Stroke Alert January 2023

Stroke Alert

On Episode 24 of the Stroke Alert Podcast, host Dr. Negar Asdaghi highlights two articles from the January 2023 issue of Stroke: “Covert Brain Infarction as a Risk Factor for Stroke Recurrence in Patients With Atrial Fibrillation” and “Subarachnoid Hemorrhage During Pregnancy and Puerperium.” She also interviews Dr. Georgios Tsivgoulis about his article “Clinical, Neuroimaging, and Genetic Markers in Cerebral Amyloid Angiopathy-Related Inflammation: A Systematic Review and Meta-Analysis.”

Dr. Negar Asdaghi: Let's start with some questions.

1) When during pregnancy is an intracranial aneurysm at the highest risk of rupture?

2) What does the presence of covert brain infarcts mean in the setting of atrial fibrillation?

3) And, finally, how is the inflammatory form of cerebral amyloid angiopathy different from the classic CAA form, and why is it important to differentiate between the two?

We'll be answering these questions and much more in today's podcast. We're covering the latest in cerebrovascular disorders, and this is the best in Stroke. Stay with us.

Welcome back to another issue of the Stroke Alert Podcast. My name is Negar Asdaghi. I'm an Associate Professor of Neurology at the University of Miami Miller School of Medicine and your host for the monthly Stroke Alert Podcast. Together with my co-editors, Drs. Nastajjia Krementz and Eric Goldstein, here's our article selection for the month of January. Symptomatic intracerebral hemorrhage is a feared complication of reperfusion therapies in acute stroke, so there's a lot of interest in looking for predictors of development of this complication, especially when you're making decisions for pursuing endovascular therapy. For many years now, we've known about some of these predictors, such as presence of a large infarct core and high blood glucose levels. But in the recent years, other radiographic markers of tissue viability, such as a poor collateral status and unfavorable venous outflow profile, have been shown to be predictors of post-reperfusion hemorrhagic transformation.

In this issue of the journal, we learn about another imaging marker that can potentially predict parenchymal hemorrhage occurrence post-endovascular therapy, which is high hypoperfusion intensity ratio, or HIR, as measured by perfusion imaging. What is HIR? It's a long name for a simple ratio that can easily be measured by dividing the volume of tissue with Tmax delay of over 10 seconds to the volume of tissue with Tmax delays of over 6 seconds. Simply put, Tmax 10 divided by Tmax 6. These volumes, as you know, are typically provided to us by almost all post-processing perfusion softwares, and so this ratio can be easily calculated in the acute setting. So, in this paper led by Dr. Tobias Faizy from University Medical Center in Hamburg and colleagues, we learned that higher hypoperfusion intensity ratios are strongly associated with parenchymal hemorrhage occurrence after endovascular therapy. So, in summary, HIR, that is a quantitative ratio, can be used as a marker to risk stratify patients that are undergoing endovascular therapy in terms of helping us predicting the risk of development of intracerebral hemorrhage after reperfusion therapies.

In a separate study in this issue of the journal, we read a very interesting paper titled "Anti-Epileptic Drug Target Perturbation and Intracranial Aneurysm Risk." How are intracranial aneurysms even related to anti-epileptic drugs? Well, first of all, it's been known for a long time based on genome-wide association studies that there are multiple common genes that are associated with increased risk of intracranial aneurysm development. Now, some of the largest genetic studies to date have shown pleiotropy between genetic causes of development of intracranial aneurysms and genes encoding targets for anti-epileptic drugs. Now that's a fascinating finding because finding commonalities between these genes may help find new treatment targets for intracranial aneurysms.

So, in this paper in this issue of the journal, the investigators from the University Medical Center in Utrecht found an association in the expression of anti-epileptic drug target gene CNNM2 and intracranial aneurysm risk. They found that certain anti-epileptic drugs, such as phenytoin, valproic acid, and carbamazepine, that are expected to lower CNNM2 levels in the blood may subsequently lead to a lower risk of development of intracranial aneurysms. And, of course, a reasonable follow-up study to this would be to investigate whether persons exposed to these anti-epileptic drugs have indeed a lower risk of unruptured intracranial aneurysms and subarachnoid hemorrhage, and how variation in CNNM2 expression can lead to development of aneurysms. Bottom line, CNNM2 may be a relevant drug target for treatment of cerebral aneurysms. As always, I encourage you to review these papers in detail in addition to listening to our podcast today. My guest on the podcast today is the Chairman of Neurology at the University of Athens, Dr. Georgios Tsivgoulis. He joins me all the way from Greece to talk about cerebral amyloid angiopathy-related inflammation, or CAA-ri. He's a remarkable researcher, and I can say with absolute confidence that we cannot find a better summary of this very tough topic elsewhere. He ends the interview with an intriguing account of the early description of dementia in Greek mythology. But first, with these two articles.

