Sapien 3 in Intermediate Risk: Death or Stroke Similar to SAVR at 5 Years

A new propensity-matched analysis looking at 5-year clinical outcomes in intermediate-risk patients with severe symptomatic aortic stenosis shows similar rates of all-cause mortality and stroke among patients treated with TAVR when compared with patients treated with surgery.

Overall, the rate of disabling stroke favored TAVR patients, although the rate of nondisabling stroke was higher among those who received the Sapien 3 (Edwards Lifesciences) transcatheter heart valve.

Presenting the analysis during the late-breaking clinical science session at the TVT Connect 2020 virtual meeting, Susheel Kodali, MD (NewYork-Presbyterian/Columbia University Irving Medical Center, NY), said their data also showed the transcatheter valve was durable and associated with low rates of paravalvular regurgitation. On the whole, Kodali said the results are “encouraging and continue to support TAVR as an alternative to surgery.”

Stephan Windecker, MD (Bern University Hospital, Switzerland), who chaired the session, pointed out that 5-year data from the randomized PARTNER 2A trial presented by Vinod Thourani, MD (Piedmont Heart Institute, Atlanta, GA), in 2019 also showed similar rates of death or disabling stroke between the two treatments. A few weeks later at the 2019 European Association for Cardio-Thoracic Surgery meeting, however, Thourani raised a few eyebrows when he presented 5-year survival rates in PARTNER 2A as a stand-alone endpoint—an analysis asked for by meeting organizers—showing that mortality was lower for TAVR early on but later favored surgery. Thourani argued at the time that the overall 5-year overall results showing parity stand, suggesting that the separation of the mortality curves might be attributable to higher rates of paravalvular regurgitation seen with TAVR.

This therapy is very durable and provides an outstanding treatment option. It makes you feel better that we have good hemodynamics and it’s not leading to any ill effects. Samir Kapadia

This new propensity-matched analysis, according to Kodali and Windecker, provides further long-term data for the more-contemporary Sapien 3 valve; Sapien XT, the valve used in PARTNER 2A, is no longer commercially available. Samir Kapadia, MD (Cleveland Clinic, OH), one of the discussants, also echoed that assessment. “This therapy is very durable and provides an outstanding treatment option,” said Kapadia. “It makes you feel better that we have good hemodynamics and it’s not leading to any ill effects.”  

No Differences in Clinical Outcomes

In the propensity-matched analysis, researchers used data from the surgical arm of the randomized PARTNER 2A trial and the PARTNER 2 Sapien 3 intermediate-risk (S3i) registry. The matched cohort included 1,566 patients, of whom 783 each were treated with TAVR and surgery, respectively. At 5 years, 353 patients in the TAVR arm were alive compared with 383 patients in the surgical arm, with complete 5-year follow-up available for 80.1% of TAVR patients and 86.8% of the surgical group.

The rate of all-cause mortality was 39.1% in the TAVR arm compared with 41.3% in the surgical group (P = 0.20). Similarly, there was no significant difference in the rate of any stroke (13.4% vs 11.4%). Nor was there any difference in the combined endpoint of death or disabling stroke between the two treatment arms (40.1% vs 42.7%). For the TAVR-treated patients, the 5-year rate of disabling stroke was 5.8% versus 7.9% in the surgical group (P = 0.05), an advantage with TAVR that was observed early and was maintained to 5 years. The rate of nondisabling stroke was 6.4% and 3.5%, respectively (P = 0.04). Here, said Kodali, the Kaplan-Meier curves showed a separation favoring surgery at 1 year, although he noted that nondisabling stroke included those with a modified Rankin Scale score of 1 or less, “so a very minor stroke.”

