- Email: [email protected]
Treatment of failed aortic bioprostheses: An evaluation of conventional redo surgery and transfemoral transcatheter aortic valve-in-valve implantation
Journal Pre-proof Treatment of failed aortic bioprostheses: An evaluation of conventional redo surgery and transfemoral transcatheter aortic valve-in-valve implantation Felix J. Woitek, Georg Stachel, Philipp Kiefer, Stephan Haussig, Sergey Leontyev, Florian Schlotter, Meinhard Mende, Jennifer Hommel, Lisa Crusius, Aileen Spindler, Friedrich W. Mohr, Gerhard Schuler, Holger Thiele, Michael A. Borger, Axel Linke, David Holzhey, Norman Mangner PII:
S0167-5273(19)32470-2
DOI:
https://doi.org/10.1016/j.ijcard.2019.09.039
Reference:
IJCA 28014
To appear in:
International Journal of Cardiology
Received Date: 13 May 2019 Revised Date:
10 July 2019
Accepted Date: 16 September 2019
Please cite this article as: F.J. Woitek, G. Stachel, P. Kiefer, S. Haussig, S. Leontyev, F. Schlotter, M. Mende, J. Hommel, L. Crusius, A. Spindler, F.W. Mohr, G. Schuler, H. Thiele, M.A. Borger, A. Linke, D. Holzhey, N. Mangner, Treatment of failed aortic bioprostheses: An evaluation of conventional redo surgery and transfemoral transcatheter aortic valve-in-valve implantation, International Journal of Cardiology (2019), doi: https://doi.org/10.1016/j.ijcard.2019.09.039. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier B.V.
Treatment of failed aortic bioprostheses: An evaluation of conventional redo surgery and transfemoral transcatheter aortic valve-in-valve implantation
Felix J. Woitek, MD*a, Georg Stachel, MD*b, Philipp Kiefer, MDc, Stephan Haussig, MDa, Sergey Leontyev, MDc, Florian Schlotter, MDb, Meinhard Mende, PhDd, Jennifer Hommel, MAa, Lisa Crusius, MDb, Aileen Spindlerb, Friedrich W. Mohr, MDc, Gerhard Schuler, MDb, Holger Thiele, MDb ,Michael A. Borger, MD PhDc , Axel Linke, MDa, David Holzhey, MD**c Norman Mangner, MD**a a
Herzzentrum Dresden, Technische Universität Dresden, Department of Internal Medicine
and Cardiology, Dresden, Germany b
Heart Centre Leipzig at University of Leipzig, Department of Internal Medicine /Cardiology,
Leipzig, Germany c
Heart Centre Leipzig at University of Leipzig, Department of Cardiac Surgery, Leipzig,
Germany d
Institute for Medical Informatic, Statistics and Epidemiology, University of Leipzig, Leipzig,
Germany * both authors contributed equally and are shared first authors ** both authors contributed equally and are shared last authors
Running title: Re-SAVR compared to valve-in-valve-TFAVI in failed aortic bioprostheses Total word count (text and figure legend): 3499 Address for correspondence: Norman Mangner Herzzentrum Dresden, Technische Universität Dresden Department of Internal Medicine and Cardiology Fetscherstr. 76, 01307 Dresden / Germany Tel. +49 351 450 25 297 [email protected]
Conflict of interest Felix J. Woitek: The author reports no relationships that could be construed as a conflict of interest. Georg Stachel: The author reports no relationships that could be construed as a conflict of interest. Philipp Kiefer: The author reports no relationships that could be construed as a conflict of interest. Stephan Haussig: The author reports no relationships that could be construed as a conflict of interest. Sergey Leontyev reports speaker´s honoraria from St. Jude Medical and Medtronic, outside the submitted work. Florian Schlotter: The author reports no relationships that could be construed as a conflict of interest. Meinhard Mende: The author reports no relationships that could be construed as a conflict of interest. Jennifer Hommel: The author reports no relationships that could be construed as a conflict of interest. Lisa Crusius: The author reports no relationships that could be construed as a conflict of interest. Aileen Spindler: The author reports no relationships that could be construed as a conflict of interest. Friedrich W. Mohr: The author reports no relationships that could be construed as a conflict of interest. Gerhard Schuler: The author reports no relationships that could be construed as a conflict of interest. Holger Thiele: The author reports no relationships that could be construed as a conflict of interest.
Michael A. Borger declares personal fees from Edwards Lifesciences, personal fees from Medtronic, personal fees from CryoLife, outside the submitted work. Axel Linke reports grants and personal fees from Medtronic, personal fees from St. Jude Medical, grants from Claret Medical, personal fees and other from Claret Medical, personal fees from Boston Scientific, personal fees from Bard, personal fees from Edwards, outside the submitted work. David Holzhey reports speaker´s honoraria from Symetis and Medtronic, outside the submitted work. Norman Mangner reports speaker´s honoraria from Edwards, Medtronic, Novartis, Sanofi Genzyme and Astra Zeneca, consultant honoraria from Biotronik, outside the submitted work.
Abstract Background: The use of bioprostheses for surgical aortic valve replacement increased substantially within the last years. In case of prosthesis failure, re-SAVR is standard of care, whereas valve-in-valve deployment of a transfemoral transcatheter aortic valve prosthesis (VinV-TFAVI) has recently emerged as an alternative. We sought to evaluate early safety, clinical efficacy, and all-cause 1-year-mortality of VinV-TFAVI and redo surgery for failing aortic bioprostheses (re-SAVR). Methods and results: Patients receiving either VinV-TFAVI (n=147) or re-SAVR (n=111) for a degenerated aortic bioprosthesis between 01/2006 and 05/2017 were included in this analysis. All-cause 1-year mortality was the primary outcome measure. Early safety and clinical efficacy according to VARC-2 endpoint definitions were evaluated at 30 days. Baseline characteristics differed significantly between both groups, including age, STSPROM, and incidence of relevant comorbidities. Re-stenosis was the predominant mode of failure in 45.9% of re-SAVR and 63.1% of VinV-TFAVI patients. The rate of “early safety” endpoints was lower with VinV-TFAVI (17.7% vs. 64.9%, p<0.01), the rate of “clinical efficacy” endpoints was lower, e.g. better with re-SAVR (53.1% vs. 32.4%, p<0.01). All-cause 1-year-mortality (VinV-TFAVI 8.8% vs re-SAVR 9.9%, p= 0.84) was not different. Treatment strategy was not associated with 1-year-mortality in a Cox regression analysis. The incidence of prosthesis-patient-mismatch was higher in VinV-TFAVI compared to re-SAVR. Conclusion: VinV-TFAVI represents a viable alternative for treatment of degenerated aortic bioprostheses in patients at increased surgical risk. However, in patients at low risk for reoperation, a better clinical efficacy and acceptable safety may favour re-SAVR.
244/250 words
Key words: Aortic valve stenosis; Heart valve prosthesis implantation; Transcatheter aortic valve replacement; Prosthesis failure; valve in valve
Abbreviations AVA aortic valve area BEV balloon-expandable valve NYHA New York Heart Association functional class PPM patient-prosthesis-mismatch SAVR surgical aortic valve replacement SEV self-expandable valve SMD standardized mean difference STS-PROM Society of Thoracic Surgeons – Predicted Risk of Mortality VinV-TFAVI valve-in-valve transfemoral aortic valve implantation VARC-2 Valve Academic Research Consortium -2
Introduction There has been a marked increase in the use of bioprostheses for surgical aortic valve replacement (SAVR) as compared to mechanical prostheses [1]. A major disadvantage of bioprostheses is a higher tendency towards structural deterioration and failure, thus frequently necessitating replacement after 10 to 20 years [2]. Standard procedure for the replacement of failing aortic bioprostheses is currently re-SAVR. However, management of these patients is challenging due to their commonly advanced age and the increased risk conferred by repeat cardiac surgery [3]. Hence, ‘valve-in-valve’ transcatheter aortic valve implantation (VinV-TAVI) has emerged in the last years, and a multicentre registry has demonstrated encouraging results concerning efficacy and safety [4] of this procedure. Direct comparisons between re-SAVR and VinV-TAVI are confined to a small number of reports consisting of small case series [5-8] with one larger trial employing propensity score matching [9] and three meta-analyses of these non-randomized trials [10-12]. Only one trial [6] demonstrated a higher 1-year-mortality after VinV-TAVI compared to re-SAVR, whereas all other trials did not show significant differences in survival between both treatment strategies despite significantly higher operative risk scores in the VinV-TAVI group. Results on the incidence of increased postoperative transvalvular gradients are ambiguous: some trials found a higher incidence in the VinV-TAVI group [5, 6], while others did not [7, 8]. Interpreting the results of these studies, it has to be taken into account that a large proportion of patients – as compared to TAVI in native valve aortic stenosis [13] - was treated via nontransfemoral access, which may have influenced the outcomes [14]. In our study, we sought to evaluate 1-year mortality as well as combined clinical efficacy and safety endpoints at 30 days [15] in patients receiving either valve-in-valve transfemoral TAVI (VinV-TFAVI) or re-SAVR.
Methods Patient Selection Between January 2006 and May 2017, 111 patients received re-SAVR for a degenerated aortic bioprosthesis at the Heart Centre Leipzig, a tertiary referral centre. VinV-TFAVI for degenerated aortic bioprostheses was performed in 147 patients between September 2007 and May 2017. After the establishment of the TAVI program at our hospital in 2006, any treatment strategy decision was made after discussion of all options in the multidisciplinary heart team. Patients receiving re-SAVR were excluded from analysis if VinV-TFAVI had not been possible, e.g. in the case of a degenerated mechanical valve, necessity of surgical treatment of other valves (other than mitral valve decalcification) or conditions not amenable to transcatheter treatment, e.g. infective endocarditis or paravalvular leaks as separate indication of the procedure. Experienced consultant cardiac surgeons performed all redosternotomies and –operations. A pre-operative computed tomography scan was performed in 100% of the patients receiving VinV-TFAVI group and 94.6% of the patients treated by reSAVR group. Baseline characteristics, procedural data and outcome data were prospectively collected. True inner diameter of the valves was determined as published [16]. Mode of failure was classified as ‘mixed’ if regurgitation ≥ grade 2 (on a scale of 3) was present in a predominantly stenotic valve and vice versa. Follow‐up was performed after 30 days and 1 year. Presence of lung disease, immunosuppressant medication, diabetes mellitus, coronary heart disease (CAD), cerebrovascular disease and peripheral artery disease were defined according to the STS-PROM [17]. The registry was approved by the Ethics Committee of the Medical Faculty of the University of Leipzig (registration no. 428/17-ek), and all patients gave written informed consent.
