Some Surprises in Low-Risk TAVI Follow-up: PARTNER 3 and Evolut Trials

(UPDATED) TAVI benefits were attenuated and mortality trended higher in PARTNER 3 at 5 years. For Evolut, TAVI maintained an edge.

Some Surprises in Low-Risk TAVI Follow-up: PARTNER 3 and Evolut Trials

SAN FRANCISCO, CA—Transcatheter aortic valve implantation for low-risk patients with severe, symptomatic aortic stenosis remains a safe, effective, and durable treatment option for at least 4 to 5 years, according to follow-up data from two landmark trials presented today at TCT 2023.

Martin Leon, MD (NewYork-Presbyterian/Columbia University Irving Medical Center, New York, NY), who presented PARTNER 3 data during the late-breaking clinical trial session, reported there was no significant difference in the risk of death, disabling stroke, or rehospitalizations between TAVI with a balloon-expandable bioprosthesis (Sapien 3; Edwards Lifesciences) and SAVR at 5 years.

Additionally, valve hemodynamics assessed by aortic valve area were “virtually superimposable” at 5 years, while the risk of bioprosthetic valve failure (BVF), defined by VARC-3 criteria, was “quite low and similar between the two therapies,” he said.

“We can tell our patients who have low-risk aortic stenosis that for these two therapies, that by the end of 5 years, there’s more than a 70% chance they will be alive and feeling well with either no or very mild symptoms,” Leon said during a morning press conference. “We can also tell them that at the end of 5 years, there’s also a more than 85% chance with either therapy they’ll be alive with a durable heart valve based on BVF assessment.”

However, the PARTNER 3 data did contain some surprises. Beyond 1 year, there was “an attenuation of the differences” in the primary composite endpoint, which had favored TAVI previously, according to investigators. Published in the New England Journal of Medicine to coincide with TCT 2023, follow-up also showed there were numerically more deaths among patients assigned to TAVI than those assigned to surgery.

“All-cause mortality [curves] did cross towards the middle of the follow-up,” Leon told TCTMD. “First, all-cause mortality is both cardiovascular and noncardiovascular, and there was a disproportionate amount of noncardiovascular deaths in the TAVR arm versus surgery. If you look at cardiovascular mortality, the difference is 0.4% over 5 years. So, it’s mainly the noncardiovascular deaths.”

Leon stressed that the difference in all-cause mortality—which was 10.0% with TAVI versus 8.2% with surgery—wasn’t statistically significant. Speaking during a deep dive session focused on PARTNER 3, Leon said he was “thrilled” with the 5-year results.

“The fact that we were able to achieve results that are essentially equivalent to our surgical colleagues tells me that that this is a meaningful alternative,” said Leon. “I think it’s superb data. We can tell our patients that if they have low risk and severe aortic stenosis, the 5-year outcomes are very reassuring.” Despite those words, some surgeons believe the numerical difference in all-cause deaths warrants further investigation.

“Statistically speaking, there was no difference in mortality, but the curves cross at that 2- to 3-year time point,” Kendra Grubb, MD (Emory University, Atlanta, GA), told TCTMD. “Personally, it’s a red flag. If you look at the 2-year data, the curves were converging and you wondered what was going to happen. I think the exuberant statements made at the 1-year point where there was superiority [with TAVR over SAVR] have been called into question. If you go back to that 1-year data, where TAVR first for everyone was the message, now we have to be a little bit more thoughtful.”

Gilbert Tang, MD, MBA (Mount Sinai Health System, New York, NY), echoed that sentiment. The difference in mortality, he commented to TCTMD, is a signal that doesn’t warrant clinical concern “right now,” but close monitoring in longer-term follow-up, as well as a deeper dive into the causes of death, will be important. 

In contrast, 4-year data from the Evolut Low Risk trial showed the primary outcome curves diverging in favor of TAVI over time. There was no statistically significant difference in all-cause mortality, but there were numerically fewer deaths with TAVI in follow-up.

“We all know that TAVR has an early advantage over surgery in terms of mortality, stroke, and recovery,” lead investigator Michael Reardon, MD (Methodist DeBakey Heart & Vascular Center, Houston, TX), told TCTMD. “The real question is, is this durable? Does it hold up?” These new data, he said, show not only that the early benefit with the self-expanding valve is sustained, but also that it increases in favor of TAVI. “When we break this down by all-cause mortality and disabling stroke, [the benefit] is being driven by better all-cause mortality with the Evolut valve versus surgery.”

