Better TAVR Outcomes With More Experienced Operators: New York Database

Casting more light on the relationship between procedural volume and in-hospital adverse events, a new study shows that individual operator experience with transcatheter aortic valve replacement is important for achieving the best possible clinical outcomes.

In a large analysis of patients undergoing TAVR in New York State, investigators showed that those treated by physicians who performed 80 or more procedures per year had a 41% lower risk of in-hospital death, stroke, or acute MI when compared with those treated by operators who performed fewer than 24 procedures per year (OR 0.59; 95% CI 0.37-0.93).

Senior investigator Mario Gaudino, MD (Weill Cornell Medicine, New York, NY), said this is the first study assessing the relationship between individual operator experience—rather than institutional volume—and clinical outcomes after TAVR. While the study has limitations, the findings are important for the future in terms of making policy and new rules for TAVR, he told TCTMD.

At the moment, the ideal number of annual TAVR procedures for optimal patient outcomes is somewhat ambiguous, said Gaudino.

“It’s difficult, and as a scientist I wouldn’t want to guess,” he said. “After the first 100, the benefit is pretty evident. We have also seen that the first 20 are crucial. So it’s something between those two numbers, and it will be a compromise between what the data tell us and patient needs. But there is no doubt, and I think anyone can understand this: that the more an operator performs a complex procedure like TAVR, the better he or she becomes at it.”

The relationship between operator experience/hospital volume and clinical outcomes is a hot topic these days, particularly since the US Centers for Medicare & Medicaid Services (CMS) is currently mulling their national coverage determination (NCD) for TAVR. CMS is evaluating whether there is sufficient evidence to support requiring hospitals and heart teams to meet prespecified volume requirements for TAVR in order to begin and maintain their programs. A decision is expected in 2019.  

Earlier this year, the American Association for Thoracic Surgery (AATS), American College of Cardiology Foundation (ACCF), Society for Cardiovascular Angiography and Interventions (SCAI), and Society of Thoracic Surgeons (STS) issued a consensus statement on the institutional recommendations and requirements for TAVR programs. These societies recommend an institutional minimum of 50 TAVR procedures annually (or 100 within the last 2 years) and 40 surgical AVRs per year (or 80 within the last 2 years).

Ashish Pershad, MD (Banner-University Medicine Heart Institute, Phoenix, AZ), an interventional cardiologist who was not involved in the study, said there is direct relationship between individual operator volume and clinical outcomes in cardiovascular surgery. “It’s no different for this type of procedure,” he said. “Up until this point, we’ve only had institutional volumes linked with outcomes in TAVR.”

Imagine doing something one time per week? How do you expect to be great at it? Ashish Pershad

Pershad praised the analysis, noting the researchers factored in the initial learning curve by excluding early cases and still saw a clear relationship between volume and outcomes. Additionally, the inverse relationship showed no signs of plateauing. For example, when treated by operators who performed 200 procedures in the year prior, there was a 59% lower risk of postprocedure stroke and a 55% lower risk of death/MI/stroke compared with operators who performed just one TAVR. The lower event rate was even more pronounced among those did more than 300 TAVRs annually.

For Pershad, the absence of a plateau highlights the continued education and learning experiences of operators.

“It’s the finesse of how good becomes great,” said Pershad. “We learn even after 200 cases. It might not reflect in mortality, but I think it would reflect in other outcomes.” Such outcomes include vascular complications, renal failure, permanent pacemaker implantation, procedural efficiency, length of stay, and costs per case. “These are all extremely important outcomes, which this study did not have the ability to assess. If looked at, it would favor the experienced operator, in my opinion, because of the fact these nuances get fine-tuned over time,” he commented.

For individual operators, Pershad pointed out that with just 50 cases, this translates into one TAVR procedure per week. “That’s not a lot,” he said. “To get a continuous exposure to doing it well, and doing it correctly, you need to be doing at least two procedures every week, or roughly 100 cases per year. Imagine doing something one time per week? How do you expect to be great at it?”  

