On Episode 22 of the Stroke Alert Podcast, host Dr. Negar Asdaghi highlights two articles from the November 2022 issue of Stroke: “Estimating Perfusion Deficits in Acute Stroke Patients Without Perfusion Imaging” and “Five-Year Results of Coronary Artery Bypass Grafting With or Without Carotid Endarterectomy in Patients With Asymptomatic Carotid Artery Stenosis.” She also interviews Dr. George Ntaios about his article “Incidence of Stroke in Randomized Trials of COVID-19 Therapeutics.”
Dr. Negar Asdaghi: Let's start with some questions.
1) What is the actual incidence of stroke after COVID-19?
2) In the setting of acute ischemic stroke, can the volume of ischemic penumbra be estimated with just a regular MRI study of the brain without any vascular or perfusion imaging?
3) And finally, can a patient with significant carotid stenosis go through coronary artery bypass graft surgery?
We're back here to answer these questions and bring us up to date with the latest in the world of cerebrovascular disorders. You're listening to the Stroke Alert Podcast, 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. The November issue of Stroke is packed with a range of really exciting and exceedingly timely articles. As part of our Original Contributions in this issue of the journal, we have a post hoc analysis of the Treat Stroke to Target, or the TST, randomized trial by Dr. Pierre Amarenco and colleagues. We've talked about this trial in our past podcast, and the main study results that were published in New England Journal of Medicine in January of 2020. TST randomized patients with a recent stroke or TIA to either a low target of LDL cholesterol of less than 70 milligram per deciliter or a target LDL of 90 to 110. The main study showed that the low LDL target group had a significantly lower risk of subsequent cardiovascular events without increasing the risk of hemorrhagic stroke. So, from this, we know that achieving a low target LDL is possible and is actually better than the LDL target of 90 to 110 post-stroke.
But in the new paper, in this issue of the journal, in a post hoc analysis of the trial, the TST investigators showed that it's not just achieving that magic low target LDL of less than 70 that's important in a reduction of cerebrovascular disorders, but it's also how we achieve it that determines the future of vascular outcomes. So, in this analysis that compared patients on monostatin therapy to those treated with dual cholesterol-lowering agents, that would be a combination of statin and ezetimibe, and showed that in the low LDL target group, only those patients treated with dual therapy had a significant reduction of subsequent vascular events as compared to those in the higher LDL category.
But the same was not true for patients on statin monotherapy, even though they had all achieved a low target LDL. Think about this for a moment. Both groups, whether on statin monotherapy or on dual anti-cholesterol treatments, achieved the same low target of LDL, but only those on dual therapy had a lower risk of subsequent vascular events as compared to those that were in the higher LDL target group. Very thought-provoking study.
In a separate paper by Dr. Shin and colleagues out of Korea, we learned that survivors of tuberculosis, or TB, are at a significantly higher risk of ischemic stroke than their age- and risk factors–matched non-TB counterparts. The authors used data from the Korean National Health Insurance Services and studied over 200,000 cases diagnosed with TB between 2010 and 2017 and compared them to a pool of over one million non-TB cases for matching. And they found that the risk of ischemic stroke was 1.2 times greater among TB survivors compared to matched non-TB cases after adjusting for the usual confounders, health behavioral factors, and other comorbidities.
Now, why would TB increase the risk of stroke? The authors talk about the pro-inflammatory state of this condition, thrombocytosis, that is a known complication of chronic TB amongst other putative and less clear mechanisms. But what is clear is that findings from a large-scale population-based cohort such as the current study support an independent association between TB and ischemic stroke. As always, I encourage you to review these papers in addition to listening to our podcast today.
My guest on the podcast today, Dr. George Ntaios, joins me all the way from Greece to talk to us about the much discussed topic of the risk of stroke in the setting of COVID-19. Dr. Ntaios is the President of the Hellenic Stroke Organization and an experienced internist who has been fighting this pandemic in the front lines since the beginning. In an interview, he talks about his recently published paper, his experience, and the lessons learned on balancing scientific rigor against the urgency of COVID-19. But first, with these two articles.
