Transcript

MIC CAVAZZINI: Welcome to Pomegranate Journal Club. I’m Mic Cavazzini for the Royal Australasian College of Physicians.

For acute ischaemic stroke due to large vessel occlusion the gold standard intervention is mechanical thrombectomy. It’s no exaggeration to say that care has been revolutionised in the last decade by imaging-guided clot-retrieval. While Australian researchers have contributed to the refinement of this approach, the expertise and technology is restricted to tertiary hospitals in the capital cities.

Outside of these centres, thrombolytic treatment is often the first option, and for almost 30 years the go-to agent has been alteplase. This is an analog of the human tissue plasminogen activator which unleashes circulating plasmin to get to work dissolving fibrin blood clots. For many years it was assumed that alteplase should be administered within 3 hours of symptom onset and it was a big deal when in 2008, research showed that that window could be safely broadened out to four and a half hours.

Today’s guest authors have been pushing the envelope much further in an effort to help the more than two thirds of stroke patients who present after that threshold. In a paper published in 2019 they demonstrated that patients receiving alteplase between four and half and nine hours still benefited from thrombolysis, so long as salvageable brain tissue was detected via perfusion imaging.

This has been recognised in the Australian and New Zealand Clinical Guidelines for Stroke Management, which since 2022 have been maintained as a living document. Indeed, best practice may be due for another update given the findings we’re discussing today, published in the July edition of the New England Journal of Medicine.

The paper reveals the outcomes of a clinical trial of a relatively novel agent known as tenecteplase. This was administered as far as 24 hours out from an internal carotid or middle cerebral artery occlusion across scores of participating healthcare centres.

To conduct the interview with me I invited Dr Duncan Austin, who is a kind member of the podcast editorial group. He’s a consultant neurologist at the Alfred hospital in Melbourne, and an Affiliate Researcher at Monash University with an interest in stroke recovery.

DUNCAN AUSTIN: Hello to everyone listening. My name is Duncan, Austin. I'm a stroke neurologist in Melbourne, Australia. It's an honour to be invited into the pomegranate virtual studio, today. I'm talking to two legends of stroke research, Professors Bruce Campbell and Mark Parsons. So, before we get started, perhaps you can just give us a two line introduction to your background in stroke research and stroke medicine. Bruce, you were the second author on this paper, so perhaps you can start.

BRUCE CAMPBELL:     Sure. So, I'm Bruce Campbell, stroke neurologist at the Royal Melbourne Health Hospital and Director of Neurology there and Professor of Neurology at the University of Melbourne. I've been doing stroke research since about 2009 and Professor Parsons was my PhD supervisor back then.

MARK PARSONS:          Mind you, Bruce, I think has holds a world record for the most publications from one single PhD of at least 11, probably more than that, Bruce? But yes, I’m Mark Parsons, I’m a professor of neurology and medicine at UNSW. I’m also a consultant neurologist at Liverpool Hospital and John Hunter Hospital in New South Wales.  Bruce and I have worked for many, many years in the area of acute stroke clinical trials.

DUNCAN AUSTIN:        Great. So perhaps before we get stuck into this recent paper, we'll just set the scene for our general audience. Since I graduated from university, the time window which we consider it's effective to give thrombolysis has gone from three hours to four and a half, to nine hours. And then in the present study, you're looking at giving thrombolysis up to 24 hours. So perhaps you can just explain to us how we got here, and particularly how your own research has contributed to that.

BRUCE CAMPBELL:     So, with thrombolysis, the idea is to open the artery in the brain and restore blood flow and therefore save tissue that's essentially hibernating. We call that the ischaemic penumbra. And people, initially, when you block an artery, have a lot of salvageable brain, and then over time, that dies off. And so, time is critical, on average. As you said originally, the trial showed it worked within three hours, and you get a still a beneficial effect, but a smaller effect out to four and a half hours and beyond that, in unselected patients, we don't have a benefit of thrombolytic therapy. But what's been able to be shown with imaging, like Mark and I have been working on for a long time, is that if you can identify people with good natural bypass vessels, collaterals, then you can identify individuals who will benefit beyond four and a half hours. So, the average doesn't work, but in individuals, there are people who can benefit.

MARK PARSONS:          For laypeople I often describe it as, you know, if the M5 had a terrible accident, is blocked, you've got to go Canterbury Road, which is incredibly slow, but you still get there. So, the blood can still get there via bypass routes, but it gets there in a delayed fashion, so that might sustain the ischaemic brain for a period of time depending on how fast the collateral flow is. This is what perfusion imaging is all about.

