Surgical Societies Question VARC-3 Endpoints for Valve Disease Trials

Several surgical associations have issued a new statement in response to the updated Valve Academic Research Consortium (VARC) definitions for aortic valve clinical studies, taking issue with a number of endpoints, among them rehospitalization, bleeding, and myocardial infarction.

Published in the Annals of Thoracic Surgery, Journal of Thoracic and Cardiovascular Surgery, and European Journal of Cardio-Thoracic Surgery, the joint position statement includes physicians representing the European Association for Cardio-Thoracic Surgery (EACTS), Latin American Association of Cardiac and Endovascular Surgery (LACES), Asian Society for Cardiovascular & Thoracic Surgery (ASCVTS), American Association for Thoracic Surgery (AATS), and Society of Thoracic Surgeons (STS). 

Patrick O. Myers, MD (Lausanne University Hospital, Switzerland), secretary general elect of EACTS, praised the academic research consortiums, including VARC, for pushing clinical research forward. The purpose of the current statement, he stressed, was to respond to the strengths and limitations of the VARC-3 definitions given the impact they will have on future clinical trials.  

“After reading the document we had a few issues with [VARC-3], and while there was some discussion on social media, we felt that since this is the basis upon which trials are designed—these are the endpoints that should be used in clinical trials if we’re following proper trial design—the endpoints are extremely important,” Myers told TCTMD. “Any small changes in these outcome measures can have a huge impact on the interpretation of a trial result.”

We depend on each other in our daily work. It’s a good working relationship, but like a good marriage, it’s not easy every day. Patrick O. Myers

The VARC-3 endpoints, he added, are not simply used by US researchers, but by all physicians involved in aortic valve research. For that reason, Myers enlisted other surgical groups around the world to develop a response to the new VARC-3 definitions. “This isn’t just individual surgeons,” said Myers. “As far as I know, this is the first time that we’ve come together, all the surgical associations worldwide, to raise a common voice.” 

Philippe Généreux, MD (Morristown Medical Center, NJ), lead author of the VARC-3 writing committee, welcomed the opinions of the surgical groups, calling the discussion healthy for aortic valve clinical research. He noted that criticism of the different endpoints—rehospitalization, thrombus, bleeding, MI, and left bundle branch block—mostly revolves around whether certain endpoints should be used in future trials of TAVI versus surgery.

“For me, it’s quite reassuring that the surgical societies wrote the paper, with fairly minor comments, and that they’re questioning the application of the endpoints,” Généreux told TCTMD. The VARC-3 definitions, he said, are intended only to create a “common language” so that future researchers can better capture and characterize clinical phenomenon. “VARC-3 is not by any means a clinical guideline and is not intended to dictate how you should conduct research or studies. It only defines a framework of definitions,” he added.

More Formal Response as Opposed to Twitter

In April 2021, as reported by TCTMD, VARC updated their standardized definitions of clinical endpoints for transcatheter and surgical aortic valve clinical trials. At the time VARC-3 was published in the European Heart Journal, some surgeons, including a co-author of the surgical position statement, disparaged the definitions by calling it an “extreme example of manipulation by big trialists” intended to benefit TAVI. The definitions were also criticized by the surgeons because they were drafted by physicians with prior involvement in industry-funded TAVI trials, including PARTNER 3.

The new joint statement, on the other hand, takes a more collegial tone, with the surgical groups applauding VARC for substantially improving the quality of studies in aortic valve disease, which is a rapidly evolving field.

“We wanted to address this in a unified surgical voice, but to also do it in a way that isn’t as sharp as we’ve seen on social media,” said Myers. “We wanted to do this in a respectful way that would be heard and possibly have an impact—having all the surgeons come together to say: ‘This seems important to us, maybe reconsider it?’”   

In terms of specifics, the surgeons are critical of the “blanket inclusion” of rehospitalization as part of a primary composite endpoint in studies comparing surgical aortic valve replacement versus TAVI. They argue that the benefit of TAVI in the low-risk PARTNER 3 trial was driven by fewer rehospitalizations in the first year and that the difference in hard clinical outcomes began to wane by 2 years. Also, at 2 years, there was no statistical difference in the risk of death or disabling stroke between TAVI and SAVR.

With respect to hospitalizations, they recommend introducing a 30-day blanking period because admitting physicians might be more likely to hospitalize some patients knowing they’d undergone an invasive procedure, such as valve surgery.

For me, it’s quite reassuring that the surgical societies wrote the paper, with fairly minor comments, and that they’re questioning the application of the endpoints. Philippe Généreux

Additionally, when defining prosthetic valve thrombosis, they criticize the VARC-3 focus on “clinically significant thrombus” as opposed to older criteria which did not require documentation of clinical sequelae. The new criteria preferences TAVI, as it overlooks thrombus that can develop without clinical consequences even though there are concerns such thrombus can affect valve durability. When using the older definition of thrombosis, there were more events with TAVI than surgery in PARTNER 3, they note. Subclinical thrombus, the surgeons believe, should be appropriately tracked for both surgery and TAVI.

“What’s really amazing is that these are the trials that have the best follow-up in the history of cardiac surgery,” said Myers. “We’ve never had protocolized, prospective, randomized follow-up of our bioprosthetic valves as has been the case of patients included in these trials. Whether it’s a surgical valve or a TAVI valve, the main idea is that all thrombi should be recorded and followed because we don’t know what this will mean later down the line. There’s a lot of bioplausibility that it might lead to earlier structural valve deterioration.”  

