Transcatheter aortic valve replacement (TAVR) has become the dominant form of aortic valve replacement for severe aortic stenosis.[1,2] There are 2 major TAVR designs: balloon-expandable valves (BEVs) and selfexpanding valves (SEVs), each with unique characteristics. Valve choice for a given patient is sometimes dictated by specific anatomic and clinical factors, but there is frequently perceived clinical equipoise and room for operator discretion.
The SOLVE-TAVI (compariSon of secOndgeneration seLf-expandable vs. balloon-expandable Valves and gEneral vs. local anesthesia in Transcatheter Aortic Valve Implantation) trial was an investigator-initiated, multicenter trial funded by the German Heart Research Foundation and Helios Health Institute that was designed to inform the decision on valve choice between the Edwards SAPIEN 3 (S3) BEV and Medtronic Evolut R SEV, as well as the decision for procedural anesthesia choice (general vs conscious).[3] Among 447 patients randomized, there was no difference in the composite endpoint of allcause mortality, stroke, moderate/severe prosthetic valve regurgitation, and permanent pacemaker implantation (PPI) between valve type at 30 days or 1 year.[4] Additionally, there was no difference in the composite endpoint of mortality, stroke, myocardial infarction, and acute kidney injury by anesthesia type.
In this issue of JACC, Feistritzer et al[5] present 5-year outcomes of the SOLVE-TAVI trial. At this longer-term follow-up interval, more clinical events have accrued, and as a result investigators were better equipped to detect differences between treatment strategies. Nevertheless, there was still no significant difference in the composite endpoint between the SEV and BEV groups (67.7% vs 63.4%) or between the general anesthesia and conscious sedation (CS) groups (61.3% vs 51.4%). Of the individual components, the SEV group had lower rates of stroke (2.2% vs 9.6%) and myocardial infarction (0% vs 2.5%) and the CS group had lower all-cause mortality (41.5% vs 54.3%).
The authors should be commended for conducting long-term follow-up of this important trial free of industry funding. The key strength of this study is its randomized nature. This is particularly important in the valve choice comparison given unique characteristics that may not be easily ascertainable from administrative or registry data sets, such as calcification of the left ventricular outflow tract or extensive coronary artery disease, that could drive both valve choice and outcomes.
Data directly comparing outcomes of patients treated with BEV vs SEV have been limited.[6,7,8] The industry-funded SMART (Small Annuli Randomized to Evolut or SAPIEN Trial) recently demonstrated that SEVs were noninferior to BEVs for the coprimary composite endpoint of death, disabling stroke, or heart failure rehospitalization.[9] While much has been made of the other coprimary composite endpoint of bioprosthetic valve dysfunction at 12 months demonstrating superiority for SEVs, it should be noted that this was driven primarily by higher echocardiographic gradients for BEVs. The validity of this finding is unclear because BEVs have been shown to demonstrate lower gradients when assessed invasively, and there was no associated difference in clinical outcomes at 1 year. While longer-term followup may reveal some benefit, similar findings were also noted during 5-year follow-up of the Bern TAVI registry in which lower echocardiographic gradients with the Medtronic SEV compared with the Edwards BEV did not translate to any clinical benefit in mortality or functional class in patients with small annuli.[7]
Interpreting the “equality” of valve choice in the SOLVE-TAVI trial requires a critical appraisal of the current findings and their consistency. With regard to PPI, for instance, the Bern TAVI registry demonstrated a significant, nearly 3-fold higher risk of new PPI at 30 days among the SEV group (20.6%) compared with the BEV group (8.3%). This finding of substantially higher PPI rates is commensurate with nearly every other trial of SEVs and BEVs, whether comparative or single platform based, sponsored trial or national registry, except the current paper. Even the SMART trial showed a 50% higher risk of PPI with SEVs that barely failed to meet statistical significance (P = 0.055).[9]
Similarly, stroke rates in the SOLVE-TAVI trial were substantially higher for BEVs than for SEVs even during early follow-up. This is also a curious finding, as most single-platform studies that randomized TAVR vs surgical aortic valve replacement consistently show either similar or lower stroke rates for the BEV than the SEV, and U.S. registry data show substantially higher stroke rates at 30 days for the SEV than was seen in the SOLVE-TAVI trial.[10] Of course, we acknowledge that the findings of different studies cannot be fairly compared, and event rates across single-platform studies may differ due to different inclusion criteria. Nevertheless, even in the comparative Bern TAVI registry, we see a numerically higher risk of stroke with an SEV than a BEV.[7] Perhaps at best we can consider both valve platforms as equivalent for stroke risk based on the totality of data. But questioning the validity of 2 components of the primary endpoint (PPI and stroke) in the SOLVE-TAVI trial leaves us on shaky ground in truly declaring the overall equality of outcomes for the SEV and BEV from the current trial.
In addition to the limitations pointed out previously, the relevance of broad composite endpoints, designed to enable investigators to achieve greater power with a smaller sample size, can be challenging to interpret in clinical practice when acute kidney injury or PPI are grouped together with an endpoint such as mortality, which has much greater significance to a patient. Additionally, although more events have accrued at the 5-year time point, the relatively small sample size continues to limit the ability to exclude a small yet clinically significant difference between treatment strategies. Finally, this was an elderly group of patients (mean age 82 years), which makes translating the findings to a younger group challenging.
So how can we consider valve choice at the current time? Perhaps we can place some faith in the SOLVE-TAVI trial that elderly patients may fare similarly well with either an SEV or a BEV. But as we consider the issue of lifetime management in treating younger and/or lower-surgical-risk patients, we must also acknowledge the very relevant differences that exist between the 2 platforms. It is a practical reality that the taller-frame SEV will exclude more patients from placement of a TAVR-in-TAVR than a shorter-frame BEV. While leaflet-modification strategies can be employed, those procedures are still performed in a small minority of TAVR centers and require commissural alignment. This is therefore an issue of paramount importance. On the other hand, operators and patients may feel a visceral negativity toward high gradients on follow-up. Among patients with a small annulus (whether in absolute terms or relative to their body habitus) or small degenerated surgical valve, perhaps use of the SEV is more reassuring, and may be considered especially if there is no anticipated need for future TAVR-in-TAVR and/or high risk for conduction deficit requiring PPI. We should also acknowledge that not all centers have access to multiple valve platforms due to hospital and health system contracting requirements, some operators have substantially more experience with one platform than another, and many operators do not assess preprocedural computed tomography scans themselves (with regard to understanding feasibility of future options). These are difficult measures to quantify but are akin to typical “learning curves” that exist in TAVR and may be relevant to the care of patients based on their treating center.
This study also reaffirms the safety of CS-based TAVR. While at time of enrollment there was wide variability in sedation choice for TAVR, clinical practice has since evolved, and the vast majority of TAVR procedures are now conducted with CS.[11] Notably, the mortality signal in favor of CS in the SOLVE-TAVI trial is consistent with results seen in the TVT (Transcatheter Valve Therapy) registry using a quasi-experimental study design. While upfront hemodynamic shifts may explain an early signal toward benefit with CS, the continued accrual of mortality benefit with CS seen in this study remains a rich area for further inquiry.
Ultimately, the SOLVE-TAVI trial may support ongoing clinical equipoise for valve choice in patients who are candidates for both types and reaffirms safety of conscious sedation in most patients. But, as with most things, it behooves us to be thoughtful about valve choice in each specific patient and question proactively, rather than rest assured that the question is solved.
J Am Coll Cardiol © 2024