The original time clinical trials only used non-contrast CT in stroke patients to exclude primary haemorrhage, which obviously you don't want to thrombolyse. So, to extend the window beyond four and a half hours, we ran a trial in Australia and New Zealand and, I think Finland.

More sophisticated CT imaging enabled us to; one, look at the arteries with CT angiography; and also what we call CT perfusion, which is where you give a bolus of intravenous contrast, and because modern scanners can rotate and cover the whole brain in a second, you can scan repeatedly as the contrast travels in and out of the brain. And then you can use some mathematical techniques to generate these pretty perfusion maps that then can give you an idea of how much tissue might have already been irreversibly injured.

And we showed that people that had what we call a favourable perfusion imaging profile, those with the relatively small infarct core or dead brain and a relatively large area of salvageable brain or penumbra, we could safely and effectively treat them with intravenous thrombolysis out to at least nine hours after witnessed onset or even beyond that, if you talk about the wake-up from unknown onset strokes with thrombolysis. A lot of patients actually present with unknown onset, so this type of imaging has enabled us to treat many more patients now because previously that was a complete exclusion.

MIC CAVAZZINI:            I think you've already answered some of my future questions, but that's interesting that it's not just about the drugs, but the imaging of the ideal candidates for this. So, this criterion for selection based on the perfusion imaging, what proportion of patients does it gate in?

BRUCE CAMPBELL:     That's a good question. More with a vessel occlusion than without, obviously, you pretty much have to have a vessel occlusion of somewhere in the arterial tree to have a perfusion mismatch. And the larger the vessel, the more likely there is to be a mismatch. In the EXTEND trial, we found about one in 10 patients in the time window had the perfusion profile. So, it's not a huge number of them, and it obviously does drop off with time, and that's why we want to treat people as early as we can, while more of them have this kind of favourable imaging. But Mark might like to comment on prevalence?

MARK PARSONS:          I guess, the population in the so-called late time window, where the proportion might be higher are those unknown-onset, particularly wake-up stroke patients many of whom actually have had the stroke around the time or just prior to awakening. And again, traditionally, they're excluded from treatment based on the standard guidelines, but the much higher proportion of those patients have a favourable profile. So, I think it's particularly helpful in that group.

MIC CAVAZZINI:            And since that EXTEND trial you’ve been involved with a trial group called TRACE to test a modified analog of the tissue plasminogen activator, and this one’s called tenecteplase. Both enzymes have been designed with recombinant DNA technology and then cultured in human cell lines. What were the theoretical selling points for tenecteplase over alteplase?

BRUCE CAMPBELL:     I think one of the most important advantages to tenecteplase is you give it as a bolus, so it goes in over 5 to 10 seconds, which is a lot more convenient than a bolus and infusion over an hour that we've done with alteplase in the past. From a theoretical point of view, it's more resistant to some of the inhibitors, like PI-1 one that take out the activity of alteplase, and it's more fibrin specific, so has less of a systemic activation, if you like.

So it's a bit more focused on the clot, a bit more resistance to inhibitors, but most importantly, longer half-life, means you can give it as a bolus and not this hour infusion, and you think about our patients, we're often shipping them between hospitals or between within the hospital to get to clot retrieval, and not having the infusion pump going during that time means that it's a lot easier and it's also a lot more guaranteed that you give the dose and you don't have these gaps between the bolus and the infusion, which muck up the pharmacokinetics of alteplase often.

DUNCAN AUSTIN:        Obviously the real question is, which one performs better clinically? And so, prior to this present study, there was the TRACE-2 trial, which I think you were both involved with as well. This was an open label phase III trial comparing tenecteplase without alteplase within the four and a half hour time window, and this found tenecteplase to be non-inferior. Can you just unpack what this concept of non-inferiority is for the non-specialists amongst us?

MARK PARSONS:          I'd love to leave that one to Bruce. Really, the only people that understand non-inferiority are  statisticians.

BRUCE CAMPBELL:     I guess normally in medicine, we like to test something that's better, so we go for superiority. But there are certain circumstances where it's more convenient, less costly, there's some other advantage, that if it's at least as good as we'd be happy. And so, with non-inferiority, we're essentially saying, as long as the confidence interval for the effect is not worse than a margin that we set, then we'll accept that it's similar, and then that's the concept behind it. You can't statistically prove something's the same, so you have to set a margin around it to say, well, it's not substantially worse than.