It reminds me to pay much more attention during long-term follow-up to arrhythmia in the patients undergoing TAVR. The nondisabling stroke might be related to unknown atrial fibrillation. Julinda Mehilli

The need for a new permanent pacemaker was significantly higher with TAVR (16.2% vs 11.7%), but rates of endocarditis, aortic valve reintervention, and valve thrombosis were low and similar between both treatments. In terms of valve performance, the mean aortic valve gradients were similar between TAVR and surgery. Rates of moderate-to-severe paravalvular regurgitation were higher with TAVR at 30 days and 1 year, but there was no observed difference compared with surgery at 5 years. Mild paravalvular leak (PVL) was more common with TAVR, said Kodali. There was no difference in the rate of hemodynamic valve deterioration or valve failure between the two treatments.

When the 5-year outcomes from the PARTNER 2A randomized trial were presented at TCT 2019, attention also focused to the rate of moderate/severe PVL, which was 6.5% with Sapien XT. In the propensity-matched analysis, the rate of moderate/severe PVL was just 0.7% with Sapien 3.

During the discussion following yesterday’s TVT presentation, Kapadia said he was struck by the similar rates of moderate-to-severe PVL with Sapien 3 and surgery at 5 years. Yet he also expressed some caution about the rates of mild PVL with Sapien 3: 28.7% in the TAVR arm compared with 6.8% among those treated with surgery. While moderate-to-severe paravalvular regurgitation has been shown to have a detrimental impact on clinical outcomes, he is uncertain that mild PVL is a “benign” finding, particularly in younger patients.

In the discussion, Kodali also highlighted another landmark analysis from PARTNER 2 showing a separation between 2 and 5 years in terms of the risk of all-cause mortality among patients with mild PVL when compared with those with no/trace paravalvular regurgitation. That, however, wasn’t observed in the present propensity-matched analysis.

“The concern is whether mild PVL would cause a longer-term impact,” said Kodali. “When we looked at this study, there was no difference [in outcomes] between mild and none/trace [PVL]. We didn’t see that separation. Maybe if you [follow patients] out 10 years, you’ll see it, but at least at 5 years, that separation we were starting to see with Sapien XT, we didn’t see.” He noted that CT assessments were mandatory in the PARTNER S3i registry but weren’t required in PARTNER 2A. “Maybe we did the procedure better—I can’t say it was the valve—or it’s ascertainment [of PVL] by different core labs. All of that plays a role. In the end, I think you have to optimize the result in the individual patient and get as little PVL as possible. Device iteration with [Sapien 3] Ultra and other valves will make leak less of an issue,” Kodali concluded.

Making Sense of Nondisabling Stroke

Regarding the higher risk of nondisabling stroke with TAVR, which increased over time starting at roughly 2 years, Kapadia that signal has never been observed before. Relatively few events, as well as the lack of systematic neurologic assessment, might explain the finding, he suggested. Unmeasured confounding variables might also be a factor. Gregg Stone, MD (Icahn School of Medicine at Mount Sinai, New York, NY), another discussant, cautioned against making too much of the higher rate of nondisabling stroke given that it was not observed in a randomized trial  

For her take, Julinda Mehilli, MD (Ludwig-Maximilians University, Munich, Germany), another of the scheduled discussants, highlighted the difference in the risk of nondisabling stroke with TAVR, which she suggested might be related to medical treatment during follow-up.

“We know that after SAVR we have more atrial fibrillation than after TAVR, and maybe during these 5 years of follow-up, patients undergoing surgery will be treated with oral anticoagulation compared with those undergoing TAVR,” she said. “It reminds me to pay much more attention during long-term follow-up to arrhythmia in the patients undergoing TAVR. The nondisabling stroke might be related to unknown atrial fibrillation.”

Kodali said the researchers looked at differences in anticoagulation use in the study but didn’t see any link to the risk of nondisabling stroke among treated patients. With respect to the repercussions of valve thrombosis—which was observed in six patients in the TAVR arm and one patient treated with surgery—none of these patients had a stroke, said Kodali.

Note: Kodali is a faculty member of the Cardiovascular Research Foundation, the publisher of TCTMD.