When published 4 years ago, both PARTNER 3 trial and the Evolut Low Risk trial were seen as big wins for TAVI in patients at low risk for death with cardiac surgery. These two trials led the US Food and Drug Administration to approve the self-expandable Evolut device and the balloon-expandable Sapien 3 and Sapien Ultra transcatheter heart valves for low-risk patients with severe aortic stenosis.

Attenuation of Outcome Differences

In PARTNER 3, which was published 4 years ago, the primary endpoint occurred in 8.5% of the TAVI-treated patients and 15.1% of the surgical patients, a statistically significant difference in both noninferiority and superiority testing.

Now at 5 years, the primary endpoint occurred in 22.8% of TAVI patients and 27.2% of the surgery group (P = 0.07). For the second primary endpoint, which was a hierarchical composite of death, disabling stroke, nondisabling stroke, and number of rehospitalization days, investigators performed a win-ratio analysis. Here, again, there was no significant difference between the two treatments (win ratio 1.17; 95% CI 0.90-1.51).

When teased out separately, the risk of all-cause mortality was numerically greater with TAVI than with surgery (HR 1.23; 95% CI 0.79-1.90), with the Kaplan-Meier curves hinting that the difference emerged just beyond 3 years. In a landmark analysis focused on 1 to 5 years, all-cause mortality was 9.1% with TAVI and 5.9% with surgery (HR 1.61; 95% CI 0.97-2.67), while rates of cardiovascular death were 4.7% and 3.1%, respectively (HR 1.58; 95% CI 0.78-3.19).

During the TCT session, Leon said that high-quality, late follow-up data in surgery are “problematic,” noting they had 5-year outcomes on just 88% of SAVR patients. When investigators performed a vital-status “sweep” analysis—using publicly available data or contacting patients and family to ascertain vital status—mortality differences narrowed such that the 5-year mortality rates were 10.2% with TAVI and 9.0% in the surgical group.

Stephan Windecker, MD (Bern University Hospital, Switzerland), focusing on the mortality  findings, noted that the mortality curves also crossed in 5-year follow-up of PARTNER 2A trial. He added, though, that the sweep analysis performed by PARTNER 3 investigators helped address possible attrition bias in the surgical arm and narrowed the observed difference in mortality between TAVI and surgery.

To address the mortality differences, Leon broke down the causes of death, stating that adjudication of cardiovascular mortality was very rigorous in PARTNER 3. From baseline to 5 years, there were 26 deaths from cardiovascular causes (5.5%) in the TAVI arm versus 21 deaths from cardiovascular causes (5.1%) in the surgical group. In contrast, there were 22 and 13 deaths from noncardiovascular causes in the TAVI and SAVR arms, respectively. Numerically, there were more deaths from cancer (9 vs 5), COVID-19 (3 vs 1), and sepsis (4 vs 1) in the TAVI arm compared with the SAVR arm.

Rates of stroke and rehospitalization remained numerically lower with TAVI at 5 years, but the difference was not statistically significant.

Regarding valve hemodynamics, there was no significant difference in mean aortic valve gradients (12.8 mm Hg with TAVI vs 11.7 mm Hg with surgery) and no difference in the risk of BVF (3.3% with TAVI vs 3.8% with surgery).

Bernard Prendergast, MD (Cleveland Clinic London, England), called the PARTNER 3 data “exciting,” noting that the community has been eagerly awaiting the 5-year outcomes.

“Importantly, for me, they reinforce the guideline recommendations on both sides of the Atlantic,” he said. “You’ll recall that fairly bold steps were made to lower the age thresholds to determine the arbitration between SAVR and TAVR. It was a very bold step in the US to go down to 65 years. Although the 10-year [PARTNER 3] data are going to be pivotal, we’re on the right track. That should be reassuring for a much wider community, not just interventional cardiologists.”

David Cohen, MD (Cardiovascular Research Foundation/St. Francis Hospital, Roslyn, NY), who moderated the press conference, noted that there have long been concerns about valve durability, particularly since TAVI was approved for low-risk patients. Pinak Shah, MD (Brigham and Women’s Hospital, Boston, MA), who wasn’t involved in the trial, noted that that valve hemodynamics with Sapien 3 “appear nice and stable” while rates of structural valve deterioration are also acceptable.

“Very encouraging data,” summarized Shah. “Of course, we’re going to have to wait and see what things are like in the next several years.”