MIC CAVAZZINI:            Now, in in a follow up analysis to the TRACE-2 trial it was revealed that it really didn’t matter how soon you started thrombolysis. Functional recovery was equivalent comparing patients that had been binned by treatment up to 90 minutes, up to 180 minutes, or up to 270 minutes. But it’s still a big step—what  gave you the confidence to up the ante, not just to 9 hours but all the way up to 24 hours?

BRUCE CAMPBELL:     I guess one point is that once you're over that initial few hours, a lot of the time we just don't know when the stroke started, so we always take it from the time the patient was last known to be well. But a lot of them have woken up with stroke or been found on the floor when their family get home from work so, we're often dealing with a lot of uncertainty. There was some pilot data from Canada that suggested going out to 24 hours looked reasonably safe. And then we had the TIMELESS trial in the states just recently, pushing out to 24 hours before thrombectomy, and it was a neutral study overall that the safety was quite reassuring. So I guess you put those things together, and again, with the imaging selection, where we think that the amount of tissue damage is probably more important than time, particularly when you're not confident about the time, that's really what gave us the confidence, I would say.

DUNCAN AUSTIN:        Good. So obviously, you're not the only group, as you've just mentioned there, that's looking at thrombolysis up to 24 hours. In February this year, the findings were published from a North American trial called TIMELESS. In that trial, patients were treated with either tenecteplase or placebo, and then most of them went on to have a planned endovascular thrombectomy, often very shortly, after receiving thrombolysis. That was a negative result, but perhaps not surprising, given how many patients got thrombectomy, was that your impression? [see also alteplase plus thrombectomy].

BRUCE CAMPBELL:     Yeah, and also how quickly they got thrombectomy. The median time from tenecteplase being administered to arterial access for the thrombectomy was 15 minutes. So, that's not a lot of time for thrombolysis to do very much. It really was like one of those direct to thrombectomy trials in the very short time between lysis and puncture at comprehensive centres. And the effect size was actually pretty similar, it was in that three to 4 percent range that was seen in the direct trials as well. So, you'd need a much larger sample size if you're going to show that to be significantly beneficial, but it would be clinically meaningful if you could prove it, just like tenecteplase over alteplase had that kind of magnitude of benefit.

The other thing about the trials themselves is that when you've got to consent and randomize a patient before you can treat them, you're building in a delay so that there's a bit of an artificial handicap, if you like. I guess, the other thing about TIMELESS is that the safety, as I said, was impressive, the efficacy was neutral overall—although if you look at the MCA m1 subgroup, it was actually quite a strong trend to benefit. And something very odd happened with their more distal M2 occlusions, which is still in the process of being worked out. And ICA carotid occlusions, we never expect thrombolysis to have miraculous effect, really,

MARK PARSONS:          To add to that, Duncan, the main the main message from TIMELESS is that thrombectomy is an incredibly powerful treatment. In fact, the number needed to treat, even in the very late time window out to 24 hours, is two or three. So, an LVO stroke, they're large clots, natural history is awful, you know, 50 percent of patients in the olden days, when I was an intern, when we had no treatment for stroke, were left with either severe disability or died. So, I think it's one of the most powerful treatments in medicine. So, to give tenecteplase prior to thrombectomy a fair chance you need to have a longer time window. Of course, if you can do thrombectomy fast, it's unethical to delay treatment. But the reality is, in most centres around the world, there's usually a significant delay between thrombolysis and thrombectomy, partly because many patients don't present to a thrombectomy capable centre.

MIC CAVAZZINI:            Right. So, TRACE-3 the trial that we’ll finally get on to discussing, was conducted in 58 institutions across China under the guidance of under of Professor Yongjun Wang from Beijing’s Tiantan Hospital. Was the choice of China rather than Australia or the USA, simply about being able to recruit enough patients who wouldn't have access to mechanical thrombectomy?

BRUCE CAMPBELL:     Yeah. I mean, the trial was initiated by that Beijing group, and they selected the TRACE-3 centres as sites that did not have 24/7 access to thrombectomy. And, you know, there's a wide range of hospital capabilities in China. They have some very advanced hospitals which do have 24/7 thrombectomy, but this was a trial network specifically to address this question, which is relevant to a lot of the developing world, and even in our country, you know, a lot of places don't have immediate access to thrombectomy.

MARK PARSONS:          We do telestroke in most states of Australia. And, for example, I was on call last night in New South Wales, and we had a patient from four hours away up the New South Wales coast. So, there's often very long delays between lysis and thrombectomy, which is where we really think thrombolysis still has a role, even if patients are getting to thrombectomy.

