Author + information
- Received August 20, 2018
- Revision received October 26, 2018
- Accepted October 30, 2018
- Published online February 4, 2019.
- Lars Søndergaard, MD, DMSca,∗ (, )@Rigshospitalet,
- Nikolaj Ihlemann, MD, PhDa,
- Davide Capodanno, MD, PhDb,
- Troels H. Jørgensen, MDa,
- Henrik Nissen, MD, PhDc,
- Bo Juel Kjeldsen, MD, PhDd,
- Yanping Chang, MSe,
- Daniel Andreas Steinbrüchel, MD, DMScf,
- Peter Skov Olsen, MD, DMScf,
- Anna Sonia Petronio, MDg and
- Hans Gustav Hørsted Thyregod, MD, PhDf
- aDepartment of Cardiology, The Heart Centre, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark
- bCardiothoracic and Vascular Department, Azienda Ospedaliero Universitaria “Policlinico-Vittorio Emanuele,” University of Catania, Catania, Italy
- cDepartment of Cardiology, Odense University Hospital, Odense, Denmark
- dDepartment of Cardiothoracic and Vascular Surgery, Odense University Hospital, Odense, Denmark
- eCoronary and Structural Heart Disease Management, Medtronic, Mounds View, Minnesota
- fDepartment of Cardiothoracic Surgery, The Heart Centre, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark
- gCardiothoracic and Vascular Department, Azienda Ospedaliero Universitaria Pisana, Pisa, Italy
- ↵∗Address for correspondence:
Dr. Lars Søndergaard, Department of Cardiology, Section 2011, Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen, Denmark.
Background Transcatheter aortic valve replacement (TAVR) is an alternative to surgical aortic valve replacement (SAVR) in patients with severe aortic stenosis and intermediate or high surgical risk.
Objectives The aim of this study was to compare the durability of transcatheter and surgical bioprosthetic aortic valves using standardized criteria.
Methods In the NOTION (Nordic Aortic Valve Intervention) trial, all-comer patients with severe aortic stenosis and lower surgical risk for mortality were randomized 1:1 to TAVR (n = 139) or SAVR (n = 135). Moderate/severe structural valve deterioration (SVD) was defined as a mean gradient ≥20 mm Hg, an increase in mean gradient ≥10 mm Hg from 3 months post-procedure, or more than mild intraprosthetic aortic regurgitation (AR) either new or worsening from 3 months post-procedure. Nonstructural valve deterioration (NSVD) was defined as moderate/severe patient-prosthesis mismatch at 3 months or moderate/severe paravalvular leakage. Bioprosthetic valve failure (BVF) was defined as: valve-related death, aortic valve reintervention, or severe hemodynamic SVD.
Results At 6 years, the rates of all-cause mortality were similar for TAVR (42.5%) and SAVR (37.7%) patients (p = 0.58). The rate of SVD was higher for SAVR than TAVR (24.0% vs. 4.8%; p < 0.001), whereas there were no differences in NSVD (57.8% vs. 54.0%; p = 0.52) or endocarditis (5.9% vs. 5.8%; p = 0.95). BVF rates were similar after SAVR and TAVR through 6 years (6.7% vs. 7.5%; p = 0.89).
Conclusions In the NOTION trial through 6 years, SVD was significantly greater for SAVR than TAVR, whereas BVF was low and similar for both groups. Longer-term follow-up of randomized clinical trials will be necessary to confirm these findings. (Nordic Aortic Valve Intervention Trial; NCT01057173)
Treatment for severe degenerative aortic stenosis prior to 2007 required surgical aortic valve replacement (SAVR), precluding patients who are at high or prohibitive surgical risk from receiving treatment. The introduction of transcatheter aortic valve replacement (TAVR) has provided a safe treatment option for many of these high- and extreme-risk surgical patients (1–4). Recent prospective, randomized clinical studies have further established the noninferiority of TAVR compared with SAVR in intermediate-risk patients (5–7). These data have led to the rapidly growing adoption of TAVR for treatment of severe, symptomatic aortic stenosis. However, the long-term durability of transcatheter prosthetic valves has become increasingly important as TAVR treatment moves to younger, lower-risk patients with longer life expectancies.