What are covert brain infarcts, or CBIs? Are these the John Wick or the James Bond of the stroke world? After all, they operate undercover. They're ominous and attack without warning. That's probably why they're also called silent infarcts. Now, whatever we call them, we need to know how prevalent they are and what does their presence actually mean. Let's dive into this topic. For at least two centuries, if not longer, we've known about covert brain infarcts. Early description of these lesions is credited to Amédée Dechambre, a medical intern at Salpêtrière Hospital in Paris who noted that there are strokes that can cause symptoms like hemiplegia, but also strokes that are asymptomatic, or so he thought at the time. In the modern times, while we agree with our pathology forefathers that CBIs are different from symptomatic strokes, we also know that they are not entirely asymptomatic. The symptoms can be subtle and tend to sneak up on the patient, but what is clear is that amassing of covert brain infarcts results in an overall decline in cerebrovascular reserve of the brain.

With the advent of neuroimaging, we now know that CBIs are age-dependent and prevalent, seen in almost 10 to 30% of even healthy adults, but much more prevalent in those with vascular risk factors, and they can be caused by nearly the entire spectrum of neurovascular disease, including large vessel, small vessel disorders, cardioembolism, and others. Now, how do these covert infarcts catch up in those with atrial fibrillation? Neuroimaging studies have shown that patients with A-fib, especially those untreated, have a higher percentage of embolic-appearing CBIs, and conversely, those with embolic formed pattern of CBIs are more likely to have undiagnosed A-fib. So the question is, what's the significance of CBI in those with confirmed A-fib? In this issue of the journal, Dr. Do Yeon Kim from Seoul National University and colleagues help us answer this question using the EAST-AF, which stands for East Asian Ischemic Stroke Patients With Atrial Fibrillation Study.

So, the paper included over 1300 patients with A-fib and first-ever stroke without a prior history of TIA or stroke. And then they categorized these patients into those who had evidence of CBI on neuroimaging and those who didn't. So, what did they find? Forty-two percent of patients with A-fib and first-ever stroke had evidence of covert brain infarcts on neuroimaging. Let's think about it for a moment. These patients presented with what was thought to be their first-ever stroke, not knowing they already had some in their brain. Now, what makes things really worse is that over a quarter of these subjects had more than just one covert infarct. Not surprisingly, those with CBI tended to be older, had higher blood pressure, and had worse white matter hyperintensity burden. This is kind of expected and also not expected was the fact that most of these covert infarcts were actually embolic in pattern.

Over 60% of them were embolic. Another 14% of cases had combined embolic and non-embolic-appearing CBIs. Now, overall, the one-year incidence of ischemic stroke and all-cause mortality was higher in those that had CBIs at baseline. When they started looking at the specific patterns of CBIs, those embolic-appearing CBIs had a threefold higher risk of recurrent ischemic stroke, whereas those with non-embolic-appearing covert infarcts had oddly a higher all-cause mortality rate but not recurrent ischemic stroke. And finally, just briefly, the authors noted that the addition of CBIs to the classic CHA2DS2-VASc score didn't meaningfully otherwise statistically improve the scoring metrics, so they left it at that. So, the take-home message is that 42% of A-fib patients presenting with first-ever stroke actually had prior strokes without even knowing based on this study. And most of these strokes were embolic-appearing, and these covert brain infarcts can be used as predictors of future clinical strokes in this population.