In the setting of a target vessel occlusion in patients presenting with an acute ischemic stroke, distinguishing the ischemic core from the ischemic penumbra is of outmost importance. The success of all of our reperfusion therapies heavily lies on our ability to differentiate between the tissue that is already dead, which would be the ischemic core, from the tissue that is not dead yet but is going to die unless revascularization is achieved. That is the ischemic penumbra.
Over the past two to three decades, there's been lots of debate over how these entities of dead tissue versus going-to-die tissue are best defined, especially when we're making these distinctions under the pressure of time. We don't even agree on the best imaging modality to define them. Should we rely on CT-based imaging? Do we stop at CT, CT angiogram? Should we do single-phase CTA or multiphase CTA? When do we perform CT perfusion, and what perfusion parameters best define core and penumbra, or should we rely on MRI-based modalities altogether?
These questions have all been asked and extensively studied, which is why, as a field, I think, we have at least some agreements today on the basics of core and penumbra definitions. And I also think that overall we are becoming better at doing less imaging to be able to predict tissue outcomes in real time. And there's definitely a growing interest in trying to estimate tissue fate based on a single-imaging modality. So, I think you're going to find an Original Contribution in this issue of the journal, titled "Estimating Perfusion Deficits in Acute Stroke Patients Without Perfusion Imaging," really interesting.
In this paper, Dr. Richard Leigh from the National Institute of Neurological Disorders and Stroke, National Institutes of Health, in Bethesda and colleagues evaluated patients with acute ischemic stroke enrolled between 2013 to 2014 in the NINDS Natural History of Stroke study. A little bit about the study: It enrolled stroke patients presenting to three hospitals in Washington, DC, and Maryland with serial MRI scans during the acute and subacute time period after ischemic stroke. For this particular paper, they included patients who received MRI and perfusion-weighted imaging and included only those who were thrombolized. That was their way of ensuring that all patients in their study were in the hyperacute stage of stroke.
They then looked at their MR imaging, specifically the fluid-attenuated inversion recovery, or FLAIR, images, for a presence of something called hyperintense vessels in the ischemic territory. Now, this is an audio-only podcast, so unless you're Googling FLAIR hyperintense vessels on MRI, to follow along, I have to take a bit of time explaining this entity. What do we mean by FLAIR hyperintense vessels? We are not just talking about the T2 hyperintense signal that's sometimes noticeable at the site of proximal occlusion. For example, in the setting of an M1 occlusion, we may be able to detect a T2 hyperintense signal at the site of M1 on FLAIR. That's not the point of this paper. The point is to look throughout the area supplied by that said target occlusion, in this case all of the MCA, and see whether there is hyperintense signal in all arteries in that potentially ischemic tissue and how the area delineated by these FLAIR hyperintense vessels could potentially correspond to the area of perfusion deficit on conventional perfusion imaging. It turns out that these hyperintense vessels actually map a pretty large area. So, this is the point of this study. The investigators developed a FLAIR hyperintense vessel scoring system and called it NIH, obviously, because this was a National Institutes of Health study, FHV, which stands for FLAIR hyperintense vessel, scoring system. And the score is based on presence of these hyperintense vessels in three vascular territories: ACA, MCA, or PCA.
Now, seeing that MCA is a larger territory, they had to further divide it into four sub-regions: frontal, insular, temporal, and parietal. So, in total, we have six regions now. Each of them would get a score of zero if there were no hyperintense vessels in them, and a score of two if there were three or more FLAIR hyperintense vessels in a single slice, or if there were three or more slices that contained FLAIR hyperintense vessels. And, of course, a score of one would be anything in between. So, we have six regions in total, each maximum getting two points, to give us a composite score of maximum 12 for this scoring system.
So, they wanted to see whether there's a correlation between the FLAIR hyperintense vessel score and the volume of perfusion deficits that is detected by conventional perfusion imaging, which is their main study result. But before we go there, it does seem like a lot of work to learn all these regions and count all these hyperintense vessels in these six regions and come up with an actual score. So, they had to do an interrater reliability to see how easy it is to score and how reliable are these scores. So, they had two independent reviewers for their study. On average, the scores of these two independent reviewers differed by one point for a κ of 0.31, which is quite a low interrater reliability.
