New VARC-3 Definitions Provide a Road Map for Future Aortic Valve Research

The Valve Academic Research Consortium (VARC) has updated their standardized definitions of clinical endpoints for transcatheter and surgical aortic valve clinical trials, including bioprosthetic valve dysfunction and failure, structural complications, conduction disturbances, and repeat hospitalizations, among others. 

Published this week in the European Heart Journal and the Journal of the American College of Cardiology, the VARC-3 update was adopted given the “rapid evolution of the field,” including the emergence of new complications, expanding clinical indications, and novel therapy strategies, according to the writing committee.

As aortic valve research moves into younger patients, as well as those with less advanced aortic stenosis, future trials—be they head-to-head comparisons of TAVI devices, TAVI versus surgery, or even TAVI or surgery versus medical therapy—will need standardized definitions for capturing appropriate endpoints, the authors stress.

“VARC-3 has, I would say, the most-contemporary definitions for aortic valve replacement-related research for both procedural complications, mid-term outcomes, and long-term outcomes, including outcomes related to valve function and patient-centric outcomes,” lead author Philippe Généreux, MD (Morristown Medical Center, NJ), told TCTMD. “I believe it’s a really good document that reflects the last 10 years of progress in the field. We put in a lot of time and effort to come up with clinically meaningful definitions to better classify outcomes and design relevant research.”

‘Allegations Need to Be Substantiated by Facts’

Some surgeons, however, took to social media to criticize the new definitions, pointing out that most if not all of the writing committee members participated in the major TAVI trials and have conflicts of interest with device manufacturers. As such, they argue, these physicians shouldn’t be involved in creating the definitions, as it amounts to drafting the rules of a game that will tilt the playing field in favor of TAVI devices for which they’ve received grant support or payments. Even the EHJ came under attack, with one surgeon saying the journal long ago “sold its soul” to be the “scientific ‘arm’ of industry.”

Others, including those not involved in VARC-3, say that line of attack is unfair.

Généreux, the lead author, said the creation of the VARC-3 definitions involved far more people than those listed as authors, including more than 75 cardiologists, interventional cardiologists, surgeons, heart failure specialists, imaging specialists, clinical trialists, US Food and Drug Administration representatives, and industry. It took nearly 5 years of unpaid weekends and nights to pull together the updated VARC-3 document, he added, and there was absolutely no involvement from industry when it came down to the writing process. To suggest that the goal of this diverse committee was to draft definitions to favor one procedure over the other is insulting, Généreux said.

Instead, the motivation was solely to develop a common language for clinical endpoints, so that the field of aortic valve research can move forward and so that patients will benefit. The smaller group involved in writing the final definitions, which included three cardiac surgeons, has been part of the major clinical trials of TAVI devices in the past decade and their expertise is reflected in their authorship status. 

“When you do a lot of procedures, when you do a lot of academic work, when you invest your time to become an expert in the field, you probably have the best understanding and knowledge base to know where the field is going and to design and inform consensus documents,” said Généreux. “Industry will always be around, and we need to be smart enough to use industry to design trials to improve patient outcomes. I believe the partnership that valve manufacturers have had for the last 10 years with the medical community has allowed us to push the field forward and to learn so much about aortic valve disease and what is best for the patient. For me, this is the most important thing.”

Sanjay Kaul, MD (Cedars-Sinai Medical Center, Los Angeles, CA), who wasn’t involved in VARC-3 but who has a long-standing interest in clinical trial endpoint definitions, said clinically meaningful and unbiased endpoints are critical for generating reliable and credible evidence for clinical practice. While he applauded the VARC-3 writing group for some of their choices, he criticized others, particularly the largely biomarker-driven periprocedural MI definition. Regardless, Kaul stressed that he did not like the tone of some of the comments on social media, specifically those impugning the character of some members of the writing group.

“The fact that many members with 'duality of interests' were invited to participate in the development of this document should not detract from the actual content, which should be appraised based on its faithfulness to evidence and clinical relevance,” he said.

Similarly, cardiac surgeon Vinod Thourani, MD (Piedmont Heart Institute, Atlanta, GA), a member of the Society of Thoracic Surgeons board of directors, believes investigators can separate their academic obligations from their past and/or present work with industry.

“Furthermore, all disclosures have been made public,” he said. “I’m personally involved with industry and have conflicts, and I disclose those very openly. In this day and age, we’re all conflicted because we do research and almost all the valve research trials are not being funded by governmental agencies. They’re provided by industry. By default, all 130 to 150 sites in the United States involved in valve trials are technically conflicted. As long as we have a separation of the [VARC-3] document and the academics from industry, and the interplay between them is reported and disclosures made, and industry isn’t personally involved in writing the manuscript, then I think it’s quite appropriate.”

Filippo Crea, MD, PhD (Catholic University, Rome, Italy), editor-in-chief of the EHJ, said the VARC-3 manuscript was reviewed by four reviewers and two editors and accepted after two revisions. He added that the role of reviewers and editors is to assess the scientific quality of submitted manuscripts, not the composition of the author group. “It is not surprising that the majority of the authors, but not all of them, have experience in TAVR trials,” he said in an email. VARC and VARC-2, both of which were also published in the EHJ and JACC, were well accepted by the scientific community, he pointed out, and “have stood the test of time.”

