Transcatheter aortic valve implantation (TAVI) has evolved significantly since its first-in-human application two decades ago.1 Initially reserved for patients with symptomatic severe aortic stenosis (AS) who were either inoperable or at high surgical risk, a series of trials demonstrated the non-inferiority of TAVI compared with surgical aortic valve replacement (SAVR) also in intermediate- and low-risk patients.2–8 Parallel to this, guideline indications for the treatment of severe AS as well as patient selection for TAVI in actual clinical practice have changed.9–11 While the durability of a TAVI prosthesis is likely to exceed the life expectancy of elderly TAVI patients, in younger patients with longer inherent life expectancy a potential repeat intervention must be considered already at the time of the first valve replacement when implanting a biological prosthesis, regardless of whether it is done via SAVR or TAVI. Additionally, patients of any age may benefit from TAVI due to the less invasive nature of the procedure and the subsequent faster recovery. This means a shift regarding future directions for TAVI as a therapy: although traditional selection criteria between TAVI and SAVR, such as age, surgical risk, anatomy and patient preference, may still be key aspects of the heart team discussion, considerations with respect to a more holistic lifetime management of AS are now becoming increasingly important.
In this review, we aim to elucidate the evolving indications for TAVI over time and to identify current key considerations in the patient selection process with regard to the most appropriate treatment strategy of SAVR or TAVI, given the context of the evolving concept of lifetime management for AS.
Surgical Risk Scores for Decision-making in AS Patients: An Historical Perspective
TAVI was initially conceived as an intervention aimed at treating patients with severe AS deemed at too high a risk for SAVR. The primary objective in this patient population was to achieve successful prosthetic valve implantation and to ensure stable discharge of the patient.1 Encouraged by the favourable outcomes observed in inoperable and high-risk surgical patients, a series of pivotal trials (namely the PARTNER trials for the balloon-expandable SAPIEN platform and the EVOLUT trials for the self-expandable CoreValve/Evolut platform) included patients with progressively lower surgical risk profiles. Finally, the PARTNER III and the Evolut Low Risk trials showed non-inferiority of TAVI compared with SAVR in a patient population at low surgical risk.2–8
Consequently, for many years, surgical risk scores, particularly the Society of Thoracic Surgeons (STS) risk score and the European System for Cardiac Operative Risk Evaluation II (EuroSCORE II), served as important tools for guiding treatment decisions and are still relevant for reimbursement considerations in many healthcare systems.12–14 These scores, which incorporate patient demographics, laboratory findings and comorbidities, estimate the perioperative mortality risk associated with isolated SAVR, categorising patients with an STS or EUROSCORE II >15% as inoperable, >8% as high risk, 4–8% as intermediate risk, and <4% as low risk. Both clinical practice and professional guidelines have consistently used these risk scores as the foundation for decision-making in the treatment of severe AS, with each guideline update integrating emerging data from randomised trials conducted in cohorts with progressively lower surgical risk (Table 1).9,10,15–18 Although these risk scores provided valuable guidance for patient selection in the early years of TAVI, particularly for high- and intermediate-risk patients, their relevance has diminished as TAVI outcomes have improved, now including rising numbers of younger, low-risk patients with fewer comorbidities.19,20
In the recent investigator-initiated, industry-independent DEDICATE trial more than 1,400 low-risk AS patients (mean age 74.4 years; STS score 1.8%) were randomly assigned to receive either TAVI or SAVR, each with the prosthesis choice being left to the discretion of the operator. This trial, reflecting current routine practice, showed at least non-inferiority of TAVI compared with SAVR after 1 year (death from any cause: 2.6% versus 6.2%; stroke: 2.9% versus 4.7%).21 The observed equivalence between TAVI and SAVR raises important questions regarding selection between SAVR and TAVI. Also, younger patients may benefit from the advantages of TAVI, such as minimal invasive access and rapid recovery, and traditional risk scores failed to reliably predict patient outcomes.
