Author + information
- Received November 3, 2016
- Revision received February 27, 2017
- Accepted March 3, 2017
- Published online May 15, 2017.
- Fadi J. Sawaya, MDa,∗ (, )
- Marcus-André Deutsch, MDb,
- Moritz Seiffert, MDc,
- Sung-Han Yoon, MDd,
- Pablo Codner, MDe,
- Upul Wickramarachchi, MDf,
- Azeem Latib, MDg,
- A. Sonia Petronio, MDh,
- Josep Rodés-Cabau, MDi,
- Maurizio Taramasso, MDj,
- Marco Spaziano, MDk,
- Johan Bosmans, MDl,
- Luigi Biasco, MDm,
- Darren Mylotte, MDn,
- Mikko Savontaus, MDo,
- Peter Gheeraert, MDp,
- Jason Chan, MDq,
- Troels H. Jørgensen, MDa,
- Horst Sievert, MDr,
- Marco Mocetti, MDm,
- Thierry Lefèvre, MDk,
- Francesco Maisano, MDj,
- Antonio Mangieri, MDg,
- David Hildick-Smith, MDf,
- Ran Kornowski, MDe,
- Raj Makkar, MDd,
- Sabine Bleiziffer, MDe,
- Lars Søndergaard, MD, DMSca and
- Ole De Backer, MD, PhDa
- aRigshospitalet University Hospital, Copenhagen, Denmark
- bDeutsches Herzzentrum Munich, Munich, Germany
- cUniversitäres Herzzentrum, Hamburg, Germany
- dCedars-Sinai Medical Center, Los Angeles, California
- eRabin Medical Center, Tel Aviv, Israel
- fRoyal Sussex County Hospital, Brighton, United Kingdom
- gSan Raffaele Scientific Institute, Milan, Italy
- hUniversity Hospital of Pisa, Pisa, Italy
- iQuebec Heart and Lung Institute, Laval University, Quebec City, Quebec, Canada
- jZurich University Hospital, Zurich, Switzerland
- kHopital Privé Jacques-Cartier, Massy, Paris, France
- lAntwerp University Hospital, Antwerp, Belgium
- mCardiocentro Ticino, Lugano, Switzerland
- nGalway University Hospital, Galway, Ireland
- oUniversity of Turku, Turku, Finland
- pGhent University Hospital, Ghent, Belgium
- qQueen Elizabeth Hospital, Hong Kong, China
- rCardioVascular Center Frankfurt, Frankfurt, Germany
- ↵∗Address for correspondence:
Dr. Fadi J. Sawaya, The Heart Center, Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen, Denmark.
Objectives The aim of this study was to evaluate the use of transcatheter heart valves (THV) for the treatment of noncalcific pure native aortic valve regurgitation (NAVR) and failing bioprosthetic surgical heart valves (SHVs) with pure severe aortic regurgitation (AR).
Background Limited data are available about the “off-label” use of transcatheter aortic valve replacement (TAVR) to treat pure severe AR.
Methods The study population consisted of patients with pure severe AR treated by TAVR at 18 different centers. Study endpoints were device success, early safety, and clinical efficacy at 30 days, as defined by Valve Academic Research Consortium 2 criteria.
Results A total of 146 patients were included, 78 patients in the NAVR group and 68 patients in the failing SHV group. In the NAVR group, device success, early safety, and clinical efficacy were 72%, 66%, and 61%, respectively. Device success and clinical efficacy were significantly better with newer generation THVs compared with old-generation THVs (85% vs. 54% and 75% vs. 46%, respectively, p < 0.05); this was mainly due to less second THV implantations and a lower rate of moderate to severe paravalvular regurgitation (10% vs. 24% and 3% vs. 27%, respectively). Independent predictors of 30-day mortality were body mass index <20 kg/m2, STS surgical risk score >8%, major vascular or access complication, and moderate to severe AR. In the failing SHV group, device success, early safety, and clinical efficacy were 71%, 90%, and 77%, respectively.
Conclusions TAVR for pure NAVR remains a challenging condition, with old-generation THVs being associated with THV embolization and migration and significant paravalvular regurgitation. Newer generation THVs show more promising outcomes. For those patients with severe AR due to failing SHVs, TAVR is a valuable therapeutic option.