But when they looked at a more liberal way of assessing interrater reliability, where partial credit was given, when the raters were at least close in their scoring, the κ improved to 0.65 for a moderate degree of agreement. So, what that means is that it's not easy to learn the score, and potentially I can give a score and another colleague can give a different score. So, we have to keep that in mind. But I want to emphasize that in the field of stroke neurology, we are kind of used to these poor interrater reliability agreements in general. For example, the interrater reliability of the ASPECTS score, a score that is commonly used in our day-to-day practice, and especially in the acute phase, we communicate the extent of early ischemic changes by using the ASPECTS score, has a pretty poor interrater reliability, especially in the first few hours after the ischemic stroke. So, we can make due with a κ of 0.65.
Now on to the results of this study. They had a total of 101 patients. Their median age was 73. The median FHV, which is that FLAIR hyperintense vessel score, in their entire cohort was four. And close to 80% of patients enrolled in their study had some perfusion abnormalities on their concurrent perfusion-weighted imaging. Now, briefly, they defined perfusion deficits as areas with delay in the relative time to peak map, or TTP maps, after applying a six-second threshold to these TTP maps. Of note, half of those patients with a perfusion deficit had a significant perfusion deficit, which meant that they had 15 cc or more of perfusion deficit.
OK, now on to the main study results. Number one, the score obtained by NIH FLAIR hyperintense score highly correlated with the volume of perfusion deficit. In fact, every one point increase on the NIH-FHV score was approximately equal to 12 cc of perfusion deficit. That's a really useful way of thinking about this score. Each score translated in 12 cc of perfusion deficit.
Number two, when they looked at the predictive ability of this score in predicting the presence of significant perfusion deficit, that is 15 cc or more of perfusion delay, the area under the curve was 0.9, which is quite high. This is quite reassuring that the FHV score was sensitive and specific in predicting the presence of significant perfusion deficit.
Next finding, how does this score do in predicting a significant mismatch? They calculated mismatch ratio by dividing the perfusion volume to that of ischemic core as measured by diffusion volume as it's done conventionally, and then did the same for the score with the exception that instead of using the perfusion volume, they actually used this score and divided it by diffusion volume. And it turns out that FLAIR hyperintense mismatch ratio had a strong predictive capability in predicting the mismatch ratio of 1.8.
So, in summary, if this score is validated in larger studies, it can potentially be used as a quick and dirty way of calculating the amount of perfusion deficit in the setting of target vessel occlusion. And, of course, it can also be used as a predictive way of presence of significant perfusion deficit, which is perfusion deficit of over 15 cc. This is all without the need to do actual perfusion imaging. Now, all we've got to do is to get comfortable with this scoring system and, of course, be able to multiply it by 12 to give us a quick guesstimate of the perfusion volume. And one final word on this is that I think the future of stroke imaging is not in doing more images, but to be able to extract more information from less imaging in the acute setting.
Stroke physicians were frequently consulted to see patients that are scheduled to undergo coronary artery bypass graft surgery, or CABG. The stroke consult would be for the optimal perioperative management of an often incidentally found carotid disease. Now, why do I say we were frequently consulted? Because at least anecdotally in my own practice, I feel that over the past decade, the number of these consults has substantially reduced. Why is that? Well, let's dive into this topic and review some of the literature.
First off, around 40% of patients who have active coronary artery disease and are scheduled to undergo CABG have concurrent carotid disease, and about 10% of CABG patients have evidence of hemodynamically significant carotid disease. And seeing that the risk factors for coronary artery disease are similar to those causing carotid disease, these high percentages are not surprising at all. But the question to ask is, can we put a patient with significant carotid disease through cardiac surgery? What is the perioperative risk of stroke in this situation? And importantly, should the carotid disease be surgically treated during carotid surgery? This is referred to as synchronous carotid endarterectomy, or CEA plus CABG. Or the carotid disease should be treated either surgically or endovascularly before CABG? We refer to this as staged carotid surgery or post-CABG. This is known as reverse staged carotid surgery. All of these questions are asked from the stroke physicians in that consult, and, like many of you, I have struggled to find the evidence to answer some of them. So, let's briefly review some of the current literature on this topic.