Outcome measures The primary outcome measure was all-cause 1-year mortality. Clinical efficacy (all-cause mortality, all stroke, New York Heart Association (NYHA) functional class III or IV or valve-
related dysfunction) and early safety (all-cause mortality, all stroke, life-threatening bleeding, acute kidney injury stage 2 or 3, coronary artery obstruction or valvular dysfunction requiring repeat procedure) were assessed at 30 days according to the definition of the Valve Academic Research Consortium -2 (VARC-2) [15]. To constitute a bleeding event in the reSAVR cohort, a source of bleeding other than the index surgical procedure (e.g. haemothorax or cardiac tamponade) had to exist in addition to the drop in haemoglobin concentration or amount of transfused packed red cell units as required by VARC-2. Data on NYHA functional class at 1 year, postoperative gradient and prosthesis-patient-mismatch (PPM) were collected. PPM and aortic regurgitation were graded as defined by the VARC-2. Haemodynamic parameters were evaluated within the whole cohort of each treatment group and according to the initial true inner diameter of the degenerated prosthesis (<20mm, 20.022.99mm, ≥23mm).
Statistical Analysis Based on the two very distinct groups in a retrospective analysis, we decided not to perform a propensity or other matching. Instead, we calculated difference measures for the baseline characteristics and adjusted for covariates associated with the primary outcome measure. Standardized mean differences (for continuous variables), odds ratios (for binary variables) and the measure A12 of stochastic superiority [18] (for ordinary variables) characterize the balance between groups. If (X, Y) is a randomly selected pair of values from both groups, A12 is interpreted as the probability that X > Y plus ½ the probability that X = Y; that is, A12 = P{X > Y} + ½ P{X = Y}. 95% confidence intervals were calculated. The study cohort is characterized by mean (standard deviation, SD) for continuous, absolute and relative frequencies for categorical variables. The groups were compared with respect to categorical variables by means of chi-squared and Fisher's exact test as appropriate. We compared means of continuous characteristics by t-test (Welch) and skew distributed traits (time spans) by Mann-Whitney U test. Median differences were estimated by the HodgesLehman method.
Survival curves were estimated and depicted applying the Kaplan-Meier algorithm. To adjust for covariates, multiple Cox regression models were built. We started with baseline traits obviously different and performed a backward selection using the AIC criterion including treatment group (VinV-TFAVI), age (per year), NYHA III or IV at baseline, sex (male), STSPROM, CAD at baseline, and mode of failure (regurgitation). A final model with treatment group forced into it was built to get risk estimates inclusive 95% confidence intervals. NYHA class was analyzed by means of a general linear model including the baseline category as covariate. We defined the significance level at 5% for two-tailed testing. Tables and text show raw pvalues. However, they were adjusted for multiple testing following the method of Bonferroni and Holm and depicted in bold for remaining significance. Data preparation, descriptive statistics and tests for group comparisons were performed by IBM SPSS Statistics (version 24). The program R (version 3.4.1, R Core Team, 2017) inclusive the software packages orddom and survival was applied for all other analyses and graph design.
Results Baseline characteristics Out of 258 patients, 111 (43%) and 147 (57%) received re-SAVR or VinV-TFAVI, respectively. VinV-TFAVI were older, had higher STS-PROM, and higher incidences of nearly all evaluated comorbidities (Table 1). Baseline NYHA functional class was worse in VinV-TFAVI, as was renal function. The incidence of CAD and number of affected vessels was higher in VinV-TFAVI. There was no significant difference in left ventricular ejection fraction but there was a higher frequency of severe mitral and tricuspid regurgitation in VinVTFAVI (Table 1). In re-SAVR, 31 patients (27%) underwent concomitant surgery on the thoracic aorta, one patient (0.9%) received additional mitral valve decalcification, 10 patients (9.0%) received a Morrow procedure, and 14 patients (12.6%) had concomitant coronary artery bypass grafting. Percutaneous coronary intervention was performed in one patient (0.9%) during the same hospital stay. Three patients (2.7%) had medically managed coronary artery disease. In VinV-TFAVI, seven patients (4.7%) received percutaneous coronary intervention within the preceding 30 days, three patients (2.1%) within the same hospital stay. CAD was medically treated in six patients (4.1%). Concerning the characteristics of the degenerated bioprosthesis, the majority of patients had been treated with stented valves, even more frequently in the VinV-TFAVI-group. Predominant mode of failure was stenosis, again with a higher incidence of stenotic and mixed pathologies in the VinV-TFAVI group (Online-only Table 1).
Procedural and short-term outcomes The combined ‘early safety’ endpoint was met significantly more often in re-SAVR compared to VinV-TFAVI (64.9% vs. 17.7%, p<0.001, Table 2) indicating an inferior safety. Conversely, significantly less patients in re-SAVR reached the composite ‘clinical efficacy’ endpoint (32.4% vs. 53.1%, p=0.001) indicating a superior efficacy. The higher incidence of the ‘early safety’ endpoint after re-SAVR was mainly driven by a higher incidence of life-threatening
bleeding and renal failure of stages 2 and 3. The number of transfused red packed blood cells was higher in re-SAVR. Repeat procedures or conversions were not necessary. Higher NYHA functional classes and higher incidences of increased gradients and PPM in VinVTFAVI (Online-only Table 1) mainly caused the difference concerning the ‘clinical efficacy’ endpoint. The incidences of postprocedural myocardial infarction, stroke and new permanent pacemaker implantation did not differ significantly between groups (Table 2). Postoperative length of stay was longer in re-SAVR compared to VinV-TFAVI. Thirty-day-mortality was 4.1% with re-SAVR and 4.5% with VinV-TFAVI (p=0.87) and the corresponding observed-topredicted 30-day mortality was 1.49 vs. 0.54.
Haemodynamic parameters after valve replacement Peak and mean postoperative transvalvular gradients were significantly higher in VinV-TFAVI (Online-only Table 1). Accordingly, AVA and AVA index were significantly larger in the surgically treated group. Consequently, prevalence of PPM was significantly higher in VinVTFAVI. Mild prosthetic aortic valve regurgitation after replacement was more frequent after VinV-TFAVI. Moderate aortic regurgitation was rare and not different between groups, severe aortic regurgitation was found in neither group (Online-only Table 1). Patients treated for stenotic/mixed degeneration had higher rates of any PPM compared to patients treated for isolated regurgitation (58.9% vs. 27.1%, p<0.001) in the whole cohort irrespective of treatment form as well as in VinV-TFAVI (66.4% vs. 38.5%, p=0.046) and reSAVR (43.5 vs. 22.9%, p=0.042). Noteworthy, the rate of stenotic/mixed mode of prosthesis failure was higher in VinV-TFAVI compared to re-SAVR who had a higher proportion of patients treated for isolated aortic regurgitation (Online-only Table 1). Haemodynamic parameters were also evaluated according the true internal diameter of the failed bioprosthesis (Online-only Table 2). The proportion of patients with a true-ID <20mm, 20.0-22.99mm and ≥23mm was 47.6%, 32.9% and 19.6% in VinV-TFAVI and 34.7%, 29.5% and 35.8% in re-SAVR, respectively (p=0.017). In re-SAVR, 21 patients (18.9%) received a mechanical prosthesis as a second valve. The mean true-ID of the second prostheses was
significantly greater in the group having an initial true-ID <20mm and 20.0-22.99mm, but was smaller in ≥23mm. Haemodynamic parameters within the three groups showed the same pattern with higher gradients and lower aortic valve area and indexed aortic valve area in VinV-TFAVI. Only in patients with an initial true-ID of 20.0-20.99mm, equivalent aortic valve area and indexed aortic valve area were evident leading to a comparable rate of PPM in this group. The rate of aortic regurgitation was similar in the groups having a true-ID <20mm and 20.0-20.99mm, but higher in VinV-TFAVI having an initial true-ID ≥23mm (Online-only Table 2).
Analysis of all-cause 1-year mortality All-cause 1-year mortality was 9.9% in the re-SAVR-group and 8.8% in the VinV-TFAVI group (p=0.76, Figure 1). Cardiovascular mortality after 1 year was 8.1% and 7.8%, respectively (p=0.84). There was no significant difference in 1-year-mortality after exclusion of surgical patients with concomitant procedures (Online-only Figure 1). In a multiple model including treatment group, only NYHA class at baseline (HR 6.35, 95% CI: 2.29 – 17.7) and sex (HR 0.38 for male sex, 95% CI 0.15 – 0.97) remained as factors associated with allcause 1-year mortality (Online-only Table 3).
Functional outcome at 1 year Concerning NYHA functional class, patients in the VinV-TFAVI group had worse functional status both at baseline and after 1 year than in the re-SAVR group (mean +0.22, p=0.002). However, both groups improved in functional status by about one NYHA class as compared to baseline (-0.95, p<0.001, Figure 2A). In the general linear model used for analysis, the interaction term time*group was excluded because it did not improve the model. Based on this, there was no evidence of differential improvement of NYHA functional class between groups (Figure 2B, Online-only Table 4).
Comparison of balloon-expandable and self-expandable valves
In VinV-TFAVI, 43 patients (29.5%) were treated with balloon-expandable valves (BEV-VinVTFAVI) and 103 (70.5%) were treated with self-expandable valves (SEV-VinV-TFAVI). One patient received a mechanically expanding valve. SEV-VinV-TFAVI had degenerated valves with smaller true inner diameter than those receiving re-SAVR (Online-only Table 5). This was not observed in BEV-VinV-TFAVI. Postoperative pressure gradients were significantly higher and aortic valve indexes were significantly lower for both valve types compared to reSAVR. Incidence of any postoperative aortic regurgitation was significantly higher after SEVVinV-TFAVI than in re-SAVR; predominantly driven by mild aortic regurgitation. There was no significant difference in the incidence of aortic regurgitation between re-SAVR and BEVVinV-TFAVI (Online-only Table 5). Comparing BEV-VinV-TFAVI with SEV-VinV-TFAVI, self-expanding valves had a tendency towards a better haemodynamic profile without reaching statistical significance (Online-only Table 6).
Discussion The main findings of this analysis are: (1) Although VinV-TFAVI constituted a group at higher surgical risk, no significant differences in survival after 1 year were observed, and treatment strategy was not associated with all-cause 1-year mortality in a multivariate model adjusted for baseline differences. (2) Incidence of the combined ‘early safety’- endpoint at 30 days, most likely driven by a higher incidence of life threatening bleeding and acute kidney injury was significantly higher in the re-SAVR-group. The incidence of the ‘clinical efficacy’ endpoint was higher, e.g. worse, in the VinV-TFAVI-group. (3) Re-SAVR resulted in a better haemodynamic profile with lower transvalvular gradients, larger effective orifice areas and a lower incidence of patient-prosthesis mismatch; however, this led neither to differences in survival nor to differences in the improvement of functional status after 1 year.