Actually, if you look at the United Kingdom, access to thrombectomy, there is very limited. There's still no 24 hour, 24/7 center in Scotland, Bruce, correct me if I'm wrong, and only a couple in in England. And if you look at countries like Brazil there's only one as far as I know. So, many densely populated countries have minimal to no access to thrombectomy. So, IV thrombolysis will remain the predominant treatment for LVO stroke, I think, for well, possibly forever, but certainly for the next couple of decades.

DUNCAN AUSTIN:        So, in the TRACE-3 trial that we're talking about today we recruited over 500 patients, all with large vascular stroke affecting either the internal carotid or middle cerebral artery. We know that strokes can affect other territories as well. Why was that considered important?

BRUCE CAMPBELL:     Well, I guess the posterior circulation, where you got vertebral basilar artery ischaemia, is clinically quite heterogeneous and different to the answer to anterior circulation. So we've generally divided those two up. And there is evidence now for thrombectomy, up to 24 hours for basilar arteries, as well. But in general, we've had an anterior circulation focus in the late window with thrombolysis, it's where we have a bit more data for our perfusion imaging as well. So, there's less validation of perfusion and working out what's penumbral in the posterior circulation.

DUNCAN AUSTIN:        So, this is probably a good point, too good time to spell out the primary outcome measure for the trial, which is absence of disability at 90 days. Which sounds very simple, but it's relying on scoring using the modified Rankin Scale and then an ordinal logistic regression to analyse the distribution of Rankin scores. This has become a fairly standard outcome measure in 21st century stroke trials, but perhaps again, you could explain to our non specialist audience how this works and what the significance of it is.

MARK PARSONS:          I can answer. So, it's an ordinal scale; zero, no disability, one, slight disability. Two, a little bit more. Three, moderate, four, pretty significant disability and you're dependent on others for your activities of daily living, and then five you’re completely dependent, typically a nursing home patient. And then later on, they modified it to out of six, which is death, which is a pretty hard outcome, hard to argue with that. Of course, you can blur sometimes the margins between a one and a two or two and a three, but it's basically a measure of day to day function, very heavily weighted by how you if you can walk unaided. That's probably the main bias of the Rankin score, but there's a pretty good training course that people have to do before they can be certified.

I presume there's a Chinese version, because the one that Bruce and I do is is narrated by an old friend of ours called Kennedy Lees, who's a Glaswegian. His accent is quite hard to follow, and I would suspect from a non-primary language English speaker being impossible. I don't know if you’ll edit this out, but I once travelled across the US with Professor Lees when I was a stroke fellow and for some reason, every time he scanned his passport he got pulled aside for tertiary screening. I think it was just after September 11, 2001. And so they started interrogating him. And of course, the Americans couldn't understand his accent at all, so I had to translate for him, which he was quite irritated by, but I found quite amusing.

MIC CAVAZZINI:            I had one of these AI transcribing softwares that had real trouble with one of my Yorkshire guests once. But just to be clear, you've described the Rankin scale and the primary outcome, absence of disability was a score of zero or one, and you're reporting the proportion of patients that had recovered to that state.

BRUCE CAMPBELL:     Yeah. And a Rankin of two is when you're not doing an activity post stroke that you were doing regularly before stroke. So that's really the key. You're back to all your usual activities.

MIC CAVAZZINI:            Thanks. That really puts it into context. Duncan?

DUNCAN AUSTIN:        So, in the TRACE-2 trial, they were comparing tenecteplase with alteplase, whereas in the trace three trial, the comparator group only received anti-platelet therapy. What was the reason behind that difference?

BRUCE CAMPBELL:     Just the time window. So, although in Australia, we've got EXTEND the guidelines, and they have it in Europe as well, it hasn't hit North America and hasn't hit the Chinese guidelines, so there's just no proven treatment for them beyond four and a half hours.

DUNCAN AUSTIN:        But the TIMELESS trial was basically comparing thrombolysis with standard of care, but they used a placebo in the control arm. But we didn't, with this trial, any reason for that?

BRUCE CAMPBELL:     I think it was just practicality. One of the problems with TIMELESS was that they couldn't get out of the comprehensive centres and go to the primary stroke centres that they wanted to recruit, because having a placebo controlled trial and the need for a clinical trials pharmacy was just too logistically complicated for non-high level academic centres. So you have to be a little pragmatic with these things.