Speaking during the press conference, Grubb said that surgery sets a very high bar for TAVI to clear in low-risk patients. While generally bullish on the 5-year outcomes with TAVI, Grubb said they’ve also seen problems with some of the newer surgical valves around 7 to 8 years. “I’m encouraged by these data, but certainly would look more at the 10-year outcomes to draw any firm conclusions.”


The Evolut Low Risk trial also included patients with a low predicted risk of death with surgery randomized to TAVI with a self-expanding bioprosthesis or to SAVR.

The primary endpoint of the trial—a composite of all-cause mortality and disabling stroke at 24 months—showed that TAVI was noninferior to surgery (5.3% vs 6.7%; P > 0.999 for the probability of noninferiority). Three-year data, which were published earlier this year, confirmed the durable benefits of TAVI in this population.

The 4-year results, which were published as a research letter in the Journal of the American College of Cardiology, include data from 94.7% of randomized patients. The primary endpoint of death or disabling stroke occurred in 10.7% of the TAVI-treated patients and 14.1% of those treated with surgery (HR 0.74; 95% CI 0.54-1.00). The benefit with TAVI appeared to increase over time, with the absolute difference of the primary endpoint increasing from 1.8% at 1 year to 3.4% at 4 years.

Overall, all-cause mortality was 9.0% and 12.1% in the TAVI and SAVR arms (P = 0.07), respectively, while disabling stroke occurred in 2.9% and 3.8% of patients, respectively (P = 0.32).

Rehospitalization for any reason related to the aortic valve occurred in 10.3% of patients treated with TAVI and 12.1% treated with surgery, but reintervention on the valve—an indicator or valve performance—occurred in just 1.3% and 1.7% of patients, respectively (P = 0.63). Rates of subclinical or clinical valve thrombosis and valve endocarditis were also low and did not differ between treatments.

TAVI patients had significantly lower aortic valve mean gradients and greater effective orifice area than the surgical patients. No/trace paravalvular regurgitation (PVR) at 4 years was better with surgery (98.4% vs 84.7% with TAVI; P < 0.001) but there was no difference in the percentage of patients with moderate or greater PVR.

To TCTMD, Tang said the Evolut Low Risk data suggest that TAVI with the self-expanding device is better than surgery when it comes to hard clinical endpoints, in addition to the sustained hemodynamic effects, at least out until 4 years. With the event curves, the separation occurs within the first 30 days, which makes sense given the up-front costs of surgery. “Once you overcome that first 30 days, essentially, your outcomes will not change versus surgical aortic valve intervention,” said Tang. “The curves basically stay parallel, meaning that there might not be outcome differences between the two therapies. That’s good because if you have a less invasive alternative to surgery, you’d certainly prefer that. It's also reassuring that in this 4-year data that the hemodynamics remain stable, that the gradients are still superior to SAVR, which I think is an important discussion with the patient.”

Reardon said that the data reassure about TAVI with the self-expandable bioprothesis in low-risk patients, but like Grubb, said that his experience as a surgeon taught him to be cautious.

“For those of us who have been doing surgery for four decades, I can name several valves I was so excited about,” he said. “Valve durability is a combination of valve design and tissue science—valve durability, which is one of the key things for younger patients, is not a class effect. It’s different for surgical valves and it’ll differ for TAVR valves. We really need to get out to 10 years.”

Why the Differences Between Trials?

As for why clinical outcomes appeared to differ between the two trials, specifically all-cause mortality, nearly everybody cautioned against making any cross-study comparisons.

“Long-term survival . . . is heavily dependent on your intrinsic risk level to begin with,” Reardon told TCTMD. “We really have different patient populations so it’s very hard to speculate. In fact, it’s scientifically invalid for me to compare my surgical valve against the PARTNER surgical valves. It’s not the way these trials work. . . . For [Evolut Low Risk], I think that it’s controlled by their intrinsic risk levels and valve performance.”

Leon also refused to engage in any comparisons between the two trials, noting that all-cause mortality was 12.1% with SAVR at 4 years versus 5.1% in surgical arm of PARTNER 3. “When the control arm is that different, it’s almost irrelevant to make comparisons,” said Leon.

Tang also emphasized that comparing PARTNER 3 and Evolut Low Risk is not appropriate given the differences in clinical trial design. For one thing, exclusion rates based on anatomic screening was higher in PARTNER 3, suggesting there was a difference in the selection of patients. Not only that, but the technology differed, with PARTNER 3 investigators using the Sapien 3 device, which wasn’t the state-of-the-art transcatheter valve at the time. The Evolut Low-Risk Trial included a broader generational range of available valves (CoreValve, Evolut R, and Evolut PRO).

