So welcome to San Francisco, everybody. It has been like drinking out of a firehose, a lot of information in the previous sessions. My talk is going to focus on how to approach status seizures and patients that might be worried about seizures in the ICU. I'll leave the fun treatment stuff to Nicole Davis. They'll come after me when it comes to picking the drugs. So most of us deal with codes. And when we walk into the room, we know what to do. And we really are guided by the monitors. So if you see someone on VTAC, you might be grabbing your pads. If someone is on the PEA, you might focus on good quality CPR. But when it comes to seizures and status, it's kind of similar. So if someone is having a lot of spikes and that's really highly metabolic in the brain, you might be thinking about treating that right away. But if you have some slow periodic spikes, maybe you should not give that patient drugs and make things more complicated. You might also use EEG to guide how much anesthesia you're going to be giving to patients in the status epilepticus. And patients with, for example, cardiac arrest who might be having myoclonus, having EEG data might really help you understand how you're going to approach that patient. And why that's important is actually pretty common to have seizures if you are showing up in the emergency room. Even if you're in the medical ICU and your patient is comatose and not doing much, you should be worried about non-convulsive seizures. Particularly in neuro patients, up to a third of patients that have seizures actually having non-convulsive seizures, which remind us that EEG is really important. So just like you walk into the room and there is a code, there's no pulse, you know to start BLS, chest compressions. And I think we need to think the same way when it comes to seizures. So someone is having a convulsion for over five minutes, you need to be thinking about getting your benzos and starting that right away. So this picture is from the 90s, and it actually happened to be in my hospital here down the street in San Francisco General. The x-axis shows time. So you see starting at 30 minutes after the seizure started. And the y-axis is the response to benzos. So the longer you wait, the worse response you get. So time is brain, that's the topic of the talk today. And that's also really important when it comes to seizure and treatment. But remember, the first bar was 30 minutes, right? So that's kind of the old way to think about status. So let's wait half hour to really treat aggressively. And that's not a good idea. So the operational definition of status, now we cut that number down to five. So if someone is seizing for over five minutes, you probably got to get started and not wait for status to really get established to start thinking about treatment. So that's something that's very important to remember. And this is more for convulsive status. And non-convulsive status might be a little different in how we approach it. So just as we think about ACLS and think about algorithms, we really need to think about this stepwise approach and think quickly. And it's really time dependent when it comes to seizures. So you might wait a few minutes to see if the seizure is gonna stop when it's on, and you've got to make your ABCs, get things ready. But you really need to be thinking about, are people ready for benzos? If the benzos don't work, what's next? And as time goes by, how we call it, it changes. So established status epilepticus, refractory status epilepticus, it's all about how you're responding to the medications. In this flowchart, we have anesthesia at six minutes, but in practice, we are often pulling propofol and other drugs much earlier, because if we give enough benzos, patients might not be breathing that well. So I think that's just to give you a general idea. And if you're really getting stuck treating seizures, I think expert guidance is really helpful because you start to stack three, four, five drugs, and you're not making much headway. So it's good to give a ring to the epileptologist. And one thing to remember is that status is bad. So a third of people will die, a third of people will have severe disability, and a third of them might do well. And I think we really need to find better strategies to treat patients. So in the code, if you know the rhythm, you know what to do. There is a very clear pathway, and you know you gotta grab your AED to shock that patient and have that handy. So when it comes to seizures, I try to think about that same way with EEG. So I'm an ICU physician, but I'm also an epileptologist, so my free time, and I'm a little biased about having EEG around. But it can really guide how you're gonna titrate patients, not to overdo it or underdo it. And again, a reminder, patients have non-convulsive seizures. And we often think about the patient that's comatose, not doing anything, as the patient that might be on non-convulsive status. But actually, when you look at the symptoms on this patient, some patients would be comatose, but someone who just have altered level of consciousness, they'll look a little confused. So having a low threshold to start EEG monitoring, I think it's really, really key. And the rule of thumb about EEG interpretation, and we're not gonna get in too much interpretation here, because we only have 15 minutes, is that if your EEG is starting to look like your EKG, that's a problem, especially if it's sinus tachycardia. So if your brain is in sinus tachy, that's bad. And I invite everybody to read or watch Tom Black's talk on this, which is really a great talk. And the Black's rules still apply. And when it comes to status, you might have the generalized type of status. That's usually what we think about when we think about status. The grand mal, the big convulsions, and your whole brain is getting activated in hypermetabolic, and we should act on that quickly. But you might also have focal status epilepticus. So someone has a stroke, might have herpes encephalitis. We are not usually as aggressive with focal status treatment, but we gotta get that under control. So it's something that might be a little different in how you approach it. It is also a reminder that you might have a little twitch. It's a lip twitch, it's a hand movement that you might miss if you're thinking about babies. The focus of the talk's not necessarily with pediatric population, but that's some very subtle things that sometimes the family will say, hey, she's doing something that's a little weird. So that's why having low threshold for EEG, I think it's very important. And this is a patient with cardiac arrest, has a very flat EEG, and those spikes, they're periodic, they come and go. And you might have a very different approach to treat that. And you might actually not treat that at all, depending on the case. So I think it is really something that EEG can really help you understand how to titrate therapy. But even though I'm saying that EEG is important, scalp EEG will miss still a bunch of seizures. And so this is a