MARK PARSONS:          We routinely do that sort of trial as well, Duncan, so called probe trial, blinded endpoint. So,we usually, typically use a central team who don't know the patient, don't know what treatment they've received, and they go through a set questionnaire to score the Rankin, typically on the telephone. So, whilst it's not as scientifically ideal as a completely double blinded trial, it's definitely a pragmatic option for large multicentre trials, particularly when they're relatively not resourced for those. And the cost too. I mean, obviously, if you're going to have a placebo-controlled trial, it substantially increases the cost of study drug as well.

MIC CAVAZZINI:            I think, if nothing else, you guys have demonstrated when people complain, “Why don't they just do another gold standard trial,” that there are so many ducks to line up.

MARK PARSONS:          It's easy for it's easy to criticize from the sidelines.

DUNCAN AUSTIN:        Great. So perhaps we'll talk a little bit more about patient selection. So, the study used the iStroke automated software to calculate the volume of ischaemic brain so perhaps it can explain how we define salvageable brain tissue in this context. And then for the Australian neurologists in the audience, how does this iStroke software differ from packages we're using here in Australia?

BRUCE CAMPBELL:     So with CT perfusion, the key way we identify irreversibly injured tissue is low cerebral blood flow, and there's a threshold of less than 30% of normal brain, which is widely used in the literature. That came from our group in Australia, actually. And it's probabilistic, and it's an estimate, but it works pretty well on average, in most patients. So, if you've got less than 30% cerebral blood flow of normal brain, we'll call that irreversibly injured. And if you've got delayed arrival of flow more than six seconds, we'll call that at risk. And the difference between it is the salvageable tissue. So, there's a number of software packages that can do that can do that, iStroke is one, that was developed in China and is quite economical for them to use. In the rest of world, there's software like RAPID or Viz or MIstar, Brainomix. There's a whole stack of packages that are commercially available, and most of them use the same thresholds, but the way they calculate it does differ a bit, and the way they segment differs a bit. But the general principle is the same of what's quite severely delayed versus what's very severely reduced in flow. And that's what you're looking for, that mismatch.

MIC CAVAZZINI: So, let’s finally get into some of the outcomes. The median interval between symptom onset and delivery of tenecteplase was 12 hours 24 minutes. And it was 24 minutes later in that group getting oral anti-platelet therapy. Hardly a critical delay and Duncan pointed out that might have been related to challenges in swallowing. But in regards to the primary endpoint, 33% of those receiving tenecteplase were free of disability at 90 days, as compared with 24.2% in the standard-treatment group. Can you give us a sense of how significant this result was, and what a relative rate of 1.37 means?

BRUCE CAMPBELL:     Yeah, so that roughly 10% difference is the difference in modified Rankin-01 that we see with alteplase versus placebo in the first three hours, and it's double what you see in the three to four and a half hour window. So, in these imaging selected patients, you're getting the same effect as you do early on in all comers, which is highly clinically significant.

DUNCAN AUSTIN:        So, a secondary outcome we looked at was major neurological improvement at 72 hours using the National Institute of Health Stroke Scale. This was four times higher in the tenecteplase place group versus the best medical therapy group. But our Rankin score was collected at 90 days, which represents very different stages of biological recovery. So, what do you take from these two results?

BRUCE CAMPBELL:     I think it's always good to see that there is major neurological improvement early on, it's a good sign of reperfusion occurring. But we are most interested in functional status. I guess the downside of a three month Rankin is other intercurrent events can occur in between. This study occurred during the COVID pandemic, so some of the patients would have got covid and had other intercurrent events. But that's the same with any stroke trial. So, we're most interested in 90 day functional outcome.

MIC CAVAZZINI:            For a non-neurologist, what does that NIH stroke score tell you, if you wanted to see and improvement of eight points within 3 days of onset.

MARK PARSONS:          A bit like the Rankin scale, it's not quite ordinal. So, it's a score from where zero is normal, and you accumulate points for more severe neurological deficit. It's a bit weighted towards language, and particularly in aphasic patients, you'll get a much higher score. So, one of the biases of the score, is that it gives you generally a much higher score for a dominant, typically left hemisphere MCA stroke compared to right MCA stroke. And partly that's the weighting, but also that people find it harder to diagnose and neglect an inattention, as opposed to usually a language disturbance is pretty obvious. So, for example, for a complete no arm movement, you'll score a four and or leg as well. For a severe aphasia, you would score a three. There are various orientation questions. So, we typically would say a moderate, moderately severe strokes score above 10, and then a severe stroke’s above 15, typically most large vessel occlusion strokes are above 10, but they're particularly right hemisphere strokes might be lower. So, to improve by a score of 8 at 24 hours is—that's a pretty high benchmark, and that's really quite rare to occur without some form of reperfusion treatment. And as Bruce said, it's basically a marker of effective recanalization with the reperfusion therapy. But traditionally FDA and other drug regulatory agencies use a three month outcome, and there's been many, many debates over the years about whether we can get away with an earlier outcome. But it's not really accepted as the definitive endpoint.