In both SAVR and TAVR studies assessing long-term outcomes, the durability of the implanted bioprosthesis has been defined in many dissimilar ways, spanning from the need for reoperation to integration of clinical and echocardiographic outcomes. Indeed, standardized definitions are needed to compare longevity of transcatheter and surgical bioprosthetic aortic valves. A recent consensus statement from the European Association of Percutaneous Cardiovascular Interventions (EAPCI), the European Society of Cardiology (ESC), and the European Association for Cardiothoracic Surgery (EACTS) was developed in response to this need for classifying bioprosthetic valve durability that accounts for bioprosthetic valve dysfunction (BVD) and bioprosthetic valve failure (BVF). Furthermore, BVD is divided into structural valve deterioration (SVD) and nonstructural valve deterioration (NSVD) as well as valve thrombosis and endocarditis (8).
To evaluate the long-term durability of bioprosthetic valves, it is required that the treated patients have survival rates that are ideally longer than what is expected for the valve prostheses. The competing risk of a nonvalve-related death will otherwise create challenges in outcome interpretation, particularly when using actuarial analyses (9). The NOTION (Nordic Aortic Valve Intervention) trial was the first, and so far, the only trial to randomize all-comer patients with severe aortic valve stenosis who were mostly at lower surgical risk for mortality (mean Society of Thoracic Surgeons score of 3.0 ± 1.7%) and therefore also expected to live longer than patients in other trials (7). Patients were randomized to TAVR using the self-expanding CoreValve bioprosthesis (Medtronic, Minneapolis, Minnesota) or SAVR using any commercially available surgical bioprosthetic aortic valve. Follow-up through 6 years is now available. A post hoc analysis of prosthetic valve dysfunction (SVD and NSVD) and BVF based on the new standardized EAPCI/ESC/EACTS definitions was completed for surgical and transcatheter valves implanted in the NOTION trial and is presented herein.
NOTION is a physician-initiated, prospective, randomized, unblinded clinical trial conducted at 3 centers in Denmark and Sweden (10). The trial was conducted per the principles of the Declaration of Helsinki and was approved by each site’s ethical review board. All patients provided written informed consent. All data was 100% monitored by an independent monitoring unit. An independent clinical events committee adjudicated all clinical events at 1 year.
Patients at least 70 years of age with severe aortic stenosis were assessed by the local heart team for inclusion in the study. Detailed inclusion and exclusion criteria have been previously described (7). Enrolled patients were randomized to SAVR (27% Mosaic, 29% Epic, 24% Trifecta, 10% Perimount, and 10% Sorin Mitroflow) or TAVR (100% first-generation CoreValve) and followed annually. All patients with an implanted prosthesis were included in this analysis. Annual echocardiographic assessments measured aortic valve mean gradient, effective orifice area (EOA), total aortic regurgitation (AR) and paravalvular leakage (PVL). There was no echocardiography core laboratory used in the NOTION trial; thus, all echocardiographic data were analyzed and reported by the individual study sites.
Valve durability definitions
BVD was defined as 1 or more of the following: SVD, NSVD, bioprosthetic valve thrombosis, or endocarditis. SVD was defined as moderate/severe hemodynamic SVD (mean gradient ≥20 mm Hg, increase in mean gradient ≥10 mm Hg from 3 months post-procedure, or > mild intraprosthetic AR either new or worsening from 3 months post-procedure). NSVD was defined as moderate/severe patient-prosthesis mismatch (PPM) at 3 months or moderate/severe PVL. Moderate/severe PPM was defined as an indexed EOA ≤0.85 cm2/m2. Endocarditis was diagnosed according to the modified Duke Criteria (11). BVF was defined as at least 1 of the following: valve-related death (death caused by BVD or sudden unexplained death following diagnosis of BVD), aortic valve reintervention (TAVR or SAVR following diagnosis of BVD), or severe hemodynamic SVD (mean gradient ≥40 mm Hg, increase in mean gradient ≥20 mm Hg from 3 months post-procedure, or severe intraprosthetic AR either new or worsening from 3 months post-procedure).
