Transcript

MIC CAVAZZINI:                Okay, so welcome to [IMJ On-Air] from the Pomegranate Studios. The Internal Medicine Journal is the flagship publication of the Royal Australasian College of Physicians. I'm Mic Cavazzini and I'm very happy to welcome Dr Duncan Austin back to the driver's seat. Duncan is one of the Pomegranate reviewers when he's not working as a stroke neurologist at Alfred Health in Melbourne, or a day clinic in Ballarat today. Listeners should bookmark his episode from last October on thrombolysis. But today we're talking about a paper from the April edition of the IMJ titled, Prevalence and Associations of Cerebral Microbleeds in an Australian Memory Clinic Cohort. Duncan, tell us why you were so keen to talk about this study.

DUNCAN AUSTIN:            Thanks Mic. So cerebral microbleeds are often an asymptomatic finding on MRI and can be due to various different aetiologies, one of which is the accumulation of amyloid-beta peptide. Now, hot off the press is a recent approval by the TGA of the amyloid-beta busting drug, donanemab, for the indication of Alzheimer's disease, which is the first disease-modifying therapy available for this condition in Australia. So, that's a really exciting development. But the treatment comes with an increased risk of bleeding and of microbleeds, which is a concerning prospect that's going to need to be panned out.

We also see these cerebral microbleeds in a condition called cerebral amyloid angiopathy, which is a leading cause of haemorrhagic stroke in older people. Like Alzheimer's disease, this condition is seen secondary to the accumulation of amyloid-beta peptide. We don't really fully understand the link between these two conditions or why amyloid-beta behaves differently in different people. So as a stroke neurologist, I'm fascinated to understand this pathology better.

So prevalent studies looking at cerebral microbleeds have typically been conducted in specific cohorts, like a post-stroke population, or people with a given neurodegenerative condition, like Alzheimer's. By contrast, the memory clinic population provides a pretty broad, not-too-biased sample that allows us to assess the correlations between demographic factors, cognitive scores, and specific aetiologies. Now studies in the general elderly population are probably too mixed because they contain people both who are likely to develop a cognitive impairment later in life and some people who will never will. So, this present study is an enriched sample which will hopefully better elaborate the specific relationship between microbleeds and different pathologies.

MIC CAVAZZINI:                Okay, now let's bring in the principal investigator behind this study, Associate Professor Paul Yates, who also sits on the IMJ editorial board now. Thanks for joining us, Paul. Can you please give us your two-line bio?

PAUL YATES:       Thanks. Yeah, so I'm Paul. I am a geriatrician at Austin Health and Honorary Associate Professor with the University of Melbourne. My clinical work is mostly these days in the memory clinic, but I also, we run a very large research and trials unit at the Austin. Do a lot of research in Alzheimer's and other forms of dementia, particularly with some of the new disease-modifying therapies such as donanemab and lecanemab and others. So yeah, we've been working with these medications for a while and trying to understand a bit more about our clinical cohorts as well.

MIC CAVAZZINI:                And you were one of eight authors on this study. Do you want to give a shout out to the most important players?

PAUL YATES:       Yeah, so the study actually began from a medical student project quite a few years ago and then has been revised over time. We've had some great contributions, so Scott Wigley, the first author, Ross Cody and Libby Galante, who did a lot of the microbleed rating and in particular, Nawaf Yassi, who's a stroke neurologist as well. He and I have collaborated quite a bit on a few microbleed and cerebrovascular disease and dementia related projects.

MIC CAVAZZINI:                Take it away, Duncan.

DUNCAN AUSTIN:            Thanks Paul. So, cerebral microbleeds are defined radiologically as between 2 and 10 mm areas of signal void on susceptibility-weighted MRI. From a stroke point of view, are two main pathophysiological processes underlying them. One is the thickening of vessel walls as a result of uncontrolled hypertension, which we call hypertensive arteriopathy. And then the other one is this cerebral amyloid angiopathy, or CAA, characterized by the buildup of amyloid-beta peptides in the cerebral blood vessels. But that's not the only difference, so can you explain the anatomical preference for these two types of microbleeds, why we see them in different areas?