With the MI definition, specifically type 5 MIs occurring within 48 hours of surgery and TAVI, the surgeons would like clinical confirmation to be part of the periprocedural MI definition given that a biomarker elevation alone would unfairly penalize the more invasive SAVR. The surgeons would also like to see left bundle branch block after TAVI—not just the need for new permanent pacemaker—be included as an endpoint in future trials. They want the VARC-3 document to more explicitly emphasize the different thresholds in bleeding, specifically when defined by a reduction in hemoglobin levels, with SAVR and TAVI.

Response to Criticisms

With respect to hospitalization, Généreux stressed that VARC-3 is not mandating it for future clinical trials or composite endpoints, only trying to define it, including its different iterations such as cardiovascular- and valve-related hospitalization.

David J. Cohen, MD (St. Francis Hospital, Roslyn, NY), a member of the VARC-3 writing committee, made a similar point, stressing that their process differs from the development of clinical guidelines. VARC-3 was developed by clinical trialists experienced in aortic valve research, the goal being harmonization of endpoint definitions across different studies.

“As such, in contrast to practice guidelines, the VARC-3 definitions are not designed to guide practice or to govern the types of clinical trials that are designed (where involvement of specialty societies is generally desired),” Cohen said in email. “Clinical trialists, whether cardiac surgeons, interventional cardiologists, or anyone else, are free to design their trials with a wide range of comparator strategies, endpoints, and analytic approaches, and VARC-3 does not attempt to dictate these choices.”

Cohen agreed that hospitalizations are more-frequent events that have the potential to “carry” a composite endpoint, but this issue can be addressed through trial design and analysis rather than by ignoring them altogether. Hierarchical endpoints, for example, are being used more often in studies to avoid counting all clinical events as equally severe, said Cohen. Additionally, if a composite is used, it should be obvious if the outcome is driven by a single type of event and this affects the overall interpretation of the study.

Hospitalizations, however, remain important, he argued.

“Hospitalizations are obvious events that can be reliably assessed and that—at a minimum—have a clear impact on patients’ quality of life and on healthcare costs,” he said. “As a clinician, I do not know a single patient who wants to go back to the hospital after a major procedure or operation. Thus, the idea of ‘blanking’ the analysis to early hospitalizations would seem to be the antithesis of a patient-centered approach.”

I do not know a single patient who wants to go back to the hospital after a major procedure or operation. David J. Cohen

With thrombus, Cohen said there is more agreement than disagreement. While there is a focus on clinically significant valve thrombosis, it remains an unproven hypothesis that asymptomatic valve thrombosis might lead to early structural valve deterioration and this warrants further study. As such, VARC-3 writing committee “recommends collecting information on all types of valve thrombosis to advance this research agenda but specify that these asymptomatic events should be reported separately from the more obvious and clinically relevant forms of valve thrombosis,” said Cohen.

He also emphasized that VARC-3 recommends diagnosing a postprocedural MI with both biomarker elevations and clinical confirmation, just as the surgeons recommend. Généreux pointed out that MI will not be part of a primary endpoint in future valve trials but might be part of a group of secondary endpoints. The relative importance of periprocedural MI in aortic valve research is less than in revascularization trials, he said.

In more general terms, Généreux said that the surgical response to VARC-3 is primarily focused on comparisons between TAVI and surgery, but they had a larger focus in mind with establishing consistent endpoints.

“It’s not only TAVI versus surgery trials, but also future clinical strategy studies, such as clinical surveillance versus another therapy, either TAVI or surgical aortic valve replacement or medical therapy,” he said. The definitions could also be used to characterize the natural history of disease or to compare two types of surgeries, such as the Ross procedure versus a mechanical valve.

Diverse Group of Stakeholders

Broadly speaking, the surgical groups lamented that the VARC-3 writing group was comprised of just two practicing cardiac surgeons. The heart team is central to daily clinical practice, said Myers, so it was “disheartening” to have only limited representation from other fields of cardiology as part of the VARC-3 writing committee.

“We depend on each other in our daily work,” said Myers. “It’s a good working relationship, but like a good marriage, it’s not easy every day. It’s good to hash things out. We’re all biased. Me, as a surgeon, I’m going to favor surgery, even if I don’t see it that way. In the discussions we have, we could be looking at exactly the same data completely differently. There’s nothing nefarious behind that—we all have these biases—but that’s why it’s important to have all stakeholders represented on these committees.

The surgical societies would like to implement a review and endorsement process for future VARC updates, one that would involve societies and individuals with minimal relationships to industry and who no direct connection to an of the relevant clinical trials.  

For his part, Généreux believes that the VARC-3 collaborators included a diverse group of physicians, noting that while the writing committee includes just 23 authors, more than 50 people, including representatives from the US Food and Drug Administration, were involved in the process. That included interventional cardiologists, imaging specialists, surgeons, clinical trialists, and a pathologist. The group of final authors on the publication does not capture the heart team-like approach they employed to reach a consensus on the definitions.  

Généreux said that the research consortium will continue to work closely with surgeons in order to push treatments for aortic valve disease forward. “We really design research and do the right thing in a collaborative manner,” he said. 

Note: Cohen is the director of clinical and outcomes research at the Cardiovascular Research Foundation, the publisher of TCTMD.