Expansion of TAVR into Low-Risk Patients and Who to Consider for SAVR

Approximately 2% of the general population have a bicuspid aortic valve, and among all AS patients undergoing SAVR up to 49% have been found to have bicuspid anatomy [10]. When stratified by age, those over 70 years had a 38% rate of bicuspid AS compared to a 62% among those ages of 50–70 years [10]. Although these data were from a study of surgical patients, it suggests that as we expand TAVR into low-risk patients, a larger proportion will be found to have bicuspid anatomy. Bicuspid valves often have heavy asymmetric calcification, associated aortopathy, and present unique anatomical challenges for TAVR due to the often non-circular implantation of the transcatheter heart valve (THV) [11‐13]. These challenges may increase the risk of periprocedural complications including annular rupture, asymmetric valve deployment, and paravalvular leak (PVL), which may all impact short- and long-term outcomes with TAVR [14].

Studies of early generation THVs in bicuspid AS found that, compared to trileaflet AS patients treated with TAVR, bicuspid patients experienced higher rates of open conversion, significant PVL, and PPM implantation [15, 16]. With the use of CT-guided annular sizing, advances in the sealing mechanism of newer THVs, and operator experience, outcomes of TAVR for bicuspid AS have improved [16‐19]. However, conversion to open surgery, peri-procedural stroke, and need for PPM continue to remain significantly higher in bicuspid patients with newer-generation THVs [18]. Computed tomography (CT) studies of implanted THVs in patients with bicuspid AS have found suboptimal, non-circular valve expansion in bicuspid patients, which may impact long-term valve performance and the rate of valve degeneration [14].

Studies directly comparing SAVR to TAVR in bicuspid AS are scant; however, one study from the National Inpatient Sample of nearly 2000 propensity-matched patients found no major difference in in-hospital mortality or stroke between TAVR and SAVR [20]. Interestingly, SAVR was associated with significantly lower mortality among patients under 65 years of age and significantly lower rates of PPM implantation [20]. A smaller study prospectively studying low-risk patients with bicuspid AS demonstrated low rates of in-hospital mortality and stroke with TAVR; however, longer-term follow-up with a larger patient population is needed [21]. For low-risk bicuspid AS patients with associated aortopathy (ascending aortic dimension > 4.5 cm), SAVR with aortic root repair should be the preferred treatment [22, 23]. Given the unique challenges of bicuspid anatomy, SAVR will continue to have a central role in the management of low-risk severe bicuspid AS patients until more data is available on short- and long-term outcomes [24].

Despite advances in THVs in patients with AS, the use of TAVR for aortic regurgitation has been limited and remains off-label in the United States. Unlike AS, the etiology of aortic regurgitation may originate at the leaflet or aortic root level [25]. Several anatomical considerations make SAVR a more suitable therapy for low-surgical-risk patients with aortic regurgitation including the lack of sufficient calcium for THV anchoring, often dilated aortic root, and risk of PVL [26]. Many patients with pure aortic regurgitation have a bicuspid aortic valve which may further increase the complexity and result in a suboptimal transcatheter result [16]. Registry data show that even with newer-generation THVs, there remained a 3.8% risk of open conversion, a 12.7% risk of needing a second THV, an 18.6% PPM rate, and a 4.2% risk of moderate or greater residual aortic regurgitation [26]. Trials with specialized THVs designed for aortic regurgitation are ongoing (NCT02732704). Until more data are available, SAVR remains the treatment of choice for pure aortic regurgitation for low-surgical-risk patients.

The prevalence of patients with AS and a small aortic annulus is as high as 17% and predominantly affects older women [27]. Although there are no universally accepted criteria for what qualifies as a small aortic annulus, many propose an annular diameter of < 21–23 mm [28‐30]. Small aortic annulus has several short- and long-term implications for patients with severe symptomatic AS principally prosthesis patient mismatch and the sequela of this condition [31‐33]. Prosthesis patient mismatch occurs at an indexed effective orifice area of less than 0.85 cm²/m², and those patients with an EOAi less than 0.65 cm²/m² are classified as severe [34]. Severe prosthesis patient mismatch has been consistently shown to predict worse long-term outcomes including mortality [34‐36]. The rate of severe prosthesis patient mismatch is higher in SAVR than for TAVR likely because of the ability to oversize a THV for a patient’s annular size [37‐39]. However, patients with small aortic annulus receiving TAVR still have a 4–20% rate of severe prosthesis patient mismatch [37‐39]. Additionally, oversizing THVs in small aortic annulus patients with heavy asymmetric annular and LVOT calcification can increase the risk of annular rupture and coronary obstruction [40, 41].

