Transcatheter valve-in-valve implantation in failed aortic bioprosthetic valves: a word of caution in times of euphoria

Listen to the podcast associated with this article, which can also befound at ESC CardioTalkhttps://www.escardio.org/The-ESC/Whatwe-do/news/ESC-Cardio-Talk This editorial refers to ‘Long-term outcomes after transcatheter aortic valve implantation in failed bioprosthetic valves’†, by S. Bleiziffer et al., on page 2731. Transcatheter aortic valve-in-valve implantation (ViV TAVI) has become an attractive alternative to redo surgery in patients with failed aortic bioprosthetic valves.1 Data are mainly available for elderly patients at increased surgical risk. Large registries demonstrated that the procedure can be safely performed in this patient group. Overall mortality at 30 days was 4.6% of 1168 patients in the Valve-in-Valve International Data Registry (VIVID),2 2.1% of 1150 patients in the Society of Thoracic Surgeons (STS)/American College of Cardiology (ACC) Transcatheter Valve Therapies Registry (TVT),3 and as low as 0.7% of 365 patients in the continued access Partner 2 ViV Registry.4 Randomized controlled trials (RCTs) are not available and comparisons between redo surgery and ViV TAVI are limited by the bias of the older patients with more comorbidities in the ViV TAVI populations. The first propensity score analysis reported similar 30-day and 1-year mortality, stroke, renal failure, and pacemaker implantation rates between treatments. While gradients were lower with surgery, hospital length of stay was shorter with ViV TAVI.5 More recently, Derrick et al.6 reported in a propensity-matched cohort including 131 pairs utilizing Canadian clinical and administrative databases lower early mortality, pacemaker rate, and blood transfusions for ViV TAVI. Length of stay was also shorter and 5-year survival was significantly higher. A larger study utilizing a US national administrative database7 has very recently confirmed in 2181 propensity score-matched pairs of surgical high-risk patients a lower 30-day mortality and morbidity, less major bleeding, shorter length of stay, and a higher rate of home discharges for ViV TAVI compared with redo surgery. Although data for low-risk patients and younger age groups are still missing, recent RCTs comparing TAVI and surgical valve replacement (SAVR) in low-risk patients with native aortic stenosis8,9 led to the hypothesis that ViV TAVI may at least not be inferior or even be superior to redo surgery in this population. The euphoria can be felt in the interventional community expecting that ViV TAVI is becoming the treatment of choice, and the question of whether the ‘panacea is approaching’ has been raised.10 There are, however, also major concerns such as the markedly higher rate of suboptimal haemodynamics after ViV TAVI, with >30% of ViV procedures in STS/ACC TVT and VIVID registries resulting in severe residual gradients,11,12 and the risk of coronary obstruction and limited later access to coronary arteries.1 In addition, data on long-term outcome beyond 3 years are still scarce, structural valve deterioration (SVD) for ViV TAVI has been studied insufficiently, and higher risk for leaflet thrombosis has been suggested.13 In this regard, Bleiziffer and colleagues14 provide an important new piece of information in this issue of the European Heart Journal presenting the updated follow-up of the VIVID Registry. In a retrospective cross-sectional evaluation, a total of 1006 patients with aortic ViV procedures performed >5 years ago were analysed. With a mean age of 78 years and a median STS-PROM score of 7.3%, the study population comprised elderly patients at high surgical risk. The study is unique in providing, for a very large patient group, the currently longest follow-up. The estimated survival at 8 years of only 38% is not surprising considering the high-risk population with advanced age and frequently severe comorbidities. Importantly, survival at 8 years was significantly lower in patients with small failed valves defined by a true internal diameter15 ≤20 mm (33% vs. 41%). Smaller failed prosthetic valves, age, and non-transfemoral access were independent predictors of mortality. Reintervention was performed in only 40 patients (4%). The most common reason (60%) was severe stenosis, and 58% of reinterventions were surgical. The estimated reintervention rate at 8 years was significantly higher for balloon-expandable Edwards compared with self-expandable Medtronic valves (6% vs. 2%). Independent correlates for all-cause reintervention were pre-existing severe prosthesis–patient mismatch (PPM), device malposition, age, and balloon-expandable valve. In the group of patients with small surgical valves, effective orifice areas (EOAs) were larger (1.51 ± 0.49 cm2 vs. 1.18 ± 0.39 cm2), and mean gradients were lower (16.5 ± 8.7 mmHg vs. 21.1 ± 10.4 mmHg) when a self-expandable compared with a balloon-expandable valve was implanted. The strongest finding of the study is the confirmation of excess mortality in patients with small failed valves at long-term follow-up. In a previous analysis of the VIVID registry, having a bioprosthesis labelled ≤21 mm was associated with increased 1-year mortality.12 In the current analysis, the investigators correctly no longer used label size but true internal diameter as this reflects much better the available space for a new valve and the eventual performance that can be achieved with a ViV procedure.15 The relationship between stent internal and true internal diameter differs markedly depending on the valve design. Valves labelled 21 mm have in general a true internal diameter <20 mm, but ranging from 16.5 mm for porcine valves such as Hancock II and Medtronic Mosaic to 19 mm for pericardial bovine valves such Edwards Perimount and Magna. For the latter, the internal diameter reaches 21 mm for labels 23 mm while it is only 18.5 mm for the porcine valves of this size.15 This makes a critical difference for the EOA and haemodynamic performance that can be achieved with a ViV intervention. Surprisingly, Bleiziffer et al.14 could not find a statistically significant difference for the survival of patients with mean gradient ≥20 mmHg vs. <20 mmHg as EOA and mean gradient are inter-related although this relationship is flow dependent and therefore, among other factors, patient size dependent. The concept of relating the EOA to the body surface area in order to define PPM (severe <0.65 cm2/m2, moderate 0.65–0.85 cm2/m2) has become widely accepted. Despite inconsistencies among studies, there is, in