Newer TAVR Devices Reassure in Failed Surgical Valves

Emerging data for two latest-generation transcatheter aortic valve prostheses in the setting of failed surgical valves are “reassuring,” according to late-breaking clinical science presentations at the TVT Connect virtual meeting.

Amr E. Abbas, MD (Beaumont Hospital Royal Oak, MI), presented a propensity-matched analysis of valve-in-valve (ViV) outcomes at 1 and 12 months with the Sapien 3 device (Edwards Lifesciences) in low-, intermediate-, and high-risk patients compared with its use in native valve disease. Guilherme Attizzani, MD (University Hospitals Cleveland Medical Center, Ohio), presented 30-day and 1-year outcomes using the Evolut R and Evolut PRO (Medtronic) for failed surgical prostheses.

Both studies were based on cases included in the Society of Thoracic Surgeons/American College of Cardiology (STS/ACC) TVT Registry, with statistical analyses done by the sponsoring companies.

Balloon-Expandable ViV

Abbas pointed out that US Food and Drug Administration approval for valve-in-valve disease was originally granted to the earlier-generation Sapien XT device for high/extreme-risk patients in 2015, with that indication extended to Sapien 3 in 2017 based on earlier TVT data. Evidence has been lacking, however, for ViV patients at lower surgical risk.

The study included all patients in the TVT Registry who had undergone transfemoral TAVR with Sapien 3 or Sapien 3 Ultra for a ViV procedure between June 2015 and January 2020. Outcomes for the ViV group, representing 3.3% of the cohort, were propensity matched to patients undergoing native valve TAVR and stratified by STS score.

At 30 days, all-cause mortality was, not surprisingly, lowest among the patients with the lowest risk by STS score, at less than 1%, but rose in tandem with STS score so that it was 2.2% among patients in the intermediate-risk category and 4.3% in patients with an STS score > 8. Cardiac deaths followed a similar pattern. What was “really interesting and exciting,” said Abbas, was the low pacemaker rate: just 2.1% at 30 days and 3.9% at 12 months.

Overall all-cause and cardiac mortality at 12 months were 10.8% and 2.6%, respectively, but differed markedly by baseline risk such that all-cause mortality in the highest STS score patients was 17.9% as compared with 5.7% in the STS < 4 patients. “You can clearly see, again, the STS predictive risk score clearly classified outcomes at 1 year in the high-, intermediate-, and low-risk groups,” Abbas said.

Long-term follow-up data on the valve-in valve population remains crucial. Amr E. Abbas

For all of the propensity-matched comparisons across the three STS score categories, ViV patients actually had lower rates of all-cause mortality than their native-valve counterparts. One-year mortality was linked with the degree of tricuspid regurgitation and baseline hemoglobin levels, but not baseline echocardiographic gradients or need for a pacemaker.

Investigators also did an additional analysis that used heart team assessment of high or low risk instead of the STS score. Here, too, outcomes for both risk categories were better in the ViV patients than in propensity-matched native valve disease patients, said Abbas.

“In this real-world study, the ViV TAVR with the Sapien 3 and Sapien 3 Ultra had good 30-day and 1-year outcomes in all STS-based surgical risk groups,” Abbas said, noting that patients across the spectrum of STS score appear to fare better following ViV procedures than in native TAVR procedures, a finding that also held true when heart team assessment was used to risk stratify patients.

“These findings do support the safety and efficacy of ViV TAVR in lower-surgical-risk patients and confirms the previously reported inability of postprocedure echocardiographic hemodynamic parameters to predict clinical outcomes in ViV TAVR,” he concluded. “However, long-term follow-up data on the valve-in valve population remains crucial.”

Missing Information

Both Abbas and Attizzani lamented some of the shortcomings of the TVT Registry database at the time it was mined for their analyses. For example, investigators had no information on whether or not prosthetic valve fracture or the BASILICA procedure was used to implant the new transcatheter valve (a data point that has since been added to the collection sheets), no details on the cause or type of valve failure, and no information on other baseline features such as coronary anatomy.

Following Abbas’s presentation, discussants J. Dawn Abbott, MD (Rhode Island Hospital, Providence, RI), and David Cohen, MD (Kansas City, MO), expressed some surprise that the ViV outcomes were better across the board than those of the native valve patients. Cohen noted that a valve-in-valve procedure is itself included as part of the STS score; as such, an STS score of 4 in a ViV patient actually denotes a “healthier” patient than a native valve disease patient with the same score.

Other unidentifiable confounders, in particular lower age, also probably played a factor in the mortality difference, Abbas conceded, but he said: “I do think that looking into the real-world data with two methods, both STS-predicted risk as well as the heart team assessment, helps to ensure or at least comfort us that that mortality benefit does exist.”

Certainly as the indications for TAVR have expanded to lower-risk populations for native TAVR, patients are more curious about their alternatives. Subhash Banerjee

Asked by discussant Subhash Banerjee, MD (UT Southwestern, Dallas, TX), if the data are enough to support the use of ViV procedures in low-risk patients, a group not currently included in the indications for use, Abbas stressed that these are not randomized controlled trial data.

“I think that certainly as the indications for TAVR have expanded to lower-risk populations for native TAVR, patients are more curious about their alternatives and I think they want more say about their preferences for what type of procedure they want,” he said. In a joint discussion, these data could “certainly reassure both the operator and the patient that the outcomes at least aren’t worse in comparison to the native TAVR, where TAVR is approved for low risk.”

Self-Expanding Supra-Annular TAVR

Attizzani presented outcomes for 5,897 patients (mean STS PROM score 7.7) who’d been treated for failed surgical valves between April 2015 and June 2019, with the majority receiving the older Evolut R device and just 836 receiving the Evolut PRO.