patient with a traumatic brain injury. We are doing advanced monitoring, and now our patient's here, so you can see a bolt to that pointy thing on the front of the head. And there's a depth electrode there, so just think a little tube with a bunch of electrodes on the tip, and we're putting that together with the ICP monitor, the oxygen monitor, the blood flow monitor. And patients who get the skull taken out, a hemicraniectomy, we will also put EEG electrodes on top of the brain to monitor for seizures. And actually in our shop, one out of four moderate to severe patients is having seizures, and about half or so, a third to a half are missed with only surface EEG. So on this picture here, just to kind of orient you, there's no bone on this side, and there's this bright spot here, which is one of the electrodes that's on the surface of the brain. And this other pointy thing here, bright, on the other side is where the bolt is going. If you look at the coronal view, you see here the electrode is on this side, and here is where the bolt is, where the ICP monitor. This is the tip of the EVD that's showing up there, just to kind of orient you. And this patient has scalp EEG, and that's why we are looking up in the top here, so the different altitudes. And this looks just slow, attenuated, but in the depth, you know, sinus tachycardia in the brain, so that's what was going on here. And this patient was having completely non-convulsive seizures. And the depth is on the good side of the brain, if you think there's a good side of the brain when you have moderate to severe TBI. So it's not even the side of the craniotomy. So that's just a reminder that someone who's comatose not doing anything, the advanced monitoring might be something to think about. So those red and blue bars, so the red bar is mortality, the blue bar is actually incidence. So if you look all the way on the bottom, on hypoxia and noxia, the red bar is really big. So these patients do very poorly, even though very few will have status epilepticus. Because I told you, you know, about a third of patients will die, and another third will have poor outcomes. But it really depends on why you're having status. So if you look here, the cases that have a bigger blue bar is really people that are not taking their medications, the antiepileptic drugs. So those patients usually tend to do pretty well. And by the way, I have QR codes on most of this paper thing, so I don't know if you can catch that from your phone, but that was the idea. Hopefully it will be available online. And then we're gonna start going to the zebra things. You know, many times we know why people are having seizures, but sometimes we don't. And the new onset refractory status epilepticus or NORS are some of the cases we need to do that are very complicated to work up. The other considerations, if you have myoclonus, myoclonic status epilepticus after cardiac arrest, most patients tend to do very poorly. Two months ago I had someone who woke up after we did the treatment, so some people might still survive that, so it's just something to keep in mind. So liver injury, renal injury, that might change how you do your drugs. Super obese patients, how you calculate your dosing, you gotta be careful with that. If a patient's pregnant, palliative care patients, you need to really individualize how you do it. So palliative care is always very tricky because you're not gonna intubate the patient to treat the seizures. You need to make them comfortable because seizing is not a comfortable thing. So it's always kind of a challenge. CSF, brain imaging, sometimes advanced imaging like PAT might be helpful, particularly on those cases that the EEG is not clear cut. This is a seizure or just something spiky, what should I do? Pen CT, ultrasound for a patient that has small tumors and you need to really look for the tumor sometimes. It can take a while to find it. And there's some biomarker panels that we do for blood CSF for antibodies and a lot of this more newly diagnosed these days type of encephalitis. And EEG is not just a diagnostic tool. It might also be how you titrate therapy. So we talk about birth suppression and making the brain flat as a way to stop seizures, but maybe seizure suppression is good enough. And we really don't know yet how deep should we go with anesthesia. But I think it is something that can be very helpful, particularly because we are very bad at it. So if we say, yes, let's make it birth suppression, most of the time we're actually not birth suppressed. So if you don't have someone looking at the EEG to really try to titrate, that can be a challenge. And sometimes the epileptologist is not gonna be sitting there and doing, it really comes back to the neurointensivist or intensivist with this type of basic training to at least see that, oh, you know what? Things are not as we planned. And this is a quantitative panel. I'm not gonna explain all the details, but the red part there is, so the red is bad usually, right? So the beginning there is when you're having a bad seizure. So there's very tall red part. And on the bottom, we have actually how flat the EEG is. So you cannot see waves in the beginning because this patient is in bad status. So we gave some propofol, the red came down a little bit, but it didn't really go away. And only after we gave ketamine, we were able to really get the brain to stop seizing. And you can see the red and blue lines going up. So this is how suppressed the EEG was. We probably overdid it a little bit after towards the end. You can see it getting closer to 100%, but sometimes it doesn't go as planned when it comes to titration with ketamine and other things. But why did I put this quantitative EEG is just to say you can see an entire picture of what happened to that patient in those 12 hours very quickly and understand the trajectory and how your treatment's going. And we actually can not only rely on the epileptologist. We have the raw EEG, those waves on the left, and on the right side, we have the quantitative panels. So in one snapshot, I can look at 24 hours of data, and I just told you the red's bad. So you start to see this flame-looking thing. So that's when the patient had a seizure. And it's not that complicated, right? I just showed this to you now in 30 seconds, 60 seconds. We actually did a workshop with our nurses in less than an hour workshop, and they were screening seizures just as well as the epileptologist did. So this is just to say that this is not like some super complicated that only someone with special powers can look at. So I think we need to be thinking about this is a bad side too, just like EKG was back in the 50s, 60s, a cardiologist. So the goal is if you walk into a room, we are in the ICU, we know this is bad, and we know what to do. And what we really want is to move towards can we have the EEG just as a bad side too, just like EKG was back in the day. Thank you, everyone. Thank you.

status seizures

seizures in ICU

EEG monitoring

status epilepticus

individualized treatment