The CABACS trial, the acronym stands for the Coronary Artery Bypass Graft Surgery in Patients With Asymptomatic Carotid Stenosis, was a randomized controlled trial that included patients undergoing CABG who are found, exactly like that consult, to have an asymptomatic carotid disease of equal or greater than 70% stenosis. The carotid disease for this trial had to be amenable to carotid endarterectomy, or CEA, and the patients were randomized to either receive synchronous CEA plus CABG or just go through with the CABG alone. The trial started in 2010 and planned to enroll over a thousand patients, but was stopped, unfortunately, prematurely in 2014 due to slow recruitment and withdrawal of funding after only 129 patients were enrolled from 17 centers in Germany and Czech Republic.
The original study was published in this journal in 2017. So, what did it find? In their intention-to-treat analysis, the primary outcome of any stroke or death at 30 days was 18% in patients receiving synchronous CEA plus CABG as compared to only 9% in patients receiving isolated CABG. Ouch, a double risk of stroke in those who had concurrent surgical treatment of their carotid disease in addition to CABG. Now, this was an underpowered study, and the results should be understood in that context, but it really didn't appear that synchronous CEA plus CABG would decrease the rate of stroke in the first 30 days.
Now, how about the long-term outcomes of these patients? We know that asymptomatic carotid disease carries a cumulative annual risk of stroke, and it's important to see if the risk of subsequent stroke was lower downstream if the carotid was already fixed early on. So, in the current issue of the journal, the CABACS trial investigators, led by Dr. Stephan Knipp from the Department of Thoracic and Cardiovascular Surgery in Essen, Germany, and colleagues are back with the five-year results of this trial. How did synchronous CABG plus CEA do as compared to CABG alone? Well, by five years, the rate of stroke or death was 40% in the combined group and 35% in the CABG-only group. This was not a statistically significant difference.
Now, when they broke down the primary outcomes, the rate of death from any cause was similar in the two groups. By five years, the mortality rate was 25% in the combined group and 23% in the CABG-only group. And the same was true for the rate of nonfatal strokes. And also the cumulative rate of nonfatal strokes from year one to year five was similar between the two groups, which meant that the higher stroke risk early on in the CABG plus CEA group was not counterbalanced by decreased rate of stroke later on during the long-term follow-up.
And finally, they looked at the rate of disability-producing stroke. First of all, after the first year, no new disabling strokes were observed in either group. That's great news. However, in the early period, unfortunately, close to half of strokes that had happened after the combined CEA and CABG were disability-producing, and about a third of strokes that happened after CABG alone were also disability-producing. So, in summary, even though this study is quite underpowered, it appears that performing synchronous CEA plus CABG increases the preoperative morbidity and mortality in patients with asymptomatic carotid disease without providing any long-term benefits to these patients.
Coronaviruses are important human and animal pathogens. By now, I think it's safe to say that most of the population of the world has heard of at least one of the members of the coronavirus's family, which was first identified in late 2019 as the cause of a cluster of cases of pneumonia in Wuhan, China. In the early months of 2020, COVID-19, the disease caused by this novel coronavirus, would rapidly spread to involve much of the world. And on March 11 of the same year, the World Health Organization declared COVID-19 a pandemic. Today, over two and a half years have passed since that day, and an avalanche of scientific papers have since been published about COVID-19, not just in medicine, but in each and every imaginable field of life. Neurology's, of course, no exception.
The clinical presentation of COVID-19 largely depends on the severity of the disease and may range from a simple asymptomatic infection to a severe, lethal, multi-organ disease. In the world of neurology, a myriad of neurological symptoms, from loss of sense of taste and smell to headache, all the way to encephalopathy and seizures, have been reported in association with this disease. Early in the pandemic, some studies suggested that COVID-19 is indeed a risk factor for stroke. Like many severe infections, COVID-19 can potentially cause a prothrombotic state and can be associated with thromboembolic events. But most of these earlier studies were smaller observational studies that were completed in an inpatient setting, including those with severe COVID.
