Hello, my name is Javier Preventio, and today we're going to be talking about seizures and status epilepticus. My disclosures are listed on this slide, but none of them should impact this conversation. Seizures have been recognized for very long times, and in olden days, they were thought to be either due to visitations from demons or spirits, the loss of phlegm from the brain, the lack of humidity in the nerves, excess blood, rapid changes in the body, and changes in season. Unsuccessful treatment strategies that were tried included rubbing your head with oil, ingesting silver nitrate, wearing an amulet or ring made of silver, wearing a bracelet made from the nail of a wrecked ship, or exorcism. As you can see, we've come a long way since that understanding. Having said that, there is still a poor understanding of what seizures are in the general medical community. So to talk about this, let's talk some definitions. First, what is a seizure? The definition of a seizure is an abnormal, hyper-synchronous, paroxysmal activity of the cerebral cortex. And the best way I think to think about this is that seizures are reentrant arrhythmias of the brain. They're basically a loop of neurons that fire over and over again and stimulate each other to fire. By contrast, epilepsy is a tendency to have repeated, unprovoked seizures because of a brain disorder. And sometimes a brain disorder is just repeated seizures. You can see in this picture that the synchronous activity that occurs in the brain, where the yellow arrow is, is a sign of an epileptiform discharge, which then in the middle leads to rhythmic activity, which is the seizure. The easiest way to understand an individual seizure is that they're like stories. They have beginnings, they have middles, and they have ends. If you look at the arrow that goes to the beginnings part, you'll see that a few of the leads in the middle of the recording begin to have hyper-synchronous activity. And then later they generalize to other parts of the brain. In the middle, the frequency increases, kind of like what you see in ventricular tachycardia. And then eventually at the end, the dotted arrow, you notice that the synchronous activity subsides, and there's a period of decreased activity, which we call the postictal phase. A little bit more vocabulary about characterizing seizures. Because most seizures in adulthood start as partial seizures, that is, the focus is in one area of the brain, which then can either stay in that area of the brain, which is called a simple seizure, or can go to other parts of the brain, enough of the brain to cause impairment of consciousness, and those are called secondary generalized seizures. A generalized seizure is actually a seizure that either initiates from the thalamus, which is a pattern generator inside the brain, and then goes to both sides of the brain simultaneously. These are typically seen in children, not in adults. And then secondary generalized seizures we've dealt with already, which is a focal seizure, which then goes on to generalize to other areas of the brain. One of the most important concepts in critical illness when it comes to seizures is the question of why it is that most seizures stop, but a few of them don't. And the reason is the brain has two built-in systems that actually decrease the risk of seizures lasting for a very long time. First is the myelin coverings. The myelin coverings of the nerves actually prevent electrical activities from going from one nerve bundle to another. The second one is refractory periods, and that's a period of silence that occurs after a neuron fires that doesn't allow it to refire quickly. Seizures are typically failures of these systems in a limited way. That is to say, the electrical signals go from one place in the brain to another in a route that's not along an axon. And more importantly, that as seizures continue, and particularly if there's inflammation involved, the refractory period can become shorter and shorter. Tens epilepticus is a complete failure of these refractory periods and myelin, which allows seizures to continue indefinitely. Finally, I would like to introduce a little bit of information about modern EEG systems. If any of you have had EEGs in your ICU, you'll notice that the squiggly lines are part of the screen, but there are also a bunch of processed images that are actually on a different part of the screen. Sometimes on the left, sometimes on the bottom, sometimes on the top. And those processed images are basically algorithms within the computer program that give you information about certain aspects of the EEG. It is not a definitive way of testing if a patient has seizures or not, but it can tell you the hotspots where looking closely at the squiggly lines will tell you if it's a seizure or not. At the bedside, it tells you a lot about whether to be worried. And typically, seizure detection is good but not great. And so if you see seizures listed in there, take them with a grain of salt. Rhythmicity is an interesting concept of how rhythmic the brain is because seizures are typically hyperrhythmic. There's also asymmetry. If something's happening on one side of the brain and not the other, you'll see differences in the blue and red pattern. And then finally, the last two things at the bottom, alpha-delta ratio and burst suppression, are typically used for signs of consciousness. But looking at these processed image parts are very critical for you to say, well, there's something going on and maybe having an epileptologist look at this more closely is worthwhile. We're going to go to an audio response question. And I realize that in this format, you can't answer the question, but we're going to ask it anyway. So the question is, how frequent are seizures or paroxysmal discharges in the medical intensive care unit patient with altered mental status? And the answer is four, 10% to 25%. There are different studies that show different rates depending on how they measure seizures or