TAVR Valve Face-off: Two French Observational Studies Suggest Inherent Differences

Randomized trials will be needed to confirm these findings, physicians said at AHA 2019.

TAVR Valve Face-off: Two French Observational Studies Suggest Inherent Differences

PHILADELPHIA, PA—Two French, propensity-matched analyses are suggesting that not all TAVR valves are created equal, with balloon-expandable devices associated with decreased mortality and rates of paravalvular regurgitation (PVR) compared with self-expanding devices.

The two commercially available transcatheter heart valves currently used in the United States are the balloon-expandable Sapien 3 (Edwards Lifesciences) and CoreValve Evolut PRO (Medtronic), and while both—or their previous iterations—have proven superior to surgical intervention in a variety of patient groups, they have never been compared in large, randomized head-to-head trial. Operators generally chose which device to use based on personal or institutional preference, although certain anatomical and hemodynamic factors might also influence decision-making.

Eric Van Belle, MD, PhD (Institut Coeur Poumon, CHU Lille, France), who presented an analysis from the FRANCE-TAVI registry at the American Heart Association (AHA) 2019 Scientific Sessions, said his study “would suggest that the two most widely used transcatheter heart valve designs may not achieve the same clinical outcomes. Overall, the present study strongly supports conducting a randomized trial in order to compare head-to-head the most recent iterations of self-expanding and balloon-expanding transcatheter heart valves on all-cause mortality.”

Likewise, Dharam Kumbhani, MD (UT Southwestern Medical Center, Dallas), who discussed the FRANCE-TAVI data during the session, said the study “points out, if nothing else, that perhaps there is an inherent difference between the TAVR valves. It may be incorrect to assume that there is a class effect.”

For Howard Herrmann, MD (University of Pennsylvania, Philadelphia), commenting for TCTMD, said the data hint at a “a potential difference in mortality, but it's relatively small and it's not in a randomized trial. So I don't know that at this point this kind of data will switch people from self-expanding to balloon expandable devices alone, but one could consider it another factor.”

Two French Data Sets

For the FRANCE-TAVI study, which was simultaneously published in Circulation, Van Belle and colleagues included 12,141 French patients with severe aortic stenosis undergoing TAVR with either a balloon-expandable Edwards (either XT or Sapien 3; n = 8,038) or self-expanding Medtronic valve (CoreValve; n = 4,103) between 2013 and 2015.

In a propensity-matched analysis of 3,910 pairs, the self-expanding device had a higher rate of the primary composite outcome (defined as at least moderate PVR and/or in-hospital mortality) and its individual components compared with those who received balloon-expandable devices.

Outcomes Based on Valve Choice




RR (95% CI)

Primary Composite



1.68 (1.46-1.91)

PVR ≥ Moderate



1.90 (1.63-2.22)

In-Hospital Mortality



1.34 (1.07-1.66)

Mortality at 2 years was also higher in the self-expanding valve group (29.8% vs 26.6%; HR 1.17; 95% CI 1.06-1.29), which was driven by cardiovascular mortality (23.3% vs 20.9%; P = 0.001). Notably, there was no difference in death rates between the devices between 3 and 6 months or after 6 months, but only from 0 to 3 months (P = 0.001).

The pacemaker rate was also higher in the self- vs balloon-expanding valve cohort (22.3% vs 11.0%; P < 0.0001) as was the rate for MI (0.4% vs 0.2%; P = 0.02) and the need for a second transcatheter heart valve (3.7% vs 1.0%; P < 0.0001). The mean gradient was lower for those who received self-expanding valves (7 vs 10; P < 0.001).

Findings were consistent across prespecified subgroups, but the difference in the primary composite endpoint was stronger for those who received transfemoral TAVR (RR 1.82; 95% CI 1.56-2.13) and those who were treated after January 2015 (RR 2.23; 95% CI 1.71-2.91).

On multivariate analysis, both PVR severity and valve design were independent predictors of all-cause mortality.