Strokes should be the last thing to worry about when we think of pregnancy. In the United States, around 30 in 100,000 women, unfortunately, experienced a stroke during pregnancy, and between 6 to 8 in 100,000 deliveries are complicated by subarachnoid hemorrhage. What's the most common cause of pregnancy-associated subarachnoid hemorrhage? In the general population, close to 80% of subarachnoid hemorrhage cases are aneurysmal. Is this true for the pregnant population as well? And importantly, what's the contemporary incidence trend, risk factors, and outcomes of pregnancy-related subarachnoid hemorrhage? In this issue of the journal, Dr. Korhonen and Petra [Ijäs] and their colleagues from the Departments of Neurology and Obstetrics and Gynecology at Helsinki University Hospital will give us the answers to some of these questions through a nationwide population-based study in Finland. So, they looked at over one and a half million pregnant women who gave birth during a 30-year time period between 1987 to 2016.

Subarachnoid hemorrhage was identified through appropriate ICD codes and then further adjudicated based on confirmatory information, including neuroimaging and data from lumbar puncture. A total of 57 cases of pregnancy-related subarachnoid hemorrhage was identified in this paper. The mean age of women was 33, ranging from 23 to 45, and the clinical presentation was typical for subarachnoid hemorrhage, including thunderclap headache and mild neurological symptoms. So, what did they find? So, first off, in terms of general observations, the overall incidence rate of pregnancy-related subarachnoid hemorrhage in this study was 3 over 100,000 deliveries. This is almost half the incidence rate reported from the nationwide registries in the United States. Seventy-seven percent of pregnancy-related subarachnoid hemorrhage cases were aneurysmal, so very similar to the general population. The other 23% were non-aneurysmal cases, but it's important to note that 40% of those non-aneurysmal cases also had vascular etiologies, so etiologies such as moyamoya syndrome, postpartum angiopathy, AVM, to name a few. Like non-pregnant patients with subarachnoid hemorrhage, the aneurysmal cases were sicker patients in general. They had a lower GCS at presentation, higher Hunt and Hess scores, and required more ICU admissions. The next finding is very important because it actually shows that development of subarachnoid hemorrhage during pregnancy significantly impacted obstetrical care. A total of 66% of women with subarachnoid hemorrhage during pregnancy ended up having a C-section and a high percentage of these cesarean sections were actually elective. This is in contrast with subarachnoid hemorrhages in the postpartum period where 67% of women had spontaneous vaginal deliveries. The other important finding of the paper was really highlighting the differences between pregnancy-related aneurysmal versus non-aneurysmal subarachnoid hemorrhages. We already talked about how, in general, aneurysmal cases had more severe neurological presentations, so, not surprisingly, they also had worse outcomes with a mortality rate of 16% for the aneurysmal subarachnoid hemorrhage cases, and only 68% of women with pregnancy-related aneurysmal subarachnoid hemorrhage reached a favorable outcome, which was defined in this study as modified Rankin Scale of 0 to 2. Other important differences included the fact that the incidence of aneurysmal subarachnoid hemorrhage increased towards the end of pregnancy and was highest in the third trimester.

This ties in with the findings from prior studies all indicating that rupture of an aneurysm is most common in the third trimester. By contrast, the incidence of non-aneurysmal subarachnoid hemorrhage peaked in the second trimester in this study. And finally, in terms of risk factors, first let's talk about age. The incidence rate of pregnancy-associated subarachnoid hemorrhage increased with age of the mother. So, in this study, there were no cases noted amongst women aged below 20 years of age to an incident rate of 12 per 100,000 deliveries among women aged 40 years or over. So that's a fourfold increase from the overall incidence rate of pregnancy-related subarachnoid hemorrhage, and very important point that we learned from this paper. Apart from age, smoking beyond 12 weeks of gestation and hypertension were also independent factors associated with pregnancy-related subarachnoid hemorrhage. So, overall, hypertension, smoking are bad and are significant risk factors for pregnancy-related subarachnoid hemorrhage. And if we have to remember just one thing from this paper, let it be this one: The rupture of an aneurysm is most common in the third trimester of pregnancy.

Cerebral amyloid angiopathy, or CAA, is an important cause of intracranial hemorrhage and refers to deposition of β-amyloid fibrils in the wall of the small- and medium-sized cerebral blood vessels, mostly involving cortical and leptomeningeal arteries. It is believed that the deposition of β-amyloid results in architectural disruption of the blood vessels, which then leads to perivascular leakage. That's the pathophysiological mechanism behind the development of cerebral microbleeds. And this process, of course, can cause frank vascular rupture resulting in cortical intracerebral hemorrhage or development of high-convexity subarachnoid hemorrhages. It is important to note that varying amounts of perivascular inflammation, that is inflammation surrounding β-amyloid-laden blood vessels, may be present in some CAA cases, rendering them the designation of inflammation-related CAA. However, frank vasculitic destruction of the vessel wall, such as what is found in amyloid-β-related angiitis, or ABRA, and primary angiitis of the central nervous system, is usually absent in most CAA-related inflammation cases.