Worse Outcomes With Less Experience

In the present study, which was published online today ahead of print in JACC: Cardiovascular Interventions with lead investigator Arash Salemi, MD (Weill Cornell Medicine), researchers analyzed the volume-outcome relationship at the operator level using data from 8,771 TAVR procedures included in the New York Statewide Planning and Research System database. Volume from 207 operators was defined by the number of TAVR procedures performed in the year prior to the index procedure, with volume classified as low (1-23 cases), medium (24-79 cases), and high (≥ 80 cases).

Among 5,912 patients who underwent elective TAVR, 1,973 were treated by low-volume operators, 1,860 by medium-volume operators, and 2,083 by high-volume operators. Compared with elective patients treated by low-volume operators, there was a trend toward less mortality (OR 0.59; 95% CI 0.32-1.08) and less stroke (OR 0.62; 95% CI 0.30-1.30) among those treated by high-volume physicians. The reduction in death/MI/stroke among those treated by high-volume operators was statistically significant compared with low-volume physicians.

When analyzed as a continuous variable, and after adjusting for patient, hospital, and physician characteristics, there was an inverse association between physician volume and the composite endpoint of death/MI/stroke, as well as with death and stroke outcomes. The relationship between clinical outcomes and operator volume was also nonlinear, with the association most pronounced among operators who performed the fewest procedures, say investigators. Sensitivity analyses were also performed to offset the initial learning curve by excluding the initial cases of every physician. When excluding the first 20 or 30 procedures, the association between operator volume and outcomes became evident, say investigators. 

In an editorial, Brian Whisenant, MD, Ed Miner, MD, Donald Lappe, MD (all from Intermountain Medical Center, Salt Lake City, Utah), say the new data informs physicians and policymakers about TAVR practice patterns in New York State, but can likely be extrapolated to other US areas. Overall, the absolute risk of in-hospital events was low, with the composite endpoint of death/MI/stroke reported to be 3.4% (mortality was 1.9%; stroke 1.6%).

“While it seems intuitive that increased experience should be associated with improved outcomes, many low-volume centers deliver TAVR with excellent outcomes,” write the editorialists. “Quality is difficult to measure, requiring detailed data with composite endpoints and risk adjustment. Measuring TAVR quality is particularly difficult as adverse events occur with a low frequency and may not be observed in low-volume centers.”

Issue of Access

At the institutional level, studies have shown an association between procedural volumes and clinical outcomes in TAVR. For example, data from the STS/ACC TVT Registry showed that increasing site volume was associated with lower in-hospital risk-adjusted outcomes, including mortality, vascular complications, and bleeding.

Those arguing against instituting mandatory volume requirements state that doing so might limit patient access to care. To TCTMD, Gaudino noted that most clinical centers in the United States would be classified as low-volume hospitals, adding that TAVR has quickly transitioned from a “quasi-experimental” procedure to standard of care. “I think the diffusion has been a little too rapid in the beginning and our study should tell people we need to look at it,” he said. “We don’t have the solution, but we may have a problem here.”

On the other hand, clearly access can’t be restricted to just a few large-volume institutions. Mario Gaudino

Gaudino agreed that limiting patient access is a possibility with volume restrictions and one of the challenges in the field is determining the “sweet spot” between volume requirements and accessibility. “Now, TAVR is not something that is really performed on an emergent basis, so it’s conceivable that patients can be referred to a high-volume center,” he said. “On the other hand, clearly access can’t be restricted to just a few large-volume institutions.” But while access is important, Gaudino stressed that “it’s also important that everybody get the best possible procedure.”

To TCTMD, Pershad said the access issue “doesn’t pass the litmus test.” Access, he said, is limited not only by location but by multiple factors unique to the US healthcare system. “A lot of it is an educational gap,” he said. “There’s also a large financial disincentive to refer patients outside a health system. There’s financial incentives to perform surgical AVRs rather than TAVR. Those are some of the competing factors that come into play. It’s not just about how far you need to go to get a procedure.” 

Gaudino stressed their analysis has numerous limitations associated with analyzing registry data, such as confounding variables that might not be taken into account. Moreover, TAVR is evolving rapidly, with the introduction of technology that is easier to use. As a result, the present study might not reflect current practice. Pershad disagreed, however, pointing out that nearly two-thirds of TAVRs in the New York database were performed by operators in 2015 and 2016, a period that included “contemporary valves, contemporary sizing, and contemporary techniques.”