Author + information
- Received September 21, 2011
- Revision received November 18, 2011
- Accepted November 24, 2011
- Published online February 1, 2012.
- Charanjit S. Rihal, MD⁎,⁎ (, )
- Paul Sorajja, MD⁎,
- Jeffrey D. Booker, MD⁎,
- Donald J. Hagler, MD⁎,† and
- Allison K. Cabalka, MD⁎,†
- ↵⁎Reprint requests and correspondence:
Dr. Charanjit S. Rihal, Division of Cardiovascular Diseases, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905
Paravalvular regurgitation affects 5% to 17% of all surgically implanted prosthetic heart valves. Patients who have paravalvular regurgitation can be asymptomatic or present with hemolysis or heart failure, or both. Reoperation is associated with increased morbidity and is not always successful because of underlying tissue friability, inflammation, or calcification. Comprehensive echocardiographic imaging with transthoracic and real-time 3-dimensional transesophageal echocardiography is key for characterizing the defect location, size, and shape. For paramitral defects, an antegrade transseptal approach can usually be guided by biplane fluoroscopy, and real-time 3-dimensional transesophageal echocardiography can usually be performed successfully. Alternative approaches to paramitral defects include retrograde transaortic cannulation or transapical access and retrograde cannulation. For oblong or crescentic defects, the simultaneous or sequential deployment of 2 smaller devices, as opposed to 1 large device, results in a higher degree of procedural success and safety because the risk of impingement on the prosthetic leaflets is minimized. Most para-aortic defects can be approached in a retrograde manner and closed with a single device. With careful anatomical assessment, procedural planning, and procedural execution, successful closure rates of 90% or more should be attainable with a low risk of device impingement on the prosthetic valve or embolization.
- aortic regurgitation
- congestive heart failure
- heart valve prosthesis
- hemolytic anemia
- mitral regurgitation
- percutaneous closure
Over 60,000 prosthetic valves are implanted surgically in the United States annually (1), and 5% to 17% of them will develop some degree of paravalvular regurgitation, usually related to tissue friability, annular calcification, or infection (Fig. 1, Online Videos 1 and 2). Surgical repair has been the standard treatment of severe cases but is usually associated with significant morbidity and mortality. Moreover, surgery may not be successful since the original anatomical problems persist. Because of these issues, there is tremendous interest in minimally invasive percutaneous techniques that may allow successful treatment of paravalvular regurgitation without another sternotomy (2–11).
Patients with paravalvular regurgitation may present with congestive heart failure or hemolytic anemia (or both). Diagnosis can be difficult because associated murmurs frequently are soft, and color flow Doppler images may be obscured by annular calcification and mechanical sewing rings. A low threshold for performing a detailed examination with transesophageal echocardiography is frequently essential for making the correct diagnosis.
Because of frequent patient morbidity and increased risk of reoperation, we have favored the less invasive approach of initial percutaneous closure, with surgical repair reserved for patients in whom percutaneous repair cannot be performed or is contraindicated (e.g., active endocarditis, significant dehiscence involving more than one-fourth of the valve ring). This paper reviews the evaluation of patients with paravalvular regurgitation, technical considerations, execution of the percutaneous procedure, and assessment of results.
Most mitral paravalvular defects are crescentic or oblong and often have serpiginous tracks, rather than being cylindrical holes. Thus, we have found that real-time, 3-dimensional (3D) transesophageal echocardiography is key for evaluating these defects because it overcomes the limitations of standard 2-dimensional imaging in capturing the full extent of the anatomy.
We use a simple triangulation method based on anatomic relationships of the atrial septum, left atrial appendage, and aortic valve with anatomically correct terminology (i.e., anterior vs. posterior, inferior vs. superior, lateral vs. medial) to facilitate communication between imagers and interventionalists (Fig. 2). By using orthogonal views on fluoroscopy, we are able to evaluate the position of the steerable catheter system with respect to the valve ring. Some operators prefer to use a clock-face approach; however, the standard clock-face view provided by transesophageal echocardiography (TEE), used in the operating room by cardiovascular surgeons, does not correspond to what is used in the interventional laboratory. For example, the left anterior oblique–caudal projection, which shows the mitral valve en face, is left–right reversed compared with the traditional left atrial surgical view, as described by Mahjoub et al. (12). Whichever method is used to localize the defect, there must be absolute clarity of communication between the echocardiographer and the interventionalist.