How these entities are best defined, diagnosed, and treated is subject of intense research. In this issue of the journal, in the study titled "Clinical, Neuroimaging, and Genetic Markers in CAA-Related Inflammation," Dr. Georgios Tsivgoulis and colleagues take us through a systematic review and meta-analysis of published studies of patients with CAA-related inflammation. I am joined today by Dr. Tsivgoulis himself to discuss this paper. He's a Professor of Neurology and Chairman of the Second Department of Neurology at the University of Athens School of Medicine. Dr. Tsivgoulis is the residency program director and the director of cerebrovascular fellowship program with extensive research and expertise in the field of stroke. Good morning, Georgios, and welcome to our podcast.

Dr. Georgios Tsivgoulis: Good morning, Negar. I'm delighted to be here and delighted to present our findings, on behalf of all our co-authors.

Dr. Negar Asdaghi: Thank you very much for being here and congrats again on the paper. So, Georgios, let's start with this interest that's going on with using clinical and radiographic features to make the diagnosis of CAA-related inflammation in contrast to moving ahead and performing brain biopsy. Can you please start us off with a brief review of the newly proposed clinico-radiographic criteria for this condition, please?

Dr. Georgios Tsivgoulis: Yes. As you mentioned, Negar, CAA-ri is a distinct, however, rare subset of cerebral amyloid angiopathy. Firstly, Greenberg and the Boston group published in Neurology in 2007 a paper highlighting that a diagnosis of a probable CAA-ri patient could be made on the basis of characteristic clinical and neuroimaging findings without requiring a biopsy. Following this observation, Chung and colleagues in 2010, in a seminal paper in JNNP, proposed the first diagnostic criteria for probable and definite CAA-ri. For the definite diagnosis, besides the typical clinical presentation with headache, encephalopathy, focal neurological signs and seizures, and the characteristic neuroimaging findings with T2 or FLAIR hyperintense asymmetric white matter lesions complicated with microbleeds and leptomeningeal or parenchymal gadolinium enhancement, and histopathological confirmation with amyloid deposition within cortical leptomeningeal vessels associated with perivascular, transmural or intramural inflammation was also required. The latest criteria developed in 2015 by Auriel and colleagues that were published in JAMA Neurology using a validation study modified the current criteria for the diagnosis of CAA-ri.

In this paper, the author supported the use of empirical immunosuppressive therapy, avoiding brain biopsy, for patients meeting the criteria proposed for probable CAA-ri. They suggested that a brain biopsy should be considered in empirically treated patients who failed to respond to corticosteroid therapy within three weeks. The criteria by Auriel and colleagues are widely applicable in everyday clinical practice, and we also use this criteria for the inclusion of studies in our current meta-analysis. I would like to highlight for our audience that the latest criteria for CAA-ri were published in 2015 by Auriel and colleagues. However, these are different for the criteria for cerebral amyloid angiopathy than the latest criteria were published in 2022 in Lancet Neurology, OK?

Dr. Negar Asdaghi: Georgios, that was a great start for this interview. You had mentioned a lot of information here. I just want to highlight what you just said. So, we are using for this meta-analysis, the latest criteria in CAA-related inflammation published in JAMA by Auriel and colleagues. That's slightly different than, we're not referring to the 2022 criteria of cerebral amyloid angiopathy. It's an important distinction. We're going to talk about this a little more as we go through the interview, but I want to come back to your current paper and start from there. Can you please tell us about the importance of this paper, why doing a meta-analysis was important in your view, and tell us a little bit about the studies that were included in your paper?

