Stroke Alert February 2021

On Episode 1 of the Stroke Alert podcast, host Dr. Negar Asdaghi highlights two featured articles from the February 2021 issue of Stroke. This episode also features a conversation with Drs. Fabian Flottmann and Matthew Maros from the Department of Diagnostic and Interventional Neuroradiology, University Medical Center, in Hamburg, Germany, to discuss their article “Good Clinical Outcome Decreases With Number of Retrieval Attempts in Stroke Thrombectomy: Beyond the First-Pass Effect.”

Dr. Negar Asdaghi:          Are women more likely to suffer from stroke than men?

Are oral anticoagulants safe in atrial fibrillation patients with a prior history of GI bleeding?

Does pregnancy increase the risk of intracerebral hemorrhage in patients with cavernous malformation?

Does the number of retrieval attempts during thrombectomy affect the outcomes of stroke patients in whom successful reperfusion is achieved?

In today's podcast, we address some of these topics and much more. You're listening to the Stroke Alert Podcast. Stay with us.

Dr. Negar Asdaghi:         From the Editorial Board of Stroke, welcome to 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 the host of the monthly Stroke Alert Podcast. We're starting our podcasts with the February 2021 issue of the journal, which also features a special section on Go Red for Women stroke, a comprehensive American Heart Association platform to improve the vascular health of women globally. I hope you enjoy it.

Dr. Negar Asdaghi:         Cavernous malformations or cavernomas are angiographically called vascular abnormalities, which can pose an increased risk for intracerebral hemorrhage. Cavernomas can have diverse neurological presentations ranging from mild neurological symptoms to seizures, but in some cases may remain entirely asymptomatic and are diagnosed incidentally as part of routine neuroimaging completed for other reasons. Earlier studies had reported higher rates of intracerebral hemorrhage from cavernomas during pregnancy, and have postulated a hormone-related increased expression of vascular endothelial growth factor or basic fibroblasts growth factors to explain this increased rate. Subsequent studies, however, have failed to demonstrate either progesterone or estrogen receptors in cavernomas. So the question is, should presence of cavernous malformation, whether symptomatic or asymptomatic, influence the reproductive choices of women of childbearing age? In the “Influence of Pregnancy on Hemorrhage Risk in Women With Cerebral and Spinal Cavernous Malformations,” Dr. Nycole Joseph and colleagues from the Departments of Neurology and Neurosurgery from Mayo Clinic Rochester in Minnesota evaluated 365 pregnancies and 160 women with brain or spinal cord cavernomas. They found that during the cumulative 402 years of study follow-up, the risk of hemorrhage amongst non-pregnant patients in the study was 10.4% per year. They found only four patients with clinical hemorrhage during pregnancy, all of which resulted in the cavernomas being first diagnosed. None of the hemorrhages occurred during delivery, and all of the four patients had functionally independent outcomes by three months. Importantly, they found that no patient who became pregnant after the diagnosis of cavernous malformation had a hemorrhage while pregnant. They had a total of 33 pregnancies in the study, including one patient who had previously bled during a prior pregnancy and also patients with brainstem lesions and those who presented with hemorrhage at diagnosis. Both of these are factors for hemorrhage in cavernomas.

So, in summary, in this prospective study, pregnancy did not increase the risk of hemorrhage in women with a known brain or spinal cord cavernous malformation. And the vaginal delivery was safe in this population. The authors concluded that the presence of cavernous malformation should not influence the reproductive choices in women or their type of delivery.

Now, speaking of hemorrhage risk, let's move on to our next paper on anticoagulation therapy in patients with atrial fibrillation. The decision to start anticoagulants for atrial fibrillation can often be challenging in those who have suffered from a prior gastrointestinal bleeding. Prior studies have shown that the non–vitamin K antagonist oral anticoagulants, or NOACs, can carry a comparable and, in some cases, even a higher risk of GI bleed than warfarin. It should be noted that patients with a prior GI bleed were generally excluded from the pivotal randomized control trials that approved NOACs.

And importantly, the risk of bleeding may also be higher in certain race/ethnic groups, such as the Asian population. In the article titled “Non–Vitamin K Antagonist Oral Anticoagulants in Patients With Atrial Fibrillation and Prior Gastrointestinal Bleeding,” Dr. Soonil Kwon from the Department of Internal Medicine, Seoul National University Hospital, in Seoul, Republic of Korea, studied over 42,000 anticoagulant–naïve patients with nonvalvular atrial fibrillation and prior GI bleed from 2010 to 2018 as part of a retrospective, observational cohort study in Korea. They evaluated the risk of ischemic stroke, major bleeding and combined outcomes in this population. What they found was that, not surprisingly, close to 60% of patients were initiated on a NOAC, with rivaroxaban leading dabigatran, apixaban, followed by edoxaban in terms of frequency of agents used. Just over 40% of patients were started on warfarin. Now, over the study follow-up, when they looked at the safety by looking at major bleeding rate and effectiveness by assessing ischemic stroke rates, NOACs generally did better as compared to warfarin, resulting in 39% risk reduction in recurrent stroke, 27% risk reduction in major bleeding and 34% risk reduction in composite outcomes as compared to warfarin.