periodic discharges, but it's a relatively common phenomenon for patients who have altered mental status in the medical intensive care unit to have epileptic discharges that are either frank seizures or very dangerous for seizures. So everything else that we've talked about up until now has largely had to do with seizures in general. And now we're going to talk specifically about ICU seizures. Because it turns out that in the ICU, because patients have encephalopathy for other reasons, it can be sometimes very difficult to tell when somebody has seizures. So there's no clear consensus on what constitutes an ICU seizure, although there's a recent attempt to name the types of EEG signals that we see and name them as high risk or low risk for having seizures. But there are some patients who clearly have frank convulsions. And those patients are probably the easiest to detect as seizures, but those are relatively uncommon. There are patients who have non-convulsive electrographic seizures, which have a beginning and an end and are very rhythmic and are not difficult to see because they're similar to EEG signals that occur in other settings. In addition, in the ICU, there are times when people have rhythmic or lateralized and generalized epileptiform discharges, as well as morphologically different epileptiform discharges that people have taken to calling angry looking PLEDs. Having said that, there is now a consensus that we should have a standardized nomenclature for ICU seizures, and that was published in the last three years in one of the epilepsy journals. It is complicated and technical, and we're not going to go over it here because it's outside the scope, but just understand that if your epileptologists don't currently use the ICU standard, it may be worthwhile doing that. The foundation for that seizure nomenclature was based on some studies that show that different types of discharges appear to have different risk for frank seizures. In addition to that, the idea that generalized periodic discharges in critically ill patients occur relatively frequently, so it makes it more difficult to say just because someone's having spiking that they're at risk for having seizures. This is the data that was a rationale for the question that we asked at the beginning of this section. So in a study done by Mauro Odo out of Columbia at the time, they looked at epileptiform discharges and electrographic seizures in medical ICU patients who had an altered level of consciousness, who did not have another known cause, and who had continuous EEG monitoring. And what they found was that somewhere between 10% and 25% of people had either frank seizures or epileptiform discharges. This was corroborated by a retrospective analysis that I took part in at the Cleveland Clinic in the intensive care unit that found that across all intensive care units and patients who had EEG monitoring, about 19% of them had some sort of epileptiform discharges, and the majority of those people had non-convulsive status epilepticus. Interestingly, the mean age and the age range and the sex was not particularly different between the two populations, suggesting that it's a relatively general phenomenon. Interesting in our sample, the more depressed the level of consciousness, the more likely the patient was to have seizures. You can see from awake, lethargic, stuporous, and comatose patients that the yellow line, which is the percentage of patients having EEG seizures, was increased substantially across those groups. Of note, the risk of having seizures was in large part associated with neurologic disturbances. And again, part of this was a bias due to the fact that this was done in all the intensive care units, including the neurological intensive care unit, where strokes, hemorrhages, and CNS infections were more common. Having said that, tumors, anoxia, as well as metabolic disturbances still had significant risks of seizures. One of the most challenging aspects of doing this type of research is trying to figure out who, in fact, we should be monitoring. In patients in the neurological intensive care unit, we know that there's an association between the amount of brain injury as well as the level of consciousness. And we know that patients who are more severely affected should probably be monitored. In other intensive care units, this becomes less clear, although my feeling about this is that if a patient has altered level of consciousness and you don't have another reason and you want to rule out seizures as a possibility, the only way to do that is by putting on a continuous EEG monitor. So let's talk specifically about status epilepticus. And status epilepticus is simply seizures that don't stop. The definition of status epilepticus is evolving. But ultimately, the take-home points are that status epilepticus is a medical emergency. The prognosis of status epilepticus has a lot to do with the underlying etiology. And the example that I typically give people is that if you develop status epilepticus because you have a glioblastoma in your brain, your prognosis is quite poor. If you develop status epilepticus because you stopped taking your epilepsy medicines, your prognosis can be quite good. Now having said that, in an anoxic brain injury, we've done some studies to look at how that one individual group does. And we know that time spent in status seems to be important, and we'll go over that data a little bit later. So making the diagnosis is critical, and it's not always very straightforward, especially in the ICU setting. But I do think that more continuous EEG monitoring would help. And the treatment should be aggressive and appropriate to the type of status encountered and the comorbid conditions present. So now let's talk about the definition of status epilepticus. The old definition of status, which many of us grew up with in medical school, was that you had to have 30 minutes of continuous seizure activity or multiple seizures without return to neurological