Comparison of mortality We did not detect a significant difference in all-cause 1-year mortality between treatment groups, which is in line with the majority of previous work [11]. This finding should be cautiously interpreted, as patients in the VinV-TAVI cohort were about 18 years older and at higher surgical risk but approximately half of the patients receiving re-SAVR had concomitant procedures. After accounting for these differences in a multivariate model, mode of replacement (re-SAVR vs. VinV-TFAVI) was not significantly associated with 1-year mortality and the observed-to-predicted 30-day mortality was in favour for VinV-TFAVI. Moreover, there was no difference in mortality after exclusion of patients receiving concomitant procedures during re-SAVR. Spaziano et al. [9] and Ejiofor et al. [7] used propensity score matching to account for differences, both coming to similar results, despite persisting differences in procedural and baseline characteristics of their cohorts and smaller patient numbers. We employed a multivariate analysis as there was only a small overlap between groups in baseline characteristics and there is no evidence for the superiority of one or the other statistical method.
In our multivariate model, mode of failure, i.e. patients presenting with stenotic failure as compared to those with severe regurgitation or mixed pathology as indication for valve replacement, was not associated with differences in mortality. In other works, most notably the VIVID registry [4], VinV-TAVI patients with stenotic failure had worse outcome. However, patients with stenotic and mixed failure experienced more often a PPM compared to patients having a regurgitation as the predominant mode of failure.
Short-term outcomes and adverse events Patients receiving re-SAVR significantly more often reached the composite endpoint for early safety, and conversely, the endpoint for clinical efficacy was more often reached in patients receiving VinV-TFAVI, most likely driven by a higher incidence of bleeding and kidney failure in the further group and elevated gradient and worse NYHA functional class in the latter. The higher incidence of bleeding and acute kidney injury in the surgical group is most likely due to the greater invasiveness of an operative valve replacement and the use of cardiopulmonary bypass [19], although all procedures were performed by experienced consultant cardiac surgeons. Acute kidney injury is known to be associated with bleeding events and consecutive need for blood transfusion [20] indicating that the lower number of bleeding events and significantly lower amounts of blood transfusion in VinV-TFAVI might have contributed to the lower incidence of acute kidney injury. The incidence of stroke was not significantly different between groups, which is comparable to the finding of others [21]. Of note, we could not prove a significant difference in the comparative incidence of myocardial infarction and is within the range reported by larger registries and seems to be substantially higher than in native aortic valve TAVI [22].
Haemodynamic profile Higher transvalvular gradients and lower aortic valve areas translated into a higher incidence of PPM after VinV-TFAVI. Noteworthy, VinV-TFAVI had higher rates of stented prostheses, more often stenotic/mixed prosthesis degeneration and in nearly half of the patients a
prosthesis with a true internal diameter <20mm. On the other hand, 18.9% of the patients in re-SAVR received a mechanical prosthesis and the true internal diameter of the second prosthesis was significantly greater in the groups having an initial true-ID <20mm and 20.022.99mm. Interestingly, the numerical values of the haemodynamic parameters were quite similar within each treatment strata irrespective of initial true-ID. These are all factors leading to a worse haemodynamic outcome in VinV-TFAVI, which is, of course, a completely different concept compared to re-SAVR and, therefore, must be associated with higher gradients compared to a new surgical prosthesis. At least up to 1 year, the haemodynamic differences between re-SAVR and VinV-TFAVI did not affect mortality or improvement in NYHA classes. The higher NYHA classes in VinV-TFAVI, both at baseline and follow-up, might be correlated to advanced age and the reported pulmonary and cardiac comorbidities. A longer follow-up is necessary to evaluate the impact of worse haemodynamic in VinVTFAVI on mortality and functional status. However, this cohort and all others treated so far by VinV might not be the right one for comparison due to the advanced age and limited life expectancy. One study comparing only patients with degenerated stentless aortic valves [8] did not see that extent of difference between re-SAVR and VinV-TAVI concerning transvalvular gradients. In our analysis, stentless prostheses were more often treated by re-SAVR than VinV-TFAVI, however, numbers were low preventing us from a solid analysis on the comparison between stented vs. stentless prostheses. One possibility to improve the haemodynamic profile of VinV-TFAVI, in particular in small prostheses, might be controlled ring fracture of the degenerated prosthesis with high-pressure balloons [23], with risk-benefitprofile of this technique remaining to be elucidated in larger studies. Overall, we could corroborate past findings concerning the comparison between VinV-TFAVI and re-SAVR. The safety profile appears to be better in VinV-TFAVI, whereas in particular the rate of patient-prosthesis-mismatch is higher in VinV-TFAVI leading to an individualized decision process to offer patients with failed aortic bioprostheses the most reasonable therapy.
Balloon-expandable versus self-expandable valves Theoretically, prostheses with leaflets located supra-annularly, in particular the Medtronic CoreValve, should have a slightly better haemodynamic profile, as, thus maximizing available orifice area. In our study, we could not prove a better haemodynamic profile of SEV-TFAVI vs. BEV-TFAVI, however there seems to be a small effect. The missing benefit might also be due to a lower true inner diameter of the initial bioprosthesis and a higher rate of stenotic failure in SEV-TFAVI.
Limitations Our analysis has limitations. First, the main limitation of this analysis is the discrepancy in baseline characteristics, which makes comparisons among groups difficult. Although it as an accepted approach to use multivariate models to account for baseline differences, these analyses are explorative, i.e. `hypothesis generating` since there is no perfect statistical method to completely account for this. Although the current study examines the largest cohort available to date, the number of patients is small, making conclusions difficult to draw, especially concerning rare events. Second, the degenerated valves are heterogeneous, and as many as 27 different valve types were used for the primary aortic valve replacement in our study population. In particular, stentless bioprostheses are challenging when it comes to both reoperation and VinV therapy. In re-SAVR, there was a higher rate of those prostheses. In an exploratory analysis (data not shown), outcome was comparable between stentless and stented valves in each treatment strata without a difference in overall mortality or within treatment group. However, we cannot exclude differences between individual valve types. Third, there might have been a learning curve leading to worse outcomes in patients treated by VinV-TFAVI in the earlier years of the study period. Fourth, an echo core lab did not evaluate echocardiography. Last, we only reported mid-term outcomes. There might be differences concerning the durability of the replaced valves or the survival over a longer period.
Conclusion In conclusion, VinV-TFAVI represents an alternative for patients with degenerated aortic bioprosthesis who are at increased risk for a surgical reoperation. However, in patients at low risk for reoperation, a better clinical efficacy and acceptable safety may favour re-SAVR. Individual decision-making within a heart team is mandatory in the treatment of patients with failed aortic bioprostheses.
References
[1] Silaschi M, Conradi L, Treede H, Reiter B, Schaefer U, Blankenberg S, et al. Trends in Surgical Aortic Valve Replacement in More Than 3,000 Consecutive Cases in the Era of Transcatheter Aortic Valve Implantations. The Thoracic and cardiovascular surgeon. 2016;64:382-9. [2] Rodriguez-Gabella T, Voisine P, Puri R, Pibarot P, Rodes-Cabau J. Aortic Bioprosthetic Valve Durability: Incidence, Mechanisms, Predictors, and Management of Surgical and Transcatheter Valve Degeneration. Journal of the American College of Cardiology. 2017;70:1013-28. [3] Maganti M, Rao V, Armstrong S, Feindel CM, Scully HE, David TE. Redo valvular surgery in elderly patients. The Annals of thoracic surgery. 2009;87:521-5. [4] Dvir D, Webb JG, Bleiziffer S, Pasic M, Waksman R, Kodali S, et al. Transcatheter aortic valve implantation in failed bioprosthetic surgical valves. Jama. 2014;312:162-70. [5] Silaschi M, Wendler O, Seiffert M, Castro L, Lubos E, Schirmer J, et al. Transcatheter valve-in-valve implantation versus redo surgical aortic valve replacement in patients with failed aortic bioprostheses. Interactive cardiovascular and thoracic surgery. 2017;24:63-70. [6] Erlebach M, Wottke M, Deutsch MA, Krane M, Piazza N, Lange R, et al. Redo aortic valve surgery versus transcatheter valve-in-valve implantation for failing surgical bioprosthetic valves: consecutive patients in a single-centre setting. Journal of thoracic disease. 2015;7:1494-500. [7] Ejiofor JI, Yammine M, Harloff MT, McGurk S, Muehlschlegel JD, Shekar PS, et al. Reoperative Surgical Aortic Valve Replacement Versus Transcatheter Valve-in-Valve Replacement for Degenerated Bioprosthetic Aortic Valves. The Annals of thoracic surgery. 2016;102:1452-8. [8] Grubitzsch H, Zobel S, Christ T, Holinski S, Stangl K, Treskatsch S, et al. Redo procedures for degenerated stentless aortic xenografts and the role of valve-in-valve
transcatheter techniques. European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery. 2017;51:653-9. [9] Spaziano M, Mylotte D, Theriault-Lauzier P, De Backer O, Sondergaard L, Bosmans J, et al. Transcatheter aortic valve implantation versus redo surgery for failing surgical aortic bioprostheses: a multicentre propensity score analysis. EuroIntervention : journal of EuroPCR in collaboration with the Working Group on Interventional Cardiology of the European Society of Cardiology. 2017;13:1149-56. [10] Gozdek M, Raffa GM, Suwalski P, Kolodziejczak M, Anisimowicz L, Kubica J, et al. Comparative performance of transcatheter aortic valve-in-valve implantation versus conventional surgical redo aortic valve replacement in patients with degenerated aortic valve bioprostheses: systematic review and meta-analysis. European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery. 2017. [11] Nalluri N, Atti V, Munir AB, Karam B, Patel NJ, Kumar V, et al. Valve in valve transcatheter aortic valve implantation (ViV-TAVI) versus redo-Surgical aortic valve replacement (redo-SAVR): A systematic review and meta-analysis. Journal of interventional cardiology. 2018. [12] Neupane S, Singh H, Lammer J, Othman H, Yamasaki H, Rosman HS, et al. MetaAnalysis of Transcatheter Valve-in-Valve Implantation Versus Redo Aortic Valve Surgery for Bioprosthetic Aortic Valve Dysfunction. The American journal of cardiology. 2018;121:1593600. [13] Pascual I, Carro A, Avanzas P, Hernandez-Vaquero D, Diaz R, Rozado J, et al. Vascular approaches for transcatheter aortic valve implantation. Journal of thoracic disease. 2017;9:S478-S87. [14] Blackman DJ, Baxter PD, Gale CP, Moat NE, Maccarthy PA, Hildick-Smith D, et al. Do outcomes from transcatheter aortic valve implantation vary according to access route and valve type? The UK TAVI Registry. Journal of interventional cardiology. 2014;27:86-95. [15] Kappetein AP, Head SJ, Genereux P, Piazza N, van Mieghem NM, Blackstone EH, et al. Updated standardized endpoint definitions for transcatheter aortic valve implantation: the
Valve Academic Research Consortium-2 consensus document. The Journal of thoracic and cardiovascular surgery. 2013;145:6-23. [16] Bapat V. Valve-in-valve apps: why and how they were developed and how to use them. EuroIntervention : journal of EuroPCR in collaboration with the Working Group on Interventional Cardiology of the European Society of Cardiology. 2014;10 Suppl U:U44-51. [17] O'Brien SM, Shahian DM, Filardo G, Ferraris VA, Haan CK, Rich JB, et al. The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 2--isolated valve surgery. The Annals of thoracic surgery. 2009;88:S23-42. [18] Vargha A, Delaney HD. A Critique and Improvement of the CL Common Language Effect Size Statistics of McGraw and Wong. Journal of Educational and Behavioural Statistics. 2000;25:101-32. [19] Mao H, Katz N, Ariyanon W, Blanca-Martos L, Adybelli Z, Giuliani A, et al. Cardiac surgery-associated acute kidney injury. Blood purification. 2014;37 Suppl 2:34-50. [20] Wang J, Yu W, Zhou Y, Yang Y, Li C, Liu N, et al. Independent Risk Factors Contributing to Acute Kidney Injury According to Updated Valve Academic Research Consortium-2 Criteria After Transcatheter Aortic Valve Implantation: A Meta-analysis and Meta-regression of 13 Studies. Journal of cardiothoracic and vascular anesthesia. 2017;31:816-26. [21] Nalluri N, Atti V, Munir AB, Karam B, Patel NJ, Kumar V, et al. Valve in valve transcatheter aortic valve implantation (ViV-TAVI) versus redo-Surgical aortic valve replacement (redo-SAVR): A systematic review and meta-analysis. Journal of interventional cardiology. 2018;31:661-71. [22] Ribeiro HB, Rodes-Cabau J, Blanke P, Leipsic J, Kwan Park J, Bapat V, et al. Incidence, predictors, and clinical outcomes of coronary obstruction following transcatheter aortic valve replacement for degenerative bioprosthetic surgical valves: insights from the VIVID registry. European heart journal. 2018;39:687-95.
