Focus Reviews on Cardiovascular Intervention
Transcatheter Aortic Valve-in-Valve Implantation for Patients With Degenerative Surgical Bioprosthetic Valves
2015 Volume 79 Issue 4 Pages 695-703
Details
2015 Volume 79 Issue 4 Pages 695-703
Bioprosthetic tissue valves are increasingly utilized during surgical aortic valve replacement. These valves have limited durability and many fail with time, resulting in stenosis, regurgitation, or both. Repeat cardiac surgery has been the standard of care for patients with failed bioprostheses. Transcatheter valve implantation inside failed surgically implanted bioprostheses (valve-in-valve) is a new less invasive alternative to repeat surgery. We review the potential and challenges of valve-in-valve implantation in patients with failing surgical aortic bioprostheses. (Circ J 2015; 79: 695–703)
Bioprostheses are less thrombogenic than mechanical valves and do not require anticoagulation. As a result, bioprosthetic valves are being increasingly utilized in patients who are elderly or at increased bleeding risk. As a consequence, the great majority of surgical valves currently being implanted in the USA are bioprostheses.1 Tissue valves, however, have limited durability and most fail within 10–20 years.2–7 With the growth in bioprosthetic valve utilization it is reasonable to anticipate an increase in patients with failed bioprostheses.
The standard of care for failed bioprosthetic valves has been open surgical valve replacement. Reoperation, however, can result in considerable morbidity and mortality.8–10 Transcatheter aortic valve implantation (TAVI) is a less-invasive alternative to open heart surgery in selected patients with aortic valve disease.11 TAVI has been shown to be superior to medical treatment for severe aortic stenosis and is associated with similar, if not better, clinical results in many patients when compared with surgical aortic valve replacement.12,13 Recently, the feasibility of transcatheter heart valve (THV) implantation within failed surgically implanted bioprosthetic valves (valve-in-valve, VIV) has been demonstrated.14–18
Tissue valves are generally fashioned from bovine pericardium or whole porcine aortic valves attached to a support structure, such as a stent or a frame. Less commonly, cadaveric human aortic valves (homografts) are utilized. Bioprostheses are generally preserved in glutaraldehyde to reduce antigenicity and various treatments to inhibit calcification.19 Surgical bioprostheses are most commonly stratified into stented and stentless valves (Figure 1A). The frame of stented bioprostheses is a composition of alloy or polymer materials and is largely responsible for its unique fluoroscopic appearance (Figure 1B). The frame is attached to a basal ring that may be circular or scallop shaped. The basal ring is covered with a fabric sewing cuff that facilitates suturing to native tissue. Stentless surgical valves are often fashioned from porcine or human aortic root tissue. Although stentless valves have better postoperative hemodynamics than most stented tissue valves, they have not been associated with better durability.20,21
(A) Common bioprosthetic surgical valves. Reproduced with permission from Webb JG, et al. Circulation 2013; 127: 2542–2550.51 (B) Fluoroscopic images of common stented bioprosthetic valves: I, Perimount (Edwards Lifesciences); II, Mosaic (Medtronic); III, Mitroflow (Sorin); IV, Carpentier-Edwards porcine (Edwards Lifesciences); V, Epic (St. Jude); VI, Trifecta (St. Jude); VII, Hancock (Medtronic); VIII, Freestyle stentless (Medtronic).
Bioprostheses are implanted in the plane of the annulus (intra-annular) or above that plane (supra-annular), allowing for a larger orifice. Most commonly surgical valve leaflets are attached to the internal aspect of the stent posts, but some valves have externally mounted leaflets: these include Mitroflow (Sorin) and Trifecta (St. Jude Medical). The majority of surgical valves in the recently reported global VIV registry were stented (80%): Perimount, Magna and porcine Edwards valves (35%), Mitroflow (24%), Hancock (15%) and Mosaic (12%) were most common.14 The most common stentless valves included in the global VIV registry were: homografts (31%), and Medtronic Freestyle (17%), Biocor (11%) and Toronto SPV (11%).14
Structural deterioration of bioprostheses has been documented in 10–30% of survivors at 10 years, and 30–60% at 15 years.4–6,22 Early failure, however, is not rare.23 The potential for earlier failure of surgical valves is evidenced by a relatively short median time to therapy in the global VIV registry of only 9 years (IQR, 6–12 years).14 Mechanisms of failure include stenosis (due to calcification, or less commonly due to pannus or thrombosis) and regurgitation (due to wear and tear, calcification, or infection). In many patients the mode of failure is combined stenosis and regurgitation. In the global VIV registry, approximately 40% of patients had predominant stenosis, while 30% had predominantly regurgitation, and 30% had both important stenosis and regurgitation.14 Stenosis is more commonly associated with pericardial prostheses and small prostheses, while regurgitation is more commonly associated with stentless prostheses.
