Five international cardiac surgery associations have banded together to address “substantive concerns” regarding the recently updated Valve Academic Research Consortium 3 (VARC-3) clinical endpoint definitions for aortic valve research.
The VARC-3 update was a multidisciplinary effort that included more than a dozen new or modified definitions for use in transcatheter and surgical aortic valve replacement (TAVR/SAVR) clinical trials, but drew criticism last year from surgeons that some of its definitions favor TAVR over surgery and that its writing committee had deep ties to industry and lacked diversity.
The new surgical associations’ position statement calls out five specific VARC-3 definitions — rehospitalization, valve thrombosis, bleeding, myocardial infarction (MI), and left bundle-branch block (LBBB).
The statement was jointly issued by the Society of Thoracic Surgeons (STS), the American Association for Thoracic Surgery, the European Association for Cardio-Thoracic Surgery, the Asian Society for Cardiovascular and Thoracic Surgery, and the Latin American Association of Cardiac and Endovascular Surgery.
It was copublished in the Annals of Thoracic Surgery, the Journal of Thoracic and Cardiovascular Surgery, the European Journal of Cardio-Thoracic Surgery, and the Asian Cardiovascular and Thoracic Annals .
“We hope that this message can be seen, even if it’s somewhat difficult to hear sometimes, as positive constructive criticism compared to some of the dialogue that we’ve had on social media,” lead author Patrick O. Myers, MD, Lausanne University Hospital, Switzerland, said in an interview. “It’s not criticizing people or the process but just trying to make these definitions better to ensure the good design of clinical trials.”
The president of each surgical association recommended representatives to help write the position statement, and once completed over Zoom meetings, it received formal endorsement from each association prior to publication, he said.
Reached for comment, VARC-3 lead author Philippe Généreux, MD, Gagnon Cardiovascular Institute, Morristown Medical Center, Morristown, New Jersey, said, “I was pleasantly surprised that their comments were actually pretty minor and that most of these comments are really more a reflection, not of the validity of the definitions, but rather their applications.”
He noted that all the potential issues with the definitions were already discussed during the making of VARC-3 and resolved by consensus of more than 50 experts including the STS president at the time, US Food and Drug Administration officials, and experts from the community.
“To be quite honest, I’m not sure they have consensus,” Généreux said. He added that the writing committee welcomes input from anyone, but “we’re not going to change the definitions to please eight individuals if we strongly believe by consensus of experts in the field that this is not the right thing to do.”
Rehospitalizations and Valve Thrombosis
The surgical associations praise VARC-3 for providing a standardized definition of bioprosthetic valve failure, but say they will not endorse the inclusion of rehospitalization as a component of the primary efficacy composite endpoint along with all-cause mortality, stroke, and quality of life.
They note that rehospitalizations outnumber mortality events, especially in short follow-up trials, and that the superiority of TAVR at 1 year in the PARTNER 3 trial of low-risk patients was driven primarily by more rehospitalizations in the surgical arm, but that this superiority was waning at 2 years of follow-up.
“The first thing we are calling for is that it shouldn’t be part of the primary composite outcome measure,” Myers said. But if it really has to be included, a 30-day blanking period for rehospitalization “would acknowledge that there’s a greater risk of rehospitalization during the acute phase of recovering from surgery.”
Généreux said that VARC-3 provides granular details for defining the different types of hospitalizations, but that a 30-day blanking period makes no sense. “If you close your eyes to anything within 30 days because you don’t like it, you’re missing the opportunity to improve your procedure, to improve your treatment, and to characterize precisely what happened with your patient.”
The new document lauds VARC-3’s focus on patient-centered and clinically relevant endpoints but questions the definition of valve thrombosis as a “clinically significant” thrombus. It points out that the incidence of valve thrombosis was significantly higher with TAVR vs SAVR in PARTNER 3 using the older VARC-2 definition, which did not require evidence of clinical sequelae (2.6% vs 0.7%; P = .02). Under the new definition, however, half of the thrombi would be relabeled as “nothing there,” Myers said.
“As we’re doing this in younger and younger patients who will survive longer, there is a question of thrombus having an effect on the valve and leading to earlier structural valve deterioration,” he added. “All this is conjecture. We don’t have the data. So mainly what we’re advocating is that all thrombi should be reported.”
MIs, Bleeding, and LBBB
The policy statement also criticizes VARC-3’s decision to define periprocedural (type 5) MI using a biomarker-only definition without need of clinical confirmation. Such definitions have been shown to have a very poor prognostic significance in surgical series compared with the Universal Definitions of Myocardial Infarction, Myers said.
“What’s interesting is that for thrombus and bleeding, they require clinical correlation, but on the perioperative MI they now use a definition that does not require clinical significance, meaning no ECG changes, no regional wall motion abnormalities or things like that,” he observed.
The decision also seems to disregard the EXCEL trial controversy that illustrated how outcomes and a trial’s message can change depending on which definition of periprocedural MI is used.
With regard to bleeding, the surgical associations agree with the VARC-3 recommendation to use different thresholds when bleeding is integrated into a composite endpoint (type 2 or greater for TAVR and types 3 or greater for SAVR) but suggest this important point should be featured in the chapter on bleeding rather than the section on composite endpoints.
The surgical associations say VARC-3 also got it right adding the need for a new permanent pacemaker to the early composite safety endpoint, but that it was a “missed opportunity” not to include new left bundle-branch block in the safety composite, despite recognizing that this may become an important endpoint to consider in the future.
Myers said that left bundle-branch block could have implications for survival as TAVR moves into lower-risk, younger patients, as some data with 1-year follow-up suggest it has a prognostic impact, even in the higher-risk older patients with more competing risks.
Finally, the surgical associations point out that only two of the 23 VARC-3 authors were practicing cardiac surgeons and say that a more diverse writing group “may help mitigate issues related to the duality of interests.”
Généreux said that the final author list is not a reflection of the rigorous work done by 11 cardiac surgeons including the two surgeon authors. The VARC-3 writing committee also had a good representation of women, unlike the surgical position statement, which was penned by eight men.
Eur J Cardiothorac Surg. Published online March 17, 2022. Full text
Myers reported no relevant financial relationships. Coauthors Nicholas Freemantle reports institutional support from EACTS; Wilson Szeto reports a financial relationship with Edwards and Medtronic; Vinod Thourani reports financial relationships with Abbott Vascular, Boston Scientific, CryoLife, Edwards Lifesciences, Medtronic, Shockwave, and JenaValve; S. Chris Malaisrie reports a financial relationship with Edwards, CryoLife, and Medtronic; and Marc R. Moon reports a financial relationship with Medtronic. Généreux reports receiving consultant fees from Abbott Vascular, Abiomed, Boston Scientific, Cardinal Health, Cardiovascular Systems, Edwards Lifesciences, Medtronic, Opsens, Siemens, SoundBite Medical Solutions, Sig.Num, Saranas, Teleflex, Tryton Medical, and having equity in Pi-Cardia, Sig.Num, SoundBite Medical Solutions, Saranas, and Puzzle Medical.