baseline for greater than 30 minutes. This was based on animal studies that showed irreversible neuronal damage after 30 minutes of continuous seizure activity, but not clinical. And this definition was not particularly helpful. What was more helpful was a working definition that was developed in Richmond in the 1990s, which looked at the probability that a seizure would stop on its own once it was seen in the emergency room. And in this study, what they found was that very early on in the first minute or so, 99% of seizures were likely to stop on their own without medications. And that probability remained relatively high until about five minutes. And at five minutes, it began to fall pretty precipitously. By seven minutes, it was unlikely that a seizure would stop on its own without any kind of medical intervention. So typically what I say is that seven minutes is the time when status is really there. But if you start diagnosing it at five minutes, it takes a few minutes to get the medications. And by the time you start giving the medications, you're probably at that probability where the seizure is unlikely to stop on its own if you don't treat it. So having a good definition of status epilepticus is important because for the majority of patients who have seizures, like I said before, they will stop on their own. And the difference between whether the patient goes home or goes to an ICU in a typical seizure is whether somebody begins treating it or not. So typically what I tell people is to look on your watch for the five minutes that it takes to call something status epilepticus. The seizure stops before then. Then you can make appropriate decisions. If it doesn't, then you begin treatment. And when you begin treatment, there are some secondary definitions we have to take into account. One is status epilepticus, which is just basically a category of seizures that go on too long. The second one is called refractory status. And we're going to talk about treatments in a second. But when patients fail their original treatment therapy, they're thought to be refractory. And then finally, there's a term called super refractory status, which is patients who fail both primary and secondary treatments and usually mean some sort of anesthetic agent to treat. So enough with definitions. So the question of why does a seizure become refractory? What is the pathophysiology? There are a couple of interesting phenomenon or interesting findings that we've had that have suggested why it is that seizures, when they are prolonged, could last for a really long time. First, there is the issue of GABA receptor desensitization. We know from studies done at the University of Virginia, where I work, that when animals have seizures, the GABA receptors, which are the receptors for the benzodiazepine medicines that we typically give in status epilepticus or in seizures in any event, downregulate very quickly, where half of the GABA receptors decrease every minute of seizure activity. The second thing is that seizures engender a great deal of inflammation, and inflammation has a paradoxical effect of decreasing the refractory period, which is a period when neurons can't fire on their own after sending off a stimulus. And that actually then leads to more rapid firing of neurons, allowing these loops to be made like reentrant arrhythmias. The other pathological effect that we have to understand is that seizures do actually have consequences and we can see those. So there is brain damage that occurs in patients who have status epilepticus, which is typically seen in the parts of the brain that are more vulnerable like the hippocampus or the part of the brain that's actually having the seizures. This oftentimes leads to chronic atrophy in that area of the brain and some associated neurologic deficits that are typically associated with those brain parts, although it's hard to make actual causal inferences. So status epilepticus has a mortality. It's somewhere between 7 and 39 percent in the United States. The 39 percent was likely a study that included, it was a study that included patients who had post-anaptic brain injury and had continuous EEG monitoring where they found a lot more patients who had EEG seizures and that may be why there's an increased percentage of mortality in that study. The etiology is most important factor in determining outcome and we talked about that already. And multiple studies have suggested a synergistic effect of seizures or status and the underlying cause of death. That is to say, what causes the status and how long it occurs are both important factors. So the general approach to treatment is first and foremost to stop the status epilepticus. If we believe that the time and status is an important predictor of outcome, then stopping the seizures is critical. And interestingly, the most time-consuming part of stopping the seizures is oftentimes a diagnosis of the status epilepticus. In patients who have generalized seizures, this can be relatively easy in some cases. In patients with previous epilepsy, it can be on the front of mind. But it's not uncommon for patients to become in encephalopathic or just moving one part of their body that's not really noticed by the by the physician team or the nursing team and that and that ultimately leads to delays in diagnosis. The protocols for initial management are relatively well described and we'll talk about them later. The second thing is to manage the underlying etiology. With a rapidly expanding GBM is different than non-compliance obviously, but there are certain things that can cause seizures where the underlying etiology treatment becomes critical. Most important for these is meningitis. In patients who have status with meningitis, the treatment of the meningitis tends to be critical and needs to be initiated early. There are also complications to seizures. When patients have prolonged periods of muscle contraction, they can have rhabdomyolysis. Patients who have prolonged