Does BEST-CLI Settle the Surgery vs Endovascular Question for CLTI?
Many hoped the trial would get everyone on the same page, but experts continue to hash out the results, push back, or double down.
When the BEST-CLI study results were presented last month at the American Heart Association (AHA) 2022 Scientific Sessions, some thought they fell short of answering the primary question that the trial was designed to address: which revascularization option is best for critical limb-threatening ischemia (CLTI) patients suitable for both surgery and endovascular therapy?
Surgery seemingly came out on top, but questions about trial design, enrollment, and equipoise, among others, have clouded the picture.
“The study is not perfect; it had many limitations. I think we need to be honest about it and lay those out so we can discuss these,” co-principal investigator Alik Farber, MD (Boston Medical Center, MA), said at a recent trialists’ meeting held in conjunction with VEITH 2022. Farber, a vascular surgeon, said his takeaway was that “bypass with adequate saphenous vein is a more effective strategy for patients who are deemed suitable for both strategies.”
The choice of surgery or endovascular therapy varies widely among providers and centers, so much so that the same type of CLTI patients may be treated completely differently depending on where they present. BEST-CLI, a head-to-head comparison of the two strategies, sought to inject fresh data into the debate and show how the variability can impact clinical outcomes.
In the main trial results, among patients with an adequate single-segment great saphenous vein (SSGSV) that could be used for bypass surgery, 42.6% of those treated surgically and 57.4% of those treated with endovascular intervention had a primary outcome event (P < 0.001), a composite of major adverse limb events (MALE) or all-cause death at a median follow-up of 2.7 years. In that group, known as cohort 1, surgery resulted in 32% fewer major adverse events related to CLTI, driven by a 65% reduction in first major reinterventions and a 27% reduction in above-ankle amputations.
In patients who did not have an SSGSV but had an alternative conduit, known as cohort 2, there was no significant difference in primary outcome event rates at a median follow-up of 1.6 years. The BEST-CLI investigators have acknowledged that cohort 2 was underpowered.
The study is not perfect; it had many limitations. I think we need to be honest about it and lay those out so we can discuss these. Alik Farber
Farber’s co-principal investigator Kenneth Rosenfield (Massachusetts General Hospital, Boston), told TCTMD that while the results confirm that CLTI patients with good saphenous veins generally do well with surgery, the trial is just as important for what it doesn't say.
“As of now, you can't say that all CLTI patients with a patent saphenous vein should have surgery. There are some in the surgical world who are saying that, and it is not correct. It's a misinterpretation, because it doesn't account for who wasn't put into the trial, and it would be irresponsible and improper and actually bad for patients to state that all patients with CLTI who have a vein should get bypass. It's equally wrong to say all patients with CLTI should have endo first, because that's not true either.”
Surgeons, however, pushed back, saying the data warrant thoughtful consideration.
“This is level-one evidence, the likes of which we don't get every day, and it should force some recalibration,” said Michael S. Conte, MD (University of California, San Francisco), speaking at the VEITH meeting. “I think everybody should take some stock over whether or not there is underutilization of treatments that work. In order to deliver those treatments effectively and to have a center of excellence, you need to be able to offer all of them with expertise and you need to have a workforce that is trained and comfortable to do that.”
Debating Endovascular Failure
For the endpoint of total number of major reinterventions, the incidence rate in the endovascular arm was more than twice that of the surgical arm, with the curves diverging early (P < 0.001), indicating that the difference was driven by early failure in the endovascular group. The initial technical failure rate was 15% for endovascular procedures versus 2% for surgery. Approximately 42% of all first major reinterventions occurred within 30 days.
That finding jumped out for Eric A. Secemsky, MD (Beth Israel Deaconess Medical Center, Boston, MA), “The technical success percentage is so much lower than what most of us see in our endovascular practice,” he told TCTMD. That may reflect the fact that BEST-CLI “didn't really represent the vascular community of providers. [Treating physicians] were overwhelmingly vascular surgeons . . . so we need to be careful about how much we generalize these findings.”
At the trialists’ meeting, Secemsky gave an official counterpoint presentation on behalf of the endovascular community, noting that the BEST-CLI results portray endovascular therapy in no better light than the BASIL trial of bypass versus angioplasty from 2005, which showed a similarly high rate of technical failures. Based on contemporary data, he said a 7% to 8% technical failure rate is generally accepted as the norm for endovascular CLTI procedures in an all-comers CLTI population.
But according to Farber, unpublished sensitivity analyses from BEST-CLI support the main findings, showing that even after excluding the technical failures there is still a statistically significant difference between the endovascular and surgical arms in increased risk of major reintervention that favors surgery. Farber said those data will be published in the near future.
As of now, you can't say that all CLTI patients with a patent saphenous vein should have surgery. There are some in the surgical world who are saying that, and it is not correct. Kenneth Rosenfield
Rosenfield noted that while there are many talented surgeons regularly doing successful endovascular procedures, the high rate of endovascular failure seen in the trial does not add up for him, either.
“But then again, you have to remember that the patients that were entered in this trial were those in whom the investigator felt that there was equipoise between surgery and endovascular. So, one might imagine that the more complex endovascular cases were randomized, while the simple, straightforward endo cases were not,” he said. “Right now we don't know any of that for sure.”
Also commenting for TCTMD, Sahil Parikh, MD (NewYork-Presbyterian/Columbia University Irving Medical Center, New York, NY), said there likely will be “anatomic and clinical confounders that need to be acknowledged and that were not in the manuscript because the investigators don't have the anatomy assessments.”