Dr. Georgios Tsivgoulis: Yes, thank you for that question. CAA-ri is an increasingly recognized entity since the recent diagnostic criteria by Auriel and colleagues published in 2015. In collaboration with the greater availability of the high-resolution MR, we can have now a reliable non-invasive diagnosis of possible or probable CAA-ri, avoiding the risk of brain biopsy. However, I need to highlight that the early diagnosis remains a great challenge for the clinicians and neurologists. Searching the literature, we observe that there is scarce data regarding the prevalence of the distinct clinical, neuroimaging, and genetic markers among patients diagnosed with CAA-ri. We believe that pooling all this information in the current meta-analysis would be very helpful for every clinician, increasing a comprehensive understanding of this rare cerebrovascular disorder. Consequently, we conducted this meta-analysis including 21 studies that recruited a total of 378 patients with CAA-ri. Our study involved only 4 prospective and 17 retrospective hospital-based cohorts of patients diagnosed with CAA-ri based on autopsy or biopsy or on the recent Auriel diagnostic criteria that do not require autopsy or biopsy. Due to limited data in the literature regarding this entity, we had to include only small cohort studies with at least five patients in our meta-analysis. We excluded case reports and case series with less than five patients. This is, by far, the largest available sample of CAA-ri patients in the literature.

Dr. Negar Asdaghi: OK, great. So, let me just recap this, more so for myself. So, we have 21 studies, and you excluded studies that included less than 5 patients. So, practically speaking, case reports.

Dr. Georgios Tsivgoulis: Yes, and single-case reports.

Dr. Negar Asdaghi: Yes. And practically speaking, of the total number of patients that are included in this meta-analysis, you have 378 cases, and basically the diagnosis of CAA-related inflammation was either based on the newly proposed criteria or based on biopsy-confirmed or autopsy cases.

Dr. Georgios Tsivgoulis: Which is the standard criteria.

Dr. Negar Asdaghi: So, now, I'm dying to ask you about these clinical and radiographic characteristics of patients with CAA-related inflammation in this meta-analysis.

Dr. Georgios Tsivgoulis: The mean age of patients in the included studies was approximately 72 years old, and there was no obvious gender predominance. Fifty-two percent of the patients were of female sex. In our study, 70% of the included patients presented with cognitive decline, which was the most common neurological manifestation, while 50% of the total sample had focal neurological signs and 54% encephalopathy presentation. Symptoms such as headache and seizures were less common, 37 and 31% respectively. With regard to the radiological findings, hyperintense T2 FLAIR white matter lesions were very, very common in 98% of our patients, and they were also complicated with lobar cerebral microbleeds, with a prevalence of 96%, and these two were, by far, the most prevalent neuroimaging findings, that white matter hyperintensities coupled with a cerebral microbleed. The pooled prevalence rates of gadolinium-enhanced lesions was 54%, and also the prevalence of cortical superficial siderosis was 51%, which is also very high in this cohort of patients with CAA-ri.

Dr. Negar Asdaghi: OK. So many of the features Georgios said, you mentioned, from presence of white matter hyperintense lesions on T2 FLAIR to presence of cortical microbleeds or superficial siderosis, these features are also seen in patients with cerebral amyloid angiopathy. What are some of the important differentiating features between the two conditions?

Dr. Georgios Tsivgoulis: Yes, this is an excellent clinical question. First of all, the lower age threshold for CAA-ri is 40 years old, whereas in cerebral amyloid angiopathy, the lower age threshold is 50 years. So, patients who are younger than 50 years can be diagnosed with CAA-ri, but they cannot be diagnosed with CAA. Another issue is that comparing the result of this meta-analysis with another recent meta-analysis focusing on CAA, on cerebral amyloid angiopathy, that our international multi-collaborative group published in Stroke in 2002, we also evaluated the presence of clinical phenotypes and radiological markers among patients with cerebral amyloid angiopathy. We have documented that transient focal neurological episodes are much more common in patients with cerebral amyloid angiopathy in contrast to patients with CAA-ri. These episodes, which are called TFNEs, transient focal neurological episodes, are attributed to cortical subarachnoid hemorrhage or cortical superficial siderosis.

So, I think this is another important clinical distinction. The most important, however, differentiating features between the two entities are neuroimaging markers, in specific, in particular, T2 FLAIR hyperintense unifocal or multifocal lesions with mass effect. These are the most prevalent neuroimaging features among patients with CAA-ri, but they're very seldomly described in patients with cerebral amyloid angiopathy, in patients with CAA. Another characteristic neuroimaging finding very indicative of the inflammation is the leptomeningeal or parenchymal gadolinium enhancement. This finding has been very rarely described in patients with non-inflammatory cerebral amyloid angiopathy. So, the clinical distinction is not so solid. However, the neuroimaging distinction would provide us with very strong information that can help us differentiate these two conditions.