With the increased experience and excellent outcomes seen with transcatheter aortic valve replacement (TAVR) for severe aortic stenosis (AS) (1–3), there has been recent interest in treating patients with pure severe aortic valve regurgitation (AR) percutaneously. Historically, treatment of native aortic valve regurgitation (NAVR) with TAVR has been relatively contraindicated because of increased risk for valve embolization and migration and paravalvular regurgitation (PVR) in the absence of aortic annular calcification. Moreover, NAVR is frequently associated with a large annular anatomy and a dilated ascending aorta, making surgical aortic valve replacement (SAVR) the treatment of choice. Data from the Euro Heart Survey on Valvular Heart Disease show that only one-fifth of patients with severe AR and left ventricular (LV) ejection fractions between 30% and 50% are referred for SAVR, and <5% of patients with LV ejection fractions <30% are referred (4). However, when left untreated, these patients face an annual mortality risk of 20% (5). Therefore, there is an unmet need to treat this patient population with a less invasive approach.
Limited data are available describing the safety and efficacy of TAVR for the treatment of patients with pure severe AR. The JenaValve (JenaValve Technology, Munich, Germany) has been the only commercially available transcatheter heart valve (THV) obtaining Conformité Européene Mark approval for the treatment of inoperable or high-risk patients with severe NAVR using a transapical (TA) approach (6). Moreover, few other reports have described successful THV implantation using the less invasive transfemoral (TF) approach (7–12) (Online Table 1). Whether patients with pure AR treated by TAVR have similar outcomes as reported for those patients with pure severe AS is not known. Moreover, TAVR in patients with failing surgical heart valves (SHVs) with severe AR is technically challenging because of slippage and lack of anchoring and represents another complex subgroup of patients.
The aim of this study was to assess the safety and efficacy of TAVR when used to treat patients with pure severe AR, both those with NAVR and those with failing bioprosthetic SHVs, on the basis of data from a multicenter, international registry.
The present study was designed as an independent international, multicenter, voluntary registry study including patients treated with TAVR for pure severe AR in both native aortic valves and failing bioprosthetic SHVs. Patients from 18 centers were included (Online Figure 1); those with mixed aortic valve disease, aortic transvalvular mean gradients ≥20 mm Hg, and failing SHVs because of endocarditis were excluded.
The local heart team at the contributing hospital selected patients for TAVR. A multidisciplinary heart team involving an interventional cardiologist, a cardiac surgeon, and an imaging specialist discussed all cases, and consensus was achieved for patients on optimal medical therapy who were deemed prohibitive for SAVR.
Pre-procedural work-up was completed according to the respective institutional guidelines. All patients underwent multimodality cardiac imaging, including echocardiography and multislice computed tomography. Multislice computed tomography was used for the accurate assessment of aortic valve anatomy and calcification, aortic root size, dimensions of the sinuses of Valsalva and sinotubular junction, height of coronary arteries, and to aid THV sizing. Valve sizing for NAVR was based on the area- or perimeter-derived mean diameter on multislice computed tomography by using the largest annular diameter in systole with approximately 10% to 25% oversizing. For valve-in-valve procedures, the THV device size was selected based on a combination of the manufacturer’s reported true internal diameter (ID) and the ID as measured on CT. The ViV Aortic app was used for most cases to ensure proper THV size selection.
TF access was the preferred route for CoreValve or Evolut R (Medtronic, Minneapolis, Minnesota), SAPIEN XT or SAPIEN 3 (Edwards Lifesciences, Irvine, California), Lotus Valve System (Boston Scientific, Natick, Massachusetts), and Direct Flow (Direct Flow Medical, Santa Rosa, California). The TA route was used for the JenaValve, as described previously (6). Rapid pacing was performed in most cases to decrease the high flow of regurgitant volume and decrease prosthesis movement during deployment.
At all centers, prospective dedicated TAVR databases were interrogated to identify all suitable cases for study inclusion treated between July 2007 and September 2016. Clinical endpoints were categorized according to Valve Academic Research Consortium 2 criteria (13). Post-procedural AR was assessed by aortography and by echocardiography at the end of the TAVR procedure and/or before discharge and at 30-day follow-up. All endpoints were assessed at the participating institution.
