Industry Corner: Silent Seizures - The Role of EEG in ICU
Okay. Ladies and gentlemen, welcome to this podcast of the European Society of of Anesthesiology and Intensive Care. We remind that, this industry corner episode is currently supported by Nihon Codden. And I'm your host today. My name is Chiara Robba.
Speaker 1:I am a professor in anesthesia and intensive care at Polyclinico San Martino in Genovahi in Italy. And today, we will be discussing a very important topic. The title is silent seizures, The Role of EEG in the ICU. And here we have invited professor Matthias Gendelblom. He's a clinician scientist and clinical attending at the Department of Neurology University Medical Center of Hamburg, Germany.
Speaker 1:Welcome, Matthias.
Speaker 2:Yaura, thank you very much. Okay. Very happy to be here and to have a chance to talk to you.
Speaker 1:Thank thank you very much for being here with us. So I have a couple of questions for you regarding the application of EEG, silent seizures, the use in the ICU. So first of all, can you please give us sort of a background regarding applications of EEG in the intensive care unit, please?
Speaker 2:I mean, it's, the EEG is unique for us when we look at the different diagnostic tools we have in a in the clinical setting. I mean, in general, as a neurologist, but also in other disciplines, we will rely very much on image. However, I mean, imaging, it's it's always it's it's kind of in certain situations, it's difficult to apply to the patients immediately and also just gives us morphological information. And, we don't get functional information, so we we are not actually assessing cortical function. Yeah.
Speaker 2:I mean, and that's why the EEG is unique because the EEG is basically a technique which gives us in a bedside, with with the possibility to have a bedside recording, like an immediate view into the cortical function inside into the cortical function of a patient. And so, I mean and that's that's what makes EEG EEG so valuable for us in a clinical setting.
Speaker 1:Yes. I I totally agree with you. I mean, there are sometimes some die diagnostic challenges in our patients in in the ICU. Like, we have difficulties in the diagnosis of either altered status of consciousness in some types of, of patients. So from a clinical perspective, how do you use this?
Speaker 1:So which are the information which EEG can give us?
Speaker 2:I mean, so the I mean, basically, I mean, you already mentioned one very important aspect or one one very important application. I mean, if you have patient with an altered consciousness, I mean, you can have various reasons for that clinical causes coming from metabolic encephalopathies up to stroke or but also seizures. And one big challenge on nonconvulsive seizures and on nonconvulsive status epilepticus because that's something I mean, sometimes these patients with seizures or status epilepticus might show very subtle neurological deficits or abnormalities like movements and so on or myoclones. However, I mean, the importance and also the clinic clinical implication of the status epilepticus in a nonconverse status epilepticus for the outcome of the patient, it's very important to have a precise and also fast diagnosis. Yeah?
Speaker 2:Because we know I mean, first of all, we know a state a nonconversive status epilepticus is much more often than we think. I mean, if you look into the literature, you see that up to ten, depending on the study, up to twenty percent of patients on a general ICU with this disturbance of consciousness have or meet the criteria of recurrent seizures or nonconvulsive status epileptic. And so first of all, we know that a delayed treatment in these patients is worsening the prognosis. I mean in general, if you have a patient with seizures or or a status epilepticus, this worsens the prognosis massively. If you do not recognize this and if you don't start the treatment early on.
Speaker 2:It's it's it's also that the the treatment gets much more difficult the longer you you you delay the start of your under converse of treatment. I mean, that's important. Also, I mean, the other thing is we can't rely on our clinical experience and on our clinical examination because, I mean, even if you're experiencing neurologists, there are no clear signs of a nonconvulsive status epilepticus. And that's why we we completely rely on the EEG.
Speaker 1:Yes. This is, I think, a very crucial point. You mentioned the the nonconvulsive status epilepticus. Us. If a patient has tonic clonic seizures, it's easy to detect it from a clinical perspective.
Speaker 1:Non convulsive status, it's much more difficult. Sometimes we wake up patients, they are in the status of coma. We do a CT and the CT is negative. So my question is when do you do the imaging and when do you do the from a practical perspective. Let's imagine a patient who had a head trauma.
Speaker 1:We wean him from the sedation because the ICP is under control and the patient doesn't wake up.
Speaker 2:I mean, the nice point is that these are two techniques which are going in parallel. You know? One technique for the imaging where you will always do an imaging because, I mean, it's also something which is accessible for us, and, so we wanna make sure that there's no stroke, etcetera. And so, I mean, we will do imaging, but at the same time, we also have to keep in mind that the EEG is closing another very or is is answering another very important question. And once the the CT scan and also your lab results etcetera do not explain the status of the patient, you wanna proceed rapidly and do an EEG because I mean it's funny.