[23] Chhatriwalla AK, Allen KB, Saxon JT, Cohen DJ, Aggarwal S, Hart AJ, et al. Bioprosthetic Valve Fracture Improves the Hemodynamic Results of Valve-in-Valve Transcatheter Aortic Valve Replacement. Circulation Cardiovascular interventions. 2017;10.
Figure Legend
Figure 1. Time-to-event curves for all-cause 1-year (A) and cardiovascular 1-year mortality (B) Figure 2. NYHA class at baseline and after 1 year of follow-up (A). Percentage of patients showing any improvement of NYHA class and categorized by in improvement in 1, 2 or 3 classes (B).
Table 1. Baseline characteristics.
Demographics age [years] sex [% male] body mass index [kg/m²] STS-PROM NYHA class, baseline I II III IV NYHA class III or IV concomitant procedures concomitant aortic surgery concomitant mitral valve decalcification concomitant Morrow procedure concomitant ACB concomitant PCI, any PCI in the last 30 days before index procedure medically treated coronary lesion Comorbidities CAD, any 1-vessel 2-vessel 3-vessel previous ACB > 30 days prior previous PCI >30 days prior previous MI atrial fibrillation chronic lung disease none mild moderate severe cerebrovascular disease prior cerebrovascular accident peripheral arterial disease diabetes, any on diet on oral therapy on insulin hypertension immunocompromising medication history of endocarditis GFR [ml/min/1.73m²] GFR <45 ml/min/1.73m² patients on dialysis Echocardiography ejection fraction [%] aortic stenosis, baseline none
re-SAVR n = 111
VinV-TFAVI n = 147
#SMD/ †OR/ ‡probability of superiority
58.5 (14.4) 66/111 (59.9%) 27.5 (4.43) 2.76 % (2.09)
76.2 (8.0) 92/147 (62.6%) 28.4 (5.14) 8.27 % (6.12)
1.58 [1.3; 1.86]# 1.14 [0.67; 1.95] † 0.17 [-0.07; 0.42]# 1.14 [0.88; 1.41]# 41.2 % [34.5%; 48.3%]‡
6/110 (5.5%) 43/110 (39.1%) 53/110 (48.2%) 8/110 (7.3%) 61/110 (55.5%)
9/147 (6.1%) 31/147 (21.1%) 86/147 (58.5%) 21/147 (14.3%) 107/147 (72.8%)
2.15 [1.27; 3.62] †
31/111 (27.9%) 1/111 (0.9%) 10/111 (9.0%) 14/111 (12.6%) 1/111 (0.9%) 0/111 (0%) 3/111 (2.7%)
3/147 (2.1%) 7/147 (4.8%) 6/147 (4.1%)
2.29 [0.24, 22.3] -1.53 [0.38, 6.27]
25/111 (22.5%) 11/111 (9.9%) 10/111 (9.0%) 4/111 (3.6%) 11/111 (9.9%) 7/111 (6.3%) 7/111 (6.3%) 21/111 (18.9%)
75/147 (51.0%) 28/147 (19.0%) 21/147 (14.3%) 26/147 (17.7%) 42/147 (32.7%) 18/147 (12.2%) 13/147 (8.8%) 65/147 (44.2%)
99/111 (89.2%) 3/111 (2.7%) 6/111 (5.4%) 3/111 (2.7%) 10/111 (9.0%) 8/111 (7.2%) 6/111 (5.4%) 18/111 (16.2%) 5/111 (4.5%) 9/111 (8.1%) 4/111 (3.6%) 96/111 (86.5%) 7/111 (6.3%) 6/111 (5.4%) 76.7 (20.5) 8/111 (7.2%) 0
73/145 (50.3%) 27/145 (18.6%) 23/145 (15.9%) 22/145 (15.2%) 26/145 (17.8%) 13/145 (8.8%) 26/147 (17.7%) 53/147 (36.1%) 12/147 (8.2%) 22/147 (15.0%) 19/147 (12.9%) 144/147 (98.0%) 14/147 (9.5%) 12/147 (8.2%) 58.8 (18.7) 37 (25.2%) 2/147 (1.4%)
7.45 [2.03, 41.2] † 1.56 [0.57, 4.74] † 1.55 [0.52, 5.22] † -0.92 [-1.17,-0.66]# 4.33 [1.93, 9.74] † --
57.4 (10.2)
54.5 (13.9)
-0.23 [-0.48, 0.02]#
24/111 (21.6%)
3/147 ( 2.0%)
34.8% [28.5%; 41.8%]‡
4.38 [2.09; 9.93] †
3.38 [1.85, 6.36] † 30.3% [24.4%, 37.1%]‡
2.18 [0.96, 5.32] † 1.25 [0.46, 3.61] † 3.74 [1.43, 11.6] † 2.91 [1.59, 5.34] †
37.3% [30.8%, 44.4%]‡
mild moderate severe aortic regurgitation, baseline none mild moderate severe mitral regurgitation, baseline none mild moderate severe tricuspid regurgitation, baseline none mild moderate severe
11/111 (9.9%) 8/111 (7.2%) 68/111 (61.3%)
10/147 (6.8%) 11/147 (7.5%) 123/147 (83.7%)
26/111 (23.4%) 29/111 (26.1%) 19/111 (17.1%) 37/111 (33.3%)
36/147 (24.5%) 57/147 (38.8%) 28/147 (19.0%) 26/147 (17.7%)
29/111 (26.1%) 73/111 (65.8%) 9/111 (8.1%) 0/111 (0.0%)
12/147 (8.2%) 109/147 (74.1%) 24/147 (16.3%) 2 /147 (1.4%)
38.2% [31.6%, 45.2%]‡
38/111 (34.2%) 66/111 (59.5%) 6/111 (5.4%) 1/111 (0.9%)
21/147 (14.3%) 101/147 (68.7%) 22/147 (15.0%) 3/147 (2.0%)
37.1% [30.7%, 44.2%]‡
57.2% [50%, 64%]‡
Values are depicted as mean (standard deviation) and standardized mean difference (#) for conGnuous variables, n (%) and probability of superiority (‡) for ordinal variables, and n (%) and odds ratio [95% confidence interval] (†) for binary variables. ACB, aorto-coronary bypass grafting; CAD, coronary artery disease; GFR, glomerular filtration rate; MI, myocardial infarction; NYHA, New York Heart Association; PCI, percutaneous coronary intervention; STS-PROM, Society of Thoracic Surgeons – Predicted Risk of Mortality; SMD, standardized mean difference; OR, odds ratio.
Table 2. Procedural outcomes at 30 days. redo-SAVR n = 111
VinV-TFAVI n = 147
Clinical efficacy
36/111 (32.4%)
78/147 (53.1%)
Early safety
72/111 (64.9%)
26/147 (17.7%)
Periprocedural myocardial infarction
2/111 (1.8%)
5/147 (3.4%)
Stroke, any
9/111 (8.1%)
8/147 (5.4%)
Life threatening bleeding
24/111 (21.6%)
9/147 (6.1%)
VARC Major access site complications
17/111 (15.3%)
17/147 (11.6%)
Acute kidney injury, stage II or III
19/111 (17.1%)
8/147 (5.4%)
Necessity of permanent pacemaker implantation Postoperative length of stay
14/111 (12.6%)
20/147 (13.6%)
11 [8, 15]
9 [6, 11]
Erythrocyte conc. packs
2 [1.5, 4.5]
0 [0, 0]
OR / probability of superiority 2.35 [1.41, 3.93] 0.12 [0.066, 0.21] 1.92 [0.37, 10.1] 0.65 [0.24, 1.75] 0.24 [0.10, 0.53] 0.72 [0.35, 1.49] 0.28 [0.12, 0.66] 1.09 [0.53, 2.27] 68.5% [61.6%, 74.5%] 81.4% [75.9%, 86.3%]
p value
< 0.001 < 0.001 0.43 0.39 < 0.001 0.38 0.002 0.82 < 0.001 < 0.001
Definitions according to VARC-2. Values are depicted n (%) and odds ratio [95% CI] for categorical variables and median [IQR] and probability of superiority for length of stay and erythrocyte concentrates. OR, odds ratio. bold: significant at 5% after correction for multiple testing.