Aortic paravalvular (i.e., para-aortic) defects that are located anteriorly can usually be imaged with transthoracic echocardiography or intracardiac echocardiography; however, if they are located posteriorly, TEE is preferred. In either instance, shadowing from the valve prosthesis or support struts may be problematic. One of the greatest challenges with para-aortic defects is estimating the severity of regurgitation (13). Frequently, there is a “garden-hose” effect as blood leaks through small, tight orifices. The resulting strong color flow Doppler signal can fill the entire left ventricular outflow tract even if the regurgitation is mild. Conversely, acoustic shadowing can give the false impression of mild para-aortic regurgitation when the regurgitation is severe. Aortography is useful in these situations, provided that there is no significant associated valvular regurgitation. Intracardiac echocardiography from the right ventricular outflow tract can provide excellent visualization of the prosthetic leaflets and the paravalvular tissues when differentiation between valvular and paravalvular regurgitation is difficult.
We prefer general anesthesia when 3D TEE is required to facilitate patient comfort and safety because these procedures may take a considerable amount of time. Conscious sedation may be used for patients undergoing closure of a para-aortic defect unless a lengthy procedure or extensive use of TEE imaging is expected.
An especially important consideration is radiation exposure during these procedures. Use of standard fluoroscopy at rates of 15 to 30 frames/s with regular dose settings can easily exceed hospital and state dose administration limits. We recommend low-dose settings, which are available on all modern x-ray equipment, and lower frame rates of 7.5 frames/s. This is particularly important if biplane fluoroscopy is used. Simple measures such as these can decrease patient and operator radiation exposure by 70% to 80%.
In our experience, paramitral defects account for over 80% of cases, which is likely related to the severely diseased annulus and friable tissues that elderly patients often have. Multiple approaches to paramitral defects are possible, including antegrade transseptal, retrograde transaortic, and retrograde transapical.
Antegrade transseptal approach
Biplane fluoroscopy facilitates guidance since one is working in a 3D structure (the left atrium) around a circular mitral prosthesis. Since fluoroscopy is a planar imaging technique, each view represents a 2-dimensional projection of a 3D object, which can lead to orientation errors.
We orient the gantries so the right anterior oblique projection shows the sewing ring tangentially (or on its side) and the left anterior oblique–caudal view shows the valve en face (Fig. 3, Online Videos 3, 4, 5, and 6.). Since most affected valves are mechanical tilting-disk valves or bileaflet valves, it is extremely important to have clear, sharp delineation of prosthetic leaflet motion, preferably in both views and with TEE. Transseptal left atrial catheterization is performed under fluoroscopic and TEE guidance with standard techniques. After the left atrium is accessed, the patient is systemically heparinized to avoid intracardiac, valvular, or intracatheter thrombosis during the procedure. The activated clotting time is checked every 30 min to ensure an adequate level of anticoagulation. The target activated clotting time varies depending on the locally available assay but should be kept in the same range as that used for percutaneous coronary intervention.
We use a telescoping coaxial system consisting of a deflectable left atrial sheath (e.g., Agilis NxT Steerable Introducer, St Jude Medical, Maple Grove, Minnesota), a 100-cm 6-F coronary guide (e.g., typically the multipurpose curve), a 5-F 125-cm multipurpose diagnostic catheter, and an exchange-length, extrasupport, angled, hydrophilic 0.035-inch wire (Glidewire, Terumo Medical Corp., Somerset, New Jersey) (Fig. 4). However, additional curved-tip catheters may be used, depending on the location of the defect (e.g., JR4 or Berenstein, Angiodynamics, Latham, New York). The system is loaded in a telescoping, coaxial fashion and steered with the deflectable left atrial sheath, which can be flexed, rotated, and translated in 3 dimensions such that the entire mitral valve ring can be quickly and accurately explored. Additional steerability is achieved using the angle-tip wire. After the quadrant in which the defect is located is selected, finely tuned movements are used to localize the defect. The catheter and wire typically appear in 3D TEE views of the defect, which can then be crossed with gentle probing of the wire supported by the 5-F diagnostic catheter. Because the wire position can be easily lost through sudden release of tension or motion, we loop the wire in the left ventricle and exit it out the aortic valve into the descending thoracic aorta. We cross the defect first with the 5-F diagnostic catheter and then steadily advance the 6-F coronary guide catheter into the left ventricle. In most cases, these catheters cross without excessive force. If either catheter does not cross, special techniques, such as guidewire snaring, may be required (as discussed later).