SAVR in low-risk patients allows for the possibility of aortic root enlargement, which is effective at reducing the rate of post-operative prosthesis patient mismatch by up to 50% and results in lower 30 days and long-term morbidity and mortality [28, 29, 42]. Importantly, in younger low-risk patients, root enlargement with SAVR as a first operation would allow for a larger THV during valve-in-valve TAVR when the first bioprosthesis fails [31]. In those small aortic annulus patients who are younger and at a low bleeding risk, SAVR with a mechanical prosthesis can also be considered to obviate the need for a second valve procedure and further limit the risk of PPM. Although randomized studies are ongoing to compare TAVR versus SAVR in small aortic annulus patients (VIVA Trial NCT 03383445), SAVR will likely play a role in low-surgical-risk patients with severe AS and a small aortic annulus.

In both the PARTNER 3 and Evolut Low Risk trials, the final decision to randomize patients rested with a case review committee analyzing all clinical and imaging data. Of the patients who failed review in PARTNER 3, 38% failed due to severe LVOT calcium and 17% failed due to hostile aortic root while 8% failed due to severe LVOT calcium in the Evolut Low Risk trial [7, 8]. Although there is no objective or universally accepted criteria for determining the severity of annular and LVOT calcification, the presence of either has been associated with worse outcomes with TAVR including annular rupture, stroke, and paravalvular leak [43, 44]. LVOT calcium, especially in oversized balloon-expanded valves (BEV) and self-expanding valves (SEV) requiring post-dilation, increase the risk of annular rupture, which is associated with over 50% mortality [40, 44, 45]. Heavy annular/LVOT calcification is also linked to increased rates of ≥ moderate PVL, which is associated with worse outcomes and was more frequently seen in the early TAVR trials [44, 46‐49]. The incidence of ≥ moderate PVL was 0.6–3.6% in the low-risk trials, likely due in part to excluding heavily annular/LVOT calcification, improvements in THV technology, and the frequent use of post-dilation (20–30%) [7, 8]. The rate of ≥ moderate PVL was approximately 15% in the low-risk subset of the Nordic Aortic Valve Intervention (NOTION) study, which did not systematically exclude patients with heavy annular/LVOT calcification [50]. Additionally, 29–34% TAVR patients had mild PVL in the low-risk trials at 1 year, and how this affects THV and left ventricular performance over extended follow-up is still unknown [7, 8]. Among high-surgical risk patients, mild-PVL was associated with increased mortality at 2-year follow-up [51]. An advantage of SAVR is the ability to directly visualize and surgically remove heavy calcification at the time of AVR resulting in < 1% rates of significant PVL [7, 8].

Calcium and atheromatous embolization also play a key role in the underlying mechanism of stroke after TAVR. The manipulation of wires and catheters in the aortic arch, pre-TAVR balloon aortic valvuloplasty (BAV), THV positioning, and post-dilation all serve to liberate calcific deposits. The ability of cerebral embolic protection devices to modify stroke risk with TAVR remains controversial given conflicting data [52, 53]. Initially, stroke was a major drawback of TAVR, however 30-day disabling stroke rates have come down dramatically from 5–6% to 0.6–2% in the low-risk trials [1‐4, 7, 8]. Post-dilation has been shown to confer a 2.5-fold increase in stroke and was performed in 20–30% of patients in the low-risk trials [7, 8, 54, 55]. Excluding patients with heavy annular/LVOT calcification may partially explain the extremely low rate of stroke in the low-risk trials. As TAVR continues to expand into low-risk patients, SAVR should strongly be considered in patients with heavy, asymmetric calcification patterns given its link to annular rupture, significant PVL, and stroke.

The vast majority of TAVRs are performed via transfemoral (TF) access. In the Evolut Low Risk trial, the use of alternative access was approximately 1%, and a lack of TF access was an exclusion criteria for PARTNER 3 [7, 8]. Alternative modes of access include transthoracic access (direct aortic and transapical) and non-femoral peripheral access (transaxillary, transcarotid, and transcaval). Patients who lack TF access are more likely to be obese men with peripheral arterial disease [56]. All modes of alternative access have been associated with higher mortality and stroke compared to TF patients with markedly worse outcomes for transapical, direct aortic, and transcaval patients [57‐63]. Newer technology such as intravascular lithotripsy to accommodate the TAVR delivery sheaths in patients with stenotic calcified iliofemoral vessels is actively being studied and holds promise [64]. However, given the morbidity and mortality associated with alternative access TAVR, and the exclusion of these patients from the low-risk trials, SAVR should remain the preferred choice in low-surgical-risk patients without TF access.