Continuous data are reported as mean ± SD and categorical variables as number of patients and percentage. Categorical data were compared using the Fisher exact test and continuous data using the Student t test or Mann-Whitney U test, as appropriate. To identify the independent predictors of New York Heart Association (NYHA) functional class III or IV dyspnea and overall mortality, all variables with p values <0.10 on univariate analysis were included in a stepwise multivariate logistic regression model. All tests were 2 sided, and p values <0.05 were considered to indicate statistical significance. All analyses were conducted using SPSS version 23.0 (IBM, Armonk, New York).
A total of 146 patients were enrolled in the study, including 78 patients in the NAVR group and 68 patients in the failing SHV group. These 146 patients represented 1.2% of all patients treated by TAVR in all contributing hospitals (Online Figure 1). Baseline characteristics of the study population can be found in Table 1. The mean age was 75 years, and 62% were men. The mean Society of Thoracic Surgeons (STS) risk score was 6.7% in the NAVR group and 7.7% in the failing SHV group. A large portion of the study population had concomitant moderate to severe mitral regurgitation and pulmonary hypertension. Failing SHV types can be found in Online Figure 2.
Procedural characteristics and outcomes
Procedural characteristics and outcomes are provided in Table 2. The majority of cases were performed under general anesthesia (n = 102 [70%]) and using the TF route (n = 103 [71%]). The TA approach was used mainly for the JenaValve (n = 23 in the NAVR group) and SAPIEN XT (n = 11 in the failing SHV group) cases.
The device success rate was 72% in the NAVR group and 71% in the failing SHV group. Device failure was primarily due to the implantation of a second THV in the NAVR group (n = 13 [17%]), which occurred less frequently in the failing SHV group (n = 7 [10%]) (p = 0.381). A second THV was needed because of THV embolization (6 of 13 in the NAVR group, 3 of 7 in the failing SHV group) and too low implantation of the first THV with moderate to severe AR (7 of 13 in the NAVR group vs. 4 of 7 in the failing SHV group). There was no procedural mortality in either group. The intended performance of the valve was affected by an elevated rate of moderate to severe AR in the NAVR group (n = 11 [14%]) and an increase in mean transvalvular pressure gradient ≥20 mm Hg in the failing SHV group (n = 11 [16%]). Plugging of residual PVR was not reported.
Early safety and clinical efficacy
The 30-day safety and clinical efficacy rates were 66% and 61% in the NAVR group and 90% and 77% in the failing SHV group, respectively. In the NAVR group, the incidence of all-cause mortality, NYHA functional class III or IV, and valve-related dysfunction was 14%, 15%, and 16%, respectively. Major vascular complications occurred in 6 patients (8%), 4 in the TF group and 2 in those treated using the TA route related to the JenaValve. Acute kidney injury (AKI) (stage ≥2) was noted in 8 patients (11%) in the NAVR group (Table 3).
The safety and efficacy profile of TAVR for failing SHV with severe AR was satisfying (Table 3, Figure 1); however, there was Valve Academic Research Consortium 2–defined valve-related dysfunction in 14 patients (22%) driven by a post-procedural mean gradient ≥20 mm Hg. A SHV ≤23 mm was shown to be a predictor of a post-procedural mean gradient ≥20 mm Hg (Online Figure 2).
Old- versus new-generation THVs in TAVR for severe pure AR
For patients with pure NAVR, 37 patients (47.4%) were treated with old-generation devices and 41 patients (52.6%) with newer generation devices (Table 4). Newer-generation THV devices (Evolut R, SAPIEN S3, Lotus Valve System, Direct Flow, and JenaValve) had a higher device success rate than old-generation THV devices (CoreValve and SAPIEN XT) (85% vs. 54%; p = 0.011). The higher device success rate seen with newer generation THV was driven mainly by a lower incidence of moderate to severe AR (2% vs. 29%; p = 0.004) and less need for a second valve (10% vs. 24%; p = 0.156). The clinical efficacy obtained with newer generation THVs was higher compared with old-generation THVs (75% vs. 46%; p = 0.017) because of lower all-cause mortality and moderate to severe PVR rates (Table 4).