A
B
80%
70%
70%
60%
NYHA IV
60%
50%
NYHA III
50%
NYHA II
40%
NYHA I
30%
40% 30% 20%
1.1%
20%
10%
10%
0% redo-SAVR re-SAVR baseline baseline
redo-SAVR re-SAVR 1 1year year
VinV-TFAVI VinV-TFAVI baseline baseline
VinV-TFAVI 1 VinV-TFAVI 1 year year
any improvement
1 NYHA class 2 NYHA classes 3 NYHA classes
3.2%
80%
22.2%
90%
46.0%
90%
71.4%
100%
21.3%
100%
44.9%
B
67.4%
A
0% redo-SAVR re-SAVR
VinV-TFAVI VinV-TFAVI
Highlights
-
VinV-TAVI is a safe and effective treatment for degenerated aortic bioprosthesis
-
data comparing this procedure to redo surgery (re-SAVR) are scarce
-
comparable 1-year-mortality between re-SAVR and VinV-TFAVI
-
incidence of bleeding and renal failure was higher with re-SAVR
-
postoperative transvalvular gradients were higher after VinV-TFAVI
S0167-5273(19)32470-2
DOI:
https://doi.org/10.1016/j.ijcard.2019.09.039
Reference:
IJCA 28014
To appear in:
International Journal of Cardiology
Received Date: 13 May 2019 Revised Date:
10 July 2019
Accepted Date: 16 September 2019
Please cite this article as: F.J. Woitek, G. Stachel, P. Kiefer, S. Haussig, S. Leontyev, F. Schlotter, M. Mende, J. Hommel, L. Crusius, A. Spindler, F.W. Mohr, G. Schuler, H. Thiele, M.A. Borger, A. Linke, D. Holzhey, N. Mangner, Treatment of failed aortic bioprostheses: An evaluation of conventional redo surgery and transfemoral transcatheter aortic valve-in-valve implantation, International Journal of Cardiology (2019), doi: https://doi.org/10.1016/j.ijcard.2019.09.039. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier B.V.
Treatment of failed aortic bioprostheses: An evaluation of conventional redo surgery and transfemoral transcatheter aortic valve-in-valve implantation
Felix J. Woitek, MD*a, Georg Stachel, MD*b, Philipp Kiefer, MDc, Stephan Haussig, MDa, Sergey Leontyev, MDc, Florian Schlotter, MDb, Meinhard Mende, PhDd, Jennifer Hommel, MAa, Lisa Crusius, MDb, Aileen Spindlerb, Friedrich W. Mohr, MDc, Gerhard Schuler, MDb, Holger Thiele, MDb ,Michael A. Borger, MD PhDc , Axel Linke, MDa, David Holzhey, MD**c Norman Mangner, MD**a a
Herzzentrum Dresden, Technische Universität Dresden, Department of Internal Medicine
and Cardiology, Dresden, Germany b
Heart Centre Leipzig at University of Leipzig, Department of Internal Medicine /Cardiology,
Leipzig, Germany c
Heart Centre Leipzig at University of Leipzig, Department of Cardiac Surgery, Leipzig,
Germany d
Institute for Medical Informatic, Statistics and Epidemiology, University of Leipzig, Leipzig,
Germany * both authors contributed equally and are shared first authors ** both authors contributed equally and are shared last authors
Running title: Re-SAVR compared to valve-in-valve-TFAVI in failed aortic bioprostheses Total word count (text and figure legend): 3499 Address for correspondence: Norman Mangner Herzzentrum Dresden, Technische Universität Dresden Department of Internal Medicine and Cardiology Fetscherstr. 76, 01307 Dresden / Germany Tel. +49 351 450 25 297 [email protected]
Conflict of interest Felix J. Woitek: The author reports no relationships that could be construed as a conflict of interest. Georg Stachel: The author reports no relationships that could be construed as a conflict of interest. Philipp Kiefer: The author reports no relationships that could be construed as a conflict of interest. Stephan Haussig: The author reports no relationships that could be construed as a conflict of interest. Sergey Leontyev reports speaker´s honoraria from St. Jude Medical and Medtronic, outside the submitted work. Florian Schlotter: The author reports no relationships that could be construed as a conflict of interest. Meinhard Mende: The author reports no relationships that could be construed as a conflict of interest. Jennifer Hommel: The author reports no relationships that could be construed as a conflict of interest. Lisa Crusius: The author reports no relationships that could be construed as a conflict of interest. Aileen Spindler: The author reports no relationships that could be construed as a conflict of interest. Friedrich W. Mohr: The author reports no relationships that could be construed as a conflict of interest. Gerhard Schuler: The author reports no relationships that could be construed as a conflict of interest. Holger Thiele: The author reports no relationships that could be construed as a conflict of interest.
Michael A. Borger declares personal fees from Edwards Lifesciences, personal fees from Medtronic, personal fees from CryoLife, outside the submitted work. Axel Linke reports grants and personal fees from Medtronic, personal fees from St. Jude Medical, grants from Claret Medical, personal fees and other from Claret Medical, personal fees from Boston Scientific, personal fees from Bard, personal fees from Edwards, outside the submitted work. David Holzhey reports speaker´s honoraria from Symetis and Medtronic, outside the submitted work. Norman Mangner reports speaker´s honoraria from Edwards, Medtronic, Novartis, Sanofi Genzyme and Astra Zeneca, consultant honoraria from Biotronik, outside the submitted work.
Abstract Background: The use of bioprostheses for surgical aortic valve replacement increased substantially within the last years. In case of prosthesis failure, re-SAVR is standard of care, whereas valve-in-valve deployment of a transfemoral transcatheter aortic valve prosthesis (VinV-TFAVI) has recently emerged as an alternative. We sought to evaluate early safety, clinical efficacy, and all-cause 1-year-mortality of VinV-TFAVI and redo surgery for failing aortic bioprostheses (re-SAVR). Methods and results: Patients receiving either VinV-TFAVI (n=147) or re-SAVR (n=111) for a degenerated aortic bioprosthesis between 01/2006 and 05/2017 were included in this analysis. All-cause 1-year mortality was the primary outcome measure. Early safety and clinical efficacy according to VARC-2 endpoint definitions were evaluated at 30 days. Baseline characteristics differed significantly between both groups, including age, STSPROM, and incidence of relevant comorbidities. Re-stenosis was the predominant mode of failure in 45.9% of re-SAVR and 63.1% of VinV-TFAVI patients. The rate of “early safety” endpoints was lower with VinV-TFAVI (17.7% vs. 64.9%, p<0.01), the rate of “clinical efficacy” endpoints was lower, e.g. better with re-SAVR (53.1% vs. 32.4%, p<0.01). All-cause 1-year-mortality (VinV-TFAVI 8.8% vs re-SAVR 9.9%, p= 0.84) was not different. Treatment strategy was not associated with 1-year-mortality in a Cox regression analysis. The incidence of prosthesis-patient-mismatch was higher in VinV-TFAVI compared to re-SAVR. Conclusion: VinV-TFAVI represents a viable alternative for treatment of degenerated aortic bioprostheses in patients at increased surgical risk. However, in patients at low risk for reoperation, a better clinical efficacy and acceptable safety may favour re-SAVR.
244/250 words
Key words: Aortic valve stenosis; Heart valve prosthesis implantation; Transcatheter aortic valve replacement; Prosthesis failure; valve in valve
Abbreviations AVA aortic valve area BEV balloon-expandable valve NYHA New York Heart Association functional class PPM patient-prosthesis-mismatch SAVR surgical aortic valve replacement SEV self-expandable valve SMD standardized mean difference STS-PROM Society of Thoracic Surgeons – Predicted Risk of Mortality VinV-TFAVI valve-in-valve transfemoral aortic valve implantation VARC-2 Valve Academic Research Consortium -2
Introduction There has been a marked increase in the use of bioprostheses for surgical aortic valve replacement (SAVR) as compared to mechanical prostheses [1]. A major disadvantage of bioprostheses is a higher tendency towards structural deterioration and failure, thus frequently necessitating replacement after 10 to 20 years [2]. Standard procedure for the replacement of failing aortic bioprostheses is currently re-SAVR. However, management of these patients is challenging due to their commonly advanced age and the increased risk conferred by repeat cardiac surgery [3]. Hence, ‘valve-in-valve’ transcatheter aortic valve implantation (VinV-TAVI) has emerged in the last years, and a multicentre registry has demonstrated encouraging results concerning efficacy and safety [4] of this procedure. Direct comparisons between re-SAVR and VinV-TAVI are confined to a small number of reports consisting of small case series [5-8] with one larger trial employing propensity score matching [9] and three meta-analyses of these non-randomized trials [10-12]. Only one trial [6] demonstrated a higher 1-year-mortality after VinV-TAVI compared to re-SAVR, whereas all other trials did not show significant differences in survival between both treatment strategies despite significantly higher operative risk scores in the VinV-TAVI group. Results on the incidence of increased postoperative transvalvular gradients are ambiguous: some trials found a higher incidence in the VinV-TAVI group [5, 6], while others did not [7, 8]. Interpreting the results of these studies, it has to be taken into account that a large proportion of patients – as compared to TAVI in native valve aortic stenosis [13] - was treated via nontransfemoral access, which may have influenced the outcomes [14]. In our study, we sought to evaluate 1-year mortality as well as combined clinical efficacy and safety endpoints at 30 days [15] in patients receiving either valve-in-valve transfemoral TAVI (VinV-TFAVI) or re-SAVR.
Methods Patient Selection Between January 2006 and May 2017, 111 patients received re-SAVR for a degenerated aortic bioprosthesis at the Heart Centre Leipzig, a tertiary referral centre. VinV-TFAVI for degenerated aortic bioprostheses was performed in 147 patients between September 2007 and May 2017. After the establishment of the TAVI program at our hospital in 2006, any treatment strategy decision was made after discussion of all options in the multidisciplinary heart team. Patients receiving re-SAVR were excluded from analysis if VinV-TFAVI had not been possible, e.g. in the case of a degenerated mechanical valve, necessity of surgical treatment of other valves (other than mitral valve decalcification) or conditions not amenable to transcatheter treatment, e.g. infective endocarditis or paravalvular leaks as separate indication of the procedure. Experienced consultant cardiac surgeons performed all redosternotomies and –operations. A pre-operative computed tomography scan was performed in 100% of the patients receiving VinV-TFAVI group and 94.6% of the patients treated by reSAVR group. Baseline characteristics, procedural data and outcome data were prospectively collected. True inner diameter of the valves was determined as published [16]. Mode of failure was classified as ‘mixed’ if regurgitation ≥ grade 2 (on a scale of 3) was present in a predominantly stenotic valve and vice versa. Follow‐up was performed after 30 days and 1 year. Presence of lung disease, immunosuppressant medication, diabetes mellitus, coronary heart disease (CAD), cerebrovascular disease and peripheral artery disease were defined according to the STS-PROM [17]. The registry was approved by the Ethics Committee of the Medical Faculty of the University of Leipzig (registration no. 428/17-ek), and all patients gave written informed consent.
Outcome measures The primary outcome measure was all-cause 1-year mortality. Clinical efficacy (all-cause mortality, all stroke, New York Heart Association (NYHA) functional class III or IV or valve-
related dysfunction) and early safety (all-cause mortality, all stroke, life-threatening bleeding, acute kidney injury stage 2 or 3, coronary artery obstruction or valvular dysfunction requiring repeat procedure) were assessed at 30 days according to the definition of the Valve Academic Research Consortium -2 (VARC-2) [15]. To constitute a bleeding event in the reSAVR cohort, a source of bleeding other than the index surgical procedure (e.g. haemothorax or cardiac tamponade) had to exist in addition to the drop in haemoglobin concentration or amount of transfused packed red cell units as required by VARC-2. Data on NYHA functional class at 1 year, postoperative gradient and prosthesis-patient-mismatch (PPM) were collected. PPM and aortic regurgitation were graded as defined by the VARC-2. Haemodynamic parameters were evaluated within the whole cohort of each treatment group and according to the initial true inner diameter of the degenerated prosthesis (<20mm, 20.022.99mm, ≥23mm).