Studies such as the AMTRAC registry, stratifying over 8,000 patients undergoing TAVI by age (above or below 70 years), as well as by their estimated surgical risk as per STS score, showed that mortality outcomes were similar regardless of the STS stratum, indicating that such risk scores in younger patients fail to accurately identify patients who should or should not undergo TAVI.22 Hence, other criteria, such as patient age and, especially, life expectancy, which indicate whether a patient may need a potential subsequent redo procedure, as well as anatomical criteria, which may indicate the feasibility of such redo procedures, become more relevant in the decision-making of TAVI versus SAVR over time and will be discussed in the following sections (Figure 1).
Key Considerations for Patient Selection
Patient Age
Recent longer-term data from low-risk trials, coupled with advances in TAVI techniques and successful re-valving procedures after failed bioprosthetic valves, enabling TAVI-in-TAVI procedures as well as TAVI-in-SAVR, mean that patient evaluation has shifted toward a more comprehensive approach with a particular focus on patient age.21–24 Although TAVI has demonstrated at least comparable outcomes to surgery across the risk spectrum, most trial participants were elderly and unlikely to outlive the lifetime of their bioprosthetic valves. Notably, even in low-risk trials patients were still relatively old (e.g. the PARTNER 3 trial had a mean STS score of 1.9% and a mean patient age of 73.4 years), leaving the question of TAVI performance in patients who are likely to require more than one bioprosthetic valve replacement and the outcome of redo procedures after TAVI, unanswered.23 However, in younger patients valve durability and the likelihood of potential redo procedures in the case of bioprosthetic valve failure need to be considered already at the time of the first aortic valve replacement (AVR; Figure 2).
Current guidelines reflect this uncertainty by implementing age cut-offs for TAVI or SAVR as the preferred treatment option, although varying interpretations of the available evidence led to different recommendations. For instance, the 2021 European Society of Cardiology (ESC) valvular heart disease guidelines currently recommend TAVI for patients ≥75 years irrespective of surgical risk. The American College of Cardiology/American Heart Association (ACC/AHA) guidelines propose a more differentiated approach, recommending a bioprosthetic valve in patients >65 years and considering a patient’s expected individual life expectancy in the choice between TAVI and SAVR. Thereby, SAVR is recommended in patients with an assumed life expectancy of >20 years.10,18
This perspective introduces the concept of lifetime management, which anticipates more than one bioprosthetic valve replacement over a patient’s lifetime, whether through SAVR or TAVI.24–27 The durability of SAVR bioprostheses is considered to be around 10–15 years.28,29 However, comprehensive long-term data after SAVR are scarce. Frequently, freedom from re-operation has been used as an endpoint to evaluate durability, which inherently introduces relevant bias.29 Elderly patients may be deemed too high a risk for re-operation several years after SAVR, hence this endpoint may fail to capture cases of clinically evident structural valve deterioration, when re-intervention was not pursued due to the assumed surgical risk. This is mirrored by the substantial increase of TAVI-in-SAVR valve-in-valve procedures in recent years with a median time from SAVR to valve-in-valve TAVI of only 9 years.30 Additionally, the assessment of bioprosthetic durability across different age groups based on the aforementioned endpoint is susceptible to bias, given that younger patients are more likely to undergo re-operation, falsely suggesting a higher and earlier rate of bioprosthetic valve failure in this group.29
Importantly, to date, there is no evidence suggesting a shorter durability of TAVI prostheses compared with surgical bioprostheses. Ten-year data from the NOTION trial, which randomised 280 patients to early-generation self-expandable TAVI or any bioprosthetic SAVR, indicated the superior performance of TAVI in terms of structural valve deterioration (SVD; moderate defined as mean gradient ≥20 mmHg, increase ≥10 mmHg on 3-month echocardiogram; severe defined as gradient >30 mmHg, increase ≥20 mmHg) and equivalent outcomes in terms of bioprosthetic valve failure (defined as: valve-related death, severe SVD, re-intervention following diagnosis of SVD).31 The different technical approach of TAVI, especially with bioprosthetic leaflets being crimped during the procedure, does not seem to translate into worse outcomes with respect to valve durability. Although these findings are encouraging, longer-term follow-up in low-risk and younger patients is needed to confirm these preliminary results – such data will be available with longer term follow-up of the pivotal trials.32