PAUL YATES:       Yeah, so we tend to see a lot more cortical, subcortical or lobar microbleeds with the amyloid angiopathy group and we often will see in sort of combinations some evidence of subarachnoid bleeding as well or sort of cortical superficial siderosis in this group in particular. And as you say, the hypertensive arteriopathy tends to be seen in the deeper structures, the basal ganglia, cerebellum at times as well. So, generally people use this topography to infer the aetiology, which is probably true for the majority of cases, but without histopathological validation, we can't say with a hundred percent certainty.

DUNCAN AUSTIN:            And you've alluded there to the difference between cortical and a subcortical pattern of microbleeds, and we also talk about a cortical and subcortical pattern of cognitive impairment. Do we know whether the anatomical distribution of microbleeds is also represented in their functional presentation? Is this something you were hoping to demonstrate with the research?

PAUL YATES:       Yeah, that's right. So, in general, yes. So, I mean, there have been quite a number of different studies looking at the cognitive presentation of people with microbleeds. And I guess one of the issues that's difficult to tease out is the specific contribution of microbleeds above and beyond the pathologies that go along with them. So, people with deep microbleeds tend to have a lot more evidence of deep white matter vascular disease, deep white matter hyperintensities for example, and lacunar infarcts. So, these people typically are more likely to present with a subcortical vascular pattern of cognitive presentation. So cognitive slowing, reduced attention, reduced sort of processing speed and sometimes depending on where the pathology is, they may also have evidence of sort of frontal lobe dysfunction, dys-executive symptoms, rigidity, those sorts of presentations.

And not unanimously, but commonly people with lobar microbleeds may have evidence of concomitant parenchymal amyloid pathology. And so these people may also present with the cognitive presentation that goes along with that in that they might have an amnestic pattern of cognitive deficits that then may progress to other domains as well. But that's probably more common in people with Alzheimer's that also have microbleeds as opposed to people with CAA as their primary presentation.

MIC CAVAZZINI: That leads perfectly to my next question. So, most listeners will recall that Alzheimer's disease features this accumulation of amyloid-beta peptide in the brain parenchyma. And just to remind people that this peptide is a cleavage product of the amyloid precursor protein, a transmembrane receptor that's concentrated at neuronal synapses, but it's actually pretty widespread throughout the nervous system. And this enzymatic breakdown is occurring all the time, but both Alzheimer's disease and cerebral amyloid angiopathy are associated with the impaired clearance of the beta peptide. What more do we know about this impaired clearance and do we think it's the same process at play in both Alzheimer's and CAA?

PAUL YATES:       I am not an expert in that the difference between the clearance of the two, A-beta 40 versus A-beta 42, I must confess. I guess I can say that we know that each generally have a normal physiological function on a good day. And for whatever reason, in different types of Alzheimer's disease there's either overproduction or reduced clearance, certainly in the familial forms of Alzheimer's disease, there's a lot greater production of pathological forms of A-beta-42 and that they tend to clump together to form oligomers, proto-fibrils, are more likely to be the cause of a sort of a toxic ongoing cascade, an inflammatory cascade.

In later onset Alzheimer's disease, you're right that the clearance seems to be the problem and people are more likely to accumulate, again, these oligamers, protofibrils, and once they sort of reach a critical mass, they start to be sequestered into plaques. At some stage, once the quantity becomes higher, this then is associated with pathological propagation of abnormal phosphorylated forms of tau protein, which coincide most closely with the symptom progression in Alzheimer's disease. But that's not really answering your question as to why some people develop CAA and why some people develop AD dementia.

MIC CAVAZZINI: I mean, the reason I ask is you write in the paper that patients with more than four cerebromicrobleeds were commonly excluded from participation in many study protocols for clinical trials of AD, is that because of a specific interaction between the diseases or just that four bleeds would already convey some significant degree of cognitive impairment?

PAUL YATES: Yeah, that's a really good question and really topical for what's going on right now. So, people who have multiple microbleeds—and the threshold is generally four—people have multiple microbleeds are excluded from participation in the majority of amyloid-modifying clinical trials. And the approved product label for donanemab excludes people with more than two, so that's an even more stringent threshold. The reason for this is that evidence of microbleeds conveys some underlying vascular fragility and so the likelihood of having a complication from amyloid therapies is higher.