Speaker 2:I mean, yeah, if you go on an ICU, they are very it's very easy to get a CT scan and it's probably harder to get an EEG. Right? And also to and also our knowledge in interpreting EEGs is probably not as good as interpreting CTs. And that's why I mean, these these these are basically both methodologies you wanna combine in a patient.
Speaker 1:Yeah. I think you have touched a very important point. Sometimes it's hard to have an EEG. You are a neurologist. For instance, I am an intensivist, and not all intensivists are able to assess properly an EEG.
Speaker 1:Sometimes in some unit, it's also difficult to ask someone who comes in the ICU and does an EEG when we need it because you mentioned an important point. So the importance of the time, the more we keep the patient in non convulsive status epilepticus the worse will be the outcome. Which are in your opinion the challenge in the use of the traditional EEG in the in the ICU?
Speaker 2:Yeah. I mean, I think I mean, it's like in in with with echocardiography. Right? I mean, nowadays, more or less an experienced anesthesiologist is pretty good in echocardiography for, like, basic questions. Right?
Speaker 2:And this this something which developed in the last, I would say, ten to fifteen years probably. Right? Also, we see development in sonography in general. And I think something similar could potentially also be applied to EEG because reading EEGs and setting up the EEG is much easier than than most people assume. You know?
Speaker 2:So, I mean, the first question is always, which EEG do you wanna use? Okay? So because, I mean, we are now I mean, in general, we use this ten twenty system with 21 electrodes, surface electrodes. That's not something you you you you have to apply in in context of an ICU. You can also look a montage with with fewer electrodes.
Speaker 2:Sometimes it's it's even okay to do a mobile EEG. I mean, there are mobile EEGs on the market which are giving excellent EEGs with only six to eight electrodes. And then you you you have a very good impression of the of both hemispheres, and you probably, you see most of the relevant of of significant pathologies. You will find with these, like, let's say, more simple approaches. However, I mean, the most important thing is and that's why you need a basic understanding of EEG is the raw signal.
Speaker 2:So first of all, the signal, when you you when you apply EEG with only like, it's with a smaller montage, It's more important to have a good raw signal than to have many electrodes. Yeah? Because you you always have the difficulty in an ICU environment. You have a lot of noise, a lot of disturbances from the patient, but also from all from the bed, from the ventilator, etcetera. So that's the first thing.
Speaker 2:You so you wanna have an impression of your EEG signal. So that's the first thing because the pathologies you are interested in, they are not that difficult to detect if the EEG signal per se is okay. Right? And that's that's that's I think that's the bigger challenge. So you have to understand how does a normal EEG look like in general, what are the basic filter settings you need to apply, where can you make changes, where what are the parameters which you shouldn't change because otherwise you kind of change the EEG signal.
Speaker 2:And then it's to to detect the status, the nonconversive status epilepticus or to have like or to also today today take significant finding for the prognosis of a patients like a a burst suppression pattern. That's not so difficult. And that's actually something which we can learn easily and and which is also I mean, also we have nowadays these trend softwares. They have spike detection. They have you can use these spectrograms also to detect noncombustive seizures in an EEG.
Speaker 2:So there are a lot of tools which you can apply which you can apply which are helping to interpret EEG. But I think, again, I think the most important thing is what is your approach? First of all, which EEG do you want to apply? Do you want to apply in, like, a trend was it 21 and EEG? Do you wanna do continuous EEG, or do you wanna do, like, on the spot, like, twenty minute EEG, which is also okay?
Speaker 2:I mean, there are studies showing that you don't necessarily need continuous EEG because if you twenty minute EEG does not show any significant pathologies in terms of epileptiform transients, you know, changes, then the the risk is massively reduced as this patient develops seizures, for example, or noncombustive status.
Speaker 1:Thank you. I think that you are right. I mean, probably it's less difficult than we think. We are just a bit scared about putting our brain in learning something that was considered just for a specialist. But I also think that the technology is helping us a lot.
Speaker 1:So I was wondering, can you go a bit deeper in new user friendly equipment or, for instance, automation? Or we are in the area of artificial intelligence. How do you think that all these tools, how are evolving, and how do you think can help?
Speaker 2:I mean, first of all, I think when you look at I mean, first, you should look at hardware. Right? So what EEG montage do you wanna use? And there are excellent new products on the market, which are giving you with a with a small montage of six to eight electrodes, a very excellent mobile Bluetooth EEGs for example. So you have no you have to so you apply these, let's say, headsets on the patient and then you have a computer like remote in the in the room and Bluetooth EEG, and this gives you excellent EEG recordings with a very good quality.