After the left ventricle is accessed with a 6-F guide catheter, a decision needs to be made about whether to use single- or multiple-device occluders. For smaller, rounder defects (i.e., typically those that cause hemolysis), a single device usually suffices. We prefer the Amplatzer Vascular Plug II (AVPII) (AGA Medical Corp., Plymouth, Minnesota). This device is low profile, multisegmented, and constructed of highly flexible, fine nitinol mesh, facilitating its deliverability through various catheters. An AVPII up to 12 mm will fit without excessive friction within a 6-F coronary guide catheter. Other devices, such as the Amplatzer Duct Occluder, Amplatzer Septal Occluder, and Amplatzer Muscular VSD Occluder (all manufactured by AGA Medical Corp.) can be used. However, these devices are woven from a larger-caliber nitinol mesh, resulting in a stiffer device with a higher profile and may be associated with a greater risk of accelerating hemolysis. For single AVPII device placement, the occluder is loaded into the 6-F guide catheter with its distal third extruded into the left ventricle, followed by careful withdrawal of the entire assembly toward the mitral valve. If normal prosthetic leaflet motion is impaired, the entire assembly should be re-advanced into the left ventricle, and a decision must be made to try a smaller device or to redeploy the device closer to the atrioventricular plane. Anterolateral defects close to the left atrial appendage require special care because they frequently have a serpiginous superior-to-inferior orientation because of massive left atrial dilation, with the inferior (or left ventricular origin) of the defect located closer to the valve prosthesis. After deployment, tilting of the device may obstruct normal leaflet motion of the anterior tilting-disk valve.
With crescentic or oblong defects, we favor a multiple-device technique (Fig. 5). The preferred approach for delivering multiple devices is a simultaneous technique. In this approach, 2 0.032-inch Amplatz ExtraStiff Guide Wires (Cook Medical, Bloomington, Indiana) are inserted into the 6-F guide catheter and advanced to make gentle loops in the left ventricle (Fig. 6). The entire catheter assembly is then withdrawn from the body, leaving only the 2 guidewires across the defect in the left ventricle. Over these separate guidewires, 2 separate delivery systems, each consisting of a 6-F multipurpose guide catheter with 5-F multipurpose diagnostic catheter, are loaded and advanced into the left ventricle. After the guidewires and 5-F catheters are removed, devices (typically 8- to 12-mm AVPII devices) can be advanced through the 6-F guide catheters simultaneously.
An alternate method is to deploy the devices sequentially. In this approach, the Terumo hydrophilic guidewire is snared and exteriorized (Fig. 7, Online Video 7). All catheters are removed and replaced with an 8-F Flexor Shuttle sheath (Cook Medical). With the arteriovenous guidewire rail in position, the first closure device can be placed through this sheath, alongside the existing guidewire rail. After placement of the first device, the sheath is removed from over the delivery cable of the first device and then replaced over the existing loop of the arteriovenous guidewire rail, leaving the first device tethered on its cable. A second, or even third, device can then be placed with use of similar techniques. Downsizing to a 6-F or 7-F Shuttle sheath may be required for crossing with more than 1 device in place. After efficacy and prosthetic valve leaflet function are assured by 3D TEE and fluoroscopy, all devices can be released. The sequential technique is particularly useful for very large paravalvular defects (about 25% of the sewing ring), which can be closed more effectively with staggered “nesting” of the devices.