Dr. Negar Asdaghi: Excellent points, I have to say, golden points, not just excellent points. I'm going to try to recap this and see if I understood it correctly. So, for our listeners, we have two conditions that potentially have many common points. One is the cerebral amyloid angiopathy, and the second one, which is obviously the subject of this interview, is cerebral amyloid angiopathy-related inflammation. The most important differentiating factors between the two are actually the neuroimaging features, as Georgios mentioned. So, the first feature that was mentioned is presence of T2 FLAIR hyperintense lesions. Some of them are large and have actually mass effects. This feature is rarely seen in patients with CAA, and it's an important radiographic factor that is seen in patients with CAA-related inflammation. The second distinguishing feature was leptomeningeal enhancement, again, rarely seen in non-inflammatory CAA, but was seen in a significant proportion of patients with CAA-related inflammation. These were the neuroimaging features. You also mentioned two other factors. The median age of CAA-related inflammation was lower than CAA. That can be helpful. And also the entity of transient focal neurological episodes, or TFNE, is rarely seen in inflammatory cases of CAA, whereas it is described in cases with cerebral amyloid angiopathy and mostly related to development of either cortical subarachnoid hemorrhage or cortical superficial siderosis. I think I got this all, correct?

Dr. Georgios Tsivgoulis: Excellent.

Dr. Negar Asdaghi: All right, so let's come now to the genetics of CAA. The apolipoprotein E gene is associated with the presence of amyloid angiopathy and development of lobar intracerebral hemorrhage, and we've learned about this in cases with cerebral amyloid angiopathy. Is there an association with ApoE, and did you find anything in this meta-analysis?

Dr. Georgios Tsivgoulis: Another very exciting question. In 2007, there was a first report that the apolipoprotein ε4 homozygosity may be considered a risk factor for CAA-ri, and there was a strong correlation reporting a high prevalence of 77% of this apolipoprotein ε4 alleles among patients with CAA-ri. To justify this correlation, the hypothesis was that an underlying pathogenic mechanism, which increases the amyloid-β deposition and has a pro-inflammatory effect, may be suspected as the cause of this disorder. The largest, however, prospective cohort of CAA-ri patients conducted by Antolini and colleagues and was published in 2021 in Neurology, reported a much lower prevalence of apolipoprotein ε4 carriers accounting for 37%, 23% heterozygotes and 14% homozygotes. So, we also documented a pool prevalence of apolipoprotein ε4 homozygosity of 34%. So, we did not confirm the initial finding of 77%. However, in our meta-analysis, the homozygosity was 34%, and we need to have a cautious interpretation of these results because data is limited, and we need larger future population-based studies and in larger cohorts to evaluate the prevalence rate of these specific genetic markers. So, we can confirm an association between apolipoprotein ε4 homozygosity, however not as strong as originally reported in 2007.

Dr. Negar Asdaghi: OK. So, Georgios, thank you. And again, very important factor to keep in mind for our clinicians listening in. Unfortunately, based on what you mentioned, we don't have yet a genetic marker to, for sure, tell us if we're dealing with CAA-related inflammation, yes or no, as you mentioned. Just to recap, earlier on, there was studies to suggest a very strong association between apolipoprotein ε4 homozygosity and CAA-related inflammation. But later on, this was not confirmed by subsequent studies, and in your meta-analysis, you found 34% ApoE ε4 homozygosity amongst patients with CAA-related inflammation and could not confirm that original high association. OK, so with all of that, it's a lot of information. I have to go to the next question regarding controversies involving the levels of Aβ40, Aβ42, and P-tau proteins in CSF in the setting of CAA-related inflammation. Can you please tell us more about these biomarkers?