MIC CAVAZZINI: We might get into those later because Duncan's got some pointed questions on that.

DUNCAN AUSTIN:            Let's get into the significance of the findings of your study. Your cohort study was conducted in a multidisciplinary service in Melbourne that involves regular neurological and cognitive examination assessments, neuroimaging imaging where necessary to establish a consensus diagnosis at case conference. So over a span of two years you had 393 patients retrospectively found to have a suitable MRI scan for analysis. From these 107 were identified as having cerebral microbleeds an overall prevalence of 27 percent, which happily seems to be in the middle of the range marked out by prior studies in international memory clinics. There's even a suggestion from data stored in neuroimaging data banks that the prevalence in the broader age population is just as high. So, can you outline the significance of cerebral microbleeds in people who may be asymptomatic?

PAUL YATES:       Yeah, really good question. So you're right, there's been a broad range of prevalence reported in population-based or healthy volunteer cohorts. And I will just comment that, and this might be a subsequent question that I'm jumping the gun on, but the prevalence does vary depending on the MRI approach used. And so, in our study, we use susceptibility weighted imaging, which is very sensitive, much more sensitive than the gradient recall echo, so the T-2 star GRE that's used in cohorts and that was used in the clinical trials. So, we do see several microbleeds in asymptomatic, cognitively healthy older people.

So, we do see evidence of asymptomatic cerebral microbleeds in otherwise healthy and cognitively healthy older people with a prevalence that goes up with age and ranges between 6 and 20 percent higher if there's known concomitant amyloid pathology as well. In terms of significance, having them does increase your risk of having a future both ischaemic and haemorrhagic stroke and does increase your risk of future cognitive decline. It's also a risk factor for these things, obviously, in people who are presenting with a stroke, whether it that ischaemic or haemorrhagic as well. So, even though they might look innocuous or may be not themselves being strongly associated with any cognitive presentation, they do rely a risk of future bad things happening.

MIC CAVAZZINI:                Now the finding from previous studies that older age was associated with higher prevalence was also replicated in your study, as you've said. When patients were stratified by severity of the cognitive diagnosis, prevalence of microbleeds was highest in those with a diagnosis of dementia at 39%, a little lower in those with mild cognitive impairment, and just 13 % in those with subjective cognitive decline.

And similarly, if you’re agnostic to diagnosis and just going on scores on the Mini-mental or the Addenbrooke cognitive examinations, you get an inverse association with respect to function and presence of multiple bleeds. In particular the domains of attention and orientation, memory, fluency and visuospatial and for the first time a negative association with verbal fluency was found in your study. Do you make too much of these findings or is it just luck of the draw which brain regions are affected by these bleeds?

PAUL YATES:       Yeah, look, I think the findings are helpful because they are consistent with the general literature in the field. We're expecting to see associations with attention and speed. And when you think about verbal fluency, guess that's a timed test that we're using. And so even though it's not been reported previously, we would expect it to be a marker of processing speed in this cohort, and perhaps some other executive function.

The interpretation, I think, must be taken with a grain of salt because we are not in all analyses adjusting for other concomitant vascular pathologies such as strokes or white matter hyperintensities and obviously we don't have a biomarker for beta amyloid or Alzheimer's disease pathology in this cohort because it's a clinical cohort. And so, if we were to adjust for some of those additional pathologies, the specific associations with microbleeds might go away because they could be a marker of these underlying processes.

DUNCAN AUSTIN:            Brilliant. So, as you say, we didn't have biomarkers in most of these patients, but they did have a formal diagnosis. Can you tell us what you found when patients were stratified by pathology, which looks like was mostly Alzheimer's disease versus cerebrovascular disease?

