How Rare Is “Is Rarely Appropriate” and Is It Important?

See Article by Inohara et al

Transcatheter aortic valve replacement (TAVR) has cemented itself as standard-of-care in the treatment of severe aortic stenosis in patients deemed prohibitive or high surgical risk for surgical valve replacement, as well as a reasonable option in patients at intermediate and low surgical risk. Given the higher associated costs of TAVR compared with surgical valve replacement, as well as the surge in utilization and ongoing dissemination of this technology, there has been a concentrated effort on refining and reinforcing its clinical indications. In 2017, the first Appropriate Use Criteria (AUC) for the treatment of severe aortic stenosis was developed by the American College of Cardiology Foundation in conjunction with several other cardiovascular societies and relevant stakeholders.¹ As with all AUC documents, its intent was to impart guidance in the management of severe aortic stenosis by rating the appropriateness of each available treatment choice for commonly encountered clinical scenarios. The AUC document cites the following definition of “Appropriateness”:

For each clinical scenario, a panel of multidisciplinary experts provided scores for the potential treatment options, which included balloon aortic valvuloplasty, surgical valve replacement, TAVR, and finally, no intervention. Following the appraisal, these treatment options were ultimately classified into 3 mutually exclusive categories labeled as “Appropriate,” “May Be Appropriate,” and “Rarely Appropriate.” It is important to recognize the use of this particular nomenclature is fairly recent, as it was introduced by an American College of Cardiology Foundation AUC Methodology Update published in 2013.² The current terminology is an evolution from the three initial categories of “Appropriate,” “Uncertain,” and “Inappropriate,” a change that was prompted by stakeholder concerns that the original categories were perceived as discrete and inflexible, rather than a continuum of treatment options. For instance, a treatment option deemed “Appropriate” should not be construed as mandatory. Instead, this term denotes that for a particular patient population, the benefits of a procedure generally outweigh the risks; however, other factors such as clinician judgment and patient preferences should still be taken into consideration during treatment decision-making. Likewise, “Rarely appropriate” does not imply that a therapy should never be utilized. Rather, treatment options that fall into this category may, in fact, be reasonable under unique circumstances or characteristics not fully captured by AUC clinical scenarios and should be outlined in detailed documentation.

Since the publication of AUC for treatment of severe aortic stenosis in 2017, the appropriateness of TAVR in a real-world setting has not yet been described. Addressing this gap in knowledge, Inohara et al³ examined a registry containing detailed clinical information regarding 2036 TAVRs performed in 14 Japanese hospitals between October 2013 and May 2017 and assigned an appropriateness rating to each case. Despite the granularity of detail included in this database, certain variables were not fully captured to allow for retrospective application of AUC in a typical fashion. As a result, the appropriateness rating for each case was evaluated under the best- and worst-case scenario. The best-case scenario was determined by the assumption that the missing variables favored a higher appropriateness rating, and likewise, for the worst-case scenario, the missing variables were assumed to favor a lower appropriateness rating. Ultimately, ≈70% of TAVRs within the registry were deemed appropriate, whereas rarely appropriate TAVRs occurred at a rate of 4.9% and 6.8% in the best- and worst-case scenario, respectively. These findings raise the question—what proportion of rarely appropriate TAVRs is considered reasonable?

Although the release of AUC for the treatment of severe aortic stenosis is relatively recent, AUC for cardiovascular imaging modalities has a more storied history. The first AUC documents in imaging were published on single-photon emission computed tomography myocardial perfusion in 2005 and have since undergone multiple iterations. This has allowed ample time to assess the rate of appropriate use in clinical practice, as well as follow trends in use after the release of each AUC document. As such, the literature regarding appropriate use in cardiac imaging offers insight into what could be the reference range for the proportion of rarely appropriate procedures. In 2015, Fonseca et al⁴ published a systematic review of various cardiovascular imaging studies over the preceding 10 years, which had published rates of appropriate and rarely appropriate testing. For rarely appropriate testing, rates were variable among the imaging modalities included in the study, which comprised of echocardiography, nuclear stress testing, and computed tomography angiography. Publications corresponding to most recent AUC edition of transthoracic echocardiography reported a 9% rate of rarely appropriate testing. In contrast, the rate of rarely appropriate transesophageal echocardiograms was lower at 2%. The rate of situations deemed rarely appropriate for stress echocardiography was the highest among all modalities at 27%, whereas single-photon emission computed tomography myocardial perfusion and computed tomography angiography were around 20%.

When considered in the context of the historical real-world rates of inappropriate cardiac imaging testing, the percentage of rarely appropriate TAVRs described by Inohara et al³ seem quite low. Given the complexity and nuances of decision-making involved in the treatment of aortic stenosis, generating a low volume of rarely appropriate TAVR cases may be justifiable. A rate approaching zero at any one center may actually be undesirable, as it would likely lead to the denial of patients who would still gain a net benefit, albeit to a lesser degree, from this procedure.

The current study also highlights the institutional variation in the proportion of rarely appropriate TAVR; in the best-case scenario, there was a median rate of 4.9% (95% CI, 3.8–6.6; P<0.001) and in the worst-case scenario, there was a median rate of 6.5% (95% CI, 5.6–8.6; P<0.001). Across this relatively narrow spectrum, does this variation across hospitals matter? Somewhat surprisingly, the authors report that the institutions with higher rates of rarely appropriate TAVRs were associated with higher 1-year mortality across all cases. Why is this?

The current study was not designed to identify the drivers that underlie this difference in institutional outcomes, but we can hypothesize as to some possibilities. It may be that in institutions with a higher rate of rarely appropriate cases, there is a less functional heart team, or perhaps a culture of more aggressive care. Although it is important to try to identify the underlying causes, regardless of the cause, the association of a higher rate of rarely appropriate cases with worse outcomes reinforces that the AUC is a reasonable metric by which to evaluate decision-making in AS. Indeed, we are hopeful that there will be more reports of AUC ratings in aortic stenosis care across other jurisdictions, especially as the indications for TAVR extend to lower-risk patients. We think that such work will provide introspection for heart teams on their performance and be a catalyst for initiating quality improvement activities, a mandatory component of a high functioning TAVR service.

Footnotes