While a number of companies in the neurovascular space have become increasingly focused on the ‘first-pass effect’ in mechanical thrombectomy treatments, many major clinical trials assessing the safety and efficacy of the technique still favour a ‘rule-of-three’ paradigm. And—although it is now broadly accepted that the risk of a poorer clinical outcome grows with every additional clot-retrieval attempt—the exact relationship between the two is still highly debatable. In the view of David S Liebeskind (University of California Los Angeles [UCLA], Los Angeles, USA), this is one of several reasons why the first-pass effect remains an imperfect measure of success in stroke care.
At last year’s Barts Research and Advanced Interventional Neuroradiology (BRAIN) conference (5–8 December 2022, London, UK), in a presentation outlining the import of imaging and angiography core laboratories, Liebeskind stated that success at the first pass is “just a result”—often indicative of a relatively easy case, and little more. He went on to note that variation between individual patients, their specific anatomies, clot types and other features, is one of many factors confounding the use of pass numbers to define success.
“Reperfusion of the downstream territory—measured by a certain threshold of the eTICI [expanded thrombolysis in cerebral infarction] classification system—in itself, is very complicated,” he expounds, speaking to NeuroNews. “That is why I am, not sceptical, but cynical, in saying that, when you get a first-pass effect, it just means things appeared like they went in the right direction.”
According to Liebeskind, discussions around the relevance of the first pass—whereby a physician successfully retrieves a clot from the neurovasculature of an ischaemic stroke patient, restoring sufficient blood flow to the brain, in a single attempt—are a fairly recent emergence. Such debates were “not even on the map” 25 years ago, he notes. This is due to “incremental changes” in neurointerventional technologies that have seen thrombectomies shift from potentially taking up to six hours, in the past, to certain cases being completed in a matter of only minutes, today.
“There was a lot of focus on time from the start in the stroke field, and [the fact that] we have to treat stroke as fast as possible,” Liebeskind says. “Therefore, conceptually, we were measuring the success of the procedure by how long it took—because we knew that time may be important in terms of [preventing] ischaemic injury.”
However, while the ‘time is brain’ mantra still rings true, the significance of the number of retrieval attempts has evolved and, over time, come to the fore. Liebeskind believes this is an “interesting concept”, not least because it is “truly an interaction between the patient-specific imaging features, operator/interventional skills, and the device they are using”.
An imperfect metric
There are a number of case- and imaging-specific factors that can contribute to the likelihood of achieving success at the first pass, he continues. The occlusion location; clot size and composition; and extent of downstream ischaemic injury; are all variables that contribute to a complex overall picture and cannot be fully accounted for when measuring success via the number of retrieval attempts alone. The fact it is not yet understood precisely what bearing many of these elements have on the likelihood of an early, successful retrieval only serves to muddy things even further.
As Liebeskind also points out, the operator themself can have a discernible impact on how many passes are needed to restore blood flow—not only because, as prior studies have confirmed, their level of skill and experience can improve the likelihood of a positive outcome, but because how ‘aggressive’ their approach is, and the devices and techniques they opt for, are relevant too.
Another current problem with the notion of the first-pass effect stems directly from its nomenclature; how does one define a single pass? Liebeskind queries what constitutes “a go” in mechanical thrombectomy—instances where a device is not fully deployed, or is used for an off-label indication, or where reperfusion is attempted prior to the device being repositioned slightly for a very quick ‘second pass’, are all examples of cases in which the definition of a single pass becomes murky.
Aspiration catheters and stent retrievers—the two most prevalent singular devices used in stroke thrombectomy treatments—are both affected by this issue. One ‘pass’ with the former could theoretically involve aspirating for two minutes, or a full 20 minutes, at a time. It is also debatable whether advancing a catheter without turning on the aspiration pump counts as a pass, because such instances fall outside of its intended, on-label use. Even for the more established stent retriever, the definition of a single pass is far from perfect, according to Liebeskind.
“There are interventionists that deploy a stent retriever for 30 seconds, and there are interventionists that deploy for 10 minutes—it is not standardly done,” he says. “And I am excusing myself from being humble here, but I have seen more angiographies of acute ischaemic stroke than anybody in history, and there is extreme variability in what is being done on routine, daily basis. Every case is a little different, and [a thrombectomy] is truly an art in the sense that it is not a technical procedure that you can automate to achieve the same results each time.”
Yet another relevant variable is the way in which technical success is subsequently measured, regardless of how much time or how many passes it takes. Recanalisation refers to the physical reopening of the blood vessel; reperfusion pertains to the actual restoration of downstream blood flow. And, while measuring the latter via eTICI scoring has generally become the gold standard following publication of five landmark stroke trials in 2015, even this system is very broad and leaves the precise definition of ‘success’ open to interpretation.
This is ultimately a “philosophical debate”, as much as anything else, Liebeskind feels, and is one of many reasons why something as ostensibly simple as measuring the number of attempts required to complete a thrombectomy can be “extremely complicated”. He also recalls that, in the wake of the ARISE II study—which saw Osama Zaidat (Mercy Health St Vincent Medical Center, Toledo, USA) et al first coin the first-pass effect—many were quick to jocularly point out that the terminology itself is an imperfect ‘first pass’ at defining the concept.
First pass or three strikes?
“[Another thing] that has come up, which is very important, is the ‘third-pass effect’—which nobody really talks about, [but] is the standard in clinical trials,” Liebeskind says.
The origin of this concept of three passes, or three ‘interventional steps’, can actually be traced back to work done by Liebeskind and his own group at UCLA involving the very first endovascular thrombectomy device: the Merci retriever.