PAUL YATES:       So when we broke the groups down by type of dementia in particular, so we had Alzheimer's disease, dementia, cerebrovascular disease, dementia, a mixed sort of AD/ vascular, DLB and FTD in this cohort. And the prevalence of microbleeds was much higher, sort of twice as high in those with vascular dementia or mixed dementia. Interestingly, we did see occasional microbleeds in people with DLB and frontotemporal lobe degeneration as well. That's probably related to co-pathology as opposed to being associated with those pathological processes.

DUNCAN AUSTIN: So it was interesting and perhaps a little surprising to see there were more microbleeds with cerebrovascular disease than with Alzheimer's. You've alluded earlier to the fact that these microbleeds convey both an increased risk of ischaemic stroke and of haemorrhagic stroke. So as a clinician, Paul, what are your thoughts here? Should these patients be on blood thinners or should they avoid being on blood thinners?

PAUL YATES: So, two things on this. In particular, I wouldn't read too much into the prevalence numbers for the vascular dementia in our cohort because there was only five people. So, three out of five had microbleeds, so it's 60 per cent, but I'm sure the confidence intervals will be very wide. Although there's a slightly larger number with probable mixed Alzheimer's and vascular, and they had a similarly high prevalence.

Should they be on blood thinners or should they not? You're right that there's a real sort of tension there between wanting to prevent the future stroke and not cause harm. I will say I'm not a stroke neurologist, so bear that in mind. But I think that in general you want to use the treatment that there is evidence for in that cohort.

And you know, there are some easier ones if someone has clearly has got ischaemic heart disease on an antiplatelet agent or if they've got previous ischaemic stroke and they're on an antiplatelet agent and then there's evidence for that. There's not sufficient evidence to withhold these agents when there's a good reason to be on an antiplatelet. Anticoagulation is a bit harder and I think the data is still emerging. But, suffice to say, I don't believe that there's strong data to withhold anticoagulation in someone who has a clear indication for it. Although, I think the jury's out on people with very large numbers of microbleeds. I think that's a space that we're actively pursuing the answers.

MIC CAVAZZINI:                Well Duncan is a stroke neurologist so maybe we'll reflect the question back to him.

PAUL YATES:       What do you do?

DUNCAN AUSTIN:            I guess one of the factors that influences our decision is whether there's any history of haemorrhage. You know, we often have patients who've had both microbleeds and have had ischaemic stroke and had haemorrhagic stroke. Although as you say, it's not been widely studied, we certainly withhold thrombolysis in people with large numbers of microbleeds frequently. And I would certainly be cautious of using anticoagulation in someone with a large number of microbleeds and think about referring those patients for something like a left atrial appendage occlusion instead if possible.

PAUL YATES:       Yep, I think that we're broadly consistent with that approach. I do also, I like to show my patients their scans and show them if they do have a very large number of microbleeds to explain to them what I'm talking about and make sure that they are aware of the risks because they can then direct that conversation if they are under the care of another specialist that's got different priorities or that isn't aware of the brain scan results. So I like to try and put the power back in the hands of the patients in that as well and make sure that they're well informed.

MIC CAVAZZINI: So, as we've mentioned, in this cohort you found that microbleeds in the cortical lobes were much more frequent than subcortical ones in all diagnostic categories, and this was also true in the Swedish and Japanese studies. But is this inconsistent with that model we talked about earlier where amyloid angiopathy is associated with lobar microbleeds whereas hypertensive angiopathy is supposed to be associated with subcortical and infratentorial bleeds.

PAUL YATES:       Yeah, this would be consistent with the likelihood that the majority of the microbleeds in this particular cohort are related to amyloid in the blood vessel walls.

MIC CAVAZZINI: Yeah, that’s a better way to put it.

PAUL YATES:       And so, so I think that, you know, it may be that in our neck of the woods that people have relatively good blood pressure control, and relatively good vascular risk factor management compared with some cohorts in other parts of the world so, the prevalence of pure vascular disease was lower. Or, I guess we’d put them into a mixed picture if we think that there's multiple pathologies, so that might be why.

I think the other thing is that a lot of people don't necessarily come to a memory clinic if they've got a lot of deep microbleeds or deep vascular pathology and they may either present with bleeds and then they often won't come our way or they present with non-memory symptoms that people don't identify as readily and might not link that with cognitive decline. They may think it's just someone who's stubborn with older age.