Extensive in vitro assessments as well as in vivo ovine and porcine models have indicated the feasibility of transcatheter VIV implantation.24–28 VIV procedures have been performed mainly with 1 of 2 THV: Edwards SAPIEN XT and CoreValve. Less frequently, VIV implantation has been performed with Portico (St. Jude Medical), Melody (Medtronic), Engager (Medtronic), JenaValve (JenaValve), SAPIEN 3 (Edwards Lifesciences) THV devices, and others (Figure 2). First in-human feasibility cases were performed in 2007 in Canada and Germany.25,29,30 A preliminary Canadian registry of VIV cases was followed by an Italian registry, 2 German registries, and numerous other small case series and isolated case reports documenting favorable outcomes.14–18,31–42 The Global Valve-in-Valve registry, an industry-independent collaboration, was introduced in 2010 to collate this increasing but widely distributed experience.43 Currently this registry includes more than 1,000 cases collected from >90 centers worldwide, including THV implantations in aortic, mitral and tricuspid failed bioprostheses. In addition, a prospective aortic VIV nested registry incorporated into the PARTNER trial has recently completed enrolment in the USA and Canada, and VIV registry of CoreValve implantations inside failed bioprostheses is underway.
Aortic valve in valve procedures utilizing different transcatheter devices. (A) SAPIEN XT (Edwards Lifesciences); (B) Jena; (C) CoreValve (Medtronic); (D) Portico (St. Jude).
An analysis of aortic VIV procedures with SAPIEN and CoreValve devices from the global VIV registry has been published.14 Patients were at high surgical risk with an average STS predicted risk of procedural mortality with reoperation of 10% (IQR, 6.2–16.1%). The rate of all-cause death 30 days after VIV was 7.6%. Most surviving patients (92.6%), were in New York Heart Association functional class I or II early after the procedure.
Degenerated bioprosthetic valve leaflets are friable and prone to tearing. Fortunately, neurological events were less frequent than anticipated, with a 1.7% rate of major stroke in the global registry; a risk similar to that associated with native valve TAVI. The rates of specific adverse events after aortic VIV procedures are different (either higher or lower) in comparison with native aortic valve TAVI (Table 1). Aortic VIV implantation is associated with lower rates of paravalvular regurgitation, pacemaker implantation, tamponade, and annular rupture. Higher rates of THV malpositioning, coronary obstruction, and elevated post-procedural gradients, however, were observed. Presumably, the sewing ring of surgical bioprostheses protects surrounding structures (aortic root, conduction system) from injury. As a result, annular rupture was not reported to be a consequence of a VIV procedure, and the frequency of pacemaker implantation after CoreValve VIV procedures (12.2%) is lower than that reported with native valve CoreValve implantation.14 In addition, the circularity of these rings seems to assist THV sealing and prevent leaks. No significant change in baseline perivalvular leakage (which was occasionally demonstrated around the surgical valve), however, is expected.
| Less common during aortic VIV |
| Significant paravalvular leak |
| Tamponade |
| Annular rupture |
| Aortic dissection |
| Conduction defect |
| More common during aortic VIV |
| Device malposition |
| Ostial coronary occlusion |
| Elevated post-procedural gradients |
NVR, native valve replacement; VIV, valve-in-valve.
Survival after aortic VIV procedures is closely associated with the characteristics of the surgical valve treated. Patients with small surgical valves (label size ≤21 mm) and those with predominant stenosis survive less long than those with large surgical valves and those with regurgitation (Figure 3). The dissimilarities between these groups in clinical outcome could be related to difference in clinical presentation, although prosthetic-patient mismatch may not be optimally treated with VIV implantation.
Time to event curves in patients undergoing valve-in-valve procedures. (A) Mechanism of failure: stenosis vs. regurgitation vs. combined. (B) Device size (label diameter): ≤21 mm; >21 and <25 mm; ≥25 mm. Reproduced with permission from Dvir D, et al. JAMA 2014; 312: 162–170.14
Misplacement of the THV device, too aortic or too ventricular, has been more common in VIV, than in native valve, procedures (Figure 4A). The rate of initial THV malposition was 15% in the early report of the global registry.14 This was associated with a relatively high rate of THV retrieval and of a need for a second THV.
(A) Malposition of SAPIEN XT in a Mitroflow bioprosthesis resulting in a need for a second SAPIEN XT implantation. (B) Coronary obstruction after a valve-in-valve procedure. (C) Severe stenosis after valve-in-valve implantation.