Categorical variables were compared with the use of the chi-square or Fisher exact test, as appropriate. Continuous variables were presented as mean ± SD and compared using Student’s t-test. Ordinal variables were compared using the Mantel-Haenszel test. BVF and SVD rates were estimated using the cumulative incidence function with death as the competing risk for SVD and nonvalve-related deaths as the competing risk for BVF. The cumulative incidence functions were compared between treatment groups using Gray’s test. All testing used a 2-sided alpha level of 0.05. All statistical analyses were performed with the use of SAS software, version 9.4 (SAS Institute, Cary, North Carolina).
Patients were enrolled in the NOTION trial from December 2009 to April 2013. For the implanted cohort, there were 139 patients in the TAVR group and 135 in the SAVR group (3 patients randomized to TAVR were converted to SAVR because of complications during the procedure and are included in the SAVR group for this analysis). Follow-up is complete through 5 years with 100% compliance in both groups (98 TAVR and 97 SAVR patients); of these, 50 TAVR and 50 SAVR patients had 6 years of data available. Baseline characteristics are shown in Table 1. Most patients were at low risk for complications as evidenced by a Society of Thoracic Surgeons score <4% in 82.1% of patients. There were no significant differences in demographics or medical history between implanted TAVR and SAVR patients. The distribution of bioprosthetic valve types and sizes is shown in Table 2.
At 6 years, the Kaplan-Meier rates of all-cause mortality were similar for TAVR (42.5%) and SAVR (37.7%) patients (log-rank p = 0.58). Forward flow hemodynamics remained at consistent levels through 6 years (Figure 1). The EOA was significantly greater for TAVR than SAVR at all time points post-procedure with an EOA of 1.53 cm2 vs 1.16 cm2 for surgical valves at 6 years (p = 0.002). Aortic mean gradient was 9.9 mm Hg for TAVR and 14.7 mm Hg for SAVR at 6 years (p = 0.001).
Bioprosthetic valve dysfunction
The rate of BVD and its components through 6 years for the TAVR and SAVR patients is shown in Table 3. There were no significant differences between TAVR and SAVR patients for BVD (56.1% vs. 66.7%; p = 0.07), NSVD (54.0% vs. 57.8%; p = 0.52), or definite endocarditis (5.8% vs. 5.9%; p = 0.95). There was no evidence of thrombosis reported in either group.
Structural valve deterioration
There was significantly more moderate/severe SVD according to the EAPCI/ESC/EACTS definition through 6 years for the SAVR group than the TAVR group (24.0% vs. 4.8%; p < 0.001) (Central Illustration), which was primarily related to differences in measures of moderate hemodynamic SVD (Table 4). A mean gradient of ≥20 mm Hg through 6 months post-procedure was observed in 2.9% versus 22.2% of patients receiving TAVR and SAVR, respectively (p < 0.0001). However, 1 TAVR and 9 SAVR patients had a mean gradient of ≥20 mm Hg at the time of the 3-month echocardiography. To account for this observation, a modified SVD definition was applied and showed a cumulative incidence for a mean gradient of ≥20 mm Hg AND an increase in mean gradient ≥10 mm Hg after 3 months post-procedure of 1.4% for TAVR and 12.4% for SAVR (p < 0.001) (Figure 2).
Nonstructural valve deterioration
There was no significant difference in NSVD between the TAVR and SAVR patients through 6 years (Table 5). TAVR patients had more moderate PVL (20.9% vs. 1.5%; p < 0.0001) post-procedure, but less severe PPM at 3 months post-procedure (12.2% vs. 28.1%; p = 0.001) than SAVR patients at 3 months. Change over time was minimal for both PVL and PPM.
Bioprosthetic valve failure
TAVR and SAVR patients experienced a similar degree of BVF through 6 years (7.5% vs. 6.7%; p = 0.89), as shown in Figure 3. There were no statistically significant differences in the frequency of the components of BVF between TAVR and SAVR groups: valve-related deaths (5.0% vs. 3.7%; p = 0.59), reintervention (2.2% vs. 0.7%; p = 0.62), and severe hemodynamic SVD (0.7% vs. 3.0%; p = 0.21).
Long-term assessment of BVD and BVF after TAVR has only been possible in recent years, and there are few reports of transcatheter bioprostheses durability directly compared with surgical bioprostheses durability. The NOTION trial is among the first to provide comparative data regarding bioprosthetic valve durability for TAVR and SAVR from a randomized clinical trial in patients with low surgical risk for mortality. Consistent clinical and hemodynamic outcomes through 6 years were observed.