In fact, to date, we still don't have an accurate and reliable estimate of stroke incidence among patients with COVID-19. On the other hand, stroke is the second leading cause of death globally and the fifth cause of death in the US. In the United States, every 40 seconds, someone has a stroke, and every four minutes, someone dies of a stroke. So, I think the question that everyone should be asking is, has COVID-19 changed this statistic?
In this issue of the journal, in the study titled "Incidence of Stroke in Randomized Trials of COVID-19 Therapeutics: A Systematic Review and Meta-Analysis," Dr. Ntaios and colleagues aim to get us a step closer to answering this very important question. Dr. Ntaios is an Associate Professor of Medicine at the University of Thessaly in central Greece, and he's the current President of the Hellenic Stroke Organization. It is my great honor to have Dr. Ntaios today in our podcast to discuss this paper and all things stroke-related COVID-19. Good afternoon, George, and welcome to our podcast.
Dr. George Ntaios: Thank you for the invitation, Negar, and for highlighting our work. It's a pleasure to be here with you today.
Dr. Negar Asdaghi: Thank you for being here, and congrats on the paper. George, can you start us off by discussing the pathophysiological mechanisms through which COVID can potentially cause a stroke?
Dr. George Ntaios: Well, one of the most attractive things about stroke, which makes it fascinating for all of us, is its complexity. So many different pathologies can cause stroke, and, quite frequently, identifying the actual cause of stroke can be really challenging. And in a similar way, the pathophysiological association between COVID and stroke seems to be, again, complex. Different pathways have been proposed. Internal, we talk about two broad mechanisms. One is the vascular inflammation and thrombosis, and the other is cardioembolism. And there are several pathways which are involved in vascular inflammation and thrombosis: activation of the complement, activation of the inflammasome, activation of thrombin, increased production of [inaudible 00:24:47] constriction, state of stress, platelet aggregation, vascular thrombosis.
So, collectively, this thromboinflammation could lead to damage of the neurovascular unit and consequently to stroke. And in a similar way, there are several cardiac pathologies which can cause stroke in a COVID patient, like acute left ventricular dysfunction, which can be caused, again, by several mechanisms, like coronary ischemia, stress-induced takotsubo cardiomyopathy, myocarditis inflammation, or also as a result of direct effect of the coronavirus at the myocardial cell. And, of course, we should not forget about atrial fibrillation, which seems to be more frequent in COVID patients.
So, we see that the proposed mechanisms behind the association between COVID and stroke, that is, vascular thromboinflammation on one hand, or cardioembolism on the other hand, are complex, but whether these derangements they have a clinically relevant effect or they are just biochemical derangements without any clinical effect is a debate. For example, the incidence of myocarditis in COVID is about 0.2%. That is, in every 500 COVID patients, you have one patient with myocarditis. But myocarditis has a very wide clinical spectrum ranging from subclinical elevation of myocardial enzymes to full and life-threatening disease. So, obviously, the incidence of severe myocarditis is even lower than 0.2%. And the same is true also for the incidence of myocarditis after COVID vaccination. The CDC estimates that one case of myocarditis occurs every 200,000 vaccinations, with the number being slightly higher in young men after the second dose. And this is extremely rare, and the huge majority of these myocarditis cases, they're mild.
So, this is about ischemic stroke. Now, with regard to hemorrhagic stroke and its association with COVID, again, it seems to be, again, very rare. The best estimate that we have comes from the Get With The Guidelines – Stroke Registry and is about 0.2% and involves mainly patients who are already on anticoagulants. So, they had already a risk factor for ICH. So, again, whether all these pathophysiologic derangements in COVID patients, they have a clinical meaningful association with stroke risk or not, I think it's a matter of debate.
Dr. Negar Asdaghi: Wow, George, it was a simple question, but it seems like the answer was not that straightforward. Let me just recap some of the things you mentioned. So, first of all, the answer is not straightforward and depends on whether we're talking about ischemic stroke or hemorrhagic stroke. There seems to be a lot of connecting points, at least, so to speak, between COVID and either forms of stroke. But you touched on two major sort of broad mechanisms. One is the idea of vascular thromboinflammation that goes along the lines of many sort of hyperacute, hyperinflammatory processes that can occur, especially in the setting of severe COVID. You touched on activation of thrombin, complement activation, platelet aggregation, sort of an activation of that microvascular or vascular unit in a sense.