Malpositioning during VIV procedures may be related to operator inexperience. A learning curve phenomenon was clearly evident in the global VIV registry data. Malpositioning may also be related to bioprosthetic valve factors, such as the absence of a fluoroscopic sewing ring (eg, Mosaic, Medtronic), minimal fluoroscopic markers (Aspire, Vascutek or Epic, St. Jude Medical), or no fluoroscopic markers (homografts, stentless porcine valves). Degenerated surgical valve leaflets may not be as calcified as stenotic native aortic valves and provide poorer THV fixation. Calcification of stentless surgical valves, such as homografts, may be dramatic within the aortic root but spare the bioprosthetic leaflets.
Optimal positioning requires understanding of the structural and fluoroscopic characteristics of the surgical valve to be treated.44,45 It is of benefit to study images from previous successful VIV procedures in similar bioprosthetic valves using an identical THV device to be used.
Coronary ObstructionCatastrophic obstruction of the coronary ostia is more common after aortic VIV, than native aortic valve, implantation.46 The reported frequency of 3.5% in early reports of the global registry experience is much higher than that associated with native valve procedures, but there is a trend to improved outcomes in more recent reports.14,43 Left main occlusion is much more common, although right coronary obstruction may rarely occur (Figure 4B). Many of these events have been associated with immediate hemodynamic collapse and fatality. Delayed presentation may occur and underestimation of this phenomenon may occur given that coronary obstruction may be incomplete or partially compensate for by bypass grafts.
The mechanism by which coronary obstruction occurs must be understood. Most commonly, the displaced leaflets of the failing bioprosthetic valve come in direct contact with the coronary ostia, or with the sinotubular junction overlying the coronary ostia. Risk factors for coronary obstruction include the underlying patient anatomy, the specific surgical bioprostheses and its manner of implantation, as well as the specific transcatheter valve and its manner of implantation. High-risk anatomical features include low-lying coronary ostia, shallow sinuses, narrow sinotubular junction with low sinus height, prior root repair/replacement, and coronary reimplantation. Factors related to the bioprostheses include supra-annular surgical implantation, tall leaflets, and an internal stent frame (eg, Mitroflow, Trifecta), or the absence of a stent frame (eg, homograft, stentless). Bulky bioprosthesis leaflets also increase the risk of coronary obstruction. Factors related to the THV device may include an extended sealing cuff, aggressive device oversizing and high implantation.
Post-Procedural GradientsGiven that bioprosthetic rings are relatively non-distensible, under-expansion of the THV device can be anticipated.26 While average mean gradients following THV implantation in the setting of native aortic valve disease are frequently 5–15 mmHg, post-procedural VIV implant mean gradients are commonly higher (10–25 mmHg). According to data from the global registry, mean gradients ≥20 mmHg were present after approximately 30% of cases.14 Importantly, high gradients following SAPIEN VIV procedures are inversely related to the size of the surgical valve (Figure 4C). THV implantation in small surgical valves was more often associated with elevated post-procedural gradients ≥20 mmHg after SAPIEN, than after CoreValve, VIV implantation (41% vs. 23%). This difference in hemodynamics, however, was not translated to a difference in 1-year survival. Elevated gradients after CoreValve VIV procedures were more related to implantation depth, with higher gradients associated with low THV implantation (>6-mm depth). It has been suggested that the difference between Edwards SAPIEN and CoreValve hemodynamics, when implanted in very small bioprosthetic surgical valves, could be related to their structural dissimilarity (Figure 5). The leaflets of the CoreValve device are located higher than are those of the Edwards SAPIEN valve. Supra-annular position of the leaflets may allow for a larger orifice than can be achieved with annular leaflets constrained within the bioprosthetic valve ring when implanted intra-annularly. The introduction of small 20-mm diameter SAPIEN XT and SAPIEN 3 THV and a 23-mm CoreValve Evolut may allow implantation in smaller surgical bioprostheses with significantly improved hemodynamics.
Analysis of post-procedural gradients after valve-in-valve procedures according to surgical bioprosthesis size: small (internal diameter <20 mm), intermediate (≥20 mm and <23 mm) and large (≥23 mm). Reproduced with permission from Dvir D, et al. JAMA 2014; 312: 162–170.14
Little is known about the long-term durability of aortic VIV implants. Data from the global registry and the PARTNER trial show that gradients, competency, and functional class are maintained at 1-year follow-up with both CoreValve and Edwards SAPIEN VIV procedures. Durability of both SAPIEN and CoreValve VIV implants has been reported out beyond 3 years.47,48 Our group has documented durability of THV implants in native aortic valves and mitral prostheses out beyond 5 years.49,50 Although encouraging, incomplete THV expansion can be expected to result in perturbation of leaflet mechanics, coaptation, flow dynamics and leaflet/frame contact. It seems reasonable to anticipate a reduction in THV durability in the setting of VIV implants, particularly when THV underexpansion is extreme.