particular, strong evidence that severe PPM is associated with an increased all-cause as well as cardiac-related mortality after SAVR.16 Although the authors claim in the Discussion that severe PPM after ViV TAVI does not appear to affect outcome, referring to a previous VIVID report of 910 patients including 25% with severe PPM in whom 1-year outcome was not affected,17 this is in disagreement with other larger studies evaluating PPM in patients undergoing TAVI. An analysis from the STS/ACC TVT Registry including 62 125 TAVI patients reports the occurrence of severe PPM in 12% which was associated with higher mortality and heart failure rehospitalization at 1 year.11 Similarly, a meta-analysis of 71 106 patients reported an increased mortality in the presence of severe PPM after TAVI.18 There is no reason to believe that PPM has less effect after ViV TAVI than after TAVI in native aortic stenosis. Taking all information together, the achievement of an adequate EOA and good haemodynamics appears critical for patient outcome. Whether this can be achieved with a ViV procedure does not depend only on the true internal diameter of the failed bioprosthesis but also on the patient’s size and the required cardiac output, as well as the pathology of the valve. It appears more difficult in the case of stenosis,19 presumably due to the greater amount of material within the stent frame than in predominant regurgitation. Therefore, careful individual evaluation is mandatory (Figure 1). Stented porcine valves with label size 23 mm or less and pericardial (bovine) valves with size 21 mm or less will most of the time require special interventional techniques to achieve reasonable haemodynamics. A higher deployment position of self-expanding valves such as Medtronic Corevalve/Evolute or Symetis Accurate with a supra-annular leaflet position may be one option,20 but limitations for future reinterventions such as limited access to coronaries must be kept in mind. The current study accordingly reports better haemodynamics in small failed bioprostheses when aiming for a supra-annular position with Medtronic self-expandable vs. Edwards balloon-expandable valves. However, lower gradients with higher positioning also need to be balanced against higher rates of regurgitation and potential for valve embolization.20 Bioprosthetic valve fracture using high-pressure non-compliant balloons has emerged as another option to increase EOA and improve haemodynamics.21 However, reported numbers of interventions are still relatively small and some prostheses such as Medtronic Hancock II and Abbott Trifecta are definitely not fracturable. If achievement of reasonable haemodynamics is unlikely in a patient at low surgical risk, ViV TAVI can obviously not be justified. Valve-in-valve procedures in small failed bioprosthetic aortic valves (i.e. true internal diameter ≤20 mm). (A–C) Examples of suboptimal haemodynamic results. (D and E) Strategies to improve haemodynamics. Failed bioprosthetic valve type and size, its true internal diameter (TID), the transcatheter heart valve used, and the residual Doppler mean pressure gradient (mean ΔP) after the procedure are indicated. Note that—depending on the valve type—valves labelled 21 mm as well 23 mm have a TID ≤20 mm and must be considered ‘small valves’ (see text). Valve-in-valve procedures in small failed bioprosthetic aortic valves (i.e. true internal diameter ≤20 mm). (A–C) Examples of suboptimal haemodynamic results. (D and E) Strategies to improve haemodynamics. Failed bioprosthetic valve type and size, its true internal diameter (TID), the transcatheter heart valve used, and the residual Doppler mean pressure gradient (mean ΔP) after the procedure are indicated. Note that—depending on the valve type—valves labelled 21 mm as well 23 mm have a TID ≤20 mm and must be considered ‘small valves’ (see text). SVD and valve thrombosis after ViV also remain of concern.20 Unfortunately, the current study cannot provide information on this. The fact that only 4% of patients underwent reintervention appears promising. However, one has to keep in mind the high mortality of this population and that in the elderly with severe comorbidities not all patients with clinically relevant SVD may have undergone reintervention. The higher rate of reintervention in balloon-expandable valves may be due to the more frequently unfavourable haemodynamic results achieved in small failed bioprostheses with this valve. In the Partner 2 registry,22 where—as recommended—19 mm failing surgical valves were not treated with the balloon-expandable valve, the reintervention rate at 3 years was indeed only 1.9%. In any case, more data on valve durability are required before expanding the indications for ViV procedures to younger low-risk patients. In conclusion, ViV TAVI has become an attractive alternative to redo SAVR in patients with failed bioprosthesis at increased surgical risk. It can be performed with a low procedural risk, but frequently achieves less favourable haemodynamics than redo surgery. This may, particularly in the case of small valve sizes, result in a worse long-term survival. Therefore, based on currently available data, the following should be considered Surgeons should aim to implant the largest possible valve size and a valve type with good pre-conditions for later ViV procedures (fracturable stents; stents with an expansion zone are being developed). When performing ViV TAVI in small failed bioprostheses, all efforts must be made to achieve good haemodynamics (choice of valve type and size, implantation techniques including valve fracture). When it is unlikely to achieve reasonable haemodynamics with ViV TAVR, the potentially resulting negative impact on outcome must be carefully weighed against the risk of surgery. Expanding the use of ViV TAVI to younger patients with low surgical risk requires supporting data on long-term durability and direct comparison with the outcome of SAVR in this patient group. When aiming at strategies of earlier use of bioprosthetic valves for SAVR in view of the option to reintervene with TAVI when the valve fails, one has to keep in mind that the number of possible reinterventions without creating severe PPM or major limitations to coronary access will be limited and the patient may lose the chance for ViV TAVI when it is eventually particular needed because of advanced age and comorbidities. Conflict of interest: H.B. has received speaker fees and congress travel support from Actelion and Edwards Lifesciences. The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.