TAVR Tussle at US Surgery Meeting Threatens to Widen Surgeon-Cardiologist Wedge
Physicians had a range of reactions to TAVR trial criticisms, but STS leaders say they stand for balanced discourse.
NEW ORLEANS, LA—Signaling a growing divide between surgeons and cardiologists that has so far focused mostly on coronary disease care, the annual meeting of the Society of Thoracic Surgeons (STS) gave the podium this week to a number of surgeons and academics critical of studies that paved the way for TAVR’s rise.
Sunday’s high-profile programming included allegations of data suppression and investigator bias, which presenters say bring needed scrutiny to this rapidly evolving field.
The presentations, however, drew sharp criticism from interventionalists and surgeons involved in PARTNER and other TAVR trials who call the talks sensationalistic, unscientific, and damaging for patients—all of this at a time when teamwork between cardiologists and surgeons is seen as a prerequisite for quality care in structural heart disease.
In the most controversial STS presentation, surgeon Anthony Furnary, MD (Providence Health System, Portland, OR, and Anchorage, AK), presented data he alleged were “withheld” from the PARTNER 1A trial when it was published in the New England Journal of Medicine in 2011 and in the years following. In another presentation, statistician Nick Freemantle, PhD (University College London, England), criticized what he characterized as “catastrophic” missing data in PARTNER 3.
Everyone who spoke with TCTMD agreed that, despite their divisions, the growing animosity and finger-pointing between specialists is not fruitful for the field, or for the patients needing care.
Looking Back to 2012
Furnary said he reviewed documents from the 2012 US Food and Drug Administration advisory panel meeting and, when reanalyzing these data, reached a number of what he deemed to be new conclusions.
In his reanalysis, there was no difference in stroke between TAVR and SAVR in the as-treated analysis over 36 months, but there were higher rates of neurological events among all TAVR procedures compared with surgery at 30 days (5.5% vs 2.4%; P = 0.04), at 1 year (8.7% vs 4.3%; P = 0.03), and at 2 years (11.2% vs 6.5%; P = 0.05).
In another analysis, TAVR did not meet the definition of noninferiority against surgery in men. Although this information is present in the FDA executive summary report, Furnary said this finding was never publicized until a 2015 meta-analysis showed that female sex independently predicted late survival following TAVR. “Because that's a dichotomous variable, what that means is male sex is an independent predictor of early death after TAVI,” Furnary said. “We have never been told that.”
“The FDA was highly concerned about this,” he added. “They mention it no less than three times in the executive summary.” Moreover, the 2-year findings, he alleged, were never published.
Furnary also had concerns about differences in reverse remodeling and hemodynamics between the two procedures; the difficulty of comparing patients who had isolated TAVR versus surgical patients also treated for coronary or other valve disease at the time of their valve implantation; and what he called “shifty nomenclature.” To TCTMD, he explained the latter concern involves “which groups make better to sense to use as intention-to-treat and as-treated” in statistical analysis.
“We can only conclude that the medical industrial complex having invested billions of dollars into the development of TAVI wanted to ensure regulatory clinical approval which would result in positive economic success and a positive return on their investment,” he concluded. “In that quest and to that end, they've withheld important clinical data [and] impaired the ability of the medical community at large to fully evaluate the comparative effectiveness of TAVI and its impact on specific patient populations. This has irreparably altered clinicians’ ability to rationally and safely deploy this excellent technology in the appropriate patient populations without concomitantly, unwillingly, and unknowingly doing harm.”
I want to make absolutely clear that I'm not trying to diss TAVI. I think it's a fantastic procedure. . . . What we want to know is how to apply it to the right patient population, and the data that was withheld didn't allow us to do that, at least between 2012 and 2016. Anthony Furnary
Of note, Furnary originally presented these results at the contentious 2019 European Association for Cardio-Thoracic Surgery (EACTS) meeting, which set off months of acrimonious debate between surgeons and cardiologists over the EXCEL trial in left-main disease.
Speaking with TCTMD after his talk, Furnary said he feels “really bad” presenting these numbers because “it’s not my data.” Moreover, “I want to make absolutely clear that I'm not trying to diss TAVI. I think it's a fantastic procedure. I think it's the right thing to do. What we want to know is how to apply it to the right patient population, and the data that was withheld didn't allow us to do that, at least between 2012 and 2016. That's the point.”
Not all of Furnary’s numbers can be found in the FDA panel review from 2012—he told TCTMD he did some of his own calculations using the data from the public FDA documents. Also, one of his slides stated that the FDA panel had voted 5:4 in 2012 to approve TAVR for high-risk surgical patients. In fact, the final vote was 11-0 (with one abstention) that the benefits of TAVR outweighed the risks, 10-2 for safety, and 12-0 for efficacy. Asked about this discrepancy, Furnary told TCTMD was unaware of his error and will retract this slide.