The traditional paradigm for doing clinical trials has to be revisited—they have to be rethought in terms of what we do in practice, and how we declare things as being a success or not.
“Decades ago now, once that device [became] available, it was realised that there was a certain threshold of passes, or attempts, of device use,” he recalls, “and it seemed like three steps would define a lot of what came after in terms of outcomes—including the final angiographic result, as well as whether you ever achieved success, and then also the clinical outcomes. That is not always the case, but it became a simplistic yet practical approach because, intuitively, three tries at something is a good set of attempts. To use a baseball analogy, you get three strikes before you are out.”
While the mantle of ‘gold standard’ in stent retrieval has since been assumed by the Solitaire (Medtronic) and Trevo (Stryker) devices, the rule-of-three ethos has endured into the present day, according to Liebeskind. Many neurovascular companies currently extol the virtues of their thrombectomy products based on an ability to achieve recanalisation or reperfusion with a single pass, but three attempts is still the more commonly used baseline in clinical research.
“Each interventional step is a procedure unto itself,” Liebeskind opines, “and what is involved in clinical trials is this sequential evaluation of success. So, for regulatory purposes, people have employed the practical approach of saying ‘three is enough’—you get three strikes at the plate, and then you are out.”
Clinical versus technical outcomes
Of course, reperfusion, recanalisation, or any other ‘technical’ outcome measure, is only half of the story in stroke thrombectomy—and, fundamentally, also the far less important half. Clinical outcomes, and the eventual impact of the treatment on the patient’s health, are of paramount significance to all other considerations.
“As the family, you would never want to have an interventionist come out of an operating suite and say, ‘the surgery was a tremendous success, but your father died’,” Liebeskind states. “That is not the success you want. And, in turn, the same concept has been applied to the first-pass effect. No interventionist is going to plant their flag, and beat their chest, and say, ‘we achieved a first-pass effect, but then the rest was a failure’.”
However, while any major clinical trial to date assessing the potential benefits of mechanical thrombectomy—as with the ‘big five’ studies in 2015, as well as more recent investigations involving basilar artery occlusions and larger infarct cores—has used clinical outcomes to measure success, there does appear to be a growing role for what Liebeskind dubs a ‘per-pass methodology’ in analyses of specific devices.
“The PROST trial, which was recently run by Phenox in the USA and Europe using the Preset device, was actually the first randomised controlled trial [RCT] comparing a stent retriever for mechanical thrombectomy to an existing device where the eTICI evaluation was used,” he avers.
In his capacity as professor of Neurology and director of the UCLA Stroke Center, Liebeskind led the core lab that provided centralised adjudication for PROST—taking on the responsibility of ensuring the procedural results reported by each physician at each centre were consistent, and could be verified on angiographic imaging. Here, Liebeskind also highlights the adoption of a per-pass methodology in the SUMMIT MAX RCT (Route 92 Medical) as another example of technical outcomes gaining more credence in contemporary thrombectomy studies.
Outlining one favourable characteristic of such success measures, he asserts that, when an interventionist is treating a patient—particularly in urgent cases, as all ischaemic strokes are—they are less likely to speculate about modified Rankin Scale (mRS) scores and 90-day functional outcomes when deliberating over which specific technique or device to use. This is because there are myriad of variables that can impact clinical results in the weeks and months following the procedure, making it nye-on impossible to determine the role played by the procedure itself in these longer-term developments. The end goal of physically removing a clot and restoring blood flow, on the other hand, can effectively be ascertained in real time.
“Put simply, devices cause a radical change in the pathophysiology—in blood flow, in this case—that is visible, and demonstrable, immediately,” Liebeskind says. “A mechanical thrombectomy is very simplistic in that you can demonstrate, via an angiography, exactly what was done and at what timepoint. Angiography tells the story of both the therapeutic intervention as well as the subsequent diagnostic information.”
The role of core labs
“From a careful and thoughtful standpoint, I think the concept of a first-pass effect is useful—it is nice, from a marketing standpoint, to say ‘in a clinical study, […] we achieved a first-pass effect in this number of cases’,” he continues, but asserts that the more objective, standardised adjudication provided by a core lab is key in substantiating these types of claims.
Liebeskind also highlights the somewhat flawed approach US and European regulators currently take to post-market medical device surveillance. The upshot on both sides of the Atlantic and Pacific is that, “right now, around the world, stroke device usage is self-reported in routine clinical practice,” he adds, “and that is a crazy system to trust and believe in”.
“The traditional paradigm for doing clinical trials has to be revisited—they have to be rethought in terms of what we do in practice, and how we declare things as being a success or not,” Liebeskind argues.
In concluding, he puts forward what he believes is a “very potent, and much more efficient” solution; proffering that the systematic evaluations delivered by centralised, core-lab adjudication—such as in the PROST RCT—could be employed more widely in post-market surveillance studies. In Liebeskind’s view, this would serve to create rigorous and objective thrombectomy datasets, against which reperfusion pass rates and other outcomes with novel products could be assessed, alleviating the “extremely high” cost burden and other imperfections carried by today’s comparative device trials.
“You cannot market your device for a treatment indication in the USA—and, probably, anywhere else in the world—if you have not done a randomised, parallel-arm study,” Liebeskind adds. “Will a single-arm vehicle get you through the regulatory approval [processes] in stroke? Probably not, and maybe never, but the way to get around that is to reengineer how you do a randomised or parallel-arm study.
“Your parallel arm does not need to be a treatment randomly assigned to the next prospective patient entered sequentially into a trial; you can use contemporaneous data from patients treated on the same day, at various sites, with approved devices. That becomes your comparator. And that is why post-market surveillance is so important, and why there needs to be a mechanism for doing it.”