Paramitral defects may also be cannulated retrogradely through the aortic valve with a retroflexed diagnostic catheter, such as an Amplatz curve or left coronary bypass curve, to seek the quadrant of the defect. Having crossed the defect retrogradely, the wire is snared in the left atrium and exteriorized, and an arteriovenous loop is formed. Over this loop, the delivery system of choice can be placed (e.g., a 6-F coronary guider or a hydrophilic Shuttle sheath) in an antegrade, transseptal manner. Closure devices can then be placed in the manner described in the preceding text.
An additional approach is to cross the defects retrogradely through a transapical puncture. This approach may be preferred when the defect is located medially, where catheter manipulation and cannulation of the defect from a transseptal approach can be quite challenging. After apical access, defects can be crossed in a retrograde manner under fluoroscopic or echocardiographic guidance. Image overlay from pre-procedural computed tomography may be very useful for crossing, as described by Jelnin et al. (14). These techniques, however, are not in widespread use and may not be available at all centers performing these procedures.
An alternative is to perform a hybrid procedure with a small thoracotomy to expose the apex and repair it under direct vision. However, this adds considerable complexity to the procedure and requires operating room time.
Most para-aortic defects can be approached retrogradely, with transthoracic echocardiographic and fluoroscopic guidance. Intracardiac echocardiography also is very helpful when the probe is positioned in the right ventricular outflow tract just adjacent to a para-aortic defect. Because many of these defects are located anteriorly, the straight lateral fluoroscopic view is used with the right anterior oblique gantry positioned to provide an orthogonal view of the prosthesis. As for paramitral defects, excellent visualization of mechanical leaflets, if present, must be ensured.
We cross these defects in a retrograde manner with a coaxial system consisting of a 6-F coronary multipurpose guide that contains a 5-F 125-cm coronary multipurpose diagnostic catheter and a 0.035-inch extrasupport, angled Terumo guidewire. After the quadrant of the defect is localized, the defect is cannulated with a wire and the 5-F and 6-F telescoped catheters are advanced sequentially into the left ventricle. Rarely, the 5-F catheter will cross, but the 6-F will not. When that happens, we replace the guidewire with a stiffer 0.035-inch wire, such as an Amplatz ExtraStiff Guide Wire (Cook Medical) or Super Stiff wire (Boston Scientific, Natwick, Massachusetts), with care to pre-form a loop on this wire to facilitate positioning in the left ventricular apex and diminish the risk of perforation. A stiffer wire loop usually allows retrograde crossing with an appropriately sized delivery catheter. Use of a 90-cm 6-F hydrophilic Shuttle sheath (Cook Medical) is another good option if crossability is an issue. We have not found sheath length to be a limiting factor in this setting, so we have not adopted a radial approach for this application.
Para-aortic defects are typically smaller than paramitral defects and can usually be closed with a single device. If >1 device is judged to be likely, we use an anchor wire technique, in which we place a stiff wire guide into the left ventricle with a curve shaped to the size of the left ventricle. Over this wire, an 8-F Shuttle sheath (Cook Medical) can be placed and devices delivered sequentially as described for the treatment of paramitral defects.
Compared with performing the procedure in structurally normal hearts, performing transseptal puncture in patients with chronic valve disease and multiple prior cardiac operations can be difficult and demanding. The interatrial septum has often been sewn or patched, and fibrosis and even calcification are frequently encountered. For the majority of paramitral lesions, the exact location of the transseptal puncture is not an issue. We typically select the thinnest portion of the septum that is visible by TEE. Use of a deflectable left atrial sheath, as described in the preceding text, generally allows for rapid cannulation of the paramitral defects, no matter where the transseptal puncture is located. The one exception to this general rule is the presence of a medial paramitral defect. These defects are immediately adjacent to the interatrial septum, and a transseptal puncture too close to the defect will not leave enough room for maneuvering catheters and wires. Therefore, we prefer a more posterior puncture in this instance—at least 4 to 6 cm from the medial paramitral defect. The left atrium in these patients is frequently huge and extends well into the right side of the chest. A transseptal puncture that is oriented toward a softer, less fibrotic portion of the interatrial septum posteriorly, or slightly to the right, is frequently required for paramitral defects. TEE imaging and, occasionally, injection of contrast dye into the pulmonary artery with levophase follow-through can help make this a safe puncture. Other ancillary techniques we have found useful in difficult transseptal punctures include using the 135-cm SafeSept Transseptal Guidewire (Pressure Products, San Pedro, California) or inserting the back end of a coronary guidewire through the Brockenbrough needle if it does not cross through the interatrial septum. Application of cautery to the transseptal needle, or use of radiofrequency ablation, can also facilitate crossing.