Statistical Analysis Based on the two very distinct groups in a retrospective analysis, we decided not to perform a propensity or other matching. Instead, we calculated difference measures for the baseline characteristics and adjusted for covariates associated with the primary outcome measure. Standardized mean differences (for continuous variables), odds ratios (for binary variables) and the measure A12 of stochastic superiority [18] (for ordinary variables) characterize the balance between groups. If (X, Y) is a randomly selected pair of values from both groups, A12 is interpreted as the probability that X > Y plus ½ the probability that X = Y; that is, A12 = P{X > Y} + ½ P{X = Y}. 95% confidence intervals were calculated. The study cohort is characterized by mean (standard deviation, SD) for continuous, absolute and relative frequencies for categorical variables. The groups were compared with respect to categorical variables by means of chi-squared and Fisher's exact test as appropriate. We compared means of continuous characteristics by t-test (Welch) and skew distributed traits (time spans) by Mann-Whitney U test. Median differences were estimated by the HodgesLehman method.
Survival curves were estimated and depicted applying the Kaplan-Meier algorithm. To adjust for covariates, multiple Cox regression models were built. We started with baseline traits obviously different and performed a backward selection using the AIC criterion including treatment group (VinV-TFAVI), age (per year), NYHA III or IV at baseline, sex (male), STSPROM, CAD at baseline, and mode of failure (regurgitation). A final model with treatment group forced into it was built to get risk estimates inclusive 95% confidence intervals. NYHA class was analyzed by means of a general linear model including the baseline category as covariate. We defined the significance level at 5% for two-tailed testing. Tables and text show raw pvalues. However, they were adjusted for multiple testing following the method of Bonferroni and Holm and depicted in bold for remaining significance. Data preparation, descriptive statistics and tests for group comparisons were performed by IBM SPSS Statistics (version 24). The program R (version 3.4.1, R Core Team, 2017) inclusive the software packages orddom and survival was applied for all other analyses and graph design.
Results Baseline characteristics Out of 258 patients, 111 (43%) and 147 (57%) received re-SAVR or VinV-TFAVI, respectively. VinV-TFAVI were older, had higher STS-PROM, and higher incidences of nearly all evaluated comorbidities (Table 1). Baseline NYHA functional class was worse in VinV-TFAVI, as was renal function. The incidence of CAD and number of affected vessels was higher in VinV-TFAVI. There was no significant difference in left ventricular ejection fraction but there was a higher frequency of severe mitral and tricuspid regurgitation in VinVTFAVI (Table 1). In re-SAVR, 31 patients (27%) underwent concomitant surgery on the thoracic aorta, one patient (0.9%) received additional mitral valve decalcification, 10 patients (9.0%) received a Morrow procedure, and 14 patients (12.6%) had concomitant coronary artery bypass grafting. Percutaneous coronary intervention was performed in one patient (0.9%) during the same hospital stay. Three patients (2.7%) had medically managed coronary artery disease. In VinV-TFAVI, seven patients (4.7%) received percutaneous coronary intervention within the preceding 30 days, three patients (2.1%) within the same hospital stay. CAD was medically treated in six patients (4.1%). Concerning the characteristics of the degenerated bioprosthesis, the majority of patients had been treated with stented valves, even more frequently in the VinV-TFAVI-group. Predominant mode of failure was stenosis, again with a higher incidence of stenotic and mixed pathologies in the VinV-TFAVI group (Online-only Table 1).
Procedural and short-term outcomes The combined ‘early safety’ endpoint was met significantly more often in re-SAVR compared to VinV-TFAVI (64.9% vs. 17.7%, p<0.001, Table 2) indicating an inferior safety. Conversely, significantly less patients in re-SAVR reached the composite ‘clinical efficacy’ endpoint (32.4% vs. 53.1%, p=0.001) indicating a superior efficacy. The higher incidence of the ‘early safety’ endpoint after re-SAVR was mainly driven by a higher incidence of life-threatening
bleeding and renal failure of stages 2 and 3. The number of transfused red packed blood cells was higher in re-SAVR. Repeat procedures or conversions were not necessary. Higher NYHA functional classes and higher incidences of increased gradients and PPM in VinVTFAVI (Online-only Table 1) mainly caused the difference concerning the ‘clinical efficacy’ endpoint. The incidences of postprocedural myocardial infarction, stroke and new permanent pacemaker implantation did not differ significantly between groups (Table 2). Postoperative length of stay was longer in re-SAVR compared to VinV-TFAVI. Thirty-day-mortality was 4.1% with re-SAVR and 4.5% with VinV-TFAVI (p=0.87) and the corresponding observed-topredicted 30-day mortality was 1.49 vs. 0.54.
Haemodynamic parameters after valve replacement Peak and mean postoperative transvalvular gradients were significantly higher in VinV-TFAVI (Online-only Table 1). Accordingly, AVA and AVA index were significantly larger in the surgically treated group. Consequently, prevalence of PPM was significantly higher in VinVTFAVI. Mild prosthetic aortic valve regurgitation after replacement was more frequent after VinV-TFAVI. Moderate aortic regurgitation was rare and not different between groups, severe aortic regurgitation was found in neither group (Online-only Table 1). Patients treated for stenotic/mixed degeneration had higher rates of any PPM compared to patients treated for isolated regurgitation (58.9% vs. 27.1%, p<0.001) in the whole cohort irrespective of treatment form as well as in VinV-TFAVI (66.4% vs. 38.5%, p=0.046) and reSAVR (43.5 vs. 22.9%, p=0.042). Noteworthy, the rate of stenotic/mixed mode of prosthesis failure was higher in VinV-TFAVI compared to re-SAVR who had a higher proportion of patients treated for isolated aortic regurgitation (Online-only Table 1). Haemodynamic parameters were also evaluated according the true internal diameter of the failed bioprosthesis (Online-only Table 2). The proportion of patients with a true-ID <20mm, 20.0-22.99mm and ≥23mm was 47.6%, 32.9% and 19.6% in VinV-TFAVI and 34.7%, 29.5% and 35.8% in re-SAVR, respectively (p=0.017). In re-SAVR, 21 patients (18.9%) received a mechanical prosthesis as a second valve. The mean true-ID of the second prostheses was
significantly greater in the group having an initial true-ID <20mm and 20.0-22.99mm, but was smaller in ≥23mm. Haemodynamic parameters within the three groups showed the same pattern with higher gradients and lower aortic valve area and indexed aortic valve area in VinV-TFAVI. Only in patients with an initial true-ID of 20.0-20.99mm, equivalent aortic valve area and indexed aortic valve area were evident leading to a comparable rate of PPM in this group. The rate of aortic regurgitation was similar in the groups having a true-ID <20mm and 20.0-20.99mm, but higher in VinV-TFAVI having an initial true-ID ≥23mm (Online-only Table 2).
Analysis of all-cause 1-year mortality All-cause 1-year mortality was 9.9% in the re-SAVR-group and 8.8% in the VinV-TFAVI group (p=0.76, Figure 1). Cardiovascular mortality after 1 year was 8.1% and 7.8%, respectively (p=0.84). There was no significant difference in 1-year-mortality after exclusion of surgical patients with concomitant procedures (Online-only Figure 1). In a multiple model including treatment group, only NYHA class at baseline (HR 6.35, 95% CI: 2.29 – 17.7) and sex (HR 0.38 for male sex, 95% CI 0.15 – 0.97) remained as factors associated with allcause 1-year mortality (Online-only Table 3).
Functional outcome at 1 year Concerning NYHA functional class, patients in the VinV-TFAVI group had worse functional status both at baseline and after 1 year than in the re-SAVR group (mean +0.22, p=0.002). However, both groups improved in functional status by about one NYHA class as compared to baseline (-0.95, p<0.001, Figure 2A). In the general linear model used for analysis, the interaction term time*group was excluded because it did not improve the model. Based on this, there was no evidence of differential improvement of NYHA functional class between groups (Figure 2B, Online-only Table 4).
Comparison of balloon-expandable and self-expandable valves
In VinV-TFAVI, 43 patients (29.5%) were treated with balloon-expandable valves (BEV-VinVTFAVI) and 103 (70.5%) were treated with self-expandable valves (SEV-VinV-TFAVI). One patient received a mechanically expanding valve. SEV-VinV-TFAVI had degenerated valves with smaller true inner diameter than those receiving re-SAVR (Online-only Table 5). This was not observed in BEV-VinV-TFAVI. Postoperative pressure gradients were significantly higher and aortic valve indexes were significantly lower for both valve types compared to reSAVR. Incidence of any postoperative aortic regurgitation was significantly higher after SEVVinV-TFAVI than in re-SAVR; predominantly driven by mild aortic regurgitation. There was no significant difference in the incidence of aortic regurgitation between re-SAVR and BEVVinV-TFAVI (Online-only Table 5). Comparing BEV-VinV-TFAVI with SEV-VinV-TFAVI, self-expanding valves had a tendency towards a better haemodynamic profile without reaching statistical significance (Online-only Table 6).
Discussion The main findings of this analysis are: (1) Although VinV-TFAVI constituted a group at higher surgical risk, no significant differences in survival after 1 year were observed, and treatment strategy was not associated with all-cause 1-year mortality in a multivariate model adjusted for baseline differences. (2) Incidence of the combined ‘early safety’- endpoint at 30 days, most likely driven by a higher incidence of life threatening bleeding and acute kidney injury was significantly higher in the re-SAVR-group. The incidence of the ‘clinical efficacy’ endpoint was higher, e.g. worse, in the VinV-TFAVI-group. (3) Re-SAVR resulted in a better haemodynamic profile with lower transvalvular gradients, larger effective orifice areas and a lower incidence of patient-prosthesis mismatch; however, this led neither to differences in survival nor to differences in the improvement of functional status after 1 year.
Comparison of mortality We did not detect a significant difference in all-cause 1-year mortality between treatment groups, which is in line with the majority of previous work [11]. This finding should be cautiously interpreted, as patients in the VinV-TAVI cohort were about 18 years older and at higher surgical risk but approximately half of the patients receiving re-SAVR had concomitant procedures. After accounting for these differences in a multivariate model, mode of replacement (re-SAVR vs. VinV-TFAVI) was not significantly associated with 1-year mortality and the observed-to-predicted 30-day mortality was in favour for VinV-TFAVI. Moreover, there was no difference in mortality after exclusion of patients receiving concomitant procedures during re-SAVR. Spaziano et al. [9] and Ejiofor et al. [7] used propensity score matching to account for differences, both coming to similar results, despite persisting differences in procedural and baseline characteristics of their cohorts and smaller patient numbers. We employed a multivariate analysis as there was only a small overlap between groups in baseline characteristics and there is no evidence for the superiority of one or the other statistical method.