And then a second mechanism you touched on is the impact of COVID on the myocardium on sort of many different pathways. Again, you talked about acute left ventricular dysfunction, stress-induced myocarditis, and the impact of COVID on perhaps increasing the rate of atrial fibrillation. Again, these are all very complex associations, and some could be already present in a patient who is perhaps of an older age, and COVID is just a modifier of that risk factor that was already present in that particular person. And you also touched on how COVID can potentially increase the risk of hemorrhagic stroke, but the study seems to suggest that those patients already had risk factors for the same. And perhaps, again, COVID is a modifier of that risk factor.
All right, so with that information, a number of studies early on, especially, in the pandemic and later, some meta-analyses, have aimed to estimate the incident rate of stroke post-COVID. Can you please briefly tell us what were their findings, and how is your current paper and current meta-analysis different in terms of methodology from those earlier studies?
Dr. George Ntaios: Yes. Well, it all started from this letter to the editor at the New England Journal of Medicine. It was published very early in the pandemic during the outbreak in New York. And in this letter, the authors had reported that within a period of two weeks, they had five young patients with COVID and large artery stroke, which they commented that it was much higher than their typical, actually their average, of 0.7 cases during a two-weeks period within the last year. And remember that back then, we knew literally nothing about COVID. So, this letter was really a huge, loud alert that something is going on here and that perhaps our hospitals would be flooded with COVID patients with stroke.
So, subsequently, several reports were published aiming to estimate the incidence of stroke in COVID. Rather contradictory with the incidence, estimates are ranging from as low as 0.5% to even 5%. However, these estimates could well be inaccurate. They were observational studies. Most of them were limited to the inpatient setting. Most of them were single-center studies. Most of them, if not all, were retrospective studies. So, there was really a high risk of registration and assessment bias, as well as reporting bias. And also remember that back then during the outbreak, people were really reluctant to visit the hospital, even if they had a serious condition like stroke, an urgent condition, which means that the real incidences could be even higher.
So, it was our feeling that these estimates were perhaps inaccurate. And there are also some meta-analyses of these studies which estimate that the incidence of stroke in COVID is about 1.5%. But, of course, any meta-analysis is as good as the studies it includes. So, we tried to find a way to have a more accurate estimate than these estimates. And we followed a different methodology. We studied randomized trials of COVID therapeutics, and we looked for strokes reported as adverse events or as outcome events. And the good thing about randomized trials is the rigorous assessment and reporting of outcomes in adverse events. So, we think, we believe, that this methodology provides a more reliable and a more robust estimate of stroke incidence in COVID patients.
Dr. Negar Asdaghi: OK. George, it's very important what you just mentioned, so I wanted to recap for our listeners some of the things you mentioned. It all started with a letter to the editor of New England Journal of Medicine on a report of five young patients that had large vessel occlusion in the setting of COVID. And then, basically, the floodgates opened in terms of all these observational studies that basically reported the same. And subsequent to that, meta-analyses that were completed containing those observational studies predominantly gave us an incident rate of 0.5 to 5%. That's much, much higher than basically the non-COVID–associated incidence rate of stroke in the population-based studies, and basically suggested that COVID-19 is indeed a major risk factor for all types of stroke.
So, that's where it all started. And, as you alluded to, these numbers had to be reverified in bigger settings, more controlled setting. And you already answered my next question, which is the difference between those studies and prior meta-analyses to the current meta-analysis is that you basically took the simple question and started looking at it in a controlled setting of randomized trials. And you already answered this question of the methodology, but I want to recap. You took basically patients included in randomized trials of therapeutics for COVID-19, various therapies for COVID-19, and you did a meta-analysis to see what were the incident rate of stroke as an outcome in these trials.
So, with that, could you please tell us a little more about the population that you had in this meta-analysis in terms of their age, the types of therapies that these randomized trials had looked at, and the duration of the follow-up, please?