DUNCAN AUSTIN:            So, you've alluded to this a bit already, but I was intrigued by the finding of microbleeds in Lewy body dementia, which is not felt to be an amyloid-related disease. So, are microbleeds perhaps just a relatively nonspecific marker of tissue damage?

PAUL YATES:       I mean, they can be caused by anything that would affect the vascular integrity. So, we see them in people with brain trauma, as well. And the other thing is that in Lewy body disease, we know that there is amyloid pathology in a proportion and so it may be that these are people with multiple pathologies. And clinical phenotype is most strongly that of a Lewy body dementia, but they may have multiple things going on, including vascular amyloid.

MIC CAVAZZINI:                Okay, well, let's now talk about possible treatments- especially in the context of the TGA’s very recent approval of donanemab for the treatment of Alzheimer’s Disease. But before we go on, can you both please put your cards on the table with respect to research affiliations involving this class of drugs and others?

PAUL YATES:       Absolutely. So, we conduct research for multiple pharma, including Eli Lilly, who produce donanemab and Eisai who produce lecanomab. I'm also, I've been a paid consultant on advisory boards for Eli Lilly, Eisai and Novo Nordisk and have conducted paid presentations for Eli Lilly and Eisai as well. So, I put my disclosures right, clearly on the table as I'm talking about these medications.

MIC CAVAZZINI:                Thanks. Duncan, what about yourself?

DUNCAN AUSTIN:            I've got no research affiliations involving this class of drugs and no industry funding.

MIC CAVAZZINI:                Okay, fabulous. So donanemab is not the first of these anti-amyloid monoclonal therapies, researchers have tinkered with a couple of dozen of these in recent years of which three have made it to approval with the US Food and Drugs Administration. Aducanumab, developed by Eisai and Biogen, was the first of these in 2021, and it was heralded as a breakthrough in being the first drug to result in disease modification vis a vis significant reduction of brain amyloid over the 76-weeks trials. But the FDA listing was considered highly controversial as there was no shift in the primary outcome, scores on Clinical Dementia Rating scale, except for some patients receiving very high doses. In January 2023, lecanemab, from the same group of firms received accelerated approval after somewhat more convincing findings from a large phase 3 trial- comparable deterioration in cogitive scores was delayed by a few months in the treatment group compared to placebo. And in August of the following year, Eli Lilly joined the fray with donanemab. Now, lecanemab was  rejected by Australia’s Therapeutic Goods Administration in December 2024 because there wasn’t sufficient evidence of safety in a sub-set of AD with the infamous APOE4 mutation. But the TGA has just given the green light to donanemab. So, Paul, how much more impressive are the outcomes of the donenemab trial, published in 2023, that have tipped the TGA over the line?

PAUL YATES:       I think there's a couple of distinctions between the two. And obviously I'm not within the sort of the decision-making group so I'm not party to the fine print of reasons for the decision. But I guess it's always going to be a balance of efficacy and safety. From a safety point of view, we know that this group of drugs, at least the currently approved group, can be associated with complications largely related to the movement of amyloid across blood vessel walls. The complications that I'm talking about are two entities that are described as ARIA. They are, which it stands for amyloid-related imaging abnormalities. And these two include vasogenic oedema, or effusion presentation and haemosiderosis or haemorrhage. So, ARIA-E and ARIA-H can have consequences, particularly if you link them with thrombolysis where some catastrophic things can happen. So, there's clear safety issues that need to be managed.

In terms of efficacy, they've both shown a slowing of cognitive decline over the trial periods in the realm of 25 to 37 per cent. Depending on which group you look at there's a trend to suggest that earlier treatment, or treatment of people with lower pathologies, has a better impact. And so, for example, in the donanemab study, people with mild cognitive impairment as opposed to early dementia, and people who had lower amounts of tau pathology actually had a 60 per cent slowing of cognitive decline, which is really impressive over an 18 month trial.