hospital stays can have ICU neuropathies. Patients with seizures oftentimes have postictal phases where they don't use their body parts very much and they're at higher risk for DVTs than most patients. And then finally, preventing seizure recurrence. How many medications are needed? What duration of therapy? And we're going to talk a little bit about this as we go forward. Again, this doesn't really allow the format to answer the questions yourselves, but we're gonna read this question anyway. What is the appropriate first intravenous dose of lorazepam for a patient with convulsive status epilepticus? The correct answer is 0.1 milligrams per kilo, which at face value seems pretty reasonable, but when you think about it more closely is actually sometimes very difficult. So the average American is now no longer 70 kilograms but 80 kilograms, which means the typical American who has comes in with a status epilepticus or I'm sorry yes status epilepticus episode should get 8 milligrams of lorazepam, that's 0.1 milligrams per kilogram, all in one dose. And typically this is not what's done in the emergency room. There was a new study that we're going to talk about later that actually did some retrospective analysis of their data that showed that on average about 75% of patients do not get an adequate first dose of lorazepam when they're first diagnosed with seizures. Study was done largely in the emergency department, but the same thing holds in the intensive care unit. So again, with most of you who have a little bit more gray hair, you've probably seen some form of this algorithm where a patient's supposed to get lorazepam and then there's and then if that doesn't work then they get phenytoin or phosphenytoin, if that doesn't work they might get a second phenytoin dose, if that doesn't work then they go on to phenobarbital, and if that doesn't work then they go to some sort of anesthetic agent. The problem with that algorithm is that for the patients who actually seize, it's a very long time to control their seizures. And I'm going to show you a little bit of data about why I don't think that this approach is actually ideal. In the best comprehensive study of first-line therapies for seizures, it was a VA cooperative trial that was done by David Tryman and looked at patients who had status epilepticus. They were given either lorazepam, phenobarbital, diazepam, and phenytoin or phenytoin alone. And what they found was that with appropriate dosing, the lorazepam patients alone had about 65% of their seizures stop within 20 minutes and not recur for 24 hours. So that turned out to be the best initial drug and that's why the recommendation of 0.1 milligrams per kilogram of lorazepam was the most effective therapy. Interestingly and far less publicized was that in this study in patients who got these very high doses of lorazepam, that hypoventilation occurred in about 10% of patients but few of them had to be had to be intubated. Hypotension occurred in about a quarter of patients which would be expected and cardiac rhythm disturbances occurred in about 7% of patients. Some of those things were subtle, some of them were more overt, but ultimately the dose, even though it seemed like a relatively large dose, the majority of patients did not need mechanical ventilation and could be supported otherwise. One of the interesting parts of this study that was not originally included in the in the paper but was subsequently published is that the way that the study was devised was that there were boxes that were blinded so the practitioners didn't know what drug the patient was getting. Unfortunately for patients who had lack of treatment success, they had to give a second drug. So they provided boxes for second and third line drugs that the practitioners were still blinded to and what they found was that in the lorazepam group where the second drug was diazepam, I'm sorry, phenytoin and the third drug was phenobarbital, that you didn't get very much of a bang for your buck when you when you added the other medicines. In fact, if you look at the numbers of lorazepam, it was 65% stopped with lorazepam alone, another 7.2% with phenytoin, and then another 2.1% with phenobarbital for a 74% total decrease. Now it turns out that in a recent study looking at lorazepam, it was closer to 71%. So basically what it says it is at 7 minutes, 35% of patients were still seizing and at 3 hours, 26% of patients were still seizing. So for those three so those three hours of treatment, there was very little bang for your buck. And this is a study that I alluded to earlier because it shows that if you look at benzodiazepine dosing in the emergency room and patients in this study, that 81% of them had inadequate dosing of lorazepam and 76% had inadequate dosing of midazolam if that was a drug that they chose to use. But that's a significant problem because it turns out that in the study by Tryman, the dosing was actually blinded so they got the whole dose. So we think that maybe we could get a lot more patients stopped in their seizures if they were able to give larger doses of the benzodiazepines early on. So as I said, there's the first-line agent which seems to be pretty effective if given in the right dose, but the second and third line agents don't seem to work very well. What should we be using a second and third line dosing? This is where a recent study became very helpful. So the study done by Jadiv Kapur, also here at the University of Virginia, showed that in patients who had, they randomized patients who had convulsive seizures for more than five minutes and had already gotten a benzodiazepine dose without it working, were then randomized to get phosphonatone, sodium valproate, or Depakote, or levotiracetam, or Keppra. And the outcomes showed that all three drugs were equally effective in stopping seizures, and in fact were actually quite effective in stopping seizures. So that when you look at the 70% that they got initial seizure stopping with the