Crea said the writing committee disclosed their conflicts of interest, which are available in the paper, and that this voluntary disclosure is a matter of trust and the responsibility of every author involved. “When postpublication concerns arise, editors can validate the issues and ask the authors specific questions and, if necessary, can also question their institution, but the allegations need to be substantiated by facts,” he stressed

Valve Thrombosis Updated

The field of aortic valve clinical research has evolved substantially in the last decade, which has led to the need for this update to VARC-3, its authors say. With VARC and VARC-2, drafted in an era when TAVI was still in its relative infancy, the primary concern for clinical research related to procedural complications. Since then, especially with the recognition that TAVI is now being performed in lower-risk populations and younger patients, lifetime management is of greater importance, and that has continued to evolve. For that reason, VARC-3 reflects not only procedural and short-term outcomes but also long-term outcomes, such as structural valve deterioration and bioprosthetic valve failure.

In their update this week, the researchers provide clinical criteria for identifying different types of bioprosthetic valve dysfunction, including structural valve deterioration, which include permanent changes to the bioprosthetic valve, nonstructural valve deterioration (valve dysfunction without any abnormalities intrinsic to the valve), thrombosis, and endocarditis. They also outline three different stages of bioprosthetic valve deterioration, with stages 2 and 3 marked by moderate and severe hemodynamic changes, respectively. Finally, they provide criteria for the different stages of bioprosthetic valve failure, a useful clinical definition that will be important for long-term follow-up, said Généreux.      

Additionally, the VARC-3 writing committee provide criteria for identifying and diagnosing hypoattenuated leaflet thickening (HALT) and reduced leaflet motion (RLM), and they tighten up their definition of valve thrombosis so that it’s clinically significant. With thrombosis, the diagnosis now requires clinical sequalae (stroke, TIA, retinal occlusion, or other evidence of systemic thromboembolism) or worsening valve stenosis/regurgitation plus hemodynamic valve deterioration (stage 2 or 3) or confirmatory imaging on CT or transesophageal echocardiography (TEE). Valve thrombosis can also be diagnosed without clinical sequelae if there is hemodynamic valve deterioration (stage 3) and confirmatory imaging.

On social media, the modification of the valve thrombosis definition was met with protest from one surgeon, who claimed that the change is intended to prop up TAVI by making it harder to document the procedure’s “Achilles’ heel.” For example, in the PARTNER 3 trial of low-risk patients, the rate of valve thrombosis defined by the older VARC criteria, which did not require documentation of clinical sequelae, was significantly higher with TAVI compared with surgery at 2 years (2.6% vs 0.7%; P = 0.02).

In response, Généreux said the change was absolutely not intended to favor TAVI in clinical trials, nor to skew the results by making it harder to diagnose. 

“One of the issues with prior VARC definitions was that any increase in gradient or leaflet morphology with thrombus was called valve thrombosis, but what we’ve learned from CT and echocardiography, after multiple studies, is that you might have increased leaflet thickening or decreased leaflet mobility that isn’t thrombosis,” he said. “We kept a very granular description of the valve morphology but enhanced the definition to have a clinical component to it.”

The new definition is more clinically relevant, he insisted, and still preserves the CT-derived criteria for leaflet attenuation and thickening. He added that CT or TEE follow-up after TAVI is not standard—it’s done only within trials, or clinically if there is a suspicion of valve thrombosis.

In contrast with the online criticism, Kaul said it’s “laudable” that VARC-3 now appropriately emphasizes clinically meaningful criteria for valve thrombosis. He cited a recent substudy of GALILEO where treatment with rivaroxaban (Xarelto; Bayer/Janssen) led to a reduction in HALT and RLM on 4D-CT at 90 days. Despite the improvement in subclinical valve thrombosis, the risks of death and thromboembolic events (and bleeding) were higher in those treated with the oral anticoagulant.

“These results question the clinical importance of these events and the clinical utility for routine imaging for the detection of RLM or HALT after TAVR,” he said.

In the next couple of weeks, the ATLANTIS trial, which is testing the use of apixaban (Eliquis; Bristol-Myers Squibb) after TAVI, will be presented at the American College of Cardiology virtual meeting. This trial includes subclinical valve thrombosis as part of the primary composite endpoint, likely to increase the number of events to power the clinical trial. Kaul said that “if the results in ATLANTIS replicate the GALILEO findings, then it would support VARC-3 recommendations to not include these subclinical events in TAVR trials.”  

To TCTMD, Thourani said the clinical consequences of HALT are stroke or structural valve deterioration. In PARTNER 3, the primary endpoint was assessed at 1 year, but there is a hint that HALT may have a potential impact between years 1 and 2 given that there was an uptick in stroke in the TAVI arm during this time period. Whether those strokes are related to HALT is not fully known. He also noted that the 2-year PARTNER 3 results were published just last month so VARC-3 doesn’t capture those findings.