Dr. George Ntaios: The follow-up included 77 randomized trials, which corresponds to more than 38,000 COVID patients. The mean age of these patients was about 55 years of age, and they were followed for an average of 23 days after study enrollment. With regard to the set strategy, I think it was not strict at all. I would rather say it was very liberal. We allowed trials of any drug in COVID patients of any age, of any severity, coming from any setting: outpatient, inpatient, either general ward or intensive care unit. And from any country. I don't think that we could achieve a wider inclusion than this strategy did. And the huge majority of patients, more than 95%, they were hospitalized patients. So, by definition, they had severe COVID disease. And the drugs studied in these trials included everything that was actually tried in COVID, including tocilizumab, IL-6R inhibitors, steroids, remdesivir, chloroquine, azithromycin, ritonavir, interferon, ivermectin, and many other drugs. So, I think we tried to include as many trials as possible.
Dr. Negar Asdaghi: OK. So, let me see if I got it. You basically included 77 randomized trials. It is a younger population of patients in these trials, median aged 55. You had a total of over 38,000 patients. It's a great sample size for this meta-analysis. And importantly, the duration of follow-up is median of 23 days. And it's just about any treatments we've heard that have been tried for COVID, from dexamethasone to remdesivir and ivermectin. And a rigorous methodology. So, I think we're ready to hear the primary results of this meta-analysis. How many strokes happened in these patients?
Dr. George Ntaios: In the overall population, that is both in the hospital and in the outpatient setting, there were totally 65 strokes in these 38,000 COVID patients, which corresponds to one stroke every 600 COVID patients or else an incident of only 0.16%, 0.16%. This is very low, much lower than the previous estimates. And, of note, all strokes occurred in hospitalized patients. There were no strokes at all in the ambulatory COVID patients. So, just to repeat the result, we just found that only one patient will have a stroke every 600 COVID patients who are either hospitalized or are ambulatory.
Dr. Negar Asdaghi: OK. So, I need to have these numbers, I think, committed to memory, especially when we speak to family members and patients in the hospital. Ninety-five percent of the patient population of this meta-analysis were inpatient COVID. So, by definition, they must be sicker in terms of the severity of their COVID disease. Out of 38,000 patients, you had 65 events of stroke. So, these are very, very important numbers, a lot basically lower than the incidence rate reported from prior studies. So, I wanted to ask you about the sensitivity analysis that was done in the meta-analysis.
Dr. George Ntaios: Yes. When we designed the analysis, we were expecting that we would find numbers was similar to those reported before. So, we thought that perhaps a sensitivity analysis would be able to increase the confidence and the robustness of the results. That's why we did this sensitivity analysis. However, it proved that the number of strokes, the number of outcome events was much lower than what expected. So, the power for those sensitivity analysis to show a meaningful conclusion was low. So, actually, that's why we don't comment at all on those sensitivity analysis because there were so few strokes to support such an analysis.
Dr. Negar Asdaghi: OK. So, basically, you had a priori design the meta-analysis based on the assumption that the incidence rate of stroke would be a lot higher, but then later on, when the incidence rates was lower, then the sensitivity analysis didn't really give any meaningful data to us. So, I mean, I think we already talked about this, but I want to ask you, why do you think that the incidence rates were so much lower in your analysis than the prior meta-analysis?
Dr. George Ntaios: I believe that our estimate is quite accurate. I think that the reports of stroke incidence published during the pandemic possibly overestimated the association. I think that the early concern that we all had in the beginning, that we would be flooded with strokes during the pandemic, was not confirmed. I think that we can support with decent confidence that stroke is a rare or perhaps very rare complication of COVID.
Dr. Negar Asdaghi: Right. That's great news. That really is great news, and we take every bit of good news in these trying times. George, something that was not touched on in the paper, but I want to ask you and basically get your opinion on this matter, is a much talked about concept in the COVID literature of how COVID could potentially modify certain risk factors. There are much talk about how people with pre-existing diabetes or obesity can potentially develop more severe COVID and, hence, have more complications of COVID, including stroke. What is your clinical experience on this matter, and do you think there are certain predictors of development of COVID-associated stroke?