MIC CAVAZZINI:            Let's now talk about adverse effects. Alteplase had always been associated with some degree of haemorrhage, both in the brain and elsewhere and it was hoped that tenecteplase would have reduce ths risk. That didn’t seem to come about in the TRACE-2 trial, both treatments were associated with a rate of symptomatic intracranial haemorrhage of 2%. In this study it was actually 3% for the tenecteplase group and under 1% for the anti-platelet therapy control. Not a deal-breaker, but has this been a bit of a disappointment?

BRUCE CAMPBELL:     Not really. You know, there were some observational studies that suggested that tenecteplase would have a lower symptomatic haemorrhage rate. One big, large, real world study from the US in particular, but most of the trials have had pretty consistent rates of symptomatic haemorrhage. And the important thing is that with the Rankin scale, there is a net benefit. So symptomatic haemorrhage can cause death, but if you look at the overall mortality rate, there was no difference, and there's no offset, if you like, at the five and six end of the Rankin scale that worries you about the benefit at the 0, 1 end of the scale.

The other point about symptomatic haemorrhage in this study and in studies in general, was that the non-contrast, CT appearance is really important. When we went through the cases that had had a symptomatic haemorrhage, quite a few of them had quite extensive hypodensity, which would make Mark and I pretty anxious about giving thrombolytics, but that's an education piece, that when we go out beyond four and a half hours, in particular, you've got to look really carefully at that noncon CT, regardless of what the perfusion is telling you, because things change. And if there's a nasty, hypodense infarct where the blood brain barrier is broken down already, and that's why it's got increased water content, and it's dark on the CT, then they're the ones that bleed.

MARK PARSONS:          Yeah, the reason that we've pursued down more advanced imaging pathways that those changes generally lag behind what's really, actually happening in the tissue. But it gives you a very good measure of the tissue clock. So, if you've got a patient, particularly with uncertain onset time, and they've already got extensive changes on non-contrast CT, that indicates it's likely to be at least six hours ago that they had their stroke, if not longer. So, that's more an issue in that late time window, slash uncertain onset patient. Typically, if we see somebody in the first, say, three hours, their non-contrast CT will be relatively pristine. They might have subtle changes, but it definitely makes you a little bit anxious when you see extensive changes on non-contrast CT.

BRUCE CAMPBELL:     It might be worth making that point that our perfusion imaging is a snapshot, so it tells you what's happening with the blood flow at the time you take the scan. And in the first few hours, that's usually reflective of what's been happening for the whole time the strokes there. But once you get out to 18 hours, there may have been much worse collaterals early on than there are now, and things might have occurred in the tissue that are not reflected by the perfusion state now. So, you need to take all the information together in conjunction. It was a big issue in the SELECT-2 trial I was involved with, with large core. When we went back, people might have quite a small CBF CT, perfusion core, but you look at the non-con and clearly the whole territory is infarcted based on hypodensity, and that kind of mismatch is something we need to take into account with our imaging.

MARK PARSONS:          Yeah, the perfusion imaging is great, but it must be—and getting back to AI, the maps are, I guess, a very basic form of artificial intelligence that give you a core volume and a penumbra volume, but you always need to look at the non-contrast CT as well. Because, for example, if you've got a very proximal say origin of the MCA m1 clot initially and in the collateral flow to the deep MCA territory, basal ganglia, et cetera, is often relatively inadequate, so that area often infarcts quickly, whereas the cortex stays alive longer because of those leptomeningeal collaterals. But if, for example, you scan the patient maybe six hours after onset, the clot might have actually migrated like a salmon downstream. Does salmon migrate downstream or upstream, I always forget?

MIC CAVAZZINI: Upstream..

MARK PARSONS: Do trout migrate migrate downstream?

MIC CAVAZZINI: They must go both ways to be fair.

MARK PARSONS: Anyway, whatever fish migrates downstream, the clot does. So, if you do perfusion imaging at a later time point, you might only see a more distal MCA perfusion deficit, which will dramatically have underestimated the original lesion. So, you have to take everything into account. You know, some people get seduced by these beautiful colour core-penumbra maps, and forget to look at the non-contrast CT as well.