There was no statistical difference in outcomes between JenaValve and non-JenaValve THVs in the NAVR group (Online Table 2). Early safety was relatively low in the JenaValve group compared with the non-JenaValve group (57% vs. 79%; p = 0.39), mainly because of a higher incidence of vascular and access complications and AKI in the JenaValve group.
Data describing outcomes with old- and new-generation THVs in TAVR for failing SHVs are reported in Online Table 3.
Predictive factors of NYHA functional class III or IV and mortality in TAVR for NAVR
TAVR for pure NAVR resulted in a high degree of NYHA functional class III or IV and mortality (Figure 1), with a mortality rate of 14% (n = 11) at 30 days. Of these 11 deaths, 6 were ascribed to cardiac mortality (2 moderate to severe AR, 1 major vascular complication, and 1 acute heart failure) and 5 to noncardiac mortality, of which 1 followed a major stroke. By multivariate analysis, independent predictors of mortality were shown to be body mass index <20 kg/m2, STS risk score >8%, major vascular complication, and moderate to severe AR at discharge. Independent predictors of NYHA functional class III or IV in the survivors were new left bundle branch block (LBBB), and moderate to severe AR at discharge (Table 5). A full overview of the univariate analysis can be found in Online Table 4.
In this large international registry, we describe the experience with TAVR when used to treat patients with pure severe AR, both those with NAVR and those with failing bioprosthetic SHVs. In TAVR for pure NAVR, THV embolization and migration and high rates of significant PVR were seen with old-generation THV devices. However, device success and clinical efficacy have significantly improved with the newer generation THVs, combined with an improved learning curve. For patients with failing SHVs due to severe AR, TAVR is a valuable therapeutic option (Central Illustration).
TAVR for pure severe NAVR
Compared with TAVR for AS, TAVR for pure NAVR has proved to be more challenging and associated with lower device success, safety, and clinical efficacy rates. The absence of aortic annular and aortic valve leaflet calcification in patients with pure AR is a known risk factor for THV embolization and migration. Consequently, the self-expanding CoreValve THV, which exerts a radial force at both the annular level and the ascending aorta, has been used predominantly because of the ability to significantly oversize the THV prosthesis without risk for annular rupture (7–9). Despite these characteristics, real-world results with this old-generation THV have not been convincing, with high THV embolization and migration and PVR rates, as seen in our study as well as in other studies reporting the need for a second THV in 20% of cases (7). The JenaValve with clip fixation of the native leaflets has been used in pure NAVR cases, with a reported device success rate of 96.5% (6,10). However, the device success rate with the JenaValve in our series was only 83% (19 of 23), with 3 instances of THV malpositioning and 1 THV with a transvalvular pressure gradient ≥20 mm Hg. Because the JenaValve is implanted using the TA approach, this was also associated with a higher risk for major bleeding and/or major access-site complications. Newer generation THVs now available to treat AS may be better suited for treating noncalcific pure NAVR. The Lotus Valve System with its high radial force and large adaptive seal could minimize the risk for THV embolization and migration and significant PVR; in addition, this device is fully repositionable and retrievable if needed (11,12). The SAPIEN 3 with its adaptive skirt may also decrease the rate of PVR, but it is not repositionable and has a higher risk for annular rupture in case of significant THV oversizing. The results reported in this study demonstrate that device success and clinical efficacy at 30 days are better with newer generation THVs compared with old-generation THVs; rates of moderate to severe AR (2% vs. 29%) and need for a second THV (10% vs. 24%) have significantly come down with the newer generation THVs.
Interestingly, AKI stage ≥2 was observed in 6 of 23 JenaValve cases but with none of the other new-generation THV devices. Although significantly more contrast dye was used in JenaValve procedures compared with the other TAVR procedures involving newer generation THVs (233 ml vs. 114 ml, p < 0.05), it is uncertain whether the acute renal failure observed in these patients should be ascribed to the higher volume of contrast dye used or other factors. Contrast dye is nephrotoxic and has been shown to increase the risk for AKI after TAVR. However, other studies have found that general anesthesia, TA access, and blood transfusions, but not contrast volume, were independent risk factors for AKI after TAVR (14,15).