Optimal Evaluation of Candidates for Aortic VIVEvaluation of a candidate for VIV is similar in many aspects to the standard evaluation for potential candidates for transcatheter valve implantation in native aortic valve stenosis. Nevertheless, there are issues specific to VIV procedures (Table 2).51 Characteristics of the prior cardiac surgery, such as supra- or intra-annular valve positioning, root reconstruction, or bypass grafts should be known; a meticulous review of the original surgical report is often helpful. The specific tissue valve and its labeled size must be identified. Unfortunately, the description of surgical valve size is not standardized. Generally the manufacturer’s labeled size approximately corresponds to the outer diameter of the stent frame and is intended to match the surgeon’s estimate of the aortic annulus at the time of operation. The internal dimension of the surgical valve, however, is the relevant measure for VIV sizing. The difference between the label size and manufacturer’s reported internal diameter varies dramatically, from <1 mm to 4 mm. In addition, the reported internal diameter may not include an allowance for leaflet attachments or cuff material, and definitely does not allow for pannus. Manufacturer estimates of bioprosthetic internal diameter can be thought of as offering a theoretical maximum internal diameter. These can be obtained from manufacturer websites, various publications.45 It can be integrated with estimates of internal diameter from transesophageal echocardiography (TEE) and cardiac computed tomography (CT).
| Before the procedure |
| Understand the failed bioprosthetic valve (model, size, structure, position, mode of failure) |
| Exclude predominant prosthetic-patient mismatch of the surgical valve |
| Exclude thrombosis, endocarditis |
| Exclude paravalvular regurgitation |
| Assess risk factors for coronary obstruction |
| Review images of as similar a VIV procedure as possible |
| Identify the fluoroscopic targets for implantation |
| During the procedure |
| Select a THV slightly larger than the internal diameter of the bioprosthesis |
| Consider a THV with supra-annular leaflets (eg, CoreValve) when the surgical bioprosthesis is small |
| No, or cautious, balloon pre-dilation when there is predominant stenosis |
| Use TEE, especially with radiolucent surgical valves |
| Identify a fluoroscopic view that is perpendicular to surgical valve ring and allows assessment of coaxial deployment |
| Implant the THV as high as possible to maximize orifice area |
| Implant the THV low enough to allow fixation on the surgical ring |
TEE, transesophageal echocardiography; THV, transcatheter heart valve. Other abbreviations as in Table 1.
Echocardiography is important to determine the mode and severity of valve failure. TEE should be relatively routine in cases in which regurgitation is predominant in order to exclude endocarditis and paravalvular leakage around the failed bioprostheses. Valve stenosis may be the consequence leaflet degeneration, but may also be the consequence of predominant prosthetic-patient mismatch. Smaller (19- and 21-mm) bioprosthetic valves commonly have much smaller orifice areas and higher gradients than generally appreciated. It is helpful to refer to published tables of expected transvalvular gradients.52 Occasionally, it is difficult to differentiate between degeneration of a small surgical valve and prosthetic-patient mismatch and some patients may have a combination of both. Suspected stenosis of a tissue valve should lead to a more detailed assessment of previous echocardiographic examinations and changes in clinical status with time.
Assessment of the risk for coronary obstruction should include meticulous fluoroscopic and cardiac CT assessment that may identify most patients at risk.53 Evaluation of the distance from the annulus to the coronary ostia, commonly performed in the setting of native valve TAVI, is less relevant when evaluating the risk for coronary obstruction after VIV. Aortic root angiography can be extremely helpful in identifying patients at risk for coronary occlusion, as can semi-selective coronary injections in left anterior oblique cranial projections. The optimal angiographic projection to assess coronary obstruction risk should be perpendicular to both the surgical bioprosthesis and the coronary ostia.