In the second study, published online ahead of print in Circulation but not part of AHA 2019, Pierre Deharo, MD, PhD (CHU Timone, Marseille, France), and colleagues included administrative hospital data from more than 31,000 French TAVR patients who received either the Sapien 3 (65.5%) or Evolut R (34.4%) between 2014 and 2018.

Among 10,459 propensity-matched pairs, there was a lower yearly incidence of all-cause death associated with balloon- vs self-expandable TAVR over a mean 358-day follow-up (RR 0.88; corrected P = 0.005). Cardiovascular death (RR 0.82; corrected P = 0.002), rehospitalization for heart failure (RR 0.84;  corrected P < 0.0001) and pacemaker implantation rates (RR 0.72; corrected P < 0.0001) were also lower for those who received Sapien 3 devices.

‘Different Than the Coronary Stent Business’

In an editorial accompanying both studies, Mohamed Abdel-Wahab, MD, and Holger Thiele, MD (both from University of Leipzig, Germany), write that they “represent the current standard of such registry-based analyses using the best available methods to adjust for baseline differences in study groups by propensity-score matching. Performing head-to-head device comparisons can sometimes resemble entering a minefield, so the authors should be commended for their courage as well.”

That said, they point to the limitations associated with the studies’ observational designs, writing that “the findings should therefore be considered thought-provoking but by no means definite.”

“Since the indication of use of balloon-expandable valve or self-expandable valve systems strongly depends on both the operators’ familiarity with the device and anatomical and clinical suitability of the patient, findings derived from observational registries are always hampered by bias and confounding, which cannot be completely corrected for by even sophisticated matching,” Abdel-Wahab and Thiele continue. “A causal relationship between device type and mortality is therefore lacking, and the probability of unmeasured patient-related confounders as the likely explanation of excess mortality in the self-expanding valve group remains high.”

The editorialists note that explaining the mortality difference observed in both studies in favor of balloon-expandable valves “remains quite complex.” Paravalvular regurgitation has previously been shown to cause higher mortality, “though it remains unclear if paravalvular regurgitation is directly acting as a cause or partly acting as a marker of a higher risk subgroup,” they write.

As TAVR expands to low-risk patients, “we are starting to consider valve durability, coronary access, and the possibility of repeat interventions in our device selection algorithms,” Abdel-Wahab and Thiele point out. “While it seems obvious that a significant proportion of patients can probably be safely treated with either a balloon-expandable or a self-expanding valve, we cannot advocate that both device types are equal and can always be used interchangeably. This is surely different than the coronary stent business.”

Calls for a Head-to-Head Trial

“This is a call for differential treatment, which can only be achieved with a certain volume and level of experience with the inherently different and continuously evolving TAVR device platforms,” Abdel-Wahab and Thiele conclude. “This is also a call for more randomized clinical evidence, particularly in anatomical situations that are believed to be equally treatable with both platforms, and with clinical endpoints relevant to the long-term expectations of a lower-risk population.”

As practitioners,” Kumbhani said, “it is important for us to match patient and anatomical characteristics to the right valve and the field urgently needs more well powered head-to-head comparison trials looking at hemodynamic parameters including device successes, hard endpoints, and also cost.”

But designing and funding such a study will be hard to do, Herrmann said. “If the companies don't participate in paying for it, and neither company has shown much interest in doing that comparative study, it then either has to be funded by an external organization . . . or it has to be a cooperative multicenter trial where people agree to randomize patients that can be randomized.”

Also, he said, “it would be nice if these devices were similar enough to be interchangeable and it would foster competition on price. But at the moment, there are still sufficient differences—some real, some perceived—between devices that make it difficult for price competition.”

For now, he surmised that “for the majority of the patients, one can get away with either valve successfully and get an excellent outcome.”

  • The FRANCE-TAVI registry was established by French Society of Cardiology and French Working Group of Interventional Cardiology with the participation of the French Society of Thoracic and Cardiovascular Surgery and with support from Edwards Lifesciences and Medtronic.
  • Van Belle, Deharo, Abdel-Wahab, and Thiele report no relevant conflicts of interest.
  • Herrmann reports receiving research funding and consultant fees from Edwards Lifesciences and Medtronic.