The rate of coronary obstruction is rare, occurring in less than 1% of TAVR cases but is associated with high procedural morbidity and mortality [41]. Coronary obstruction typically occurs due to displacement of the native valve leaflets and any associated calcium. Identification of patients at risk of coronary obstruction has markedly improved, and patients with coronary ostia less 10–11 mm above the nadir of the associated sinus, effaced sinuses, and a narrow STJ are at higher risk of occlusion [41]. This often affects the left coronary ostium and is seen more frequently in women, patients with a small aortic annulus and effaced sinuses, bicuspid aortic valves, and during valve-in-valve TAVR [41]. Over 90% of these cases can be rescued with percutaneous coronary intervention, however one study found an 8.3% mortality rate [41]. The use of a chimney-stent to protect a coronary artery at risk of post-TAVR obstruction has shown good mid-term survival rates, however long-term data and coronary re-access issues remain in these patients [65]. Studies of the bioprosthetic or native aortic scallop intentional laceration to prevent iatrogenic coronary artery obstruction (BASILICA) technique has been shown to prevent coronary obstruction [66]. However, these studies were primarily done in intermediate and high-surgical-risk patients. Therefore, SAVR should remain the mode of treatment for low-surgical-risk patients with severe AS at significant risk of TAVR-induced coronary obstruction.

There are ample data on the long-term durability of bioprosthetic surgical valves, with approximately 23% primary bioprosthetic valve failure rates at 15 years [83]. Importantly, the age at which bioprosthetic SAVR is performed impacts the rate of bioprosthetic valve failure. Patients under 65 years of age who undergo bioprosthetic SAVR have a ~ 26% rate of bioprosthetic valve failure while those 65 years and older have a ~ 9% rate of this complication 15 years after valve implantation [83]. The method of defining bioprosthetic valve failure is vital to understanding valve durability. The surgical literature defines bioprosthetic valve failure as patients who actually undergo surgical re-operation [84]. In the era of valve-in-valve (ViV) TAVR, many patients with prohibitive surgical risk who would previously be declined for surgery, and thus not “counted” as bioprosthetic SAVR failure, are now being treated with ViV-TAVR. Data from the ViV International Database (VIVID) show that the average time from bioprosthetic SAVR to ViV-TAVR is 9 years [85]. As more patients undergo ViV TAVR, our understanding of the durability of surgical bioprosthetic valves is likely to evolve. For THVs, long-term durability beyond 5–8 years is poorly understood and short-term durability appears similar between SEV and BEV [86‐88]. Unlike the definition of bioprosthetic valve failure in SAVR studies, there are numerous proposed methods of defining structural valve degeneration (SVD) and bioprosthetic valve failure after TAVR [89]. SVD defined by the European Association of Percutaneous Cardiovascular Interventions (EAPCI) is divided into mild, moderate, and severe based on the mean gradient, change in mean gradient from the initial post-procedural assessment, and degree of intra-prosthetic regurgitation [90]. Bioprosthetic valve failure is defined as a THV with severe hemodynamic compromise either requiring intervention or causing death [91]. The Valve Academic Research Consortium (VARC) criteria define severe structural valve dysfunction as a mean gradient > 40 mmHg (moderate to severe stenosis), EOAi < 0.6–0.65 cm²/m² (severe patient prosthesis mismatch), DVI < 0.35, or moderate-to-severe prosthetic valve regurgitation [91]. At 5–8 years THVs have a 3.6–10.8, 0–2.5, 0.6–7.5% for moderate SVD, severe SVD, and bioprosthetic valve failure, respectively [92]. Numerous factors unique to TAVR can affect long-term durability including THV crimping, post-dilation, noncircular implantation, PVL, and leaflet thrombosis [93]. Hypo-attenuating leaflet thickening (HALT) and hypo-attenuation affecting motion (HAM), both likely caused by leaflet thrombus formation, have been associated with SVD and post-TAVR stroke [89, 94]. While less commonly seen in SAVR patients, HAM and HALT have been reported in up to 40% of TAVR [95, 96].