“I think that, at this point, the area of valve thrombosis is still in its infancy in terms of understanding the clinical short-term impact, ie, neurologic events, and long-term [impact], ie, structural valve deterioration,” said Thourani. “It’s important for us to continue to monitor how valve thrombosis will impact longer structural valve deterioration.” To truly understand the impact of HALT, longer-term follow-up is necessary, but those data are limited at this point, he added.

MIs and Hospitalizations

For future clinical trials, the VARC-3 writing group propose a number of composite endpoints for assessing device success, early patient safety, and clinical efficacy. As part of the clinical efficacy endpoint, they suggest a composite endpoint that includes all-cause mortality, stroke, rehospitalization, and quality of life. For that reason, they provide criteria for defining the different types of hospitalizations, including cardiovascular complications related to the procedure or valve, other cardiovascular hospitalizations not related to the index procedure or untreated native aortic valve, and noncardiovascular hospitalizations.

Kaul questioned the inclusion of rehospitalizations as a recommended component of the primary endpoint, noting that of all the major TAVI trials to date, only PARTNER 3 included it in the primary endpoint. The success of TAVI in that trial was driven primarily by rehospitalizations, which is arguably the least important outcome, and it was clearly biased against surgery, said Kaul.

Thourani, likewise, suggested that rehospitalization is better as a secondary endpoint than as a primary endpoint, a line of thinking in sync with most surgeons, he suspects.

“Rehospitalization is critically important in patient care. . . . However, we’d all probably agree that it’s not at the same level as death or stroke,” said Thourani.

Asked about this choice, Généreux said rehospitalization is an important indicator of patients’ quality of life and to dismiss it would be a mistake. In aortic valve-related research, it is necessary to have well-defined hospitalization endpoints, such as heart failure-related hospitalizations, given the relationship with future mortality, he argued. Moreover, the composite endpoint is intended for longer-term follow-up out beyond 1 year, and while early hospitalizations might favor TAVI over surgery, that’s not necessarily true over the long-term.

Généreux stressed, too, that VARC-3 is only intended to provide enhanced granularity for reporting repeat hospitalizations, which will be important in future strategy trials, such as those comparing aortic valve replacement against medical therapy in patients with severe aortic stenosis without symptoms. Here, including repeat hospital visits as a primary endpoint may help clinical event committees, as well as registries, speak a common language.

Finally, some surgeons expressed dismay at the definitions of MI, which were adapted from the Fourth Universal Definition, the SCAI, and ARC-2. Specifically, they questioned why periprocedural MI measured less than 48 hours after SAVR or TAVI (type 5) used the same changes in peak CK-MB or cardiac troponin for both procedures to define clinical events. In contrast, VARC-2 required ischemic signs or symptoms in addition to biomarker elevations to diagnose periprocedural MI. 

Here, Kaul said the new periprocedural MI definition is at odds with the more-stringent approach the VARC-3 writing committee took with valve thrombosis, where they emphasized the importance of clinical criteria. The type 5 periprocedural MI definition, he said, is heavily weighted towards biomarker criterion without mandatory requirements of clinical corroboration with ECG, wall motion abnormalities, or angiographic changes. Like in the EXCEL trial, another study that elicited a great deal of criticism from the surgical community, a biomarker-driven MI definition will always favor the less invasive procedure, be it TAVI or PCI, over surgery, said Kaul.

Thourani, on the other hand, said periprocedural MI is an important endpoint in any cardiovascular procedure, but that the rate is exceedingly low in patients undergoing aortic valve procedures. “It’s normally a footnote in tables, more or less, and is not really relevant as a primary endpoint. Nor should it be,” he said. For that reason, Thourani said, the criticism around the MI definition is largely overblown.

Généreux said they believe their MI definition is clinically meaningful, one that has been adapted from already validated definitions, but acknowledged that periprocedural MI is a difficult endpoint to capture. Nonetheless, like Thourani, he stressed that aortic valve research is not the same as the coronary field, where periprocedural MI is more relevant.      

“Periprocedural MI is way less frequent in aortic valve stenosis research,” he said, “so no study is ever going to include MI as part of the primary endpoint.” Where documenting periprocedural MI is more relevant is in the realm of combined procedures, such TAVI/PCI or CABG/AVR, he said. “It’s very difficult to come up with a definition that will please everyone, but at least we have a starting point. We’re not going to blindly assume that one strategy is better than another because you have more type 5 periprocedural MIs. That’s not the way it’s going to work. We’re not going to power a trial based on MI or hospitalizations.”

As for any overarching concerns about the relationships between structural cardiologists and industry, Généreux rejected those who “throw stones” and cast aspersions on the motives of the VARC-3 writing committee, insisting that the endgame has always been to advance the field of aortic valve research and clinical care. “My role is to use [industry’s] momentum, use their platform, [and] use their resources, to answer important questions that will lead to improved patient outcomes,” he said.