A second useful technique to facilitate paravalvular leak closure is the creation of an atrioventricular loop through snaring the guidewire. Standard gooseneck snares (Amplatz GooseNeck snare, Microvena, St. Paul, Minnesota) facilitate this because they can be rotated, and this technique can be quickly learned by experienced interventional cardiologists. Creation of an atrioventricular loop allows complete control of the catheters, which can be positioned either antegradely or retrogradely, and allows for sequential device deployment as described in the preceding text.
Left ventricular apical puncture can be used for primary access for closure or for wire exteriorization should 2 mechanical valves be in place (15). A left coronary angiogram is useful to delineate the course of the left anterior descending coronary artery and its branches because these diseased hearts are frequently dilated, spherical, and rotated so the left anterior descending coronary artery and its branches are not in the usual position in the chest. If a small-caliber catheter or sheath (i.e., 4-F or 5-F) is used, the sheath can typically be removed without risk of intrathoracic hemorrhage. However, if a larger catheter or sheath (i.e., 6-F or 7-F) is required for delivering a device or snaring a wire for exteriorization, removal of the catheter can be complicated by bleeding. Jelnin et al. (14) have reported their successful experience with apical closure using Amplatzer devices, and this technique should be considered to minimize the risk of bleeding and hemothorax from the apical puncture site.
Finally, we have found that a large-bore 20-F venous sheath (DrySeal Sheath, W. L. Gore & Associates, Flagstaff, Arizona) with an inflatable diaphragm is useful to facilitate multiple simultaneous catheters while minimizing bleeding from the venous access site. The large-bore sheath can be removed at the end of the procedure and the access site managed with a subcutaneous figure-of-8 suture (16).
Obstruction of mechanical tilting-disk valve leaflets is a feared complication. Tilting-disk valves may be pulled shut during device deployment, which can be recognized immediately if the fluoroscopic planes are set up correctly (as discussed earlier). Obstruction may be relieved by reversing the deployment maneuvers and pushing the delivery catheter and device back into the mid left ventricle. Devices may also obstruct prosthetic valve leaflets during systole and prevent proper closure. This complication can be more difficult to recognize, and frequently the only evident sign is an abrupt increase in valvular regurgitation that is greater than the normal closing volume.
Rarely, a device may tilt after deployment and block the prosthetic leaflet. In this instance, 2 options are available. The first is to use a long, flexible, bioptome through the steerable left atrial sheath. The second option is to snare the distal portion of the device and pull it antegradely through the defect. If it cannot be removed, surgery is the only alternative.
Devices may also embolize from the heart; typically, they go past the carotid arteries and lodge within the common iliac arteries. These devices can be removed by placing an 8-F to 10-F sheath in the common femoral artery and either snaring the device or retrieving it with a bioptome. The device may also be stretched by a proximal and distal snare to allow it to be removed through smaller French catheters.
Coronary artery obstruction may occur when para-aortic defects are treated, because devices may protrude over the ostia of the coronary arteries. Various imaging planes may be necessary to demonstrate occlusion or clearance because the 2-dimensional fluoroscopic image may cause significant overlap between devices and coronary ostia (Fig. 8). Because the circumflex coronary artery runs in the left atrioventricular groove, placing devices around the posterolateral aspect of the mitral annulus could theoretically result in coronary obstruction, although we have not experienced this complication. It has occurred in our experience, though, during closure of a large left ventricular pseudoaneurysm.
