Statistics from Altmetric.com
Surgical aortic valve replacement (SAVR) has historically been the treatment of choice for patients with severe symptomatic aortic stenosis (AS). More recently, transcatheter aortic valve implantation (TAVI) has become a well-established alternative option for patients considered at high and intermediate preoperative surgical risk. TAVI provides effective relief of valve obstruction with excellent immediate and short-term outcomes. However, long-term (ie >10 years) durability of TAVI, relative to the known durability of surgical bioprosthetic valves,1 also needs to be considered. Indeed, there is, up to now, very few data addressing durability of TAVI beyond 5 years,2 which is particularly important in younger patients and those with few comorbidities who have many remaining expected years of life. Hence, long-term structural valve degeneration/deterioration (SVD) has become an important issue for patients and physicians making an informed decision between the choice of TAVI and SAVR bioprostheses.
SAVR structural valve deterioration
Two different mechanisms have been described for bioprosthetic valve failure. Structural valve deterioration/degeneration refers to acquired intrinsic changes of the prosthesis and includes leaflet calcification, leaflet tear, stent fracture or creep, and suture line disruption.3 4 Non-structural deterioration/degeneration includes prosthesis-patient mismatch (PPM), valve thrombosis, pannus ingrowth, paravalvular leakage (PVL) and endocarditis (figure 1). SVD due to progressive deterioration of valve leaflet tissue is the most common cause of surgical bioprostheses failure. A recent systematic review1 of 93 studies including 53 884 (age 53–92 years) patients with AS that focused on prognosis after SAVR with bioprosthetic valves showed that the median survival among those aged ≤65, 65–75, 75–85 and ≥85 years was 16, 12, 7 and 6 years, respectively. Importantly, the freedom from reoperation for valve dysfunction at 10, 15 and 20 years was 94%, 81.7% and 52%, respectively.1
Engineering differences between TAVI and SAVR
TAVI leaflets are made of either glutaraldehyde-treated bovine pericardium or porcine pericardium; however, TAVI leaflets are thinner (~0.25 mm vs ~0.4 mm in SAVR) to allow transcatheter-device delivery.5 Furthermore, TAVI leaflets experience higher stresses and strains than SAVR leaflets, particularly in the presence of native aortic annular calcification or an oval-shaped annulus, resulting in non-circular, asymmetric stent-frame deployment.5 6 Therefore, one may expect that TAVI devices will have reduced durability as compared with SAVR. In fact, data from a computational, tissue-fatigue model suggests that, even when a TAVI is properly deployed, durability is predicted to be about 7.8 years less than SAVR.5 The durability of TAVI might be even shorter if implanted in elliptical annulus or if under-expanded (due to TAVI oversizing).5
Incidence of transcatheter bioprosthetic SVD
A systematic review2 that focused on prognosis after TAVI, including 13 observational studies with a total of 8914 patients (mean age 79–83 years), showed that the pooled incidence rate of SVD was 28.08 per 10 000 patient years, and, in those that did experience SVD, 12% underwent valve reintervention. These data are consistent with the publication of long-term data from trials not captured in the review.7–9
Inconsistent definitions of SVD among TAVI studies
A limited number of studies have reported on SVD after TAVI, with inconsistencies among the definitions of SVD and thus the reported incidence rates. The incidence of SVD is usually based on increased mean transprosthetic gradient and less often to the need for valve reintervention for TAVI failure.2 The thresholds in the Valve Academic Research Consortium-2 definitions10 have been inconsistently used thresholds for SVD definition.7 8 11 12 Moreover, some patients exhibiting severe SVD may have died and thus not undergone redo-TAVI or simply been denied for a second procedure because of multiple-comorbid conditions and/or limited life expectancy in the very elderly. An analysis from Placement of Aortic Transcatheter Valve (PARTNER) Trials8 looked at 207 patients with TAVI who had at least two consecutive echocardiograms before they died,; among these patients, eight had an increase of >10 mm Hg on consecutive echocardiograms before death (five in year 1, one in year 2 and two in year 3). Also, six patients had a Doppler velocity index decrease to <0.25 before death (five in year 1 and one in year 3). Finally, three TAVI patients developed new severe aortic regurgitation on the echocardiograms preceding death (two in year 1 and one in year 2).