And the rates of upper and lower GI bleed were similar in NOACs versus warfarin users. NOACs still did better as compared to warfarin amongst patients who suffered from GI bleed as they had a lower transfusion rates and shorter hospital stay. NOACs were also associated with lower risks of fatal clinical outcomes except for fatal GI bleed. So the authors concluded that contrary to some of the prior reports, NOACs may be a better option than warfarin for stroke patients and atrial fibrillation patients with prior GI bleed.

Dr. Negar Asdaghi:         Moving from secondary prevention to acute stroke therapy, our last article discusses how the technical details of endovascular thrombectomy may affect the outcomes in patients with ischemic stroke. So, achieving a successful reperfusion is the cornerstone of improving clinical outcomes in patients undergoing endovascular therapy, but how many retrieval attempts should be made by the interventionist to obtain that desired successful reperfusion is still unclear. Importantly, if successful reperfusion is ultimately achieved, it's still not clear if there's a relationship between the number of retrieval attempts and favorable clinical outcomes.

 Joining me now are doctors Fabian Flottmann and Matthew Maros from the Department of Diagnostic and Interventional Neuroradiology, University Medical Center, in Hamburg, Germany, who are the first and senior authors of the study titled “Good Clinical Outcome Decreases With Number of Retrieval Attempts in Stroke Thrombectomy: Beyond the First-Pass Effect.”

Good morning from Florida, and good afternoon, Fabian and Máté, in Germany. Thank you for joining us.

Dr. Fabian Flottmann:    Thank you very much, Negar, for the nice introduction. Good afternoon from Hamburg. At the moment, it's really, really cold here. It's -4 degrees Celsius. I can't translate it to Fahrenheit, but it's pretty cold, let me assure you. And thank you very much for having us today.

Dr. Negar Asdaghi:         It's great to have you. So I start with Fabian. This is a very interesting and timely study as we're learning more that the way in which we achieve a goal in acute stroke reperfusion therapies is almost as important as the goal itself. Can you tell us a bit about the background of your study, Fabian, and why you felt the need to look at these granular details, which unfold inside the angio suite during endovascular thrombectomy?

Dr. Fabian Flottmann:    Of course, that's a question that's highly relevant for a neurointerventionalist. This research topic developed from our clinical practice, because quite often we have the situation in the angiography suite, where we try to open a vessel, a patient with a large vessel occlusion, and everything is very easy if the vessel opens after one retrieval attempt, because everybody is happy and you can end the procedure. But what happens if the vessel doesn't open? Then you try again. And what happens if the vessel doesn't open? You try it again, and so on and so on. So the question is, when should you stop? And we ask ourselves, are these maneuvers that we do, like three or four or five maneuvers, are they as successful or as beneficial for the patient as the first maneuver?

We did an analysis of our data in Hamburg, and that led to the first paper that we published in Stroke in 2018. And our finding was that the third or fourth retrieval, they were successful in achieving recanalization, but the clinical outcome of those patients was not as good as those patients that you opened with just one retrieval attempt. That was the first finding that we had with our data and our center. And then in the same year, the first pass effect was described. The first pass effect, being the finding that the first retrieval attempt is the most important for the patient. This data was very interesting. And then there were other publications that said, no, there's no connection between the number of retrieval attempts and the clinical outcome. So, as always, in science, when there's more than one opinion, things begin to get interesting. And we said we want to investigate this further. And we decided to do a multicenter study with more patients. And we decided to look at each retrieval attempt separately, to not look just at a first retrieval attempt versus the others, but at each retrieval attempt.

Dr. Negar Asdaghi:         So interesting indeed. Please tell us, before you tell us about the study findings, about the German Stroke Registry. How many years has the registry been active, and how many centers are involved, and please walk us through your study population and the selection process of your study?

Dr. Fabian Flottmann:    Germans Stroke Registry. It's a systematic observational registry study from Germany. It's academic, it's independent, prospective, multi-central, there are 25 centers who participate in this registry. And its goal is to have a systematic evaluation of endovascular stroke treatment in Germany. There are stroke centers from all around the country who consecutively enroll their patients. All patients with an intention to treat in the angiography suite are included. All the patient data are collected at the center and all these data are then centralized and we have a central quality check. And what is important that we also try to include the clinical follow-up information for every patient at day 90. So, the modified Rankin Scale at day 90 is also included. And in our work, we did an analysis of the first 2,600 patients of this German Stroke Registry, and our goal was to eliminate bias. So, for example, we wanted to include data on the stroke severity, the NIHSS score, the amount of early infarction, the ASPECTS score and the location of occlusion, the age of the patient. We selected all the patients that had these data entered. So, we were able to select about 1,200 patients from the German Stroke Registry that fulfilled our inclusion criteria for the present study. To our knowledge, this is the largest multicentric, retrospective study that investigated this effect of retrieval attempts on clinical outcome.