Finally, stroke or transient ischemic attacks may result from systemic thromboembolism, either during the procedure (0 patients in our experience) or after the procedure and related to transition of anticoagulation regimens (2 patients in our experience). Careful attention to catheter flushing and adequacy of heparinization are key in preventing these complications.
Assessment of Results
The therapeutic endpoint should be determined before initiating the procedure. If the procedure is being done primarily for treatment of heart failure, any marked reduction in regurgitant volume is likely to help the patient. However, if the procedure is done primarily to treat hemolysis, complete obliteration of regurgitation is usually necessary for efficacy. Complicating matters is the impossibility of determining which specific leak is responsible for the hemolysis. Should residual regurgitation be present, color flow Doppler assessment can be extremely difficult because of splaying of the jet through the residual leak and around the first device. Measuring left atrial pressure does not usually help because the pressure typically does not decrease immediately. We typically do not perform left ventricular angiography to assess results.
Our results are based on 141 defects that were closed in 115 patients (mean age: 67 ± 12 years; 53% men) (17). Most defects were paramitral (78%); the rest were para-aortic. Severe heart failure was present in 93% of patients, and some degree of hemolysis in 37%. Nineteen patients had multiple treated defects. Snaring and wire exteriorization was used in 29 patients (25%); left ventricular apical puncture was required in 13. The technical success rate for device deployment was 89%, which is similar to that reported by Ruiz et al. (18) for 44 treated patients (89%). Mean procedure time was 147 ± 54 min, but it decreased considerably with increasing experience. In 90% of cases, regurgitation severity decreased considerably to 2+ or less; in 77% of cases, severity decreased to 1+.
Several procedural failures occurred early in our experience and were related to prosthetic leaflet impingement in 5 patients and the inability to cross defects in 3 patients. Severe residual regurgitation persisted despite device placement in 11 patients. In 3 patients, devices had to be retrieved: 1 device was malpositioned in a prosthetic valve strut in the left ventricle, and 2 devices (both duct occluders) had embolized to the common iliac arteries. Various devices were used; the majority were AVPII occluders (63% of devices placed). The total 30-day complication rate was 8.7%. Complications included 1 sudden death 28 days after a successful procedure; 1 emergency surgery for prosthetic leaflet impingement; 3 strokes, including 1 related to transition of anticoagulation; 2 device embolizations without sequelae but necessitating retrieval; 1 unknown cause of death; and 4 hemothoraces after left ventricular apical puncture. Similar types and incidence of complications were reported by Ruiz et al. (18).
These procedures involve a significant learning curve. Procedural time, radiation dose, and technical results have improved with more experience. Importantly, acute technical success is a determinant of long-term outcome. In our study of 126 patients, 72% of survivors were free of severe symptoms or need for surgery at 3-year follow-up (19). However, those with moderate or severe residual regurgitation had poorer survival free of death and severe symptoms (30.3%) compared with those who had no (63.3%) or mild (58.3%) residual regurgitation (p = 0.01). Moreover, symptom improvement occurred only in patients who had no or mild residual regurgitation. As expected in this high-risk population, noncardiac morbidity was responsible for one-third to one-half of deaths at follow-up. Nevertheless, long-term clinical efficacy is highly dependent on residual regurgitation.
For accompanying videos, please see the online version of this article.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- Amplatzer Vascular Plug II
- transesophageal echocardiography
- Received September 21, 2011.
- Revision received November 18, 2011.
- Accepted November 24, 2011.
- American College of Cardiology Foundation
- Hourihan M.,
- Perry S.B.,
- Mandell V.S.,
- et al.
- Nietlispach F.,
- Johnson M.,
- Moss R.R.,
- et al.
- Zamorano J.L.,
- Badano L.P.,
- Bruce C.,
- et al.
- Sorajja P.,
- Cabalka A.K.,
- Hagler D.J.,
- Rihal C.S.
- Ruiz C.E.,
- Jelnin V.,
- Kronzon I.,
- et al.
- Sorajja P.,
- Cabalka A.K.,
- Hagler D.J.,
- Rihal C.S.