Notably, patients may show an increase in transprosthetic gradients during echocardiographic follow-up, but not necessarily experience symptoms or have severe prosthetic valve stenosis.4 7 11 In such cases, the scenario for SVD is characterised as ‘isolated haemodynamic dysfunction’ of the prosthesis.4 Patients also may present with early anatomical changes in leaflet thickness, calcification, mobility and completeness of leaflet closure without significant haemodynamic dysfunction; these patients are characterised as having ‘morphological’ SVD.4 Finally, the term ‘bioprosthesis valve failure’ includes severe SVD and its clinical consequences (symptomatic patients); hence SVD now is the preferred main outcome of interest in studies assessing the long-term performance of TAVI and SAVR.4
Predictors of SVD
Numerous observational surgical series have identified risk factors for accelerated bioprosthetic SVD including younger age at valve implantation, impaired phosphate and calcium metabolism, renal insufficiency and metabolic risk factors for atherosclerotic and native valve disease. Likely, these same clinical factors will affect long-term TAVI SVD but there is concern that other factors may also be important. Unlike surgical valves, first generation TAVI valves did not have anticalcification treatment, although subsequent generations now incorporate an anticalcification process that removes both glutaraldehyde and phospholipids molecules at calcium-binding sites. Potential structural damage related to crimping, expansion and compression of TAVI leaflets13 may impact valve durability but this issue has received little attention to date. Moreover, paravalvular leakage after TAVI, mainly related to the non-circularity of the aortic annulus in combination with asymmetric calcium distribution, may result in higher turbulent flow as compared with SAVR prosthesis. In turn, this disturbed flow may predispose to accelerated SVD.
Predictors of early SVD after TAVI identified to date include lack of anticoagulation therapy, a smaller TAVI size (23 mm), a valve-in-valve procedure and a larger body mass index.12 The likelihood of PPM is higher when implanting smaller-size TAVI devices, either as a first implant in small aortic annuli or as a valve-in-valve procedure, although PPM was not a predictor of SVD in the multivariate analysis in this study.12 In line with these findings, another study14 did not find an association between severe PPM and SVD using the first generation of balloon-expandable TAVI. Nonetheless, PPM is a well-recognised predictor for SVD after SAVR15–17 and also was associated with worse outcomes in high-risk patients undergoing TAVI or SAVR in the PARTNER trial.18 The potential risk of PPM is an important consideration as we extend this technology to intermediate or even low-risk patients with longer life expectancies. In addition, SVD after SAVR is more likely to occur in younger patients; thus, we need to consider that this may also be true after TAVI.
How does valve thrombosis relate to SVD?
TAVI thrombosis presents either with increased transprosthetic gradients or as an imaging findings without significant valvular haemodynamic dysfunction. Current definitions of SVD exclude subclinical leaflet thickening and reduced leaflet motion due to valve thrombosis because this process is extrinsic to the structure of the prosthesis itself. However, there is a theoretical concern that valve thrombosis might increase the incidence or progression of SVD. Alternatively, the presence of valve thrombosis might result in overestimation of the prevalence of SVD.
In patients with surgical bioprosthetic aortic valves, the presence of atrial fibrillation, left ventricular dysfunction, previous thromboembolism and hypercoagulable states have been associated with an increased risk of valve thrombosis. However, different mechanisms for TAVI valve thrombosis have been proposed. During the TAVI procedure, the native valve leaflets are compressed between the prosthetic valve and aortic wall. This environment is associated with an extended zone of flow stagnation zone at the base of the sinus that may promote thrombus formation resulting in valve leaflet thrombosis and/or embolic events.19 Whether thrombus formation might lead to further progressive leaflet structural changes and eventually SVD is yet unknown.
