‘Failure to Rescue’ Doesn’t Fully Explain Higher Mortality at Lower-Volume TAVR Centers

The observational study contributes to the ongoing debate over the optimal TAVR volumes needed for institutions and operators.

‘Failure to Rescue’ Doesn’t Fully Explain Higher Mortality at Lower-Volume TAVR Centers

Hospitals that perform more TAVR cases achieve lower in-hospital mortality, but the volume-survival link doesn’t seem to stem from their ability to handle periprocedural complications, according to new observational data.

It is well known that increased procedural experience leads to improved outcomes with TAVR, but the exact reason why hasn’t been explored much in contemporary data so far.

“There are a lot of new hospitals starting TAVR programs,” lead study author Tomo Ando, MD (Detroit Medical Center, MI), told TCTMD. “That’s one of the reasons that volume outcomes are important.” The concept of “failure to rescue”—when a patient dies after major complications—has been thoroughly explored in the surgical literature, he said, and could potentially explain the differences in TAVR outcomes by hospital volume.

However, that did not turn out to be true, at least in the current study.

For the study, published online, June 14, 2018, ahead of print in the American Journal of Cardiology, Ando and colleagues looked at almost 49,000 TAVR cases from the Nationwide Inpatient Sample (NIS) between 2011 and 2015. After dividing the cases into three groups depending on annual hospital case volume—low (≤ 30), intermediate (31-130), and high (≥ 130)—the team found that the adjusted rate of inpatient mortality decreased with hospital volume but the failure to rescue rate remained similar.

Outcomes by Hospital Volume





P Value

Inpatient Mortality





Failure to Rescue





‘Hard to Interpret’

Commenting on the study for TCTMD, Isaac George, MD (NewYork-Presbyterian/Columbia University Irving Medical Center, New York, NY), said the findings are interesting and somewhat unexpected. “High-volume centers have historically pointed to volume-mortality literature in other specialties to justify regionalization, arguing that higher-volume centers with more experience and resources have fewer complications and can also ‘rescue’ complications more effectively,” he said in an email. “However, this data . . . seems to dispute this concept.”

Although the time period encompassed the era of third-generation valves, the fact that no differences in failure to rescue were noted is “surprising,” George said. “The list of complications was comprehensive enough to expect differences to be found, yet they were not. I suspect that complication rates and failure to rescue were low enough that the study was not powered to detect small differences in failure to rescue—as you can see, there was a nonsignificant trend to lower failure to rescue in higher-volume centers, and that may become significant with higher numbers.”

There are limitations to using ICD-9 codes and the NIS database, he added, but “regardless, it does indicate that TAVR is getting safer and safer and catastrophic complications are fewer and fewer. If this holds up in larger analysis, it can generate very interesting discussion as to the need for perfusion support, cardiac surgery backup, cardiac surgery on location, and the utility of regionalization versus unabashed expansion of TAVR into the rural community.”

For Rahul Sharma, MD (Virginia Tech Carilion School of Medicine, Roanoke), who also was not involved in the study, the main message of the results is that “there is a difference between someone who has a TAVR and dies in the hospital randomly, either after the procedure or right before they leave the hospital, versus someone who has a TAVR but dies of a periprocedural complication that the doctors were unable to resolve.”

But making the leap to using these findings in an attempt to explain outcome differences by hospital volume “is really hard to interpret,” he told TCTMD. “We have felt for quite some time that high-volume TAVR centers have better mortality outcomes and based off of that conclusion, there's been a push to try to have TAVR centers of excellence where patients go to have their TAVRs done at hospitals that do a lot of them. . . . I think this paper just sort of supports that.”

Ando and colleagues suggest a couple reasons for why there may not be differences in failure to rescue rates including the fact that most TAVRs are done at teaching institutions with the resources available to handle periprocedural complications regardless of their case volume.

Attacking Low-Volume Centers?

Beyond pure fascination with the topic, the only reason to continue conducting research on TAVR outcomes and hospital volumes going forward, Sharma said, “is if you're going to try to push toward limiting where TAVR can be done. The bottom line is that not everybody can do 200, 300, or 400 TAVRs a year. The majority of hospitals are going to be somewhere in that 30-100 range.

“We're certainly not going to find that low-volume centers have better outcomes, so it's just going to be an attack on low-volume centers, essentially,” he continued. Also, “it's going to be impossible to convince health systems and hospitals that they shouldn't be doing TAVR when they feel like they have the resources to do it.”

Even if there was a rule in place by the Centers for Medicare & Medicaid Services, for example, limiting the institutions that could perform TAVR with a certain case threshold, it would be difficult to enforce. There is also the question of whether that rule should apply to institutions generally or to operators specifically, Sharma noted.

“To limit someone because they only do 30 or 40 a year is obviously politically and kind of economically very challenging,” he said. “And what if there is a hospital that does 200 TAVRs a year, but they have six operators? And maybe one or two of the operators are high-volume, but the other four guys are only doing 15 or 20 a year just to kind of stay in the game? Should we put an individual limit and say if you do less than 50 a year as an individual operator then you shouldn't be doing it?”

Moreover, it is difficult to apply a blanket rule to all high- or low-volume centers because some low-volume centers might have better outcomes than others, either because they are “cherry-picking” low-risk cases or truly possess the expertise, Sharma said.

For Ando, there is merit to both centralizing TAVR care to expert centers and supporting low-volume centers in more remote setting. Many patients who wait for their TAVR procedures end up being admitted to the hospital for heart failure, he said. “Hospitals have difficulty catching up with the numbers,” and that is one “very reasonable” motive for wanting to new, low-volume TAVR programs.

  • Ando, George, and Sharma report no relevant conflicts of interest.