Determining the optimal plane perpendicular to the bioprosthesis can usually be accomplished by finding a fluoroscopic projection where the radiopaque components of the circular bioprosthetic basal ring appear as a straight line or the radiopaque components of the valve posts appear to be at the same height. Finding a projection perpendicular to the coronary ostia is more complex. A simple maneuver that provides perpendicularity to the coronary ostia is the “1-2” technique (Figure 6). The fundamental principle is that surgeons typically implant aortic bioprostheses in a fashion that avoids positioning the commissural posts directly in front of the coronary ostia. The coronary ostium is typically located halfway between 2 posts; consequently a projection perpendicular to a coronary ostium is usually achieved when the 2 adjacent posts are perfectly superimposed. CT angiography is an important tool for assessing the risk of coronary occlusion as well. The anticipated distance of the THV to the coronary ostia can be estimated (virtual THV-coronary distance, VTC). This is best performed by superimposing a virtual ring simulating the diameter of the anticipated, fully expanded THV centered along the geometrical center of the surgical prosthesis followed by a caliper measurement from the ring towards the coronary ostium. Smaller VTC (<3 mm, high risk; 3–6 mm, intermediate risk) may confer an increased hazard for coronary occlusion.
(A) Bioprosthesis posts are aligned in 1-1-1 fashion. Even though the projection is perpendicular to the bioprosthesis, coronary obstruction risk is difficult to define. (B) Semi-selective injection in the left main ostium after aligning 2 posts together (arrow) in 1-2 fashion. (C) Reconstruction of the bioprosthesis position in the root shows that coronary flow will be maintained after valve-in-valve (arrow). Reproduced with permission from Dvir D, et al. Circ Cardiovasc Interv 2015; 8: e002079, doi:10.1161/CIRCINTERVENTIONS.114.002079.53
Aortic VIV procedures have been successfully performed using femoral arterial, axillary arterial, apical, and direct aortic access. Apical access has been advocated due to the proximity of the aortic valve, but there is no evidence to support improved outcomes with this approach and data from the global VIV registry indicates similar malpostioning rates and higher mortality with the non-transfemoral approach. It seems that a least invasive approach allowing for good coaxial catheter engagement and control should be favored. Pre-dilation is usually not necessary in the setting of VIV implantation, particularly in the presence of regurgitation. The argument against pre-dilation is that degenerated bioprosthetic valves are often friable, with an attendant risk of embolization and stroke or disintegration and acute regurgitation. Difficulty in crossing a severely calcified, bulky, stenotic valve, however, can sometimes be encountered. Of note, balloon pre-dilation was performed in almost one-third of cases in the global VIV registry, with very few complications reported.14 Cautious pre-dilation with an undersized balloon may be considered in the presence of a severely calcified and bulky stenotic valve, particularly when a retrograde approach or a self-expanding valve is utilized. Rarely this may be helpful for sizing poorly documented bioprostheses or where pannus may be present.
Optimal positioning requires a thorough understanding of the structural and fluoroscopic characteristics of the specific bioprosthetic device.53 TEE during implantation can be extremely helpful and used for device positioning, particularly with stentless valves and when the surgical valve leaflets are non-calcified or severely regurgitant. Similarly to conventional THV implantation, it is helpful to obtain an appropriate fluoroscopic view perpendicular to the bioprosthetic annular plane. That could be accomplished relatively easily by lining up the bioprosthetic fluoroscopic ring. TEE can be helpful for positioning; particularly with stentless valves, stented valves where the bioprosthetic basal ring is radiolucent, or when the leaflets are non-calcified or regurgitant. Repositioning with self-expanding valves adjusting the depth of implantation may not be possible after initial contact with the bioprosthetic frame. Device depth should be controlled from the onset and rapid pacing should be considered, particularly with regurgitant valves.
There are numerous strategies that should be utilized in patients considered high-risk for coronary occlusion after excluding the option for redo cardiac surgery. These may include utilizing a retrievable THV device (eg, Evolut-R, Lotus), a device with clipping mechanism (eg, Jena, Engager) or simply undersizing/balloon-underfilling of the THV device. Coronary protection with a wire and undeployed stent stationed in the coronary vasculature, and intervention, when needed, can be undertaken in high-risk cases in order to maintain coronary flow and improve clinical outcome. TEE may be helpful in evaluating the presence of coronary obstruction or other causes of procedural hypotension.
Transcatheter treatment of failed aortic bioprostheses in future may become an accepted standard, in preference to repeat surgery. With the increasing use of bioprosthetic aortic valves, VIV implants may become increasingly common. The procedure, however, includes several efficacy and safety concerns, such as elevated post-procedural gradients in the setting of small bioprostheses, a high malposition rate in inexperienced hands especially in stentless regurgitant bioprostheses, and the potential for coronary obstruction. Improved operator understanding of the factors that contribute to these complications may improve patient selection and increase the safety of the procedure.
Danny Dvir is a consultant for Edwards Lifesciences and received research grant and honoraria from Medtronic. John G. Webb is a consultant for Edwards Lifesciences.