The frequency and severity of leaflet thrombosis appear to be lower after SAVR compared with TAVI.20 Indeed, subclinical leaflet thrombosis was associated with reduced leaflet motion and higher transvalvular gradients,20 and the absence of anticoagulation therapy was associated with persistence or progression in reduced leaflet motion.20–22 Thus, leaflet thrombosis might lead to overestimation of the incidence of SVD.
The prosthetic valve durability to life expectancy ratio
The rapid evolution in TAVI technology for the treatment of severe AS and the lack of a robust evidence base regarding long-term TAVI durability leave patients and physicians with some difficult decisions. Hence, we propose that the key element in the choice of aortic valve prosthesis for the treatment of severe AS is the ratio of the expected valve durability to the expected remaining years of life in each individual patient (figure 2). Of course, there are many other considerations including procedural risk, comorbid conditions, avoidance of PPM or anticoagulation, functional status and patient preferences and values. Older patients often place high value on avoiding open-heart surgery and are willing to consider a repeat transcatheter procedure, if feasible and needed, in the future. However, younger patients may place a particularly high value on avoiding valve reintervention23 and thus, prefer a valve with established long-term durability.
Guidelines continue to provide a useful framework for choosing which type of aortic valve prosthesis is best in an individual patient. In younger patients (<50 years), a mechanical valve should be considered given the known durability of these valves unless there are other considerations related to anticoagulation or lifestyle in that patient.24–26 Once the decision to use a bioprosthetic valve has been made, the choice between SAVR and TAVI is guided by estimated surgical risk and patient age. When surgical risk is high or prohibitive for SAVR, TAVI durability is of limited concern and this approach is appropriate, unless intervention is futile when palliative care is recommended.
However, relative durability of SAVR versus TAVI becomes the primary consideration in patients at intermediate or low risk for SAVR. The BMJ Rapid Recs 27 provide age-stratified evidence tables and decision aids to assist in selection of a TAVI versus SAVR for treatment of severe AS in intermediate-risk patients. We have limited evidence for TAVI durability beyond 5 years2; thus, the expected remaining years of life and patient expectations become critically important in decision-making. The problem is illustrated in a graph of expected remaining years of life (shown as lines for men and women) with the points where valve durability time (shown in coloured boxes) equals life expectancy shown by the vertical lines (figure 2). This graph allows a quick estimate of the ratio of valve durability to average life expectancy at any age, keeping in mind that 50% of patients live longer than average. For example, an 85-year-old man can expect to live only an additional 6 years, thus a TAVI is likely to be durable enough to last the patient’s lifetime. A valve durability to life expectancy ratio close to, or greater than 1, helps support a strong recommendation for TAVI over SAVR in this age group, indicating that most patients and physicians would make this choice. In contrast, a 60-year-old man has a life expectancy of 22 years, so, a longer known valve durability is preferable. The BMJ Rapid Recs 27 make a strong recommendation for SAVR in patients under age 65 years because it is probable that SVD will occur with TAVI when the patient is still in good health otherwise. Although a transcatheter valve-in-valve procedure might be possible should SVD occur, we should not be prematurely optimistic about an approach with unknown long-term outcomes. For patients between age 65 and 85 years, the relative benefits risk of undergoing transfemoral TAVI versus SAVR and the issue of valve durability are more evenly balanced, with a weak recommendation for SAVR for those aged 65–75 years and a weak recommendation for TAVI in those aged 75–85 years, with individual patient preferences and other factors playing a larger role in this decision.27
Ideally, durability of an aortic bioprosthetic valve should outlast longevity of the patient. A patient-centred approach is of paramount importance, always considering the ‘valve durability to life expectancy ratio’.
The authors specially thank Dr Arvand Barghi (Programme of Internal Medicine, Western University, London, Ontario, Canada) for his assistance in drawing figure 1.
Competing interests None declared.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.