Predictors of clinical failure in TAVR for pure NAVR
As illustrated in Figure 1, the composite rate of death or NYHA functional class III or IV at 30 days was high in the NAVR group, resulting in an unsatisfactory result in one-third of the patients. Even for those patients treated with the newer generation THVs, the composite rate of death or NYHA functional class III or IV was still more than 20% (Table 4).
This result most likely reflects the selection of NAVR patients referred for TAVR, namely, highly symptomatic patients with AR who had LV function already reduced. In a study of 246 patients undergoing SAVR for severe AR, 10-year mortality was 25% in symptomatic patients versus only 3% in asymptomatic patients. In addition, patients with LV dysfunction had a 3-fold increase in mortality compared with those with preserved LV function (16). Improved survival has been reported when patients undergo early SAVR after the onset of mild symptoms, mild LV dysfunction, or LV end-systolic dimension <55 mm rather than delaying surgery (17). Perhaps patients with NAVR and contraindications to surgery could benefit from a lower threshold for intervention instead of a wait-and-see approach, particularly if more efficacious THVs become available.
Patients with severe AS included in the PARTNER (Placement of Aortic Transcatheter Valves) trial had a mean STS risk score of 11.8%, and all-cause mortality was 3.4% at 30 days (18). In the present and previous studies of TAVR for pure NAVR, mortality at 30 days was 10% to 20% despite STS risk scores of 5% to 10%. Not surprisingly, patients with pure NAVR have worse outcomes, as they have more often dilated left ventricles with reduced LV function, concomitant moderate to severe mitral regurgitation, and pulmonary hypertension (Table 1). In a large Italian CoreValve registry, 1,557 patients undergoing TAVR, 1.6% of whom for pure severe NAVR, were prospectively followed. Patients with NAVR were younger (73 years vs. 82 years), had more advanced heart failure (NYHA functional class III or IV in 95% vs. 73%), and had a higher incidence of pulmonary hypertension (31% vs. 10%) as compared to patients with AS. Despite similar STS scores, device success was lower (79% vs. 96%) and 30-day mortality higher (23% vs. 6%) in the NAVR group compared with the AS group (19); our results corroborate these data.
Besides patient selection, procedural outcomes were also associated with increased 30-day mortality. Major vascular complications (2 of these related to TA access for the JenaValve) were an independent predictor of mortality in patients with NAVR. A prior study on the use of the JenaValve for NAVR reported a similar access complication rate (10%) and a 30-day mortality rate of 13% (6). In addition, patients with moderate to severe AR at discharge had increased risk for mortality and NYHA functional class III or IV at 30 days (Table 5).
Finally, to minimize the risk for THV embolization and migration and post-procedural AR, it is recommended to oversize the THV for TAVR in pure noncalcific AR (7). However, this should be balanced against the increased risk for annular rupture and cardiac conduction abnormalities that comes with oversizing. In this study, post-procedural development of new LBBB was an independent predictor of NYHA functional class III or IV after TAVR for pure NAVR. Whether a thin-walled, dilated left ventricle in case of severe AR tolerates LBBB-induced LV dyssynchrony less well than a hypertrophic left ventricle in case of severe AS is not known. However, this finding is important and should be studied in more depth in future studies. Clearly, LBBB-induced dyssynchrony on a background of pre-existing LV dysfunction could be a potential mechanism for the increased NYHA functional class III and IV symptoms in patients with new LBBB.
TAVR for failing SHV with severe AR
Early safety and clinical efficacy was markedly better in patients with severe AR due to failing SHVs compared with NAVR. The SHV frame provides better visualization and (in some cases) better anchoring for TAVR, which led to significantly lower rates of PVR or need for a second valve. In contrast, THV implantation in the small orifice area of an SHV carries the risk for post-procedural elevated transvalvular pressure gradients ≥20 mm Hg, as observed in 12 patients (19%) in this study (6 of 45 with the CoreValve or Evolut R vs. 6 of 22 with the SAPIEN XT or SAPIEN 3; p = 0.29). Besides these elevated transvalvular gradients in one-fifth of the failing SHV patient population, clinical outcomes at 30 days were very good, with a mortality rate of only 2% and NYHA functional class III or IV symptoms in only 5% of cases.