Last, in younger patients, apart from bioprosthesis considerations, mechanical valves also need to be considered. While such choice may, per definition, end any discussion regarding lifetime management, in patients who are likely to outlive at least two or three bioprosthetic valves it might be preferable to choose a solution that eliminates the ever-increasing risk of re-interventions with ever-increasing technical and anatomical challenges. The potential advantage of a single operation providing a potentially lifelong solution has to be counterbalanced against the inherent risks of necessary continuous oral anticoagulation using vitamin K antagonist (VKA). This renders mechanical valves contra-indicated in patients who do not tolerate VKA, cannot be reliably monitored under VKA, or who do not want to undergo VKA treatment. US guidelines incorporate these considerations, stating that the choice of a mechanical over a bioprosthetic valve in patients <50 years is a reasonable approach, and suggesting informed shared decision-making regarding the valve choice in patients 50–65 years of age (class 2a recommendation, each).18
In summary, TAVI should be considered the first choice in elderly patients with limited life expectancy. In younger patients it must be determined, under consideration of further criteria, whether TAVI or SAVR or even implantation of a mechanical prosthesis is the most suitable approach based on individual patient characteristics.
Anatomical Factors
A comprehensive assessment of a patient’s valvular and vascular anatomy is critical to evaluate the feasibility and potential challenges of TAVI (Figure 3). Transfemoral access is generally preferred because it maximises the advantages of TAVI over SAVR, both technically and in terms of clinical outcomes.5,33 Downsizing of TAVI delivery systems and the use of adjunctive techniques such as intravascular lithotripsy has broadened the population amenable to transfemoral access.34
The anatomy of the aortic valve complex is another critical consideration. Although major trials have primarily enrolled elderly patients with tricuspid anatomy, a significant proportion of younger patients have bicuspid valves.35 Registry data have suggested similar short-term outcomes for TAVI in patients with bicuspid or tricuspid anatomies, however, recent data from the randomised NOTION-2 trial, enrolling low-risk patients ≤75 years of age, raised concerns because TAVI and SAVR had comparable outcomes only in tricuspid patients, while for patients with bicuspid anatomy the incidence of adverse outcomes was notably higher in the TAVI patient group (14.3% versus 3.9% in SAVR).36–38 High calcium burden and the semilunar orifice of bicuspid valves may lead to eccentricity and underexpansion of TAVI prostheses, which in turn lead to an association with adverse outcomes.39 Hence, in bicuspid anatomies TAVI should be reserved for selected patients considered unsuitable for SAVR.10 This especially applies to those bicuspid anatomies with a prominent calcified raphe and excessive leaflet calcification, which has been shown to be associated with an increased risk of procedural complications.39
Furthermore, the amount and distribution of calcium in the aortic valve area, the left ventricular outflow tract (LVOT) and aortic root are factors that may affect the outcomes of AVR. Severe LVOT calcification can increase the risk of annular rupture or paravalvular regurgitation during TAVI.40 Conversely, extensive calcification at the level of the aorta, that is, porcelain aorta anatomies, may advocate against surgery because the risk of clamping is substantially increased.41 According to recent data the impact of calcification at these levels may also affect the risk of conduction disturbances as well as paravalvular leakage following TAVI.42,43
Last, the dimension of the aortic root, the size of the aortic annulus and the sinuses as well as the distance to the coronaries are important parameters to consider in the CT evaluation of a severe AS patient. The size range of available TAVI prostheses is limited compared with SAVR prostheses. Consequently, patients with very large anatomies or with concomitant aortic pathologies may benefit from a SAVR-first strategy. In smaller anatomies (defined in recent trials as ≤430 mm²), TAVI may be beneficial given the superior haemodynamics of TAVI prostheses compared with SAVR prostheses.
