Long-term PARTNER 2A Results Put Spotlight on Paravalvular Leak in TAVR

SAN FRANCISCO, CA—Patients with severe symptomatic aortic stenosis and an intermediate surgical risk appear to fare just as well with TAVR as with surgery through 5 years of follow-up, with no evidence that the valves are breaking down, long-term data from the PARTNER 2A trial show.

But moderate-to-severe paravalvular regurgitation (PVR) occurred at a higher rate with TAVR versus surgery throughout the 5-year follow-up period and was associated with a heightened risk of all-cause death.

For the primary results, the rate of death or disabling stroke at 5 years was 47.9% in patients treated with the Sapien XT transcatheter valve (Edwards Lifesciences) and 43.4% in those who underwent surgery (P = 0.21), Vinod Thourani, MD (MedStar Heart & Vascular Institute, Washington, DC), reported here at TCT 2019.

That finding was consistent among patients treated via transfemoral access, but those whose procedures were done via transthoracic access had significantly worse outcomes compared with surgery (59.3% vs 48.3%; HR 1.32; 95% CI 1.02-1.71). Valve-related rehospitalizations and reinterventions were more frequent in the overall TAVR group.

Much of the reaction to these results, both in a morning press conference and later in the Main Arena, centered on the paravalvular leaks. Speaking to the press, Thourani noted that progress has been made to bring those down with routine use of CT sizing (sizing in PARTNER 2A was based on transthoracic echocardiography) and with device improvements. The newest-generation Sapien 3 valve (Edwards Lifesciences) has been associated with rates of moderate-to-severe PVR of less than 1% in low-risk patients, much lower than the 6.5% rate seen with Sapien XT in these 5-year data.

“We have to be vigilant in not having moderate-to-severe PVR and if . . . for whatever reason we leave moderate-to-severe [PVR], we have to do other adjunctive therapies to decrease that,” Thourani said. “We’ve already shown that that’ll be beneficial.”

Commenting for TCTMD, Dharam Kumbhani, MD (UT Southwestern Medical Center, Dallas, TX), observed, “When this trial was ongoing, we didn’t have an appreciation that leaving these patients with paravalvular leaks, large leaks, was actually going to be bad for them. That’s what we’ve learned actually from the PARTNER trials, that leaving them with a lot of leaks is a bad thing.”

Evidence has accumulated since then to show that moderate-to-severe PVR does have a detrimental impact on outcomes, and the issue of leaks has been largely addressed through improvements in devices and sizing techniques, he said. “It is desirable in the current generation to have valves that do not have significant [PVR],” Kumbhani said.

A major take-away for him, however, was that the transcatheter valves held up well in terms of hemodynamics through 5 years, with no difference between the trial arms in mean aortic valve gradient. “It gives us . . . reassurance that you’re not seeing this dramatic drop-off in terms of long-term performance,” Kumbhani said. “For me, this is actually a very big addition of this trial.”

Thourani noted that the investigators will be following patients out to 10 years, and Kumbhani said that additional information on long-term durability is “really what’s needed as we now have gravitated toward a lower-risk pool of patients.”

PARTNER 2A

The PARTNER 2A trial randomized 2,032 patients—after stratification according to eligibility for transfemoral access or the need for transthoracic access—to TAVR or surgery. Patients had an average STS score of 5.8%, and the mean age was 82.

The main 2-year results reported in April 2016 showed that TAVR was noninferior to surgery in terms of all-cause death/disabling stroke, with an advantage for TAVR in the transfemoral subgroup that approached statistical significance.

Thourani presented the 5-year results, which showed no overall differences between TAVR and surgery for all-cause death/disabling stroke or the individual outcomes of all-cause death, CV death, any stroke, and disabling stroke.

A landmark analysis starting at 2 years, though, showed a higher risk of death/disabling stroke with TAVR both in the overall cohort (36.3% vs 29.5%; HR 1.27; 95% CI 1.06-1.53) and in the transfemoral subgroup (34.0% vs 28.4%; HR 1.23; 95% CI 1.00-1.52).

Moderate-to-severe paravalvular leak was higher in the TAVR group at all time points. And in an association that has been observed in prior analyses, all-cause death occurred in 41.1% of patients with no or trace PVR, 48.7% of those with mild leaks, and 64.8% of those with moderate-to-severe leaks.

Thourani reported that in patients treated transfemorally who had no or only trace PVR, outcomes were similar at 5 years with TAVR and surgery.

TAVR carried greater risks of procedure- or valve-related rehospitalizations and aortic valve reinterventions but a lower risk of postoperative A-fib compared with surgery. The rate of reinterventions in particular was very low for both groups, Thourani noted.

Aortic valve area remained smaller in the surgery group throughout the 5-year follow-up, although mean aortic valve gradient did not differ between groups.

Cardiac symptoms and quality of life improved in both the TAVR and surgery arms, with the benefits sustained through 5 years.

Reduce Leaks as Much as Possible

In a panel discussion following Thourani’s presentation in the Main Arena, experts focused on the need to reduce PVR as much as possible, given its prior, documented association with reduced survival.

“Leaving people with moderate-severe PVR, no matter which device you’re using, becomes exquisitely a bad idea,” Thourani said, noting that many patients in the trial had that degree of leak for years. Because it negatively affects measures like ejection fraction, PVR should be fixed with another valve and potentially paravalvular leak closure, he advised.

Michael Borger, MD, PhD (Leipzig Heart Center, Germany), introduced a note of caution when interpreting the findings regarding moderate-to-severe PVR, “because that’s a marker that the operator did not want to reintervene, which means that they were probably the more high-risk patients,” like those who are elderly or fragile. “At least, that’s reflected in the clinical practice,” he said.

Borger pointed out that mild PVR is much more common with TAVR than with surgery and that there was a nearly significant difference in 5-year mortality between those with mild leaks and no or only trace leaks (P = 0.07). He asked whether that’s something that should be considered as TAVR moves into low-risk patients.

Thourani responded that multiple studies have indicated that mild PVR does not have a major impact on outcomes. “That will be our goal—to have less than mild,” he said. “But so far mild, when we’ve looked at this data, has not made a major impact in mortality or other clinical outcomes.”

Martin Leon, MD (NewYork-Presbyterian/Columbia University Irving Medical Center, New York, NY), on the other hand, advocated for minimizing leaks as much as possible. “As much as we can with valve sizing, potentially valve choice, with postdilatation, with careful analysis, we should try to eliminate as much as we can even mild [leaks] and get as close to a surgical result as we can.”