Interestingly, in the VIVID (Valve-in-Valve International Data) registry, patients with small SHVs (size ≤21 mm) and those with predominant stenosis had worse 30-day mortality than those with large surgical valves and those with pure AR (20). This lower mortality rate in the AR group (8.8% vs. 23.4% in the AS group) was driven by lower patient-prosthesis mismatch.
The results of this voluntary observational multicenter study should be considered given the limitations inherent to the study design. In particular, the absence of an independent clinical events committee and echocardiographic core laboratory has the potential to introduce bias. Moreover, given the heterogeneity of devices and degree of oversizing used by the different centers and operators, it is difficult to determine a specific threshold for oversizing that leads to less THV embolization and migration or residual AR. Another limitation is the limited follow-up duration of 30 days; future long-term follow-up studies are warranted.
A TAVR-NAVR registry to collect outcome data of similar procedures performed worldwide is needed. We anticipate that this registry will provide insights to further improve patient selection, procedural success, and the overall clinical outcomes obtained in this patient population not eligible for standard aortic valve surgery.
TAVR for pure NAVR remains a challenging condition, with old-generation THVs associated with THV embolization and migration and significant PVR. Newer generation devices show more promising outcomes. However, further study of this patient subset, and likely new device technology, is required before TAVR can be routinely recommended as a state-of-the-art treatment option for pure NAVR. For those patients with severe AR due to failing SHVs, TAVR is a valuable therapeutic option.
WHAT IS KNOWN? Excellent outcomes are seen with TAVR to treat severe AS. Treatment of NAVR with TAVR has been relatively contraindicated because of increased risk for THV embolization and migration and PVR. As a result, limited data are available about the “off-label” use of TAVR to treat pure severe AR.
WHAT IS NEW? TAVR for pure NAVR remains a challenging condition, with old-generation THVs associated with THV embolization and migration and significant PVR. Newer generation devices show more promising outcomes. TAVR is a valuable therapeutic option to treat pure severe AR in patients with failing SHVs.
WHAT IS NEXT? Further study of this challenging patient subset, and likely new device technology, is required before TAVR can be routinely recommended to treat pure severe NAVR. A TAVR-NAVR registry to collect outcome data of similar procedures performed worldwide is warranted to better understand its safety and efficacy in a larger patient population.
For supplemental figures and tables, 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
- acute kidney injury
- aortic valve regurgitation
- aortic stenosis
- left bundle branch block
- left ventricular
- native aortic valve regurgitation
- New York Heart Association
- paravalvular regurgitation
- surgical aortic valve replacement
- surgical heart valve
- Society of Thoracic Surgeons
- transcatheter aortic valve replacement
- transcatheter heart valve
- Received November 3, 2016.
- Revision received February 27, 2017.
- Accepted March 3, 2017.
- 2017 American College of Cardiology Foundation
- Thyregod H.G.,
- Steinbruchel D.A.,
- Ihlemann N.,
- et al.
- Iung B.,
- Baron G.,
- Butchart E.G.,
- et al.
- Vahanian A.,
- Alfieri O.,
- Andreotti F.,
- et al.
- Seiffert M.,
- Bader R.,
- Kappert U.,
- et al.
- Roy D.A.,
- Schaefer U.,
- Guetta V.,
- et al.
- Praz F.,
- Windecker S.,
- Huber C.,
- Carrel T.,
- Wenaweser P.
- Franzone A.,
- Piccolo R.,
- Siontis G.C.M.,
- et al.
- Wohrle J.,
- Rodewald C.,
- Rottbauer W.
- Wendt D.,
- Kahlert P.,
- Pasa S.,
- et al.
- Schofer J.,
- Nietlispach F.,
- Bijuklic K.,
- et al.
- Dujardin K.S.,
- Enriquez-Sarano M.,
- Schaff H.V.,
- Bailey K.R.,
- Seward J.B.,
- Tajik A.J.
- Tornos P.,
- Sambola A.,
- Permanyer-Miralda G.,
- Evangelista A.,
- Gomez Z.,
- Soler-Soler J.