MIC CAVAZZINI:            I think we’re going to have to release a trainees version of the podcast that's an hour and a half long. That's really interesting, though, and well explained. Go, Duncan.

DUNCAN AUSTIN:        So, we know that with thrombolysis, small amounts of haemorrhage afterwards are not uncommon, but don't necessarily alter the natural history. But in this trial, in common with previous stroke trials, we've defined symptomatic haemorrhage as an increase in the National Institute of Health Stroke Scale of four points or more, which, as you've alluded to, is equivalent to complete paralysis of one limb. Do you wonder if we're using too high threshold for definition of symptomatic haemorrhage?

BRUCE CAMPBELL:     Personally, I don't. I think you know, the ECAS definition that was used in this study, where it's just the four point increase with any type of haemorrhage in the brain, is possibly a little too generous. You know, there are lots of people who deteriorate for other reasons and have a little bit of blood in the brain where the blood has nothing to do with the deterioration, they've got a large infarct or a swelling or whatever. So the SITS definition that we've often used in our trials, where you have to have a large lump of blood that's taking up space that would otherwise have been occupied by normal brain, that's a criteria that I favour, personally, to ensure that the blood is actually likely to be the cause of the deterioration.

MARK PARSONS:          I think, Duncan going back to the original NINDS trial in the mid-90s that was done in the US that that led to alteplase being started to being used for stroke. They their definition of symptomatic haemorrhage was basically any haemorrhage that was deemed to be might have caused deterioration. So, when we go back and look at some of these cases, they defined as symptomatic haemorrhage. They were massive, hemispheric infarcts with a tiny dot of blood. And clearly the deterioration was nothing to do with the blood, so Bruce is spot on. You need to be confident that the haemorrhage is the cause of the deterioration. Of course, sometimes it's a bit difficult to be sure.

And now with thrombectomy, which wasn't used in this study, we're seeing a lot more small subarachnoid haemorrhages around the area of where they pull the clot out. Because, without going into too much detail, the main way of removing the clot these days is with the stent retriever. So, they poke the stent downstream from the clot then open it up and drag the clot out so it actually can damage the endothelium of the vessels quite a lot, and can lead to subarachnoid haemorrhages as a result of the thrombectomy. Which I guess really doesn't have that much to do with having thrombolysis beforehand or not.

BRUCE CAMPBELL:     And most of those subarachnoids are asymptomatic and seem to wash away and don't have any consequence. Baut there's certainly a lot of traction on the vessels when they do try to retrieve  a clot, and sometimes that does lead to a subararach.

MARK PARSONS:          Yeah. And my understanding, from talking to neurosurgeons that once you get beyond out to the distal M2 the vessels look quite translucent, then the walls are pretty thin. They're pretty easy to perforate.

MIC CAVAZZINI:            Sticking with thrombolysis, without doing another podcast on thrombectomy—so the tenecteplase is tolerable—the rate of this adverse event is tolerable as it was with alteplase. There was a post hoc analysis of the LACE-2 trial where it was observed that patients over 80 years of age were shown to be at no greater risk than younger patients. I'm not sure what the current guidelines are for administering alteplase to the elderly. Are we going to see some changes in response to these findings?

BRUCE CAMPBELL:     We don't have an upper limit for thrombectomy or thrombolysis. It's really about functional status and every time people have looked at treatment effect, the elderly benefit at least as much as the younger ones. So, obviously age is a prognostic factor, as it is with any condition in medicine, but if you treat people and get the artery open, they do a lot better than if you leave it blocked. And that's at least as much of a benefit in the elderly as it is in the young people.

MARK PARSONS:          There's probably a slightly higher risk of bleeding in people over 80, but doesn't negate the benefit great.

DUNCAN AUSTIN:        Great, so putting all these findings together, what do you think the implications are going to be for regional centres in Australia that don't have access to thrombectomy on site, if a patient presents the Royal Darwin hospital with a large artery occlusion, should they have alteplase or tenecteplase or neither and be flown down to Adelaide for clot retrieval four hours away.