Cardiologists, Surgeons React
Reactions to Furnary’s presentation have ranged from surprise to anger. Twitter, with all its typical subtlety, has erupted with a mix of outrage, blame, and counter-accusations. To TCTMD, both surgeons and cardiologists emphasized TAVR’s track record to date, with several noting that the device studied in PARTNER 1A is no longer being used.
Vinod Thourani, MD (Piedmont Health Institute, Atlanta, GA), a cardiothoracic surgeon who was on the STS panel for Furnary’s talk and served as an investigator on the PARTNER 1A study, vehemently rejected Furnary’s conclusions. Given the extent to which TAVR has progressed since the first trials, he told TCTMD, “PARTNER 1 data at this point makes no sense for us to really rehash.”
Thourani noted that papers were in fact published regarding the sex-specific differences in PARTNER 1A, including 2-year data, in the Journal of the American College of Cardiology in 2014. “It actually said that among male patients, there was no overall survival benefit of TAVR versus SAVR [whereas] there was in females. . . . In the PARTNER series we have published multiple papers, if I remember correctly three papers, on sex differences between TAVR and SAVR within our randomized family. So this is something that we have actually investigated,” he commented.
“Maybe a more thorough literature review of the published data instead of the FDA data would be beneficial for people to do,” Thourani said.
Speaking during the panel discussion, Thourani noted that surgeons must also shoulder some blame for assuming that the addition of concomitant procedures would benefit their patients, one of the things that Furnary noted has muddied the water. “It’s actually the people in this room who should be criticized,” Thourani said.
Moreover, breaking out patients according to whether they did or did not undergo additional surgical procedures at the time of their SAVR was not something that could not have gone in the paper, he added, since the New England Journal of Medicine does not allow for subanalyses. In fact, as reported by TCTMD, Thourani himself has explored the impact of concomitant surgeries in PARTNER 2A.
To TCTMD, panelist and surgeon Gilbert Tang, MD (Mount Sinai Health System, New York, NY), stressed that it is important to keep in mind the evolution of TAVR. In PARTNER 1A, patients were at very high risk for surgery, so “if they survived, that was a win for the patient,” he said, adding that concerns over stroke and paravalvular leak came later.
With regard to the concomitant surgeries, however, “it is generally understood [that] you're going to bias the results against surgery because it's like comparing apples to oranges versus isolated TAVR,” Tang continued. “So you'll never be able to have an accurate comparison between the two groups, and I think what will be interesting to see, in PARTNER 3 and EVOLUT Low-Risk, is [what happens] if you exclude all the patients who had concomitant procedures in both arms and redo the analysis to see if there's any difference in outcomes. That would be a better matched comparison.”
Craig Smith, MD, the cardiovascular surgeon who with Martin Leon, MD (both from NewYork-Presbyterian Hospital/Columbia University Irving Medical Center, New York, NY), was the co-principal investigator for PARTNER 1A, had strong words in response to Furnary’s analysis.
“Dr. Furnary relies on innuendo without a grain of proof to support his reckless accusation that data was hidden or withheld in pursuit of profit,” Smith said in an email. “Instead, he disassembles a randomized trial post hoc, using various unpowered subgroup analyses and cohort debridements to undermine a result he doesn’t like. He’s correct that TAVR isn’t perfect. SAVR, an enormously beneficial and sophisticated operation that I love to do, isn’t perfect either.”
He’s correct that TAVR isn’t perfect. SAVR, an enormously beneficial and sophisticated operation that I love to do, isn’t perfect either. Craig Smith
In a statement to TCTMD, Edwards Lifesciences said the company unequivocally denies Furnary’s accusations: “We proudly stand behind the Executive Committee of the PARTNER trials and the integrity of the data as presented and published for the entire PARTNER series of trials, including six publications in the New England Journal of Medicine. These data have been further validated through independent registries such as the TVT Registry in the US, as well as various country registries around the world.”
Additionally, “the protocol, endpoints, and prespecified analyses of the PARTNER 1A trial were developed with the sponsor (Edwards), the executive committee, and FDA, and data from the trial were published and presented many years ago,” they write. “We are confident that the rigor applied to the review of the PARTNER 1A trial and other subsequent trials’ data provided a fulsome and transparent evidence base supporting the appropriateness of this lifesaving therapy for hundreds of thousands of patients worldwide.”