BEST-CLI began enrollment in 2014 and was funded by the National Heart, Lung, and Blood Institute, but the investigators experienced slow enrollment and ran out of money. This forced them into what Farber said was an “unprecedented” situation of having to seek additional funding from industry and vascular societies to complete the study.
They ultimately enrolled 1,830 patients from five countries. An investigator credentialed in surgery had to agree with one credentialed in endovascular therapy that each CLTI patient was an appropriate candidate for the trial. While three-quarters of enrolled sites had multidisciplinary collaboration representing all specialties that treat CLTI, Farber acknowledged that the investigators were not successful in ensuring this at all sites. They also fell short when enrolling women, who accounted for less than 30% of the study population.
Enrollment and Equipoise
Like Farber, Rosenfield said the BEST-CLI investigators are well aware of the multiple questions surrounding how patients were enrolled, noting that the “the threshold for putting patients into the trial may have been defined by the investigators’ concept of whether the patient was appropriate at all for surgery or appropriate at all for randomization.”
Rosenfield said the investigators are working on obtaining all angiograms and case report forms from trial participants to get further clarification about the types of patients who did and did not end up being randomized.
A related question, said Parikh, is how many CLTI patients were assessed over the nearly 10 years that it took to do the trial versus the number actually enrolled and followed. Rosenfield told TCTMD that the investigators don’t know the total denominator of CLTI patients from all the BEST-CLI sites, but he said he suspects that it was “quite large” compared with the actual number of patients enrolled. “Beyond that, we don't know what the characteristics were of the numerator compared to the denominator and the characteristics of the patients,” he added.
Secemsky said knowing the proportion of all CLTI patients that have true equipoise for endovascular therapy and surgery really is one of the most important aspects to consider when trying to dissect the BEST-CLI results and apply them to clinical practice.
“If we're talking about 5% of the PAD population, that's great. But we can't make these global recommendations that everybody needs a surgical consult or a vein mapping,” he added. “If it's that small, that's not a cost-effective mechanism. It also delays time to therapy.”
He agreed that while screening for suitable veins should be done more frequently than it is in some CLTI patients, a better understanding is needed about the cost, time considerations, and feasibility of screening the majority of patients prior to deciding on a treatment plan.
Rosenfield concurred, telling TCTMD that “if CLTI patients are approaching equipoise, it's good to know what venous conduits are available.” However, in the absence of equipoise and if an operator feels it is a straightforward, short endovascular case as opposed to a multi-hour surgery, “then I don't think you necessarily have to [vein map] those patients,” he said.
Rosenfield also noted that the exclusion of CLTI patients considered high risk for surgery from BEST-CLI further muddies the picture when generalizing the results to real-world decision-making and equipoise. “We can’t say that we know anything about those types of patients, because we don’t based on this trial,” he added.
“Every one of us are going to have to figure out how this is going to impact our practice,” Menard noted. He said that in navigating enrollment, the BEST-CLI investigators tried to challenge physicians from both the surgery and endovascular sides to “put down their bias” with regard to treatment. He also defending the importance of the vein mapping, noting that “nowadays with CT and MR we're getting the vein assessment as part of those studies and you can do an on-table assessment of the vein pretty easily.”
Another question that some in the endovascular community have expressed relates to optimal medical therapy in the trial following the endovascular procedures, because that information was not included when the results were published in the New England Journal of Medicine. At the trialists’ meeting, Farber presented data showing that 75% of cohort 1 were on at least one statin at 30-day follow-up compared with 70% at baseline. At the same time frame, 83% were on at least one antiplatelet agent compared with 72% at baseline; 36% were on dual antiplatelet therapy compared with 18% at baseline, and 6.7% were on at least one direct oral anticoagulant compared with 4% at baseline.
‘Recalibration’ and Responsibility
Farber and his co-investigators noted that despite the controversy the trial had engendered, the BEST-CLI results support a complementary role for open and endovascular strategies and highlight the need for expertise in both.
In terms of the “recalibration” Conte was urging on his colleagues, Rosenfield said identifying where change is needed “is actually a little bit of our responsibility . . . it's the responsibility of our entire vascular community.”
Also weighing in, Konstantinos Katsanos, MD, PhD (Patras University Hospital, Rion, Greece), told TCTMD that he sees parallels in the controversy over BEST-CLI and the firestorm that emerged from his meta-analysis of paclitaxel-coated devices for PAD, which suggested increased long-term mortality compared with uncoated devices.
Although Katsanos agreed that the majority of real-world CLTI patients most closely resemble those in BEST-CLI’s cohort 2, he said the study “sets the bar in terms of what should be the benchmark for delivering CLTI outcomes while at the same time likely being a driver for the endovascular community to up their game.” He suggested that “upping” their game should include in-depth study of more varied device-drug options, including sirolimus, and “doing proper quality control of our endovascular procedures, which we have never done.”
Katsanos said he believes the BEST-CLI investigators are also to be congratulated for showing improved quality-of-life (QoL) trends with endovascular therapy, something that he said no drug-coated balloon studies in PAD have ever shown and were not designed to do. In the prespecified QoL analysis of BEST-CLI, endovascular therapy was statistically superior on several patient-reported QoL instruments in cohort 1 despite being below the threshold of clinically meaningful difference.
“The BASIL and the BEST-CLI studies, whether we like the results or not, are exemplary studies because of the sample sizes, the statistics, and also the fact that the endpoints matter to the patient, to the healthcare providers, to the hospital, etc,” he added. The BEST-CLI investigators have acknowledged, however, a certain amount of “missingness” among the QoL responses, the degree of which was similar in both treatment arms.