Slight Uptick in Mortality With TAVI in Recent Years: TVT Registry

Following years of steady improvements in outcomes, more recent observational data suggest that mortality among TAVI patients may—though only slightly—be increasing, according to an analysis of the Society of Thoracic Surgeons/American College of Cardiology TVT Registry.

The results, which showed a year-over-year rise when death risk was adjusted for patient and procedural characteristics, were published online this week in JAMA Cardiology.

Senior author Suzanne V. Arnold, MD (Saint Luke’s Mid America Heart Institute and University of Missouri-Kansas City), told TCTMD that this new study emerged from a prior report using the TVT Registry to look at which patient, procedural, and device factors were driving improvements in TAVI outcomes from 2011 to 2018. In the midst of that work “was when we first kind of noticed this plateauing of outcomes after [around] 2018,” she said.

This time around, the researchers returned to the same database to dig deeper into what happened in the ensuing years, said Arnold. They found, surprisingly, significant increases in both 30-day and in-hospital adjusted mortality.

The reason for this trend wasn’t obvious, so the researchers performed numerous exploratory analyses to see if they were missing any potential causes—they found none. It’s possible there are some nuances that simply can’t be captured by a registry, especially when it comes to things like the impact of complications, Arnold noted. Still, “based on the data set that we had, we did the best we could.”

What’s worth remembering, though, is “the absolute differences were very small, so it's not something that I think is alarming in itself,” she stressed. Instead, “this is something that we're going to need to follow and to make sure this was perhaps just a blip or the natural variation in outcomes that can definitely happen over time, particularly when the event rates become so low. The mortality rates are quite low after TAVR.”

Dharam Kumbhani, MD (UT Southwestern Medical Center, Dallas, TX), who serves on the TVT Registry steering committee, noted that the study’s authors have “been, I think, very good and careful about sort of making sure people don't overinterpret these results to mean that there's a lot of harm happening, because the absolute differences are very, very small.” The registry is large, he pointed out, so even quite small differences are detectible.

The results, though not alarming, are important, said Kumbhani.

For one, the study period coincides nicely with the US Food and Drug Administration’s approval of TAVI in low-risk patients, which came in August 2019. “You would think that with lower-risk patient profile, that it would actually be getting better,” he said. Counterintuitively, it seems the opposite might be the case.

The absolute differences were very small, so it's not something that I think is alarming in itself. Suzanne V. Arnold

John Carroll, MD (University of Colorado School of Medicine, Anschutz Medical Campus, Aurora), past chair of the TVT Registry, also characterized the change in 30-day risk-adjusted mortality as small. Emphasizing the report “has no impact on clinical care,” he pointed out that data from 2023 would have made any findings more robust and that about 11,000 patients (about 4% of the total TVT Registry cohort) were excluded from analysis because of missing 30-day vital status.

“While the importance of the Registry and the need to continue collecting data are clear, I have concerns about prematurely reporting short-term trend data that may represent statistical gyrations rather than clinically meaningful and actionable findings,” he told TCTMD in an email. The only action at this point, said Carroll, is to wait for more data.

Study Spans 2019 to Early 2022

Arnold and colleagues culled data from 786 hospitals participating in the TVT Registry between January 2019 and March 2022. In the cohort of 210,495 TAVI patients, the median age was 79 years, 43.4% of patients were female, and the median STS PROM score was 3.3%.

There was no difference in 30-day mortality, the study’s primary endpoint, between the first quarter of 2019 and first quarter of 2022 (2.4% vs 2.2%; unadjusted OR 0.98; 95% CI 0.94-1.01). Nor were there any changes in 30-day composite adverse events (death, stroke, stage 3 acute kidney injury, moderate or severe paravalvular regurgitation, and major, life-threatening, or disabling bleeding) or in-hospital mortality.

But after adjusting for patient characteristics, 30-day mortality was significantly higher in the latter versus earlier time period (OR 1.05 per year; 95% CI 1.02-1.08). Adjustment for procedural characteristics showed an even greater increase in risk (OR 1.09 per year; 95% CI 1.04-1.13).

In-hospital mortality was significantly increased with adjustment, as well, though 30-day adverse events were in fact reduced when accounting for patient-related factors (OR 0.95 per year; 95% CI 0.93-0.97) and not significantly different when adjusting for procedure-related factors.

Analyses that excluded numerous groups—newer and low-volume sites, as well as patients who were low risk, had a bicuspid valve, died in hospital, or experienced a major vascular complication—did not erase the statistical significance of the associations for mortality.

If it keeps going up, year over year, then we definitely need to do a deeper dive. Dharam Kumbhani

In short, “no site-level, patient-related, or process-related factors were identified that could explain these findings,” the researchers report.

The COVID-19 pandemic is “an attractive explanation,” Arnold and colleagues admit, but they point out that the increase in adjusted mortality preceded that era.

Kumbhani wasn’t ready to rule out that theory. “It's a little muddy because there is a spike in 2020 during COVID. [In that era] we were still doing TAVRs, and there was a lot written at that time about patients getting lost or not being able to make it to the finish line, et cetera. But, we were doing the more urgent ones.” he observed. “I imagine that a small blip—during the 2020, early 2021 horizon—could be part of it.”

One possibility not yet explored is how changes in the healthcare workforce, such as the nursing shortage that grew acute in the pandemic, might have affected outcomes. “Perhaps it's a staffing thing. Because you would think that [for] periprocedural care and those aspects, we're learning over time. Those things should get better,” Arnold observed.

Big-picture drivers like this, though, as well as things like social determinants of health, are very hard to pin down, she added. This is especially true when dealing with an outcome like death that happens so rarely, said Arnold, “because you have to look at large numbers of patients to be able to really see these trends and outcomes, but then to drill down on these very detailed things that are difficult to collect. I think that becomes a challenge for sure.”

With these new data bringing awareness to a possible mortality signal, Arnold said TVT Registry leadership will keep an eye out going forward. It’s good news that there’s interest in doing so, she said. “We tend to want to highlight our wins and the things that we're doing really, really well. And when you're highlighting something that is a little bit not as good, that's to be commended.”

Kumbhani agreed that the findings speak to the value of the TVT Registry, noting, “These data would be very hard to get if we didn't have uniform capture of all these procedures.” He, too, called for ongoing follow-up, so that any persisting trends would be detected. “If it keeps going up, year over year, then we definitely need to do a deeper dive and understand what's driving this mortality,” Kumbhani concluded.