In our multivariate model, mode of failure, i.e. patients presenting with stenotic failure as compared to those with severe regurgitation or mixed pathology as indication for valve replacement, was not associated with differences in mortality. In other works, most notably the VIVID registry [4], VinV-TAVI patients with stenotic failure had worse outcome. However, patients with stenotic and mixed failure experienced more often a PPM compared to patients having a regurgitation as the predominant mode of failure.
Short-term outcomes and adverse events Patients receiving re-SAVR significantly more often reached the composite endpoint for early safety, and conversely, the endpoint for clinical efficacy was more often reached in patients receiving VinV-TFAVI, most likely driven by a higher incidence of bleeding and kidney failure in the further group and elevated gradient and worse NYHA functional class in the latter. The higher incidence of bleeding and acute kidney injury in the surgical group is most likely due to the greater invasiveness of an operative valve replacement and the use of cardiopulmonary bypass [19], although all procedures were performed by experienced consultant cardiac surgeons. Acute kidney injury is known to be associated with bleeding events and consecutive need for blood transfusion [20] indicating that the lower number of bleeding events and significantly lower amounts of blood transfusion in VinV-TFAVI might have contributed to the lower incidence of acute kidney injury. The incidence of stroke was not significantly different between groups, which is comparable to the finding of others [21]. Of note, we could not prove a significant difference in the comparative incidence of myocardial infarction and is within the range reported by larger registries and seems to be substantially higher than in native aortic valve TAVI [22].
Haemodynamic profile Higher transvalvular gradients and lower aortic valve areas translated into a higher incidence of PPM after VinV-TFAVI. Noteworthy, VinV-TFAVI had higher rates of stented prostheses, more often stenotic/mixed prosthesis degeneration and in nearly half of the patients a
prosthesis with a true internal diameter <20mm. On the other hand, 18.9% of the patients in re-SAVR received a mechanical prosthesis and the true internal diameter of the second prosthesis was significantly greater in the groups having an initial true-ID <20mm and 20.022.99mm. Interestingly, the numerical values of the haemodynamic parameters were quite similar within each treatment strata irrespective of initial true-ID. These are all factors leading to a worse haemodynamic outcome in VinV-TFAVI, which is, of course, a completely different concept compared to re-SAVR and, therefore, must be associated with higher gradients compared to a new surgical prosthesis. At least up to 1 year, the haemodynamic differences between re-SAVR and VinV-TFAVI did not affect mortality or improvement in NYHA classes. The higher NYHA classes in VinV-TFAVI, both at baseline and follow-up, might be correlated to advanced age and the reported pulmonary and cardiac comorbidities. A longer follow-up is necessary to evaluate the impact of worse haemodynamic in VinVTFAVI on mortality and functional status. However, this cohort and all others treated so far by VinV might not be the right one for comparison due to the advanced age and limited life expectancy. One study comparing only patients with degenerated stentless aortic valves [8] did not see that extent of difference between re-SAVR and VinV-TAVI concerning transvalvular gradients. In our analysis, stentless prostheses were more often treated by re-SAVR than VinV-TFAVI, however, numbers were low preventing us from a solid analysis on the comparison between stented vs. stentless prostheses. One possibility to improve the haemodynamic profile of VinV-TFAVI, in particular in small prostheses, might be controlled ring fracture of the degenerated prosthesis with high-pressure balloons [23], with risk-benefitprofile of this technique remaining to be elucidated in larger studies. Overall, we could corroborate past findings concerning the comparison between VinV-TFAVI and re-SAVR. The safety profile appears to be better in VinV-TFAVI, whereas in particular the rate of patient-prosthesis-mismatch is higher in VinV-TFAVI leading to an individualized decision process to offer patients with failed aortic bioprostheses the most reasonable therapy.
Balloon-expandable versus self-expandable valves Theoretically, prostheses with leaflets located supra-annularly, in particular the Medtronic CoreValve, should have a slightly better haemodynamic profile, as, thus maximizing available orifice area. In our study, we could not prove a better haemodynamic profile of SEV-TFAVI vs. BEV-TFAVI, however there seems to be a small effect. The missing benefit might also be due to a lower true inner diameter of the initial bioprosthesis and a higher rate of stenotic failure in SEV-TFAVI.
Limitations Our analysis has limitations. First, the main limitation of this analysis is the discrepancy in baseline characteristics, which makes comparisons among groups difficult. Although it as an accepted approach to use multivariate models to account for baseline differences, these analyses are explorative, i.e. `hypothesis generating` since there is no perfect statistical method to completely account for this. Although the current study examines the largest cohort available to date, the number of patients is small, making conclusions difficult to draw, especially concerning rare events. Second, the degenerated valves are heterogeneous, and as many as 27 different valve types were used for the primary aortic valve replacement in our study population. In particular, stentless bioprostheses are challenging when it comes to both reoperation and VinV therapy. In re-SAVR, there was a higher rate of those prostheses. In an exploratory analysis (data not shown), outcome was comparable between stentless and stented valves in each treatment strata without a difference in overall mortality or within treatment group. However, we cannot exclude differences between individual valve types. Third, there might have been a learning curve leading to worse outcomes in patients treated by VinV-TFAVI in the earlier years of the study period. Fourth, an echo core lab did not evaluate echocardiography. Last, we only reported mid-term outcomes. There might be differences concerning the durability of the replaced valves or the survival over a longer period.
Conclusion In conclusion, VinV-TFAVI represents an alternative for patients with degenerated aortic bioprosthesis who are at increased risk for a surgical reoperation. However, in patients at low risk for reoperation, a better clinical efficacy and acceptable safety may favour re-SAVR. Individual decision-making within a heart team is mandatory in the treatment of patients with failed aortic bioprostheses.
References
[1] Silaschi M, Conradi L, Treede H, Reiter B, Schaefer U, Blankenberg S, et al. Trends in Surgical Aortic Valve Replacement in More Than 3,000 Consecutive Cases in the Era of Transcatheter Aortic Valve Implantations. The Thoracic and cardiovascular surgeon. 2016;64:382-9. [2] Rodriguez-Gabella T, Voisine P, Puri R, Pibarot P, Rodes-Cabau J. Aortic Bioprosthetic Valve Durability: Incidence, Mechanisms, Predictors, and Management of Surgical and Transcatheter Valve Degeneration. Journal of the American College of Cardiology. 2017;70:1013-28. [3] Maganti M, Rao V, Armstrong S, Feindel CM, Scully HE, David TE. Redo valvular surgery in elderly patients. The Annals of thoracic surgery. 2009;87:521-5. [4] Dvir D, Webb JG, Bleiziffer S, Pasic M, Waksman R, Kodali S, et al. Transcatheter aortic valve implantation in failed bioprosthetic surgical valves. Jama. 2014;312:162-70. [5] Silaschi M, Wendler O, Seiffert M, Castro L, Lubos E, Schirmer J, et al. Transcatheter valve-in-valve implantation versus redo surgical aortic valve replacement in patients with failed aortic bioprostheses. Interactive cardiovascular and thoracic surgery. 2017;24:63-70. [6] Erlebach M, Wottke M, Deutsch MA, Krane M, Piazza N, Lange R, et al. Redo aortic valve surgery versus transcatheter valve-in-valve implantation for failing surgical bioprosthetic valves: consecutive patients in a single-centre setting. Journal of thoracic disease. 2015;7:1494-500. [7] Ejiofor JI, Yammine M, Harloff MT, McGurk S, Muehlschlegel JD, Shekar PS, et al. Reoperative Surgical Aortic Valve Replacement Versus Transcatheter Valve-in-Valve Replacement for Degenerated Bioprosthetic Aortic Valves. The Annals of thoracic surgery. 2016;102:1452-8. [8] Grubitzsch H, Zobel S, Christ T, Holinski S, Stangl K, Treskatsch S, et al. Redo procedures for degenerated stentless aortic xenografts and the role of valve-in-valve
transcatheter techniques. European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery. 2017;51:653-9. [9] Spaziano M, Mylotte D, Theriault-Lauzier P, De Backer O, Sondergaard L, Bosmans J, et al. Transcatheter aortic valve implantation versus redo surgery for failing surgical aortic bioprostheses: a multicentre propensity score analysis. EuroIntervention : journal of EuroPCR in collaboration with the Working Group on Interventional Cardiology of the European Society of Cardiology. 2017;13:1149-56. [10] Gozdek M, Raffa GM, Suwalski P, Kolodziejczak M, Anisimowicz L, Kubica J, et al. Comparative performance of transcatheter aortic valve-in-valve implantation versus conventional surgical redo aortic valve replacement in patients with degenerated aortic valve bioprostheses: systematic review and meta-analysis. European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery. 2017. [11] Nalluri N, Atti V, Munir AB, Karam B, Patel NJ, Kumar V, et al. Valve in valve transcatheter aortic valve implantation (ViV-TAVI) versus redo-Surgical aortic valve replacement (redo-SAVR): A systematic review and meta-analysis. Journal of interventional cardiology. 2018. [12] Neupane S, Singh H, Lammer J, Othman H, Yamasaki H, Rosman HS, et al. MetaAnalysis of Transcatheter Valve-in-Valve Implantation Versus Redo Aortic Valve Surgery for Bioprosthetic Aortic Valve Dysfunction. The American journal of cardiology. 2018;121:1593600. [13] Pascual I, Carro A, Avanzas P, Hernandez-Vaquero D, Diaz R, Rozado J, et al. Vascular approaches for transcatheter aortic valve implantation. Journal of thoracic disease. 2017;9:S478-S87. [14] Blackman DJ, Baxter PD, Gale CP, Moat NE, Maccarthy PA, Hildick-Smith D, et al. Do outcomes from transcatheter aortic valve implantation vary according to access route and valve type? The UK TAVI Registry. Journal of interventional cardiology. 2014;27:86-95. [15] Kappetein AP, Head SJ, Genereux P, Piazza N, van Mieghem NM, Blackstone EH, et al. Updated standardized endpoint definitions for transcatheter aortic valve implantation: the
Valve Academic Research Consortium-2 consensus document. The Journal of thoracic and cardiovascular surgery. 2013;145:6-23. [16] Bapat V. Valve-in-valve apps: why and how they were developed and how to use them. EuroIntervention : journal of EuroPCR in collaboration with the Working Group on Interventional Cardiology of the European Society of Cardiology. 2014;10 Suppl U:U44-51. [17] O'Brien SM, Shahian DM, Filardo G, Ferraris VA, Haan CK, Rich JB, et al. The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 2--isolated valve surgery. The Annals of thoracic surgery. 2009;88:S23-42. [18] Vargha A, Delaney HD. A Critique and Improvement of the CL Common Language Effect Size Statistics of McGraw and Wong. Journal of Educational and Behavioural Statistics. 2000;25:101-32. [19] Mao H, Katz N, Ariyanon W, Blanca-Martos L, Adybelli Z, Giuliani A, et al. Cardiac surgery-associated acute kidney injury. Blood purification. 2014;37 Suppl 2:34-50. [20] Wang J, Yu W, Zhou Y, Yang Y, Li C, Liu N, et al. Independent Risk Factors Contributing to Acute Kidney Injury According to Updated Valve Academic Research Consortium-2 Criteria After Transcatheter Aortic Valve Implantation: A Meta-analysis and Meta-regression of 13 Studies. Journal of cardiothoracic and vascular anesthesia. 2017;31:816-26. [21] Nalluri N, Atti V, Munir AB, Karam B, Patel NJ, Kumar V, et al. Valve in valve transcatheter aortic valve implantation (ViV-TAVI) versus redo-Surgical aortic valve replacement (redo-SAVR): A systematic review and meta-analysis. Journal of interventional cardiology. 2018;31:661-71. [22] Ribeiro HB, Rodes-Cabau J, Blanke P, Leipsic J, Kwan Park J, Bapat V, et al. Incidence, predictors, and clinical outcomes of coronary obstruction following transcatheter aortic valve replacement for degenerative bioprosthetic surgical valves: insights from the VIVID registry. European heart journal. 2018;39:687-95.