At 30 days, all-cause mortality was 2.5% and the rate of pacemaker/ICD implantation was 4.9% overall, but significantly lower for the Evolut PRO device (3.0%) than for the Evolut R (5.3%). At 1 year, mortality had risen to 9.7% and pacemaker implantations to 7% overall, and both again were higher for the Evolut R. Mean gradients out to 1 year were also better for the Evolut PRO.

Both paravalvular regurgitation (PVR) and overall aortic regurgitation were better for the Evolut PRO postprocedure and at 30 days, but by 1 year there was no difference between the two valve types.

“‘TAV in SAV’ with self-expanding supra-annular valves showed favorable results with regards to survival, complications, and pacemaker use,” Attizzani concluded. “Evolut PRO was associated with small but significant improvements in 30-day and 1-year mean gradients, postprocedure aortic valve area, and 30-day total aortic regurgitation compared with Evolut R. TAV in SAV using Evolut R or Evolut PRO valves is a safe and effective treatment for patients with a failing surgical bioprosthesis.”

‘TAV in SAV’ with self-expanding supra-annular valves showed favorable results with regards to survival, complications, and pacemaker use. Guilherme Attizzani

In the discussion following Attizzani’s presentation, Cohen zeroed in on what he characterized as a surprisingly high rate of mild PVR—more than 17% at 30 days in the combined cohort (and 14.5% for Evolut PRO), falling to 13.6% at 1 year. That’s particularly surprising given that the “skirt” as part of this design was expressly intended to cut PVR rates, Cohen noted.

One explanation, Attizzani said, is that these numbers actually include “trace” PVR as well as mild, whereas other analyses typically treat trace as “zero” PVR.

Abbott pointed to the lack of any signs of “trade-off,” saying that the “most important thing” was the low rate of new pacemakers as compared with earlier valve iterations, including a further drop in pacemaker use with the Evolut PRO. But that, she noted, hasn’t appeared to come at the expense of higher PVR.

In fact, the most worrisome number, said Banerjee, was the signal of increased myocardial infarction with the newer valve at 1 year: 1.8% versus 0.8% for the new versus old device. “Are you in any way concerned about unintended coronary obstruction?” Banerjee asked.

Attizzani, in reply, stressed that he thinks many of the findings of this analysis reflect both increased operator experience as well as elements intrinsic to the valves themselves. “People with their [improved] understanding of the procedure are implanting the valve in a shallower position, therefore delivering better gradients and lower pacemaker rates—that’s my hypothesis,” he said.

“We don’t know anything about the baseline anatomy of these patients,” Attizzani continued, reiterating that the information included in the TVT Registry “is very limited. I’m definitely largely speculating, but I think that it could have been . . . that we are implanting them in a shallower position and therefore this might have, over time, impaired the flow of the coronaries. But again, this is speculation, and I do not have a way to precisely answer your question.”

Reassurances and New Questions

Sunil Rao, MD (Duke University Medical Center, Chapel Hill, NC), who moderated the virtual session, told TCTMD that he thought they data were “interesting, and the associations seen are reassuring for the clinical use of the valves.”

That said, Rao continued, “it begs the question about the utility of large registries for comparative effectiveness in the valve space, and what the role of these data are in extending the labels of approved valves. Obviously, we’d like to see more randomized trials, but since industry has access to the registry data and are able to do their own analyses, what is the future of clinical trials in this space?”

This was a point of discussion during the live segment that followed the late-breaking clinical science session today and explored the benefits of the TVT Registry—including the consistency, demographic breadth, and oversight of the data collection, as well as its linkage with Medicare outcomes—but also its shortcomings. Adding new fields to the data collection forms that could cover technical and anatomic details, for example, is a cumbersome process. The registry is also not designed to be able to compare different devices, a top-of-mind question for many in the field, and indeed head-to-head comparisons are expressly prohibited in the TVT Registry, an agreement made with industry partners when it was first funded and launched.

One option, suggested Cohen, would be a nested randomized clinical trial within the registry such that data prospectively collected could be repurposed for a head-to-head comparison. The SAFE-PCI trial led by Rao, for example, was done within the CathPCI Registry. Rao cautioned, however, that this process had been more cumbersome and complicated that it might look from the vantage point of hindsight, particularly since any changes made to case collection forms would then also need to be addressed in the software aggregating that information for the randomized trial.

Obviously, we’d like to see more randomized trials, but since industry has access to the registry data and are able to do their own analyses, what is the future of clinical trials in this space? Sunil Rao

But panelists today agreed that, even in the absence of comparative data on the two market-leading devices, the data coming out of the TVT Registry are reassuring for ViV use.

“I’m quite pleasantly, not surprised I would say, but glad to see the consistency of results across platforms,” Banerjee said. “There are high adopters of one platform more than another, so to see that both platforms perform well in ViV therapy is actually reassuring.”

At the end of their discussion the panel also broached the fact that both late-breaking presentations today had their analyses done in-house by the manufacturer of the device in question. These analyses are not independently verified or validated by the STS, ACC, or any other third party, confirmed Cohen, who has worked with the TVT Registry since its inception. Industry has the ability to make their own requests of the data set and run their own analyses, he said, and that does open some potential for problems.

“There are some things that are very simple for industry to do, but there are a lot of issues that come up in [the TVT Registry] like in any real-world data set, and frankly in randomized controlled trial data sets as well, about things like data completeness that are tricky to deal with and you really have to be sophisticated to deal with them,” Cohen said. “I do worry with some of the industry studies that they are not as used to that, because they are used to working with randomized controlled data that are super clean and the TVT data are not exactly that.”