Statistics in Low-risk Trials
In his presentation, Freemantle focused on PARTNER 3 and argued there was “catastrophic loss to follow-up” in the 1-year outcomes studied for this low-risk trial.
“When we are looking at a low-risk population with a low event rate, you can see that we're losing more patients than are experiencing many of the events in the trial and that is a problem because we don't know what happened to the patients who are missing,” he said. “Frankly more work in the trials before randomization in the run-in period could’ve left us in the position where these trials didn't have those patients who weren't suitable to be randomized because they weren't going to have surgery if they happened to be randomized to that group.”
In a similar vein, Michael Reardon, MD (Houston Methodist Hospital, TX), in a presentation rehashing the Evolut Low Risk TAVR trial, said that one thing he wished they had done in the study protocol was follow the patients who were excluded from randomization in a registry. “My biggest regret is we don’t know what happened to those people and we're never going to know what happened to those people because of the way we designed the trial,” he said.
Calling the PARTNER trials “poorly designed,” Freemantle argued that “TAVR may or may not be a safe and effective alternative to surgery” and that “the regulators are asleep at the wheel” for using different standards for Class III devices than they do for pharmaceuticals.
He called for a randomized controlled trial in the TAVR space sponsored by “somebody other than a manufacturer with a vested interest.” With that, “I would feel just so much more comfortable,” Freemantle added.
Cherry-picking the Data?
Other physicians contacted by TCTMD bristled at the idea that anything had been intentionally withheld. David Cohen, MD (Kansas City, MO), who was an investigator for the PARTNER trials, observed that “to the extent these data were hidden, they were hiding in plain sight.” Two-year stroke and TIA rates, he noted, are in fact published in the 2-year New England Journal of Medicine paper. Also, he asserted that Furnary “cherry-picked data, . . . which in and of itself would be okay, but then he inappropriately interpreted much of it.” By doing so, these claims are undermining public confidence in biomedical research, Cohen added. “I find that to be very concerning.”
Similarly, Philippe Généreux, MD (Morristown Medical Center, NJ), who has also been involved in the PARTNER series of trials, called Furnary’s presentation “ridiculous.” Referencing the ongoing drama unfolding between cardiologists and surgeons regarding the 5-year EXCEL data, he asked, “What's happening with all these conspiracy movements thinking that we hide data? What's going on?”
“To insinuate that there's an issue in transparency is really not knowing and understanding the process of the last 10 years and what the key opinion leaders have been trying to do,” Généreux added. “[Furnary] has to be careful because his message can hurt more people than anything else.”
Domenico Pagano, MD (University Hospitals/Queen Elizabeth Hospital, Birmingham, England), is the secretary general of EACTS who was behind the decision by the society to pull out of the current treatment recommendations for left main CAD. He has not previously responded to multiple requests for comment from TCTMD about the EXCEL controversy. Approached by TCTMD after giving a plenary talk at STS, Pagano again said he would not comment on EXCEL but advised: “I think certainly people should work together and resolve the issue.”
Physicians who participated in Sunday’s session also had a range of reactions. Toby Rogers, MD (MedStar Washington Hospital Center, Washington, DC), was the sole interventional cardiologist on Sunday’s STS panel, something that he told TCTMD he was made to feel “very aware of.”
Overall, he said, “we've got to be careful not to let one negative talk sort of overshadow the rest. . . . The positives in the session were that there was clearly a desire from some of the parties for ongoing engagement between cardiologists and cardiothoracic surgeons in the field of TAVR, and that even with approval of low-risk we should continue to work together. To me that was encouraging.”
I think certainly people should work together and resolve the issue. Domenico Pagano
Edward Chen, MD (Emory University, Atlanta, GA), who co-chaired Sunday’s session and served as a member of the STS program planning commission, told TCTMD he stands behind the presentations selected.
“It was felt that it was important that we present not a pro-TAVR session and that we needed to have a balanced viewpoint. We didn't want it to be a situation where it was going to be all ‘TAVR is here as a definitive therapy,’ because I don't think the question is fully answered in certain patient populations,” he said. “The point was to give a broader overall perspective of the trials, the strength of the trials, the weaknesses that were outlined in the session, and just to give an overall perspective and that way we should ultimately hopefully be open and transparent about what is actually best for our patients.”
Reardon’s key advice for surgeons is to “learn the data for these trials.” He pointed out that, for example, the Evolut Low Risk trial excluded patients with bicuspid valves and coronary disease, calling those knowledge gaps. “You need to be able to converse with your patients and your cardiologists in a really granular manner,” he advised.
