Hospitals With Higher SAVR Mortality Likely to Have Higher TAVR Mortality, Too

The quality of an individual hospital’s cardiac surgical care may be a strong determinant of TAVR procedure outcomes, a new analysis suggests.

“The primary takeaway of our findings is that strong cardiac surgical programs are likely one of the building blocks for high-quality TAVR programs and may actually be an important ongoing measure for hospital quality assessment,” said lead author Harun Kundi, MD (Beth Israel Deaconess Medical Center, Boston, MA), in an email.

“There’s been active debate as to the requirements that hospitals must demonstrate to qualify for [Centers for Medicare & Medicaid Services (CMS)] coverage, with much of the focus devoted to ensuring minimum volume standards,” he told TCTMD. “While sufficient volume is important, other measures of quality—in this case, historical surgical AVR mortality—may be just as important features to assessing hospital performance.”

In Kundi’s study, mortality rates at both 30 days and 1 year following initiation of a TAVR program tracked with historical risk-adjusted outcomes of SAVR at the same institution, suggesting that surgical program quality might be a new and useful metric for hospitals looking to offer TAVR services.

In an editorial accompanying the study, John D. Carroll, MD (University of Colorado School of Medicine, Aurora), notes that the findings are timely in light of a national coverage determination update from CMS, which is expected sometime next year. That decision is expected to lay out the requirements necessary for both clinicians and hospitals to qualify for TAVR reimbursement. According to Carroll, the “new and underappreciated association of TAVR outcomes” seen in Kundi and colleagues’ analysis may have important policy implications.

“This study from Kundi et al provides further insights into factors that may impact the results of the TAVR procedure,” Carroll writes. “It provides support for the heart team concept. These two important forms of aortic valve replacement, SAVR and TAVR, will continue to evolve and will require ongoing research for clinicians to understand how to optimize care for patients with valvular heart disease.”

The study was published online, December 5, 2018, in JAMA Cardiology.

Mortality Closely Aligns With Surgery Outcomes

Kundi and colleagues used the CMS Medicare Provider and Review database to identify 51,924 TAVR procedures performed in 519 hospitals between 2010 and 2015. Each hospital was classified into one of four quartiles, ranging from lowest to highest risk-adjusted SAVR mortality rate, for a total of about 130 hospitals per quartile.

At 30 days, risk-adjusted mortality for TAVR increased with rising risk-adjusted mortality for SAVR: 4.6% for quartile 1, 5.0% for quartile 2, 5.1% for quartile 3, and 5.6% for quartile 4 (P < 0.001 for all comparisons). The same pattern was observed for 1-year risk-adjusted mortality: 17.5% for quartile 1, 18.4% for quartile 2, 19.0% for quartile 3, and 19.4% for quartile 4 (P < 0.001 for all comparisons). The results were unchanged after exclusion of hospitals performing fewer than 10 TAVRs during the study period.

In multivariable analysis, and when modeled as a continuous variable, undergoing TAVR at a hospital with higher baseline SAVR mortality continued to be associated with increased risk of death at 30 days and 1 year. The risks were again borne out in subgroup analyses regardless of whether the procedure was performed via femoral or transapical access. The association between baseline hospital SAVR mortality and TAVR mortality was stronger for transapical patients than transfemoral patients at 30 days (P = 0.048 for interaction) but not at 1 year (P = 0.11 for interaction).

Volume and Beyond

Kundi and colleagues say hospitals with higher-quality surgical programs “may have had better patient selection, more highly functioning operating theaters (where TAVRs are commonly performed, particularly early after approval), and better quality cardiac surgical care units.”

While the authors acknowledge limitations placed on the study by its retrospective, database design, Carroll notes that excluding hospitals performing fewer than 10 TAVR procedures is not the same as “performing a more rigorous sensitivity analysis that excludes the initial 10, 25, or 50 TAVR cases from each hospital included in the analysis.” This is potentially important, he says, since TAVR rolled out in the United States “to hundreds of new and inexperienced sites” during the period of the study.

In an email, the study’s senior author Robert W. Yeh, MD, MSc (Beth Israel Deaconess Medical Center, Boston, MA), said Carroll’s suggestion is a “reasonable” one, but that “in general, these types of findings are rarely sensitive to small changes like this. My suspicion is that the results would be unchanged.”

Carroll also points out that in comparing SAVR outcomes with TAVR outcomes important between-group differences may be obscured, including social determinants of outcomes unrelated to skills and experience of the surgical team. Moving forward, he says, it will be “very important to sort out the potential influence of these factors on the results of outcome-focused studies such as this one to avoid unintended consequences of changing site and clinician requirements.”

To TCTMD, Yeh said he agreed that researchers need to look further into identifying characteristics beyond volume that may clarify how high-quality cardiac centers should be defined. 

“We have some ongoing work now trying to look more systematically at just these types of issues to help inform policymaker, clinicians, and patients,” he said. “It may well be that an optimal [number] of TAVR centers that enables full access to this important technology while preserving quality is either less or more than what we currently have. With the field moving so fast, that number is also a moving target but still worth trying to discover.”