BRUCE CAMPBELL:     I think it's optimistic that you could get them from Darwin to Adelaide in four hours. To be honest, a medical transfer always takes a lot longer than a commercial flight, unfortunately. But yeah, I think these data are very encouraging that we should be giving tenecteplase to patients with large vessel occlusion with the imaging profile. And we're fortunate in Australia that pretty much all of our stroke centres do perfusion imaging. That's different, to say, the United States, where there's still some difficulty in obtaining this imaging, for whatever reason. So, I think it should be implemented. We've already got guidelines out to nine hours and nine hours from the midpoint of sleep in the patients with wake-up stroke. And nine hours from the midpoint of sleep is often about 16 hours from the last-known well. So, to go from 16 hours with EXTEND to 24 hours with trace three is actually not a big leap. And I think it's I think the evidence is there now.

DUNCAN AUSTIN:        So do you think it's likely that the Australian guidelines may change in the near future in response to this?

BRUCE CAMPBELL:     I think that's very likely.

MARK PARSONS:          I'd be interested to see if the American guidelines change based on a Chinese trial.

MIC CAVAZZINI:            Not under Donald Trump.

MARK PARSONS:          As you probably know, we also recently completed a sub four and a half hour study of tenecteplase versus alteplase called TASTE. And it also showed that tenecteplase was non inferior to alteplase for a Rankin 0 to 1 at three months, which basically means almost full recovery, or near full recovery. And if you look at the overall effect size, I think 59% of patients receiving tenecteplase achieved excellent recovery, and alteplase was 56%. So, there was actually a numerical 3% increase with tenecteplase, but that was not superior, but it was shown to be non-inferior.

But ironically enough, if you add all of the tenecteplase versus alteplase trials together, which includes several Bruce and I have done in the past, when we completed TASTE we finally tipped the study level meta-analysis of tenecteplase versus alteplase to superiority. So, there was an overall 4% effect size, so 4% increase in Rankin 0 to 1 at three months, which basically means if you treat 25 patients with tenecteplase instead of alteplase, one more, will get to an excellent recovery.

MIC CAVAZZINI:            I think that's a way that most clinicians can latch on to.

MARK PARSONS:          Yeah. And I think that that's important on a, on a, on a global scale, if you're treating, you know, 1000s patients, yeah.

DUNCAN AUSTIN:        So, research into and the use of thrombolysis has come a long way in quite a short space of time. What are the big questions that remain to be answered around intravenous thrombolysis, and what trials can we expect to see coming through in the next couple of years?

BRUCE CAMPBELL:     The data now do indicate that tenecteplase is superior to alteplase in the disability outcomes and obviously in the convenience as well. So, I think we will see a switch to tenecteplase as soon as we can get our hands on it. There is a global shortage outside North America at the moment, but when supplies return, I think people will just switch wholesale to tenecteplase throughout the 24 hour window.

The next challenge is really to have a step up in our reperfusion success. Fibrinoytics, which is what we've already used, is one component of the clot, fibrin, but there are other components. There's DNA, which I'm interested in, and there's von Willebrands factor and there's platelets. And there are trials going on at the moment looking at targeting each of those components and hopefully in the future in combination, and really hoping that we could use intravenous strategies to get the artery open before you even need to get to a thrombectomy. But that's a little way off still.

MARK PARSONS:          Obviously the main concern depending on what adjunctive agent you use, particularly if it's has an effect on platelets, is whether there's additional risk of bleeding, but there are various agents that show promise. I guess, probably, the other possible trial to do worldwide is whether we can safely expand the treatment window to 24 hours without the fancy perfusion imaging. Because, for example, in many countries, say, in India, most patients only get a non-contrast CT, maybe a CTA. So, can we do it safely to 24 hours without the whizz-bang perfusion imaging? There is some data to suggest you can, but the outcomes aren't as good as if you select with perfusion imaging.

MIC CAVAZZINI:            Many thanks to Mark Parsons and Bruce Campbell for sharing their expertise for this episode of Pomegranate Health. And thanks also to our guest host, Duncan Austin, and all the members of the podcast review group who provided feedback.

You can hear more from Bruce Campbell on diagnosis and management of acute stroke at the College Learning Series. His is just one of thirty neurology lectures based on the basic training curriculum that can be found via the landing page elearning.racp.edu.au.

There’s a transcript of today’s conversation with links to the NEJM paper and all other literature cited via the address racp.edu.au/podcast. Just click on the thumbnail for this episode, and from there you can also follow a link that prefills your MyCPD record.

I’d love to hear your feedback and suggestions, whether it’s for another Journal Club, Case Report, Guest Lecture or any other theme on the culture of medicine. Just send an email to podcast@racp.edu.au.

This podcast was produced on the lands of the Gadigal clans of the Yura nation. I pay respect to those who have watched over this country for thousands of years. I’m Mic Cavazzini. Thanks for listening.