Overall, there were no differences in NSVD between groups, although more PVL was seen after TAVR and more PPM after SAVR. SVD and NSVD rates remained relatively unchanged over time and are more likely the result of suboptimal valve sizing for both groups than actual deterioration of the prostheses. The PVL rate in the NOTION trial is higher than contemporary reports of PVL after TAVR, probably due to the introduction of computed tomography (CT) scanning for sizing of the aortic annulus after NOTION was initiated and the use of the original CoreValve prosthesis without outer sealing skirt and possibility for resheathing. Rates of endocarditis were low and similar between groups, and no thromboembolic event or clinically detectable valve leaflet thrombosis was reported in either group. Four-dimensional volume-rendered CT was not used for detection of subclinical leaflet thrombosis (12). Importantly, there was also no difference in the patient-oriented components of BVF, including valve-related death, reintervention, and severe hemodynamic deterioration.
Causes of BVD include leaflet tear, calcification, pannus deposition, flail or fibrotic leaflet, endocarditis, and thrombosis. In surgical stented bioprostheses, SVD has been most commonly caused by calcification of the valve, which leads to stiffening and aortic valve restenosis or, more common in nonstented bioprostheses, leaflet tears and aortic valve insufficiency. Because SVD develops slowly over time, rates remain low for 5 to 10 years after SAVR (13). However, most long-term studies enroll elderly patients, and results are confounded by high noncardiac mortality rates. Furthermore, many of these studies report explantation rates due to SVD, but they do not employ a clear definition of hemodynamic parameters of SVD (13,14). A recent article from Bourguignon et al. (15), provides clearer criteria for the assessment of severe SVD (mean gradient > 40 mm Hg or severe AR) and reports severe SVD in 153 of 2,559 patients (5.9%) at a mean follow-up of 6.7 ± 4.8 years after Carpentier-Edwards Perimount bioprosthetic valve (Edwards Lifesciences, Irvine, California) implantation (15). The late linearized mortality rate was 6.0%/valve-year. Calcification and leaflet tears were the predominant causes of SVD.
Detailed analysis of transcatheter and surgical bioprosthetic aortic valve durability found significantly more moderate/severe SVD in the SAVR group, primarily related to higher aortic valve mean gradients, compared with the TAVR group. To account for the number of patients with a mean gradient ≥20 mm Hg already at the time of the 3-month post-procedure echocardiography, we looked at the proportion of patients with both a mean gradient ≥20 mm Hg AND an increase in mean gradient ≥10 mm Hg from the 3-month measurement. Future modifications of the EAPCI/ESC/EACTS consensus statement may account for observations that small surgical bioprostheses may have a mean gradient ≥20 mm Hg early after valve implantation.
SVD is typically detected by echocardiography and reflects irreversible valve dysfunction (15–17). A recent study of 152 patients implanted with a CoreValve bioprosthesis using the same standardized definitions proposed by Capodanno et al. (8) reported no severe SVD with follow-up to 8.9 years (18). There were 5 reinterventions related to PVL resulting in an estimated actuarial rate of BVF of 7.9% at 8 years. Stable hemodynamics and 5-year rates of SVD have also been reported by the Italian (4.2%) and Canadian (3.4%) registries, although again, definitions vary (19,20). However, these studies included patients with limited life-expectancy who were therefore likely to die before the occurrence of SVD or BVF. Furthermore, all studies were without an SAVR control group.
In surgery, bioprosthetic aortic valves are more commonly used over mechanical valves, partly due to patient preferences to avoid lifelong anticoagulation therapy. However, particularly younger patients face the risk of bioprosthetic valve deterioration and need for reintervention. Although this can be performed as a less-invasive transcatheter valve-in-valve procedure, the risk-benefit ratio needs to be acceptable. Similar, durability is an important factor for TAVR, particularly when expanding the therapy to younger, lower-risk patients with longer life-expectancy. Ideally, the durability for transcatheter bioprosthetic valves should be at least as good as for surgical bioprostheses. To compare different bioprostheses used for SAVR and TAVR, standardized definitions of prosthetic valve dysfunction and failure need to be applied. Furthermore, prospective and preferably randomized patient cohorts with a life-expectancy allowing long-term follow-up are needed.