Dr. Georgios Tsivgoulis: Yes. The overlap of Alzheimer's disease and CAA can be attributed to the coexistence of some degree of cerebrovascular amyloid deposition and amyloid plaque pathology, which is very common. And, of course, the evaluation of amyloid and tau proteins in CSF is of high significance for the prognosis and the evolution of CAA patients. In our previous review, we have summarized the literature and noticed that CSF concentrations of Aβ40 and, secondarily, Aβ42 were much lower in patients with cerebral amyloid angiopathy compared with Alzheimer's disease. Total tau and phospho-tau CSF levels were comparable to healthy controls in CAA and lower than patients with Alzheimer's disease. Moving now to CAA-ri, there were scarce data about these biomarkers amongst CAA-ri patients. The majority of the relevant studies have found relatively low levels of Aβ42 and Aβ40 in the CSF and high levels of P-tau. In the present meta-analysis, the pooled means of biomarker levels were based on the findings of only two studies with heterogeneity, and these limit substantially the validity of our observations. However, they confirm the previous reports indicating, as I said before, but I would like to repeat, low levels of Aβ42 and Aβ40 in the CSF and high levels of P-tau.

Dr. Negar Asdaghi: Perfect. So, thank you, Georgios. I'm going to recap what you said. So, we're talking about CSF biomarkers, and first what you mentioned is going back to the original studies concentrated on using these biomarkers as ways of differentiating between cerebral amyloid angiopathy and Alzheimer's disease. And very briefly, to recap what you said, in general, the levels of Aβ40 and, secondarily, Aβ42 was found to be much lower than the Alzheimer's levels in patients with CAA. Now coming to the inflammatory form of CAA, what you mentioned and what you found in this meta-analysis, practically speaking, confirmed that the levels of Aβ40 and Aβ42 in CSF are low and the levels of P-tau are high in this condition as well. So, one thing I want to ask as a secondary question to that is, that it sounds like these biomarkers are more or less similar in CAA and CAA -related inflammation, not that different. Is that correct?

Dr. Georgios Tsivgoulis: It's absolutely correct. And I would also like to highlight a major limitation of the meta-analysis that we had available data from only two studies to pool the mean of these CSF biomarker levels. So, these results need to be acknowledged with caution, and we would love to repeat our meta-analysis after the publication of more studies and prospective cohorts measuring the CSF biomarkers in patients with CAA-ri.

Dr. Negar Asdaghi: OK. So, again, important to note, as you mentioned, that there's heterogeneity in data because of just paucity of information on this, but as we stand today, the biomarkers won't really help us in terms of differentiating between the two conditions that are CAA or CAA-related inflammation. And so, I think I've learned a lot from this interview myself, but I think we have to just talk briefly about the available therapies for CAA-related inflammation.

Dr. Georgios Tsivgoulis: Yes. In our meta-analysis, we sought to summarize the available information regarding different therapeutic strategies and outcomes among CAA-ri patients. Our results supported our clinical experience indicating that corticosteroids represent the first-line treatment in these patients' outlook. Steroids have been associated with clinical and radiological improvement of the primary disease episode and decreased risk of subsequent relapses in patients with CAA-ri. Additional immunosuppressive therapies, including cyclophosphamide, mycophenolate mofetil, azathioprine, IVIG, or rituximab, have been also reported as adjunct therapies in selected cases with a more severe course of the disease. However, this is another limitation that needs to be acknowledged. That data regarding the treatment and the outcomes are limited and heterogeneous, which prevented us from drawing robust conclusions using a meta-analytical approach. And we believe that we need future cohort studies with prospective data validation in order to generate a proposal for a therapeutic algorithm management in these cases.

Dr. Negar Asdaghi: Thank you, Georgios. So, we have a condition that is now being more and more recognized. We now have criteria based on clinical and radiographic presentation features of patients that might help us with this diagnosis to differentiate it from cerebral amyloid angiopathy. And in terms of therapies, the idea is that the most studied drug is really just first-line therapy, that's corticosteroids. And then there's positive data regarding use of all other forms of immunosuppression, including, as you mentioned, cyclophosphamide, rituximab, and oral agents such as mycophenolate mofetil or azathioprine. We have limited information about those, but I want to highlight something you actually mentioned earlier on in the interview, which is the field is moving towards making these diagnoses based on clinical features and radiographic features that you had highlighted and actually giving patients immunosuppression early on and only move on to a biopsy if the patient had failed these therapies for a period of time, which you mentioned three weeks. So, I think it's important for us as clinicians to keep this evolving criterion and recommendations in mind. And before we end, I want to ask you a hypothetical question, Georgios. In your opinion, what's an ideal randomized trial for CAA-related inflammation in the future?