Speaker 2:So first of all, yeah, you have a very good signal to noise ratio, etcetera. So that that's that's actually the first thing. The the next point is what you mentioned. When you do continuous EEG, you need tools to to monitor your EEG because in general, it's like, you know, your consultant comes, like, one or two times a day on the ward, looks at the EEG, but he can't screen through twelve hour EEG monitor. That's that's impossible.
Speaker 2:You know? I mean, that would take him, like, an hour. Yeah. He can't he will not be able to do that. So what that's why you need trend trend software because this trend this this new trend software, and there are also like a couple of options on the market now.
Speaker 2:They give you very good indications of pathologies which are evolving over time in the long term recording, and they give you clear indications when you look at spike detection or seizure detection, for example. Yeah? These tools are pretty good to to then screen certain episodes in your EEG and to see, okay, is there something interesting? Is there do do I find it like a specific pathology? Yeah.
Speaker 2:Do I find seizures, for example? And that's why these trends these trends are just very important or very, very helpful, actually, and facilitates EEG. However, I mean, you always have to keep in mind the raw signal has to be good. Okay? I mean, if you I mean, that's that's just I mean, I can just emphasize that because just I mean, just looking at these trends is dangerous.
Speaker 2:So you should always have the raw EEG signal just look because, I mean, if some electrodes are dislocating whatever, you know, are getting lost, then you can't work with that anymore. So that's why but, I mean, that's that's the same in every other technique also. You know? Then we do echocardiography. I mean, the image has to be reasonable in the first place.
Speaker 2:You know? And that's why that that's also the same is true for EEG. The the signal has to be reasonable. And if you have a good signal, you can work with all these new with all these news trend software. And, I mean, when you talk about AI and machine learning, I think we will I mean, I I think it's too early to comment on that.
Speaker 2:You know? I think right now, there are lots of people working on new algorithms to to improve the analysis of EEG, but there's nothing which we can apply nowadays and which really helps us. I think what helps us is this are these trends softwares which which on the market, which are very good, which can be combined with continuous EEG because these the software is actually there to facilitate, to make the analysis of your EEG easier. Yeah. I I think I mean, it will be interesting.
Speaker 2:I think in five years, this will be completely different.
Speaker 1:Well, five years is not a long time, so we can, we can we will see. And, again, as you were mentioning, expanding technologies is improving the the the the noise resistance. It ensures more DAP accuracy. Correct?
Speaker 2:Yeah. Yeah. I mean I mean, that's a point. You know? And that's why I mean, if you want to implement EEG, I think the most important point is you have to make a decision what what machine do you want to use on your ICU.
Speaker 2:Right? And that's something you should invest some time and because they as I said, they're excellent systems on the market, and this this makes your life much easier.
Speaker 1:Yep. True. Now as the last part of this video post custom, let's talk a bit about the prospects in the clinical application. In your ICU, in your dream ICU, how will EEG be applied both for early diagnosis, but also I was thinking about expanding application like, EGS monitoring tool, like, monitoring of long long term neurological recovery, etcetera. How do you see this?
Speaker 2:Well, I think, I mean, once you are able to apply EEG on an ICU, it will be much more I mean, it will go far beyond seizure detection or also control of status epilepticus, for example. I mean, just think about subarachnoid hemorrhage, like cortical dysfunctions in vasospasm. I mean, that's a huge clinical challenge we have today. You know? And all the techniques we have are only giving giving an incomplete picture of what's going on in these patients.
Speaker 2:And I think that's, for example, one patient subpopulation or population of patients in which we need much more precise monitoring of cortical function or, let's say, dysfunction. And I think these are typical applications in which EEG could be extremely helpful to guide therapy. You know? I mean, because, I mean, what we are doing is obviously is is obviously not always sufficient to to to guide the therapy in these patients or to to decide what's the next step or how do where do we have to escalate our therapy? You know?
Speaker 2:Who's going to the angiography for for local nemodependium application, etcetera. Right? And I think in this this is just one one example of patients in which I think that EEG with a view with a direct view on cortical function, insight into cortical function would be extremely helpful because we are not relying on flow and doppler or on morphological parameters because then then we would really have an insight into does, you know, does the basal spasm affect cortical function? And that's actually something we are we are interested at.
Speaker 1:Yes. This is this is so true. Matthias, I really would like to thank you because we have been very clear, and your message have been really straightforward. I think that we are at the end of this episode. I really would like to thank everybody for listening this episode, but mostly mostly Matthias for being with us.
Speaker 1:And I just would like to remind you that the Ezek releases monthly podcast on the Ezek website and on various streaming platforms. So we hope you will join us for the next one. Thank you very much to all of you and for listening.
Speaker 2:Thank you, Kiana. Thank you very much. This episode is sponsored by Nihon Kodan.