A critical implication of bioprostheses durability is how the life-span of the valve compares to the life expectancy of the patient and the possibility of needing a second valve replacement. Indeed, in younger low-risk patients who have a > 15-year life expectancy, it is reasonable to consider implantation of a mechanical valve given the lower rate of mechanical aortic valve failure compared to surgical bioprosthetic valve failure [81]. For failed bioprosthetic surgical valves, there is a growing body of data demonstrating the safety and short-term clinical efficacy of implanting a THV (TAV-in-SAV) [97, 98]. Among patients in the TAV-in-SAV substudy of PARTNER 2 and CoreValve Pivotal Extreme risk trials, TAV-in-SAV 30-day mortality was approximately 2.6% with stable THV hemodynamics out to 3 years [99‐102]. TAV-in-SAV patients in the VIVID registry had a 5.3% overall mortality at 30 days [103]. When TAV-in-SAV outcomes are compared to re-do SAVR, limited available data show comparable 30-day and 3-year mortality [104, 105]. On the contrary, limited data exist on the treatment of a failed THV valve. Surgical removal and replacement of a failed THV can be a complex surgery potentially requiring aortic root replacement while the implantation of a second THV (TAV-in-TAV) has not been widely performed. A recent registry study found that 0.3% of TAVRs were TAV-in-TAV procedures done predominantly for valvular regurgitation with 2.9% overall mortality, low rates of stroke and coronary obstruction, but a 9% rate of ≥ moderate PVL [106]. As TAVR patients continue to live longer, our understanding of long-term bioprosthetic valve durability will improve so we can better inform our decision on the best “first implanted valve” in younger, low-risk AS patients.

Given the lack of direct visualization of the aortic root during implantation of a THV, the interplay of the implanted valve commissures, aortic root geometry, displaced native aortic valve leaflets, and coronary sinus calcification can limit access to the coronary ostia after TAVR. The rate of coronary artery disease (CAD) in TAVR patients ranges from 40 to 75% and as TAVR expands to low-risk patients, it is likely that these patients may require future coronary angiography and percutaneous coronary intervention (PCI) for either urgent or elective indications [107]. One study reported a 10% rate of acute coronary syndrome after TAVR with 47% of cases occurring within the first year [108]. A MDCT study of TAVR patients found that in 51% of cases, the ostia of one or both of the coronary arteries was blocked by the neo-commissure of the THV [109]. This is particularly problematic in patients needing a TAV in TAV procedure where there are two THVs with neo-commissures that could both block access to the coronary arteries. In patients receiving SAVR, there is direct visualization of the aortic annulus during implantation and re-accessing the coronary arteries is done routinely. As TAVR expands into younger low-risk patients, the ability to re-access the coronary arteries may prove to be a major limitation of TAVR, particularly with the SEV [110].

A key secondary endpoint studied in both the PARTNER 3 and Evolut Low Risk trials was QOL measured as the change in Kansas City Cardiomyopathy Questionnaire Overall Summary (KCCQ-OS) score. In both trials, there was a marked improvement in the KCCQ-OS score for TAVR patients at 30 days and 1 year [7, 8]. Importantly, TAVR performed better than SAVR in regard to KCCQ-OS at 30 days, however by 12 months there was minimal difference between the two groups in both trials [7, 8]. Prior analyses of the PARTNER 1 and CoreValve extreme risk trials, as well as analysis of the STS/TVT registry, have shown similar improvements in QOL extending out to 3 years [111‐113]. These findings are reassuring for low-risk patients who are weighing the recovery from these two treatment modalities as they will likely enjoy the same, marked improvement in QOL by 1 year regardless of the strategy they pursue.

Women have consistently been shown to have differential short- and long-term outcomes after TAVR compared to men. Early studies of the earlier generation THVs in high and intermediate surgical risk patients showed that women had improved survival compared to men at 1 year [114‐116]. However, treatment of severe aortic stenosis in women studies of the newer-generation THVs have shown no significant long-term differences in mortality between men and women [117]. Notably, women have consistently been shown to have higher rates of periprocedural complications including major vascular complications, stroke, and major life-threatening bleeding [116, 118, 119]. Additionally, women tend to have lower coronary heights and smaller annuli, which increases their risk of coronary obstruction, patient prosthesis mismatch, as well annular rupture [119‐121]. Women represent a larger portion of patients requiring alternative access TAVR likely due to smaller diameter femoral arteries in women [119]. Given these unique anatomic and procedural outcomes, women with severe AS who are at low surgical risk should be given special consideration for SAVR given the consistently worse periprocedural complications seen in this group.