Dr. Negar Asdaghi:         This is really nice because we are really not used to getting granular details and radiographic details in such large numbers. So, the multicenter nature and the large number of patients included in your study are certainly important strengths of your paper, and that should be noted. Now, Matthew, over to you. Please tell us the main findings of the paper.

Dr. Matthew Maros:     So, one specialty of our applied methodology is that we used a generalized mixed-effects models, if we didn't know logistic regression framework. That means that our target variable was the mRS90 and the good functional outcome, defined by zero to two scores by mRS. We also implied this framework to be comparable to the HERMES meta-analysis by Goyal et al. And we investigated, in our primary analysis, the effect of age, the baseline stroke severity NIHSS score, ASPECTS score, and also the main explanatory variable that we investigated was the successful reperfusion at N-th retrieval attempt. And we found that, so as one would expect, a younger age and the less severe stroke clinical manifestation, like NIHSS score, was inversely associated with a good functional outcome. So, younger patients and less severe stroke were associated with a favorable outcome. And also, a less severe ischemic changes on a non-contrast head CT, so ASPECTS score eight, nine or ten, were also independent predictors for a good function outcome at 90 days. Our main finding was that the success at the first, second, or third retrieval attempts were significantly and independently associated with a good functional outcome. And interestingly, the effect of the consecutive retrieval attempts were gradually diminishing from an odds ratio from six (around) to three.

Dr. Negar Asdaghi:         This is interesting. So, basically, what you found is that you go in with the first attempt, second and third, you don't achieve that successful recanalization. If you achieve your successful reperfusion after the third attempt, it's good, but not so good, meaning that it doesn't translate to that beautiful, favorable outcome at 90 days as it did for the first three attempts.

Dr. Matthew Maros:      So, for four or more retrieval attempts, this positive effect on the outcome has flattened, so the curve is more like a sigmoid curve that was asymptotic to a virtual threshold.

Dr. Negar Asdaghi:         Understood. So, I find it very interesting that this decline in the odds of favorable outcome, despite successful reperfusion, was not simply a factor of time, meaning that, if you tried once and you achieve reperfusion right away, it's so much faster. And of course, time is brain, but if you try five times, it would take longer. It is interesting in your results and your multivariate analysis that even if you adjusted for the factor of delay in time, the results persisted. Could you please tell us about your multivariate analysis and what other factors and co-founders you adjusted for?

Dr. Matthew Maros:      Exactly. So, as a sensitivity analysis, we also included the time from groin puncture to flow restoration and also sex, and also to be almost identical or highly similar to the model applied in the HERMES meta-analysis. We also included the site of the intracranial occlusion and better intravenous thrombolysis was administered or not. And in the sensitivity analysis, we had almost 90% of our dataset. So almost a thousand one hundred patients. And we found that all the effects of age and NIHSS score stayed significant, and also the effect of the first, second and third retrieval attempts associated with good functional outcome at 90 days were also significant. While interestingly, the effect of intravenous thrombolysis, and also the ASPECTS score, had diminished, but also just narrowly escaped a significant threshold. And interestingly, the effect of time, so time from groin puncture to flow restoration, seemed to be not relevant or be interpreted that way, that the number of retrieval attempts and the effect that we see is not a surrogate of time, that it simply takes longer to perform the interventions, but it's the true effect of achieving recanalization at a certain attempt.

Dr. Negar Asdaghi:         So, what should be our takeaway from your study, Fabian? Is three that magic number? Are we asking the interventionalist to stop the procedure after the third retrieval attempt if they're unsuccessful, and what should the future hold in terms of studies on this project?

Dr. Fabian Flottmann:    That's the most important question. Of course, we have to keep in mind that every patient and every intervention is different. The decision to continue or stop the thrombectomy procedure is a very important decision, which is taken by the neurointerventionalist together with his team. And they will take into account multiple factors, including patient's biography, medical history at time from symptom onset, image data, and so on. Our study can provide some guiding information when making this decision. And yes, three could indeed be called a magic number in the following sense. We would like to encourage interventionalists to make at least three attempts in case of persistent occlusion, because we can see a clear benefit even when reperfusion is achieved after the third attempt. Then, in patients with younger age and/or, for example, a good ASPECTS score, even more retrieval attempts are probably warranted regardless of IV thrombolysis, site of occlusion and potentially increased procedure time.

Of course, in all these retrospective studies, a bias remains. We don't know why the procedure was stopped in each case. The best thing would be a randomized controlled trial with the following design. You could, in case of persistent occlusions, after two retrievals, randomize to continue or to stop the procedure. And then we would know what the right answer is. So, taken together, our study suggests that in EVT for anterior circulation strokes, at least three retrieval attempts should be performed in cases of persistent occlusion, and up to five attempts of beneficial association with good clinical outcome is expected.

Dr. Negar Asdaghi:         Doctors Fabian Flottmann and Matthew Maros, thank you very much for joining us and congratulations on this work. And this concludes our podcast today. Don't forget to check the February table of contents for the full list of publications, including original contributions, brief reports, editorials, and our special section on Go Red for Women stroke. Until our next podcast, stay alert with Stroke Alert.