The other thing in terms of a distinction between the two, their mechanism of action is different. So, lecanemab targets protofibrils, which is one of the early aggregates of beta amyloid, whereas donanemab targets the plaque. As such the proposed treatment courses is different between the two, so donanemab in their trial they were able to treat to clearance and then come off the drug if they were amyloid negative during the trial which is an attractive prospect for a funded drug because you can stop

Whereas in lecanemab they continue to treat because they proposed that that these proto-fibrils are always being produced and so that their rationale was that you continue to treat for the entire 18 months. So, there are distinctions between the two which may have sort of weighed the hand of the TGA in approving one versus the other. But I think in terms of the overall efficacy at this point they're quite similar in many respects.

DUNCAN AUSTIN:            I think you've answered quite a bit of my next question there but you'll be familiar with the 2024 meta-analysis in Annals of Family Medicine which highlighted a high rate of adverse events. You've outlined different forms of ARIA, which are imaging abnormalities, some of which are asymptomatic or with just mild clinical symptoms. But the meta-analysis found that the numbers needed to harm for any type of ARIA was around 10, and for a symptomatic event it was 86. So, are we, perhaps being a bit too conservative in our concerns about adverse events.

PAUL YATES:       I think that we shouldn't shy away from the fact that at least these drugs do have effects on the brain. But I think if you look at the incidence of symptomatic ARIA, it is quite low and there's no evidence that having had an ARIA event causes longer term impairment. People can have their ARIA—generally—g their ARIA resolve, particularly if it's mild or asymptomatic, and continue treatment with no ill effects. In fact, people often come through with preserved or better cognition from an ARIA event, anecdotally. The evidence certainly isn't there that a mild ARIA event causes any longer-term detriment over and above the illness that they already have.

I think that the main take homes for me are that we need to be really careful to manage safety and make sure that we're choosing the right people for these drugs and that we have the right systems in place to avoid and prevent these the worst of the possible side effects. So, what I think is really important is that we have systems and processes in place to manage the safety for these patients. It's critical that we identify the right people, that they are monitored carefully for the emergence of treatment-related effects, and that we have systems that can manage these and most importantly avoid doing harm.

What I would say is that compared to the trial population, the approved product label is a lot more stringent in terms of narrowing down who can get the drugs. Compared to the trial, which allowed any APOE4 status in Australia, donanemab has only been approved for E4 heterozygotes or non-carriers. As well, the number of microbleeds that for someone to get donanemab in Australia is only two, whereas in the trials they could have four. And the presence of superficial siderosis was also permitted at the beginning of the donanemab trial. And that's also an exclusion, an absolute exclusion. So, the population that are eligible for the drugs in Australia are a lot a lot narrower population than what was in the trial and we would expect that this should result in even better safety outcomes.

Another aspect of our monitoring for safety outcomes that has evolved since the beginning of the original TRAILBLAZER study are the timing of MRIs. So, safety MRIs now are much earlier and more frequent at the beginning of the treatment, which would theoretically have picked up some of the people that had complications within the studies much earlier and could potentially have prevented some of the adverse events that did happen within the trial.

MIC CAVAZZINI:                And what do we suppose the prospect of having something like a donanemab treat cerebral amyloid angiopathy?

PAUL YATES:       Really good question. So, there's been a few approaches to CAA by itself and there are some disease-modifying therapies that are currently in trial for CAA. Some of the earlier monoclonal antibodies seem to promote preferentially bind to CAA such as ponezumab—I don't believe that that's still in active investigation. There are other novel approaches including RNA therapy, including intrathecal investigational products that are being used to try to reduce CAA specifically. So, there are certainly some centres that are recruiting for these trials as we speak.

MIC CAVAZZINI: And speaking of the progression of the pathology, it seemed from your study that people with a single microbleed, that there was no impact on cognitive examination results. So, are we thinking of this as a progressive pathology or is there a sense of how long that subclinical window might be that would be open for intervention or just observation?

PAUL YATES:       Yeah, so I think overall the microbleeds are likely to be related to whatever other pathologies are going on as we've discussed before. So, if there's a single microbleed, it's either that there's a very small, a less severe amount of amyloid angiopathy, which is common in people with Alzheimer's disease, both preclinical prodromal and in dementia. What we know from large cohort studies is that there is a very long time that Alzheimer's disease pathology accumulates in the brain, maybe 15, 20 years before the onset of clinical symptoms. And there are trials at the moment that are looking at whether we can intervene at that point, much earlier, in the cascade of pathology to try and further delay or possibly even prevent the cognitive impairment we see with Alzheimer's disease.