Dr. Georgios Tsivgoulis: I think before going to the randomized, the ideal randomized trial for CAA-ri, and designing this trial, we need much more information regarding the underlying pathophysiological mechanisms. There are many unanswered questions. What is the diagnostic value of CSF biomarkers such as amyloid, we discussed earlier, and tau protein? And, of course, what is the value of CSF and the amyloid-β autoantibodies, if there is any? What is the value of genetic markers such as apolipoprotein E genotype and a correlation with the co-existing inflammation in CAA-ri? However, I don't want to defer this question. So, a typical answer would be that with regard to the ideal patients, we would want a young patient without comorbidities after the first manifestation of CAA-ri who has shown a good clinical and radiological response to corticosteroids in order to define the best second-line therapy. However, before answering all these questions in a clinical trial, if we can, I think that we need to understand the CSF and genetic biomarkers in order to uncover mechanisms regarding pathophysiology that can help us to design more targeted clinical trials studying novel disease-modifying treatments.

Dr. Negar Asdaghi: Thank you.

Dr. Georgios, it's been a pleasure having you on the podcast, and I can say we've learned a lot. We look forward to having you back here and talk about that hypothetical randomized trial, and I'm sure one day hopefully will happen in our lifetime. Thank you for being here.

Dr. Georgios Tsivgoulis: Thank you. Thank you for having me. It was a pleasure.

Dr. Negar Asdaghi: Thank you.

Homer, the legendary Greek poet, described a case of dementia in his seminal work, The Odyssey, in the late eighth century before Christ. He described the cognitive decline of Odysseus's father, King Laertes. The detailed account of the king's mental decline, loss of short-term memory with retention of long-term memory combined with his depression and despair over the loss of his son, is dramatically accurate for a nearly 3,000-year-old description of dementia. Before I ended the interview, I had to use this opportunity to ask Georgios about lessons learned from ancient Greeks and this seemingly timeless disease.

Dr. Georgios Tsivgoulis: Thank you for this question. King Laertes was indeed Odysseus's father, and it's a great paradigm describing dementia. However, the ancient history of dementia may be separated according to the Greek philosopher Posidonius in two periods. The first period is called dementia appearing due to old age, which is called in Greek, eros. And the second one is dementia appearing in other ages and mainly due to other reasons, called morosis. Posidonius of Rhodes was a Greek stoic philosopher of the second first century BC who strongly believed and suggested that morosis, which is that dementia appearing in younger ages due to other disorders, should be treated immediately after its onset. So, if I would like to end this podcast, I would just suggest that CAA-ri could be classified as morosis according to Posidonius. And what we could learn is that the early diagnosis is essential since the prompt initiation of corticosteroids should not be unreasonably delayed.

Dr. Negar Asdaghi: And this concludes our podcast for the January 2023 issue of Stroke. Please be sure to check this month's table of contents for the full list of publications, including a series of Focused Updates on post-stroke neurological recovery, from management of post-stroke attention deficit, neglect and apraxia to post-stroke memory decline. And with this, we end the start of our 2023 podcast series. Like all new things, a new beginning can come with new directions, and sometimes a new direction is all that we need. After all, as the legend has it, it was a direction of that falling apple back in the year 1666 that gave Isaac Newton the idea of the universal law of gravitation. Now, Isaac Newton has, without a doubt, given science some of its biggest discoveries in mathematics, physics, and astronomy. But most may not know that Newton had a pretty rough start in life.

A January-born premature baby, he was thought not to survive the first few days of life. Newton had a difficult childhood, and at the age of 16, he was pulled out of school by his family and forced to become a farmer, a job he didn't like and he was miserably bad at. So, as we start a new year, let's remember that even the smartest people are not good at everything, and it does take time to find one's passion in life. Now, while things may not always be clear, what is clear is that a great way to find that center of gravity is, as always, staying alert with Stroke Alert.

This podcast is produced by Wolters Kluwer and supported by the editorial team of Stroke. Our Stroke Alert podcast and production staff includes Danielle Cross, Eric Goldstein, Nastajjia Krementz, Ishara Ratnayaka, Erinn Cain, Rebecca Seastrong, and Negar Asdaghi. This program is copyright of the American Heart Association, 2023. The opinions expressed by speakers in this podcast are their own and not necessarily those of the editors or of the American Heart Association. For more, visit AHAjournals.org.