Dr. George Ntaios: That's a very good point. For the last two years, I was involved in the hospitalization management of COVID patients. So, what we see is what is also described in the literature, that there are certain patient characteristics that predispose them to severe COVID. For example, obesity, for example, older age, pregnancy. Perhaps our analysis was not designed to respond to this question. The data available on the studies that were included, they could not support such an analysis. So, I cannot provide information from our study. But the fact that all strokes in our study, they occurred in hospitalized patients and none of them occurred in ambulatory patients, confirms what is known, that those strokes occurred in patients who, by definition, they have severe COVID disease. So, they confirm this putative association that perhaps severe COVID is associated with stroke rather than just mild COVID.
Dr. Negar Asdaghi: All right. Thank you. And I just want to end with this simple question that I get asked often, and I want to see how you respond to patients or their loved ones when you're asked this question: "Doctor, did COVID give me a stroke?" How should we answer that question?
Dr. George Ntaios: Yes. As we discussed, I think that stroke is a rather rare or perhaps very rare complication of stroke and certainly less frequent than we initially thought. And in those stroke patients who had already other pathologies which can cause stroke, I would be rather reluctant to attribute it to COVID. I would be perhaps more willing to do so in younger patients, but again, only after exhaustively looking for another cause, like PFO, dissection, etc.
I mean, the concern is that if we as the treating stroke physicians assume that the stroke is caused by COVID, then we might discourage patients from doing the necessary diagnostic workup to find the actual cause of stroke. And if it happens, then perhaps an underlying pathology may be missed, which means that the patient will remain vulnerable to stroke recurrence. So, in general, I'm rather very reluctant to say that the stroke is caused by COVID unless a really thorough diagnostic workup shows nothing else at all.
Dr. Negar Asdaghi: All right. Very important message now to all practicing clinicians is don't stop at COVID. Don't just say simply, "Oh, this is COVID. COVID gave you a stroke." Keep looking for potential causes of stroke. Still do put that patient in the category of potentially ESUS or cryptogenic stroke if no other causes are found. And keep in mind that stroke is rare or, as George said, a very rare complication of COVID. Dr. George Ntaios, this is an exceedingly timely topic and a very important contribution to the field. Congratulations again on your paper, and thanks for taking the time to chatting with us today.
Dr. George Ntaios: Thank you for the wonderful discussion, Negar, and for the focus of our work.
Dr. Negar Asdaghi: Thank you.
And this concludes our podcast for the November 2022 issue of Stroke. As always, please be sure to check out the table of contents for the full list of publications, as we can only cover a fraction of the incredible science published in this journal each month. And don't forget to check our fantastic Literature Synopsis. In this month's issue, we read a short summary of the ACST-2 trial published in Lancet on the results of a randomized comparison of stenting versus endarterectomy in asymptomatic carotid disease patients with over 60% of carotid stenosis.
We also have the results of the CASSISS randomized trial, which was published in JAMA earlier this year, and it studied the effect of stenting plus maximal medical therapy versus maximum medical therapy alone on the risk of subsequent stroke and death in patients with symptomatic intracranial stenosis, either in the anterior or in the posterior circulation. CASSISS did not show that stenting was superior to maximum medical therapy, and sadly, these patients remain at a substantial risk of recurrent stroke despite being on best medical therapy.
But I wouldn't be too despondent about the future of interventional therapy for intracranial atherosclerotic disease. After all, we've come a long way since Dr. Charles Thomas Stent, an English dentist, started experimenting with products to advance the field of denture making around 1865. The work that Dr. Stent had started would be continued by his two sons, both dentists, to eventually make its way to products to create surgical tools. But it would be another 100 years before the first percutaneous coronary procedure was completed in 1964. And in honor of Dr. Stent's pioneering work, the device used to keep the coronaries open was named, you guessed it, stents.
Today's stroke care cannot be imagined without the use of various stents, and there's no doubt the future is promising for ways in which we will be able to safely treat intracranial atherosclerotic disease amongst all other vascular disorders. And what better way to keep our enthusiasm than staying alert with Stroke Alert.
This program is copyright of the American Heart Association, 2022. 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.