“I think we have to treat these two trials separately,” said Tang. “I don't think it's fair to say one device is worse than another.” 

Right now, there is no clear reason why mortality event rates were higher with TAVI in PARTNER 3, said Tang, adding that the data did reassure when it came to risk of stroke, hospitalizations, and BVF.

To TCTMD, Grubb said when it comes to counseling patients, which is what these data are meant to inform, she’ll tell them there’s no difference in hemodynamics or clinical outcomes at 5 years between surgery and Sapien 3. “Am I concerned about the [mortality] curves? I am, but right now it’s not statistically different. I have to be honest with what the data shows,” she said.

Given the differences in clinical trials, Grubb agreed that a head-to-head comparison between Sapien 3 and CoreValve/Evolut based on these data is impossible. Her role as a physician is to determine if a patient in her office is “similar enough” to those in the trials so that she can advise them on the best possible treatment. “My patient might not fit into either one of those trials, but I do think I can safely say that if a 73-year-old patient chooses TAVR, and their anatomy is suitable for a Sapien valve, then their results are going to be the same as surgery at 5 years based on what we know,” Grubb explained

Isolated Aortic Valve Replacements

Vinay Badhwar, MD (West Virginia University, Morgantown), who wasn’t involved in the studies, said the evidence from both trials suggests physicians will need to be cautious when advocating for TAVI in low-risk patients. “I haven’t seen anything here today that is overly convincing, at least scientifically, that supports a wholesale to move to TAVI in young, low-risk patients,” he commented. “We know already from other national large-scale evidence of the excellent long-term survival we have with isolated SAVR.” 

Last week, Badhwar and colleagues published a real-world analysis of patients undergoing isolated SAVR in the Society of Thoracic Surgeons (STS) Adult Cardiac Surgery database that showed survival at 5 years is almost 93% and nearly 90% at 8 years. Survival was even better in patients under 75 years and in those with an STS-PROM score of less than 1%.

He also pointed out that aortic valve replacement is largely an isolated procedure in transcatheter clinical practice, but up to 25% of the surgical patients in PARTNER 3 and Evolut Low Risk trial underwent concomitant procedures, including CABG surgery. 

“We know clearly that concomitant operations are associated with worse operative outcomes over isolated AVR procedures,” Badhwar told TCTMD. “Patients with ischemic disease are different that those with isolated valve disease. That, for example, is a fundamental difference. Now, we all have the utmost respect for randomized evidence—but it’s only as good as the comparator groups.” 

For this reason, Badhwar said the analysis that physicians want and need, particularly in low risk, is a comparison of isolated TAVI with isolated SAVR. 

“For all valve practitioners, we must have equipoise for both options as we use all evidence to judge the optimal low-risk therapy,” he said. “Therefore, what would be helpful would be to look at the STS benchmark for surgical survival of over 42,000 patients following low-risk isolated SAVR, and then do an apples-to-apples-to-apples comparison: isolated SAVR [from the STS registry], isolated TAVR from both trials versus isolated SAVR from both trials. That’s the data we really want to see. Then we can truly make sure we have data to make clinical decisions. We want a truly informed heart team, not for one specialty or another, or one therapy over the other, but for the best long-term option for our patients.”

Vinod Thourani, MD (Emory University, Atlanta, GA), who presented mortality data during the deep dive session, shed light on patients undergoing isolated SAVR. In that analysis, there was “no appreciable difference” in all-cause mortality at 5 years between those undergoing isolated and concomitant SAVR and this rate was “strikingly similar” to the STS benchmark data, said Thourani. Reardon, speaking during the press conference, said they plan to perform an analysis of TAVI versus isolated SAVR and will publish this with their 5-year outcomes data.

Michael O’Riordan is the Managing Editor for TCTMD. He completed his undergraduate degrees at Queen’s University in Kingston, ON, and…

Read Full Bio
  • Mack MJ, Leon MB, Thourani VH, et al. Transcatheter aortic-valve replacement in low-risk patients at five years. New Engl J Med. 2023;Epub ahead of print.

  • Forest JK, Deeb GM, Yakubov SJ, et al. 4-year outcomes of patients with aortic stenosis I the Evolut Low-Risk Trial. J Am Coll Cardiol. 2023;Epub ahead of print.

  • Leon reports grant support from Abbott, Boston Scientific, Edwards Lifesciences, and Medtronic. He reports consulting for Anteris and Foldax.
  • Forest reports grant supper/research contracts/consultant fees/honoraria from Edwards Lifesciences and Medtronic.