The VIVA trial compared SAVR and TAVI in patients with small aortic annulus: the primary haemodynamic endpoint at 60 days was similar in the two groups with, however, numerically higher rates of severe prosthesis–patient mismatch after SAVR, a finding that has been previously associated with adverse mid- and long-term outcome after SAVR but not after TAVI.44–48 Of note, SAVR carries the possibility of surgical root enlargement (used in the VIVA trial in only approximately 1 out of 10 patients), enabling implantation of a larger bioprosthesis also in smaller valves, simplifying later redo procedures and improving short- as well as longer-term haemodynamics.44,45 The randomised SMART trial, comparing self-expanding (supra-annular) with balloon-expanding (intra-annular) TAVI platforms in small anatomies, suggested superior haemodynamics of self-expanding TAVI in this patient population, although its impact on hard clinical endpoints has yet to be shown.49 Given the lack of data associating individual prosthesis haemodynamics and outcomes in TAVI patients, from a lifetime management perspective the superior haemodynamics of self-expanding prostheses need to be counterweighed against the higher risk of coronary access impairment or sinus sequestration in future procedures.50
In the future, refined imaging methods as well as more precise analyses of acquired imaging data, enhanced by 3D planning, artificial intelligence (AI)-assisted anatomical measurements and digital lifetime management simulations, may further improve patient selection and identification of the most ideal individual treatment pathway.51–53
Further Key Aspects to Consider: Patient History and Comorbidities
Besides anatomical parameters, a holistic assessment regarding patient selection for different AS treatment modalities needs to consider patient history and comorbidities. Comorbidities such as heart failure with reduced ejection fraction, peripheral artery disease and severe renal impairment are well-known predictors of increased surgical risk and hence are incorporated into classic risk scores. Apart from the perioperative risk, these factors are inherent predictors of reduced long-term mortality irrespective of treatment modality.
In randomised TAVI trials the prevalence of concomitant coronary artery disease (CAD) ranged from almost 70% in high-risk to around 30% in low-risk cohorts.2,23,54 For diagnosing CAD in patients scheduled for TAVI, invasive coronary angiography, currently, is the mainstay; however, in younger patients with a low cardiovascular risk profile coronary CT may represent a reasonable resource-saving alternative. Coronary access may become challenging after TAVI, especially after implantation of tall-frame prostheses or after TAVI-in-TAVI and TAVI-in-SAVR procedures, in which the pinned-up leaflets of the index prosthesis create a static neo-skirt.55,56 The complexity of concomitant CAD has to be taken into account, given that operable low-risk patients with complex CAD (unprotected left main disease with a SYNTAX score >32 or multivessel disease with a SYNTAX score >22) may benefit from a SAVR plus coronary artery bypass graft strategy.57 In patients scheduled for TAVI the timing and modality of treatment for concomitant CAD is a matter of ongoing debate.
In the randomised NOTION-3 trial, percutaneous coronary intervention (PCI) was associated with a lower risk of a composite of death from any cause, MI, or urgent revascularisation after 2 years compared with conservative treatment.58 However, in the REVASC-TAVI registry complete revascularisation as well as PCI before TAVI was not associated with superior outcomes.59,60 Hence, conservative treatment of concomitant CAD may be reasonable in older TAVI patients depending on symptoms and coronary lesion characteristics. Of note, the presence of concomitant CAD may not only influence the choice between SAVR or TAVI, it will also influence the choice of a specific TAVI platform, given that either short-frame balloon-expandable or tall-frame self-expandable devices have different advantages and disadvantages regarding future coronary access.61
Patients with additional concomitant valvular pathologies also require special consideration. Moderate or greater mitral regurgitation (MR) is present in up to 30% of AS patients, however, significant MR regression due to ventricular remodelling is seen in approximately 50% of patients after TAVI.62,63 Hence, in cases of severe primary MR and severe AS, a combined surgical approach may be preferred if the surgical risk is reasonable.10,18 However, given that a combined surgical replacement of two valves may come at a substantially increased operative risk, for patients with intermediate or high surgical risk and relevant MR, a TAVI-first strategy with subsequent monitoring may be appropriate, allowing for stepwise transcatheter treatment if still necessary.64
Redo Procedures: Considerations for Aortic Valve Replacement in Patients with Failed Bioprosthetic Valves
