Online Exclusive

Evaluation of Endothelialization After Percutaneous Closure of Paravalvular Leaks

Mehmet Özkan, MD, FESC, M. Ali Astarcıoğlu, MD, M. Ozan Gürsoy, MD

Mehmet Özkan, MD, FESC, M. Ali Astarcıoğlu, MD, M. Ozan Gürsoy, MD

To view the videos from these cases on the multimedia page, visit

ABSTRACT: Background and aim of the study. There is limited data regarding the duration of endothelialization following paravalvular leak closure. We aimed to observe the endothelialization process in 2 patients who underwent surgery 6 and 16 months after failed percutaneous mitral paravalvular leak closure, respectively. Methods. Two-dimensional transesophageal echocardiography (2D-TEE) and real-time 3-dimensional transesophageal echocardiography (RT-3D TEE) were utilized to demonstrate mitral paravalvular leaks. The status of endothelialization was explored in the surgery. Results. Two patients underwent percutaneous closure of mitral paravalvular leaks both with 2 occluder devices. The first patient was admitted with dyspnea 6 months later. RT-3D TEE demonstrated a defect around the proximal part of one of the occluder devices. The residual mitral regurgitation was considered moderate to severe by 2D TEE and RT-3D TEE. The patient was referred to surgery in which failed endothelialization of both devices was observed. In the second patient, 2 occluder devices were implanted. He underwent surgery at 16 months due to progressive increase in the severity of mitral regurgitation, which disclosed partially endothelialized closure device. Conclusion. These cases suggest that endothelialization of closure devices may be significantly delayed or even absent for a long time following implantation.

J INVASIVE CARDIOL 2012;24(4):E72-E74

Key words: endothelialization, percutaneous closure, paravalvular mitral regurgitation, real time 3-dimensional transesophageal echocardiography


The use of percutaneous closure devices is growing. On the other hand, our recent knowledge concerning the long-term safety and efficacy of these devices is limited. It is accepted that 6 months of antiaggregation is a long enough period to prevent thrombus formation in patients with atrial septal defect (ASD) and ventricular septal defect (VSD) occluder devices. However, currently there is no consensus of opinion about the healing process following device implantation and the time required to achieve an effective endothelialization of the paravalvular leak closure devices.

Two-dimensional transesophageal echocardiography (2D-TEE) and real-time 3-dimensional transosephageal echocardiography (RT-3D TEE) were utilized prospectively and clinically at 1, 3, 6, 12 months, and every 12 months thereafter to assess and follow-up mitral paravalvular leaks (PVLs). Post-interventional treatment included oral aspirin (100 mg/day) for lifetime and clopidogrel (75 mg/day, at least 1 year) in addition to warfarin. Prophylaxis of bacterial endocarditis was recommended for 6 months according to the guidelines of the AHA/ACC. Endothelialization status was explored during surgery performed after failed percutaneous closure intervention.

Case Report 1

A 33-year-old woman presented to our clinic with a history of progressively worsening dyspnea at rest and anemia. The patient had rheumatic heart disease in childhood for which she underwent mitral valve replacement with a mechanical prosthetic mitral valve at 14-years-old and 2 years later she required reoperation because of endocarditis. 2D-TEE and full volume RT-3D TEE revealed 2 separate severe PVLs. She underwent percutaneous closure of PVLs with 2 occluder devices (Amplatzer vascular plug III, AGA Medical Corporation). However, moderate mitral regurgitation persisted adjacent to one of the devices.

Six months later the patient admitted to our institution with further dyspnea and New York Heart Association (NYHA) Class III Heart Failure. In RT-3D TEE, a dehiscence around the proximal part of one of the medially placed occluder devices was detected. The mitral regurgitation was considered moderate to severe by 2D-TEE. The patient underwent successful reoperation by direct suture without explanting the devices. However, non-endothelialization of both of the occluder devices with bare nitinol wires was observed during operation (Figure 1A, Videos 1 and 2). A serial 1, 3, 6, and 12 months of TEE studies after surgery revealed normal findings of mechanical prosthetic mitral valve and occluder devices. However, serial RT-3D TEE failed to depict the endothelialization of the closure devices.

April 2012 Ozkan video_edited.mp4

Case Report 2

A 42-year-old man underwent combined mitral and aortic valve replacement due to rheumatic heart disease 2 years earlier. He presented 1 year later with pulmonary edema and heart failure with NYHA Class III. On admission, 2D-TEE showed severe mitral regurgitation due to paravalvular leaks as assessed by subsequent RT-3D TEE. Following stabilization of the patient’s clinical status, an Amplatzer vascular plug III was implanted due to a high risk of reoperation. A second Amplatzer vascular plug III was required and placed adjacent to the first under RT-3D TEE guidance, achieving a significant decrease in the degree of regurgitation with a residual mild to moderate mitral regurgitation. A series of TEE follow-up studies were performed 1, 3, 6, 8, 10, 12, 14, and 16 months after device employment, which showed gradual progression of mitral regurgitation (Figure 2). RT-3D TEE was not satisfactory in depicting the endothelialization of closure device. TEE disclosed nonobstructive thrombus measuring 6 mm in length at the edge of closure device at 10-month follow-up, which was successfully lysed with warfarin (INR levels between 3 and 4) in 2 months. Finally, the patient underwent surgery at the 16 months; the left ventricular part of the device was explanted (Figure 1B) and the residual defect was sutured.


