patients (no recurrences reported).
Recent registry review suggests device oversizing and deficient anterosuperior rims as risk factors for CP and recommends abolishing intentional oversizing, stop-flow method for SBD, and careful follow-up of “high-risk” cases (1).
Technique-related CP during catheterization is inherent to the procedure, typically occurs before hospital discharge and is amenable to intervention. In contrast, device-related CP occurs after a technically adequate procedure, frequently after hospital discharge (66.6% in this review), and has the potential for a fatal outcome (1).
The rounded design and flexibility of the ASO is speculated to minimize risk of CP, even when oversized (2). An occlusion device within confines of the anteroposterior septal length is subjected to deformation forces (9). Fatigue fractures are not reported with the ASO, and the disks retain preformed shape; we therefore hypothesize that the ASO transmits deformative forces to the tissues at the point of contact between the aorta, the device, and the anterosuperior atrial walls resulting in CP.
This review shows that the anterosuperior atrial walls and/or adjacent aorta are uniquely vulnerable, the side of the larger ASO disk does not predict site of CP (Table 4), trauma over multiple cardiac cycles can damage even the thicker, more resilient aortic wall, and larger devices are not disproportionately represented (13 devices ≤25 mm, 11 devices >25 mm) in patients with CP. The CP associated with other devices also involves the free/anterosuperior atrial walls and/or adjacent aorta (3,4).
Closure of ASDs with deficient anterosuperior rims and ability to intentionally oversize by splaying over the aorta are among published merits of the ASO (2). Registry review now suggests otherwise (1). Comparison among patients receiving intentionally oversized ASOs with and without complications is necessary. Registry recommendations, although important, need validation. Several points deserve scrutiny. First, CP developed in registry patients (10 of 28) with devices sized equal to the SBD (1). Second, CP did not develop in all patients receiving intentionally oversized devices (2). Third, recurrences are not reported in patients in whom CP developed where devices remain implanted. Forth, unpredictable timing for developing CP makes careful follow-up difficult.
All postmortem reports concluded CP to be non–device-related. With information now available, we believe that our patient clearly had device-related CP. It is possible that cases #27 and #28 (Table 1) with unexplained fresh blood in the pericardial sac, may now be seen in a different light.
Prior to November 2004, there are three reports of ASO-associated perforation or fistula formation (5–7). During the same interval, the MAUDE database reports several cases. Therefore, physicians should be aware of and utilize resources other than traditional literature. Ideally an international registry and periodic end-user notification is necessary.
We acknowledge the limitations of this retrospective review. The MAUDE database is not intended for this purpose. The study was not designed for statistical analysis or causation but focuses on a poorly understood complication.
Informed consent should highlight device-related CP. Symptoms consistent with CP warrant prompt evaluation including urgent echocardiography. Widespread awareness may allow for timely recognition. Prompt pericardiocentesis before surgical exploration may minimize morbidity and mortality. Device removal should be considered in patients presenting with CP. Occluder size relative to anteroposterior septal dimension may be important and needs to be studied.
Understanding the pathophysiology of CP is crucial to the ongoing success of transcatheter therapy.
The authors thank Colin Phoon, MD, NYU Medical Center, New York, New York, and Reeni Soni, MD, HSC, Winnipeg, Canada, for editorial support.
- Abbreviations and acronyms
- atrial septal defect
- Amplatzer septal occluder
- cardiac events
- cardiac perforation
- Food and Drug Administration
- Manufacturer And User facility Device Experience database
- patent foramen ovale
- stretched balloon diameter
- transesophageal echocardiography
- Received October 23, 2004.
- Revision received December 18, 2004.
- Accepted December 20, 2004.
- American College of Cardiology Foundation
- Amin Z.,
- Hijazi Z.M.,
- Bass J.L.,
- Cheatham J.P.,
- Hellenbrand W.E.,
- Kleinman C.S.
- Pedra C.A.,
- Pihkala J.,
- Lee K.J.,
- et al.