Friction in the Field
Moving forward, Furnary said he doesn’t expect his talk to influence current TAVR practice. “What I think it points out is that industry-sponsored trials are inherently biased,” he said. “But it's incumbent upon the industry to let us know all the data so that we can make important decisions. . . . Because all we want to do is do the right thing for the patient.”
Asked whether controversial sessions like this one at STS are driving an insurmountable wedge between cardiologists and surgeons, Furnary said, “I certainly hope not.” While it’s a “valid concern” that surgeons have their own conflicts in this matter given that some are losing patients to interventional colleagues, he acknowledged, the fact that many surgeons are employed by hospital systems tempers the issue.
“I'm not overly concerned about ‘My gosh someone just moved my cheese and I'm going to have to fight to get the cheese back,’” said Furnary. “That's not the deal. The deal here is what is right for the patient and what is open and transparent, because that's what helps us make good decisions.”
Rogers agreed that he sees good collaboration between surgeons and cardiologists at his institution. “In a way, I was a little surprised but not [totally] surprised that at a surgical meeting . . . the slant [here] is very much ‘We need to emphasize the relevance, the importance of surgery,’” he said. “I don't think they should be seen as competition or as one exclusive of another. To me the ongoing collaboration is critical.”
Because patients generally prefer a less-invasive option, “that presents a conundrum” for surgeons, Tang said. “There’s always been disagreements between cardiologists, interventional cardiologists in particular, and surgeons because we work on the same conditions and, to some degree, we compete for the same patients,” Cohen added. “There’s always a tension there.”
Surgeons may have “insecurity in terms of volume, but when you look since TAVR was initiated, we saw an increase in the volume of surgery just by disease awareness,” Généreux said.
“Some people lose the big picture,” he added. “We have a surgical strategy, we have a transcatheter strategy, and both are complementary.”
With regards Furnary’s PARTNER 1A criticisms, Cohen said he would want to ask Furnary which procedure he would advise for “the next 86-year-old man with a prior CABG, moderate [chronic obstructive pulmonary disease], and a creatinine of 2 who walks into his office. . . . I don't think it's going to be surgery.”
Furnary’s take on what he calls the surgeon-cardiologist “firestorm” that grew out of the EACTS meeting is that Europeans tend to more open to transparency delivered in a more aggressive fashion compared to how US physicians handle things. “In the United States, we're a little more conservative, a little more guarded about what we say,” he observed.
Here Cohen is in agreement. “Much of this questioning seems to have originated outside of the United States, led by physicians who feel excluded from the heart team, who feel threatened by these procedures,” he said. “I feel like physicians in the US have been more accepting of the heart team concept, especially in the case of TAVR because all procedures are performed by a cardiologist and a cardiac surgeon working together.”
Note: TCTMD was contacted by Furnary on February 3, 2020, regarding some errors related to data sources, copy editing, and context for specific quotations. In particular, some but not all of Furnary’s numbers can be found in the FDA executive summary, and he clarified that the only statistical analysis he calculated was that involving the valve implant group with concomitant surgery removed and with crossovers appropriately applied.
Furnary AP. Eyes wide shut: the unpublished data from PARTNER 1A. Presented at: STS 2020. January 26, 2020. New Orleans, LA.
Freemantle N. Low-risk transcatheter aortic valve replacement: what do the latest data tell us? Presented at: STS 2020. January 26, 2020. New Orleans, LA.
- Furnary reports holding equity in Edwards Lifesciences, Intuitive Surgical, and National Health Investors, intellectual property in Portland Protocol, and multiple patents in the United States, Australia, United Kingdom, and Canada.
- Freemantle reports directing the EACTS Methodology Support unit at his institution.
- Thourani reports receiving research grants from and serving as a consultant to Edwards Lifesciences, WL Gore, Boston Scientific, and Abbott Vascular.
- Généreux reports serving as a consultant for and receiving speaker fees from Abbott, Boston Scientific, CSI, Medtronic, Edwards, and Abiomed.
- Cohen reports receiving research grants and consulting income from Edwards Lifesciences, Medtronic, Boston Scientific, and Abbott.
- Chen reports serving on the speaker’s bureau for Medtronic.
- Smith reports serving as the surgical PI on Partner 1 and Partner 2, the executive PI on Partner 3, and receiving reimbursement from the Edwards Lifesciences for travel and customary expenses associated with trial management.
- Reardon reports serving on the surgery advisory board for Medtronic.
- Tang reports receiving other research support from Intuitive Surgical and serving as a consultant for Auris.
- Rogers reports serving on the speakers bureau for and as a consultant to Edwards Lifesciences and Medtronic as well as being a co-inventor on patents assigned to the NIH.