This observational study confirmed by the surgical findings suggests that the time course of endothelialization of closure devices, which were used for percutaneous closure of paravalvular defects, may be significantly delayed or even absent for a long time following implantation.

In animal studies and human explants, endothelial cells were observed as soon as 30 days after implantation.1 Protruding parts of the metal framework like the screw thread of the Amplatzer device are the last parts of the septal occluders to endothelialize as shown by scanning electron microscopy and methacrylate resin embedding (after 3-5 months in vivo). According to previous literature reports, chronically persisting inflammatory response has been observed, both in animals and humans. Endothelialization of septal defect occlusion devices after implantation is clinically relevant because superficial thrombus formation may be a possible source for embolism as observed in patient 2. Antiplatelet or anticoagulant therapy is mostly given for 6 months as a primary prevention until neoendothelialization is completed. However, advocated duration of anticoagulation therapy after transcatheter device placement is variable and controversial. In a review of the literature,2 thrombosis was diagnosed at a mean of 5 months after septal closure device deployment and the patients with thrombosed closure device were receiving different treatment regimens: aspirin and clopidogrel, warfarin, heparin, and even aspirin plus warfarin. Therefore it is uncertain whether there is any therapeutic regimen that can effectively prevent the formation of thrombosis. It was suggested that all patients should have early echocardiographic surveillance for device thrombosis. As expected, thrombosis is likely to develop, particularly in the first months because of incomplete endothelialization of the devices that could be disclosed by intensive early surveillance with TEE, as we did in our patients.

The healing response to ASD and VSD occluder devices in humans is largely known. Initial formation of thrombotic material between metal wires and around polyester fibers were demonstrated to be transformed to connective tissue in septal defect occlusion devices. Although cellular organization of fibrin deposits were advanced after 10 weeks already, they were completed only after 24 months.3 In one clinical vignette, a patient who had undergone mitral valve replacement 26 months after closure of an ASD with an Amplatzer device, surgical inspection of both sides of the septal occluder device revealed bare nitinol wires suggesting failed endothelialization.4 In another one, incomplete endothelialization of the explanted Amplatzer ASD occluder was demonstrated even 15 months after device implantation.5 Based on these and our observations, we may suggest giving antiplatelet therapy for 12 months or longer.

No difference in pattern or time course of immunohistochemistry based findings was observed comparing Amplatzer and Cardioseal / Starflex occlusion devices or different sites of implantation (ASD versus VSD).3 However, atrioventricular mechanics are different from that of interatrial flow dynamics; percutaneous closure of paravalvular leaks may respond in a different fashion regarding endothelialization patterns of implanted ASD or VSD occluder devices. The low rate of device explants in humans hinders the potential to evaluate the exact duration and mechanisms of endothelialization in paravalvular closure devices. One of the specifications for an ideal closure device may be replacement of the synthetic covering fabric with a biomaterial that promotes healing response. Decrease in the amount or complete elimination of the metal in the closure devices is another way to enhance them. Acellular, nonimmunogenic, biodegradable, and collagen-based biomaterials may be utilized for facilitation of improved healing process as in recently introduced PFO closure device in a swine model.6

Although the intense monitoring of patients by TEE enables depiction of the devices and thrombi, the 2D- or RT-3D TEE studies were not able to provide information about the endothelialization of the devices.


Our findings raise questions as to the stability of such devices, duration of echocardiographic surveillance for device thrombosis, and documentation of endothelialization of devices, and time required to achieve an effective endothelialization, which necessitates strategies to step up endothelial coverage of such occluder devices for percutaneous PVL implantation. Based on our observation, antiplatelet therapy may be given at least 12 months in patients who underwent PVL closure device implantation. Frequent 2D-TEE monitoring may be necessary in the first 12 months follow-up to disclose device thrombosis. Disappointingly, serial RT-3D TEE was unable to depict the endothelialization of the closure devices.

Acknowledgments. We gratefully acknowledge the technical contributions of Mustafa Yıldız, MD, PhD, and Mehmet Balkanay, MD.


  1. Sigler M, Jux C. Biocompatibility of septal defect closure devices. Heart. 2007;93(4):444-449.
  2. Sherman JM, Hagler DJ, Cetta F. Thrombosis after septal closure device placement: a review of the current literature. Catheter Cardiovasc Interv. 2004;63(4):486-489.
  3. Foth R, Quentin T, Michel-Behnke I, et al. Immunohistochemical characterization of neotissues and tissue reactions to septal defect-occlusion devices. Circ Cardiovasc Interv. 2009;2(2):90-96. 
  4. Astroulakis Z, El-Gamel A, Hill JM. Failed endothelialization of a percutaneous atrial septal defect closure device. Heart. 2008;94(5):580.
  5. Bauriedel G, Skowasch D, Peuster M. Pathology of explanted ASD occluder. Eur Heart J. 2007;28(6):684.
  6. Pavcnik D, Takulve K, Uchida BT, et al. Biodisk: a new device for closure of patent foramen ovale: a feasibility study in swine. Catheter Cardiovasc Interv. 2010;75(6):861-867.


From the Department of Cardiology, Koşuyolu Kartal Heart Training and Research Hospital, Istanbul, Turkey.
Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.
Manuscript submitted September 2, 2011, provisional acceptance given September 27, 2011, final version accepted October 24, 2011.
Address for correspondence: M. Ozan Gürsoy, MD , Department of Cardiology, Koşuyolu Kartal Heart Training and Research Hospital, Istanbul, Turkey. Email: