
Optimal treatment of patients with chronic limb-threatening ischemia (CLTI) has been an ongoing debate within the vascular community, fueled by growing numbers of patients, evolving technologies, provider bias, and a lack of high-quality evidence.
Few randomized controlled trials (RCTs) have focused on this patient population. Now, within a six-month period, the primary results of the long-awaited BEST-CLI and BASIL-2 RCTs have been reported. At the top level, their results seem wholly discordant with respect to limb-based versus survival outcomes. A deeper dive into the designs of these two trials, their populations and outcomes reported to date is needed to interpret their meaning.
The BEST-CLI trial was funded by the U.S. National Heart, Lung and Blood Institute and involved 133 centers in the U.S., 12 in Canada, and five in other countries. It was designed to compare the effectiveness of endovascular intervention versus open bypass surgery among CLTI patients with infrainguinal occlusive disease who were deemed acceptable candidates for either treatment. The study was conducted as two parallel RCTs based on preoperative assessment of an adequate great saphenous vein (GSV) for bypass. The primary endpoint was major adverse limb event (MALE)-free survival, defined in accordance with the published SVS Objective Performance Goals (MALE: major amputation or major reintervention [new open bypass, major open graft revision, thrombectomy or thrombolysis]). A total of 1,434 subjects were randomized in cohort 1 (adequate GSV) and 396 in cohort 2.
Analyzed by intention-to-treat (ITT), patients assigned to open surgery in cohort 1 (median follow-up time 2.7 years) experienced significantly lower rates of MALE or death (32% relative risk [RR]), major amputation (27% RR), any major reintervention (65% RR), and less than half the total number of major reinterventions in the treated limb over time.3 These findings were robust across virtually all pre-defined patient subgroups. In cohort 2, there was no significant difference in the primary endpoint after a median of 1.6 years of follow-up. Importantly, perioperative mortality, long-term survival, and overall major adverse cardiovascular event (MACE) rates were not different by assigned treatment in either cohort. The BEST-CLI investigators concluded that both procedures were equally safe, and that CLTI patients deemed suitable for either approach who had an adequate GSV experienced better overall clinical outcomes after open bypass surgery.
BASIL-2 was funded by the UK National Institute of Health Research and involved 39 centers in the UK, one in Sweden and one in Denmark. It was designed to compare the effectiveness of endovascular intervention versus open bypass with vein among CLTI patients requiring treatment of infrapopliteal (IP) disease, with or without a more proximal infrainguinal intervention. Patients had to be deemed suitable for either procedure, with an anticipated life expectancy of at least six months. The primary endpoint was major amputation-free survival (AFS). Among 345 patients enrolled and followed for a median of 40 months, those assigned to bypass surgery experienced a 35% greater incidence of major amputation or death. This result was driven entirely by excess long-term mortality in the surgical group.
The BASIL-2 investigators concluded that while limb-based outcomes were similar, patients treated by an endovascular-first strategy were more likely to be alive with an intact limb compared to those who underwent surgery first.
What gives with these seemingly opposite results?
Let’s start with similarities between the two RCTs. Both trials were designed to compare revascularization strategies among patients deemed suitable for either, with equipoise determined locally by site investigators. Immediate technical success rates reported for endovascular interventions were strikingly similar (85% in BEST-CLI, 87% in BASIL-2). Approximately 70% of subjects in both trials had diabetes. More than 70% in both trials were treated with an antiplatelet agent, and similar proportions were treated with a statin. Both trials had to deal with the major impacts of a COVID pandemic on healthcare delivery and clinical research. And once again these trials highlight that, despite concerted efforts at guideline-based medical therapy, mortality rates in CLTI approximate 10% or more per year. However, differences in the designs, enrolled populations, treatments received and trial execution are all likely related to the disparate outcomes reported. Key characteristics of the two RCT cohorts, most importantly the anatomic complexity of disease treated, remain incompletely described at this time.
Comparison of the populations is important for understanding both the contrasting results and the generalizability of each trial. BEST- CLI was more than five times the size of BASIL-2 and had broader inclusion criteria. BEST-CLI enrolled more women (28% vs 18%) and more non-white (28% vs. 9%) subjects. The BASIL-2 study population was defined as requiring an IP intervention. In BEST-CLI an IP procedure was performed in just over half of all subjects, and only a minority had treatment of IP disease alone (16% in cohort 1). The severity and management of femoropopliteal disease in BASIL-2 is unclear and likely different from BEST-CLI. Patients in BASIL-2 were older (72 vs. 67 yrs), and, notably, those randomized to open surgery had a higher prevalence of prior MI (24% vs. 13% for the endo arm). Together with the six-month life expectancy entry criterion for BASIL-2, these factors may explain worse long-term survival among surgical patients in BASIL-2. Perioperative mortality after bypass surgery was twice as high in BASIL-2 (6% vs 1.7% in BEST-CLI cohort 1).
Both RCTs were “pragmatic” and, thus, an array of endovascular and open interventions were applied. Befitting the infrapopliteal design of BASIL-2, there were twice as many bypasses originating from the popliteal artery in BASIL-2 (41%) versus cohort 1 of BEST-CLI (16%). Data reported to date on endovascular devices and techniques are hard to compare but plain balloon angioplasty was the most common procedure below the knee in both trials. In BEST-CLI, vascular surgeons performed 73% of endovascular interventions, compared to BASIL-2 where interventional radiologists performed 84%. These differences reflect the healthcare systems of the United Kingdom and U.S. in relation to vascular care.
AFS is an important outcome in CLTI but is dominated by mortality rather than limb events, and does not reflect the burden of reinterventions and unresolved or recurrent symptoms that are experienced. As such, it is an insensitive measure of the quality of limb revascularization. Just as in coronary disease RCTs, need for repeat revascularizations and other major limb events are highly relevant to compare treatment strategies. BASIL-2 was likely quite underpowered for limb-specific outcomes, a problem amplified by extending follow-up time rather than enrolling more new subjects to reach a targeted number of events, since most MALEs occur within the first year. Reinterventions in CLTI are largely driven by ongoing symptoms during follow-up surveillance.
The impacts of the COVID-19 pandemic on healthcare delivery and trial execution in these trials were likely disparate. BASIL-2 investigators noted that face-to-face follow-up assessments were substantially affected after March 2020. In looking at some of the published top-line results, major amputation rates were higher in both arms of BASIL-2 (20% open and 18% endo vs. 10.4% and 14.9%, respectively, in BEST-CLI cohort 1); reintervention rates were lower in the endovascular arm of BASIL-2 (32% vs. 43% in BEST-CLI cohort 1); and mortality after open bypass was significantly worse in BASIL-2. It is presently unclear if these differences between study outcomes reflect the patients enrolled, the care delivered, or both.
All RCTs have limitations of which generalizability is paramount. CLTI patients are heterogeneous in systemic risk, severity of limb threat, and anatomic complexity of occlusive disease. The “endo-first vs. open-first” debate is and has always been an oversimplification, and a disservice to this clinical complexity. The results of BEST-CLI clearly demonstrate advantages of surgical bypass in a subset of CLTI patients. To better understand the relative effectiveness of these complementary strategies, trials and registries need to incorporate staging such as WIfI (Wound, ischemia and foot infection) and GLASS (Global Limb Anatomic Staging System) in their design and reporting. We anticipate much further and more granular data in regard to important patient subsets in these trials, as well as secondary endpoints and economic analyses. It remains for the trial investigators to further unravel the puzzle by comparing like-to-like patients and identifying predictive factors for outcomes. The good news is that we now have a large trove of high-quality data in CLTI from these two RCTs, thanks to the great efforts of the investigators, the sponsors, and most importantly the patients who volunteered to participate. These data will undoubtedly advance evidence-based practice in CLTI and improve future study designs in this arena.
Michael S. Conte is Edwin J. Wylie, MD, Chair in Vascular Surgery at the University of California San Francisco (UCSF).