In clinical practice, patient selection for AS treatment increasingly involves not only patients with native AS but also a growing number of patients with degenerated bioprosthetic valves (defined as per the Valve Academic Research Consortium, with a pragmatic aggregation of these criteria implemented in the NOTION trial as detailed above), both surgically implanted or delivered via transcatheter approaches.65 Given the elevated surgical risk associated with redo procedures, both SAVR and surgical explantation following TAVI, valve-in-valve (ViV) and TAVI-in-TAVI have become the preferred strategies in these populations, given the substantially lower in-hospital mortality compared with repeat SAVR.66 However, this necessitates a foresighted choice of the initial TAVI device, to allow for a successful redo procedure, and also an informed choice of the second valve as well as its implant position, because this may have an impact on aspects such as the resulting haemodynamics. Furthermore, careful consideration of potential complications, particularly coronary access after TAVI, is essential in planning and executing these procedures. Techniques such as coronary protection through chimney stenting or the BASILICA procedure (bioprosthetic or native aortic scallop intentional laceration to prevent iatrogenic coronary artery obstruction) need to be included in the armamentarium of interventional techniques.67 From a long-term management perspective, the BASILICA technique may be favoured because it not only reduces the risk of acute coronary obstruction but also facilitates coronary access for future PCI. With the development of specialised devices that are designed to enable leaflet laceration with a dedicated easy-to-use mechanism, such as the novel ShortCut device (Pi-Cardia), this technique could see a broader application in the future.68 While early data on BASILICA and dedicated devices facilitating a streamlined leaflet laceration look promising, long-term data on such approaches and its implementation in lifetime management strategies are still awaited.
Additionally, the need for redo procedures underscores the importance of optimising the initial intervention, whether through surgical approaches (such as root enlargement in cases of small annuli to avoid the necessity of undersized prostheses during future ViV procedures) or through meticulous planning of TAVI procedures, including commissural alignment to facilitate future TAVI-in-TAVI interventions.
Apart from the clinical aspects, it is reasonable to also consider the socioeconomic perspective regarding the use of TAVI and its role in lifetime management of AS. Currently, it remains a rather costly procedure and is not fully available across all healthcare systems worldwide. New devices stemming from emerging markets, which may have the potential to reduce cost, have been introduced.69 Using established devices as well as such novel alternatives, cost-effectiveness studies will be needed to support the role of different treatment strategies across the globe.
Finally, the significance of shared decision-making with the patient has been granted a class 1 recommendation in the 2021 ESC valvular heart guidelines. Hence, after consideration of the aforementioned aspects by the heart team, the patient’s informed choice needs to be taken into account and plays a critical role in the decision process between TAVI and SAVR, especially in populations amenable to both strategies.
Conclusion
Advances in technology and growing experience have made TAVI the preferred treatment modality for most patients with severe AS. However, this shift, especially towards younger patients, emphasises the need for a comprehensive assessment that considers several aspects including age, life expectancy, comorbidities, aortic anatomy and patient preference to select the most appropriate treatment strategy for each individual patient, that provides a durable solution over the patient’s lifetime, including consideration of potential redo procedures. For patients in whom a transcatheter approach, including future redo procedures, is feasible, TAVI may be preferred also in younger populations. Conversely, in patients presenting with anatomical challenges such as bicuspid aortic valves or small aortic roots, or those with significant comorbidities, such as severe multivessel CAD or relevant multi-valvular disease, SAVR may be the preferred initial approach, enabling a safe valve-in-valve TAVI procedure in the future. Further research is needed to evaluate safety, effectiveness and long-term outcomes of different lifetime management pathways, and also to compare different strategies against each other in a head-to-head fashion. These studies need to be accompanied by additional research on the technical particularities of redo procedures as well as on the socio-economic impact and relevance of these treatment pathways.
Regardless of TAVI or SAVR, this framework underscores the importance of optimising the initial procedure through meticulous implantation techniques. It also highlights the need to refine redo procedures through the development and implementation of new techniques and devices that address the challenges inherent in valve-in-valve and TAVI-in-TAVI scenarios.