[23] Chhatriwalla AK, Allen KB, Saxon JT, Cohen DJ, Aggarwal S, Hart AJ, et al. Bioprosthetic Valve Fracture Improves the Hemodynamic Results of Valve-in-Valve Transcatheter Aortic Valve Replacement. Circulation Cardiovascular interventions. 2017;10.
Figure Legend
Figure 1. Time-to-event curves for all-cause 1-year (A) and cardiovascular 1-year mortality (B) Figure 2. NYHA class at baseline and after 1 year of follow-up (A). Percentage of patients showing any improvement of NYHA class and categorized by in improvement in 1, 2 or 3 classes (B).
Table 1. Baseline characteristics.
Demographics age [years] sex [% male] body mass index [kg/m²] STS-PROM NYHA class, baseline I II III IV NYHA class III or IV concomitant procedures concomitant aortic surgery concomitant mitral valve decalcification concomitant Morrow procedure concomitant ACB concomitant PCI, any PCI in the last 30 days before index procedure medically treated coronary lesion Comorbidities CAD, any 1-vessel 2-vessel 3-vessel previous ACB > 30 days prior previous PCI >30 days prior previous MI atrial fibrillation chronic lung disease none mild moderate severe cerebrovascular disease prior cerebrovascular accident peripheral arterial disease diabetes, any on diet on oral therapy on insulin hypertension immunocompromising medication history of endocarditis GFR [ml/min/1.73m²] GFR <45 ml/min/1.73m² patients on dialysis Echocardiography ejection fraction [%] aortic stenosis, baseline none
re-SAVR n = 111
VinV-TFAVI n = 147
#SMD/ †OR/ ‡probability of superiority
58.5 (14.4) 66/111 (59.9%) 27.5 (4.43) 2.76 % (2.09)
76.2 (8.0) 92/147 (62.6%) 28.4 (5.14) 8.27 % (6.12)
1.58 [1.3; 1.86]# 1.14 [0.67; 1.95] † 0.17 [-0.07; 0.42]# 1.14 [0.88; 1.41]# 41.2 % [34.5%; 48.3%]‡
6/110 (5.5%) 43/110 (39.1%) 53/110 (48.2%) 8/110 (7.3%) 61/110 (55.5%)
9/147 (6.1%) 31/147 (21.1%) 86/147 (58.5%) 21/147 (14.3%) 107/147 (72.8%)
2.15 [1.27; 3.62] †
31/111 (27.9%) 1/111 (0.9%) 10/111 (9.0%) 14/111 (12.6%) 1/111 (0.9%) 0/111 (0%) 3/111 (2.7%)
3/147 (2.1%) 7/147 (4.8%) 6/147 (4.1%)
2.29 [0.24, 22.3] -1.53 [0.38, 6.27]
25/111 (22.5%) 11/111 (9.9%) 10/111 (9.0%) 4/111 (3.6%) 11/111 (9.9%) 7/111 (6.3%) 7/111 (6.3%) 21/111 (18.9%)
75/147 (51.0%) 28/147 (19.0%) 21/147 (14.3%) 26/147 (17.7%) 42/147 (32.7%) 18/147 (12.2%) 13/147 (8.8%) 65/147 (44.2%)
99/111 (89.2%) 3/111 (2.7%) 6/111 (5.4%) 3/111 (2.7%) 10/111 (9.0%) 8/111 (7.2%) 6/111 (5.4%) 18/111 (16.2%) 5/111 (4.5%) 9/111 (8.1%) 4/111 (3.6%) 96/111 (86.5%) 7/111 (6.3%) 6/111 (5.4%) 76.7 (20.5) 8/111 (7.2%) 0
73/145 (50.3%) 27/145 (18.6%) 23/145 (15.9%) 22/145 (15.2%) 26/145 (17.8%) 13/145 (8.8%) 26/147 (17.7%) 53/147 (36.1%) 12/147 (8.2%) 22/147 (15.0%) 19/147 (12.9%) 144/147 (98.0%) 14/147 (9.5%) 12/147 (8.2%) 58.8 (18.7) 37 (25.2%) 2/147 (1.4%)
7.45 [2.03, 41.2] † 1.56 [0.57, 4.74] † 1.55 [0.52, 5.22] † -0.92 [-1.17,-0.66]# 4.33 [1.93, 9.74] † --
57.4 (10.2)
54.5 (13.9)
-0.23 [-0.48, 0.02]#
24/111 (21.6%)
3/147 ( 2.0%)
34.8% [28.5%; 41.8%]‡
4.38 [2.09; 9.93] †
3.38 [1.85, 6.36] † 30.3% [24.4%, 37.1%]‡
2.18 [0.96, 5.32] † 1.25 [0.46, 3.61] † 3.74 [1.43, 11.6] † 2.91 [1.59, 5.34] †
37.3% [30.8%, 44.4%]‡
mild moderate severe aortic regurgitation, baseline none mild moderate severe mitral regurgitation, baseline none mild moderate severe tricuspid regurgitation, baseline none mild moderate severe
11/111 (9.9%) 8/111 (7.2%) 68/111 (61.3%)
10/147 (6.8%) 11/147 (7.5%) 123/147 (83.7%)
26/111 (23.4%) 29/111 (26.1%) 19/111 (17.1%) 37/111 (33.3%)
36/147 (24.5%) 57/147 (38.8%) 28/147 (19.0%) 26/147 (17.7%)
29/111 (26.1%) 73/111 (65.8%) 9/111 (8.1%) 0/111 (0.0%)
12/147 (8.2%) 109/147 (74.1%) 24/147 (16.3%) 2 /147 (1.4%)
38.2% [31.6%, 45.2%]‡
38/111 (34.2%) 66/111 (59.5%) 6/111 (5.4%) 1/111 (0.9%)
21/147 (14.3%) 101/147 (68.7%) 22/147 (15.0%) 3/147 (2.0%)
37.1% [30.7%, 44.2%]‡
57.2% [50%, 64%]‡
Values are depicted as mean (standard deviation) and standardized mean difference (#) for conGnuous variables, n (%) and probability of superiority (‡) for ordinal variables, and n (%) and odds ratio [95% confidence interval] (†) for binary variables. ACB, aorto-coronary bypass grafting; CAD, coronary artery disease; GFR, glomerular filtration rate; MI, myocardial infarction; NYHA, New York Heart Association; PCI, percutaneous coronary intervention; STS-PROM, Society of Thoracic Surgeons – Predicted Risk of Mortality; SMD, standardized mean difference; OR, odds ratio.
Table 2. Procedural outcomes at 30 days. redo-SAVR n = 111
VinV-TFAVI n = 147
Clinical efficacy
36/111 (32.4%)
78/147 (53.1%)
Early safety
72/111 (64.9%)
26/147 (17.7%)
Periprocedural myocardial infarction
2/111 (1.8%)
5/147 (3.4%)
Stroke, any
9/111 (8.1%)
8/147 (5.4%)
Life threatening bleeding
24/111 (21.6%)
9/147 (6.1%)
VARC Major access site complications
17/111 (15.3%)
17/147 (11.6%)
Acute kidney injury, stage II or III
19/111 (17.1%)
8/147 (5.4%)
Necessity of permanent pacemaker implantation Postoperative length of stay
14/111 (12.6%)
20/147 (13.6%)
11 [8, 15]
9 [6, 11]
Erythrocyte conc. packs
2 [1.5, 4.5]
0 [0, 0]
OR / probability of superiority 2.35 [1.41, 3.93] 0.12 [0.066, 0.21] 1.92 [0.37, 10.1] 0.65 [0.24, 1.75] 0.24 [0.10, 0.53] 0.72 [0.35, 1.49] 0.28 [0.12, 0.66] 1.09 [0.53, 2.27] 68.5% [61.6%, 74.5%] 81.4% [75.9%, 86.3%]
p value
< 0.001 < 0.001 0.43 0.39 < 0.001 0.38 0.002 0.82 < 0.001 < 0.001
Definitions according to VARC-2. Values are depicted n (%) and odds ratio [95% CI] for categorical variables and median [IQR] and probability of superiority for length of stay and erythrocyte concentrates. OR, odds ratio. bold: significant at 5% after correction for multiple testing.
A
B
80%
70%
70%
60%
NYHA IV
60%
50%
NYHA III
50%
NYHA II
40%
NYHA I
30%
40% 30% 20%
1.1%
20%
10%
10%
0% redo-SAVR re-SAVR baseline baseline
redo-SAVR re-SAVR 1 1year year
VinV-TFAVI VinV-TFAVI baseline baseline
VinV-TFAVI 1 VinV-TFAVI 1 year year
any improvement
1 NYHA class 2 NYHA classes 3 NYHA classes
3.2%
80%
22.2%
90%
46.0%
90%
71.4%
100%
21.3%
100%
44.9%
B
67.4%
A
0% redo-SAVR re-SAVR
VinV-TFAVI VinV-TFAVI
Highlights
-
VinV-TAVI is a safe and effective treatment for degenerated aortic bioprosthesis
-
data comparing this procedure to redo surgery (re-SAVR) are scarce
-
comparable 1-year-mortality between re-SAVR and VinV-TFAVI
-
incidence of bleeding and renal failure was higher with re-SAVR
-
postoperative transvalvular gradients were higher after VinV-TFAVI