NOTION enrolled patients from 2009 to 2013 when echo-based sizing of the aortic annulus was standard of care instead of CT sizing, which could have affected patient outcomes. Surgeons in the trial did not use annular enlargement techniques or an algorithm to avoid patient-prosthesis mismatch. In addition, echocardiographic measurements were not adjudicated by an echocardiography core laboratory. Valve durability was assessed through 6 years of follow-up, but not all patients were eligible for 6-year follow-up at the time of this analysis. TAVR was performed with the self-expanding CoreValve bioprosthesis; thus, results may not be generalizable to balloon-expandable transcatheter bioprostheses. CoreValve may have a favorable design with regard to the supra-annular leaflet position, which provides a larger effective opening area than transcatheter heart valves with an intra-annular leaflet position.
Through 6-year follow-up of the NOTION trial, adequate hemodynamic valve performance was maintained for both groups. SVD was significantly greater for SAVR compared with TAVR mainly due to higher mean valve gradients present shortly after the procedure. Overall, NSVD was similar for both groups, although rates of PVL and PPM differed between TAVR and SAVR. Bioprosthetic valve failure rates were low and similar for the self-expanding transcatheter and surgical valves implanted, with no valve thrombosis and similar rates of endocarditis for both groups. Although follow-up is medium-term, the low rates of SVD and BVF after implantation of the self-expanding CoreValve bioprosthesis are encouraging. Longer-term follow-up of randomized clinical trials will be necessary to confirm these early results.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: In patients with severe aortic stenosis at low surgical risk, the rate of structural deterioration of bioprosthetic valves is greater within the first 6 years after SAVR than TAVR, while the rates of bioprosthetic valve failure, endocarditis, and all-cause mortality were low and similar with both methods.
TRANSLATIONAL OUTLOOK: Further studies are needed to clarify the pathophysiological mechanisms responsible for the different rates of structural valve deterioration after SAVR compared with TAVR.
The authors thank study nurses Line M. Kristensen, Lisette L. Larsen, Ane L. Johansen, Ida Rosenlund and Eva-Lena Pommer and Colleen Gilbert, PharmD, of Medtronic Inc. for copyediting and formatting of the manuscript; and all patients participating in the trial.
This work was supported by the Danish Heart Foundation (grant numbers: 09-10-AR76-A2733-25400, 12-04-R90-A3879-22733, and 13-04-R94-A4473-22762) and Medtronic. Dr. Søndergaard has received consultant fees and institutional research grants from Abbott, Boston Scientific, Edwards Lifesciences, Medtronic, and Symetis. Dr. Ihlemann has received educational compensation from Medtronic. Dr. Jørgensen has received a research grant from Edwards Lifesciences. Dr. Kjeldsen has served as a proctor for Edwards Lifesciences. Ms. Chang is an employee and shareholder of Medtronic. Dr. Steinbrüchel has received research contracts from Medtronic and St. Jude Medical. Dr. Petronio is a consultant for Medtronic, Abbott, and Boston Scientific. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Listen to this manuscript's audio summary by Editor-in-Chief Dr. Valentin Fuster on JACC.org.
- Abbreviations and Acronyms
- surgical aortic valve replacement
- structural valve deterioration
- transcatheter aortic valve replacement
- Received August 20, 2018.
- Revision received October 26, 2018.
- Accepted October 30, 2018.
- 2019 American College of Cardiology Foundation
- Popma J.J.,
- Adams D.H.,
- Reardon M.J.,
- et al.
- Reardon M.J.,
- Van Mieghem N.M.,
- Popma J.J.,
- et al.
- Thyregod H.G.,
- Steinbruchel D.A.,
- Ihlemann N.,
- et al.
- Capodanno D.M.,
- Petronio A.S.,
- Prendergast B.,
- et al.
- Capodanno D.
- Durack D.T.,
- Lukes A.S.,
- Bright D.K.
- Sondergaard L.,
- De Backer O.,
- Kofoed K.F.,
- et al.
- Lancellotti P.,
- Pibarot P.,
- Chambers J.,
- et al.
- Holy E.W.,
- Kebernik J.,
- Abdelghani M.,
- et al.
- Barbanti M.,
- Petronio A.S.,
- Ettori F.,
- et al.
- Toggweiler S.,
- Humphries K.H.,
- Lee M.,
- et al.