DUNCAN AUSTIN: You've mentioned the fairly tight eligibility criteria for donanemab. Do you have any idea roughly how many Australians would be eligible under those criteria?

PAUL YATES: Yeah, there's modelling, which looks at the prevalence of, Alzheimer's within the dementia population. So, we think of four hundred-odd thousand people with dementia right now. perhaps between 60 to 80 per cent is related to Alzheimer's Disease, although certainly, mixed pathologies are common, particularly later on in life. And concomitant cerebrovascular disease and other things is very common which may exclude a large number of people.

We don't really know the prevalence of MCI. So, within the current approved label and also the trials, people could have mild cognitive impairment or early dementia due to Alzheimer's disease, and we don't really know the prevalence of MCI. Prevalence community-based sort of cohorts have varied quite widely. I think from memory clinic cohorts the proportion of people coming through who are eligible are probably around about 10 to 20 percent and that's a generous estimate.

MIC CAVAZZINI: And as you’ve mentioned donanemab is more selective than lecanemab for plaques, lecanemab goes more for the fibrils. It’s been suggested in the reason that Eli Lilly’s first offering solanezumab failed was that it instead went for the soluble monomeric form that precede pathology by more than a decade. But there are other pharmaceutical companies investigating interfering RNAs at that stage. So yeah, what do we know about the molecular cascade in angiopathy and whether it would be donenemab or any other agent that's better focused?

PAUL YATES: I think that we probably can't make too much comment about the relative impact of either of these drugs on CAA itself. I think that a proportion of CAA will lower as a result of global brain amyloid being reduced. And that's probably about as much as we can say. I think that the markers of CAA or the imaging markers that we see that we would associate with CAA in an untreated person, so these are the ARIA, these we see with slightly lower incidence in lecanemab than donanemab, but broadly they're comparable. And the TRAILBLAZER 6, a new protocol of donanemab dose titration, actually showed a lower incidence of ARIA with a more gentle dose up-titration than the published TRAILBLAZER 2 study which is what the majority of the data has been based on.

MIC CAVAZZINI: Duncan, what do you think about the last two radiology questions, should we go there?

DUNCAN AUSTIN: That’s probably not a bad idea because Paul mentioned the increased sensitivity of SWI. So, microbleeds CMBs were defined following a formalised rating scale that dates back many years, the MARS scale, and were checked against the MARS recommendations for potential mimics with two raters reviewing all MRI scans. You found there was good inter-rater reliability with a kappa of 0.75. This was much better when a 3 Tesla magnet was used compared to a 1.5 Tesla magnet, but the prevalence of microbleeds was also higher in those evaluated with the more powerful field. So, does this just mean that a bigger magnet equals higher resolution and more discriminating power?

PAUL YATES:       Broadly, yes, although I'll acknowledge I'm not a MRI physicist or a neuroradiologist. But broadly, so the more powerful the magnet, the greater capacity to detect lesions will be and the choice of imaging approach. So that susceptibility weighted imaging, as I said earlier, compared with the T2 STAR gradient recall ECHO certainly does affect the yield of detection of microbleeds and superficial siderosis.

MIC CAVAZZINI: Hang on, which one is more sensitive?

PAUL YATES:       SWI is more sensitive than GRE. About three times more sensitive depending on the citation, and so we're able to detect microbleeds more easily than using the older T2-STAR GRE. So, if you're using the SWI as a screening tool to exclude people then people will miss out on donanemab compared with what was used in the trial.

MIC CAVAZZINI:                Right, so you're talking in the context of excluding patients with CNBs from these trials. was sort sticking with the bleeds as a primary diagnosis.

PAUL YATES:       Yeah, this is relevant because for most of Australia we use susceptibility weighted imaging or SWI or something similar. So, this is something that I think that the international field is still grappling with. I know in the States, sometimes they do both sequences and report both and allow the clinician to interpret based on what was done in the trials but the current guidelines are not clear as to what the best approach is.