benzodiazepine alone, they were able to get more than 50% of the patients under control with one more medicine given relatively quickly. Interestingly, valproate was most likely to be a bit superior, not significantly, and had fewer side effects. So I don't think that I would choose one of these drugs over another generally based on this data, but if I had to choose based on all the experience we have, valproate might be just a little bit more effective and a little bit safer than the other two. So now we'll go to our third audience response question, which in the light of the paper I just showed you has to be changed a little bit. So a patient with status epilepticus, refractory to lorazepam and a second-line medication is intubated and midazolam high-dose IV is chosen as a sedative agent. What statement about EEG monitoring best helps you manage the patient? Is it one, EEG burst suppression has been demonstrated to be necessary to control refractory status epilepticus? Two, achieving burst suppression means the patient will not have seizures. Three, a burst suppression pattern of five bursts per minute is optimal for treatment. And four, cessation of seizures is the goal of therapy. And the answer here is four, that the idea of burst suppression, which is largely a phenomenon seen with pentobarbital and not seen with other sedative agents, is just in sheathed in a number of myths. One myth is that you have to have burst suppression to control things, that you have to achieve burst suppression, that burst suppression ensures that patients won't have seizures. I've seen patients in burst suppression where the bursts are seizures, and that burst suppression pattern is really up for debate. But really the goal of all these therapies is just to stop the seizures. And this quickly basically goes to the same thing, is that burst suppression has not been has been demonstrated, not been demonstrated to be necessary to control seizures. Their burst suppression pattern is not easily recognized or easily taught to the practitioner. So it's very, very dubious whether it's a worthwhile endeavor to try to get burst suppression. So if you look at a standard algorithm that was published recently, pre-emergency room there are some evidence that IV and rectal or buccal benzodiazepines are helpful, particularly in the pediatric situation, but can be used. In the emergency room, I favored lorazepam, 0.1 milligrams per kilogram. Midazolam has been tested and seems to be worthwhile. IV phosphenetone is probably not a good first-line choice when benzodiazepines are available. And then for refractory seizures or seizures that are a little bit further along, then IV valproate, levotiracetam, or phosphenetone are good choices. Although like I said before, I favor the IV valproate. And then finally in patients who continue to have seizures after that, then intensive care management with a potent sedative, which is either propofol, midazolam, pentobarbital, and the new kid on the block, ketamine, have been shown to be effective against status epilepticus, although the data for this is quite new. And also in keeping with what we talked about before, the valproate should be listed in the emergency room, not in the intensive care unit. I mean typically IV lorazepam takes about four to eight minutes to work, so within the first 30 minutes you can actually give the IV lorazepam, as well as the intravenous valproate, levotiracetam, or phosphenetone, and then know relatively quickly within half an hour whether a patient needs to be intubated and put on a deeper sedation. Recently status epilepticus has become an interesting area of research, particularly for drug treatment trials. There was a recent trial called for SAGE 547, which was a drug that was a GABAergic, or synaptic alloselic modulator of GABA. Unfortunately this trial failed to meet their primary endpoint and the drug was never FDA approved. There are other attempts at new drugs, but none of them are ready for primetime and I'm not going to discuss them here. One of the things that I think is actually important is that what we have found is the need for continuous EEG monitoring is becoming more and more important. And in fact, remote monitoring done by a central location is now available throughout most of the United States. But it is very clear that having some sort of telemetry or having some sort of a continuous EEG monitor with a an experienced and trained epileptologist to read through the EEG is becoming a critical part of the management of these patients. And just a quick plug for what's happening to patients who have unresponsiveness in intensive care unit settings. I've already told you that somewhere between 10 and 25 percent of those patients are having frequent seizures. But of the patients who don't have seizures, who are in a coma, there's this recent work that was done by Jan Clausen's group in New York that have shown that in patients who have coma and not from seizures, that there's a proportion of them, around 15 percent of them, who if you give them a very complicated algorithm, it's not really ready for bedside, are actually able to follow commands or able to show consciousness. And that is to say that their responses by EEG, that the changes in their EEG, are most coordinated with sentences asking them to do certain commands and that are not replicated in patients who are getting sort of sham sentences. This suggests that there's a certain amount of consciousness in a large proportion of patients in whom status epilepticus is not suspected but who are in a coma. And again, EEG is a critical monitoring device for these people and should be considered. This went a little bit longer than what it would have gone had we been in person, but I think that explaining a little more slowly since you can't see my face was important. If you have any questions, please email me.

Seizures

Status epilepticus

Cerebral cortex

Epilepsy

Partial seizures

Generalized seizures

Continuous EEG monitoring