Self-Expanding vs Balloon-Expandable TAVR? Some Insights From CHOICE

PARIS, France—High-risk patients with aortic stenosis undergoing transcatheter aortic valve replacement fare just as well when treated with either the balloon-expandable or self-expanding transcatheter valves, but there do appear to be some nuanced differences in valve performance over time, according to the long-term results of the CHOICE study.

At 5 years, study investigators observed comparable rates of mortality, stroke, and repeat hospitalization for heart failure among patients treated with early-generation balloon-expandable (Sapien XT; Edwards Lifesciences) or self-expanding (CoreValve; Medtronic) valves.  

There was a difference in terms of how well the valves held up, however, with structural valve deterioration (SVD) documented in six patients treated with Sapien XT, including four patients with moderate SVD and two with severe SVD. In contrast, valve deterioration was not observed in any patient treated with CoreValve.

“Actually, we were expecting when we did longer follow-up to 5 years as we planned the initial differences in paravalvular leak favoring the balloon-expandable device would somehow translate into improved outcomes,” lead investigator Mohamed Abdel-Wahab, MD (Heart Center Leipzig, Germany), told TCTMD. “We were actually surprised to see the opposite at 5 years. Obviously, leaks were an issue with the old CoreValve—we know this, and this is why the device has improved and the reason why we have other devices that are repositionable—but there are obviously other downsides, at least hypothetical, for the balloon-expandable device that offset the better sealing.”

Alain Cribier, MD (Hôpital Charles Nicolle, Rouen, France), who moderated the late-breaking clinical trial session at EuroPCR 2019 where the results were presented, said the findings are reassuring given the lack of substantive differences between the two valves. “Of course, we’re talking about the first-generation devices,” he said. Regarding the difference in the rate of SVD between CoreValve and Sapien XT, he cautioned against making too much of the finding, particularly since nearly 50% of patients had died at 5 years. With a smaller number of patients in each treatment group, just one clinical event in either arm can swing the results, said Cribier.

Better Sealing, but Worse Hemodynamics

The CHOICE study was an investigator-initiated trial in 241 high-risk patients with severe aortic stenosis. The earliest results were published in JAMA in 2014, and at that time investigators showed use of the balloon-expandable Sapien XT was associated with greater device success. This finding was attributed to better sealing and less paravalvular leak compared with CoreValve, but these early hemodynamic differences did not meaningfully impact hard clinical endpoints at 1 year.

By 5 years, the rate of all-cause mortality among patients treated with Sapien XT and CoreValve was 53.4% and 47.6%, respectively (P = 0.38). Roughly one in six patients had a stroke and approximately one-quarter of study participants were rehospitalized for heart failure, but there were no significant differences in rates between the balloon-expandable and self-expanding valve study arms. The need for a new permanent pacemaker was significantly higher with the self-expanding CoreValve (40.4% vs 25.4%; P = 0.01).

Forward flow hemodynamics favored CoreValve throughout the study, with the mean pressure gradient significantly higher and mean effective orifice area significantly smaller with the balloon-expandable Sapien XT device. Differences in paravalvular aortic regurgitation—which was significantly lower with Sapien XT at 30 days and 12 months—were not evident at 5 years.  

The assessment of bioprosthetic valve dysfunction was a post hoc analysis. Overall, valve dysfunction, which includes SVD, non-SVD, valve thrombosis, and endocarditis, was not significantly different between the two valves. In addition to the six patients with SVD, there was a slight trend toward a greater incidence of clinically-diagnosed valve thrombosis with Sapien XT (7.3% vs 0.8%; P = 0.06). Rates of valve failure, a composite that included valve-related death, valve dysfunction requiring reintervention, and severe hemodynamic structural valve deterioration, did not differ between the two valves.  

The higher gradients and smaller orifice area with the balloon-expandable device might be contributing to the higher risk of SVD with Sapien XT, said Abdel-Wahab. These “subtle” differences in valve thrombosis and SVD need further confirmation given the small numbers, however. The seven cases of valve thrombosis—six with Sapien XT and one with CoreValve—occurred in patients not taking oral anticoagulation.

What About Next-Gen Devices?

Although the first-generation devices have been replaced by more sophisticated products, Abdel-Wahab said the CHOICE findings would likely hold up if the next-generation products were studied. The reason, he said, is that the hemodynamic performance of Evolut R is comparable with CoreValve but the Sapien 3 device doesn’t quite perform as well as its predecessor, which he attributes to the change in sizing practices.

“With Sapien 3, it has additional sealing characteristics so you can use smaller valves in the same anatomy and get the same sealing or even better,” said Abdel-Wahab. “And because you’re afraid of annular rupture with balloon-expandable devices, that’s a good thing. Use a smaller valve, no annular rupture, and the same sealing, but it’s at the expense of higher gradients and smaller orifice areas.”

For this reason, he suspects that differences in flow hemodynamics would likely be more pronounced when comparing Evolut R with Sapien 3.

From a clinical perspective, Abdel-Wahab said that if one valve is more durable than another, physicians are inclined to go with that device, but there are other equally important considerations aside from durability. For example, redo procedures are easier and access to the coronary arteries is better with the shorter balloon-expandable valve. In the end, physicians will personalize the decision based on the individual patient before them, he said.

For example, in an elderly patient with a small annulus, in whom the goal is to optimize hemodynamics and durability, CoreValve might be a better option. “You want to give her a valve that’ll last 10 or 15 years, or as long as it could, [so] I’d give her a self-expanding device,” he said. “If I’m doing a young man with coronary artery disease and he has a large anatomy, I’d probably give him a balloon-expandable device. I’d be able to access the coronaries at any time, and even if it degenerates a little bit sooner than a self-expanding device, say 1 or 2 years, it would be easier to treat.”

Chairing the session with Cribier, Juan Granada, MD (Cardiovascular Research Foundation, New York, NY), spoke generally about the difficulties in comparing the performance of different valves. Device failures, he said, are multifactorial, affected by patient selection and physician skill/comfort with specific devices, among other things. So, he said, CHOICE, with its high mortality rate and small sample size, should be interpreted cautiously.

“I think we should be grateful to see these data, and we should learn from the data, but it’s still, I think, very early to talk about durability and differences between devices,” Granada said.