MIC CAVAZZINI:                As a take-home message on your paper, what is the most valuable finding from that, from a clinical management point of view or trials management point of view?

PAUL YATES:       I think from the paper, the main take home for me is that these lesions are very common. They may be a marker of underlying vascular or amyloid pathology in people if you don't have another biomarker. So, it's important to pay attention to them. And the prevalence, I think, was quite high, which means that when we're talking about what we think is a relatively typical memory clinic population thinking about whether they would be eligible for monoclonal antibody therapies, a high proportion of people would be excluded due to microbleeds alone. So, this is important when we're thinking about planning our services and also finding the right patients who'll benefit most from future therapies—making sure we're looking in the right places and that we can provide an approach for the people who are eligible, but also making sure that we don't lose sight of the fact that many of our clinic patients are not eligible and we need to have treatments and approaches for them as well.

MIC CAVAZZINI:                Fabulous Duncan, anything left over that you want to clean up?

DUNCAN AUSTIN:            No, I think we'll leave the health economics question. It's too incendiary a topic.

PAUL YATES: One last thing I'd like to add as well. So, as we are implementing these new therapies, I think it's really critical that we do it in a safe, evidence-based and sort of monitored way. And so, you know, it's really important that we continue to collect post-marketing data, that we collect safety data, that we can participate in things like registries to contribute to ongoing understanding of the risk and benefit of these drugs as we bring them to our patients. And, I think it's clear that they're not the be all and end all, they're not the, they're not possibly as dramatic an improvement as we'd like them to be. But they’re certainly a step in the right direction, and it's really great that we've got something to offer to our patients.

And I think the improvements, although they're modest, it's hard to show a massive difference in a short trial. They're disease-modifying therapies so the effect size within the trial isn't all that they do, right? They should continue to have an effect beyond the duration of the trial compared with sort of the previous symptomatic therapies that we've seen 20 something years ago. So, I think that there's a lot of optimism that's warranted and there are more and more exciting therapies in the pipeline. So, I think this is a really exciting time for dementia research and clinical practice in Alzheimer's Disease.

MIC CAVAZZINI: It certainly is an exciting time for the field of Alzheimer’s research and treatment, given that it’s taken almost thirty years for the amyloid hypothesis to translate into a disease-modifying drug. But that excitement still needs to be tempered by the fact that the protective effect of drugs like donanemab and lecanemab, hasn’t yet been shown to meet the MCID, the minimal clinically important difference.

That’s to say that while the relative cognitive decline, as measured on test scores, was around 30 per cent less in treated patients was than those on placebo over the duration of the trials, the absolute difference is scores wouldn’t practically make any difference.

I wanted to better understand where this MCID came from and what other resourcing would be needed to rollout such therapies. How many staff and beds to administer the infusions and collect CSF biomarkers and how many hours on the MRI machines to conduct all the follow up? After three pages of notes I decided that this would be better explored in a follow up interview representing different schools of thought. So stay tuned for that one.

In the meantime, I want to thank Associate Professor Paul Yates for participating in this episode of [IMJ On-Air]. His paper can be found in the April edition of the Internal Medicine Journal titled, Prevalence and Associations of Cerebral Microbleeds in an Australian Memory Clinic Cohort. And many thanks also to Dr Duncan Austin, who is an invaluable reviewer of this podcast and very relaxed behind the microphone too. They’ve both volunteered their time to support your continuous learning.

If you liked the podcast please share it with your colleagues, your friends, even your mother. They just need to search for Pomegranate Health in any podcasting app such as Spotify or Apple podcasts. There’s an alerts mailing list you can subscribe to at our website, racp.edu.au/podcast, and a comments section below each episode to keep the conversation going.

Please email any feedback or ideas to podcast@racp.edu.au. I’m always keen to hear what you’re keen to hear. My name is Mic Cavazzini, and I produce this podcast on the lands of the Gadigal clans of the Yura nation. I pay respect to their elders past and present. Thanks for listening.