So, all right, these are my disclosures. I do have some funding from NIH and will be presenting some of the work, but the others are not related to this talk. So, I'm going to start by continuing the case that Neha initially presented. Remember, this was a gentleman who presented with traumatic brain injury, had bilateral contusions, and post-admission they had deterioration of their clinical exam. And so, we were performing optic nerve sheet diameter to assess whether they have raised intracranial pressure. And while performing that, he was noted to have forced gaze deviation as well as facial and eyelid twitching, and then sort of subsequently continued to have recurrent episodes of that. So, who here would consider continuous or EEG monitoring for this patient? Raise of hands. Okay, so majority. And that's really what we are seeing nationally as well. There has been an increasing trend in the use of continuous EEG monitoring, not just for neurocritical care in the neuro-ICU population, but also in general critical care with an increasing recognition of non-brain injured patients, particularly sepsis, acute liver failure patients also being at risk for seizures and other seizure-like patterns. And with, of course, the increase in use of EEG monitoring, we're recognizing not just seizures but other similar patterns, and there's a lot of work going on and research around the diagnostic and prognostic significance of these patterns and how to approach treatment. So, to start off, just I wanted to provide a quick review of when we should consider continuous EEG monitoring. The American Clinical Neurophysiology Society has put out consensus recommendations for, to provide guidance around when to consider monitoring. And, of course, the first most common indication is persistently altered mental status after generalized or convulsive status epilepticus or recurrent seizures. So, most patients, after having a generalized seizure, should regain, you know, should have improvement in their level of consciousness and alertness within 30 minutes to 60 minutes after a seizure. But if someone comes in with recurrent seizures or has recurrent seizures and they do not have improvement in their mental status beyond that sort of one-hour time frame, I would consider EEG monitoring in them because they're at a higher risk for nonconvulsive seizures. Second, most common indication is an acute brain injury with altered mental status. And this is really altered mental status that is out of proportion to what the imaging would explain. So, if you have a small contusion with matching focal neurologic deficits, of course, I would not consider EEG monitoring. But if there is a small contusion and the exam is out of proportion for what I would expect, then in those situations, I would consider EEG monitoring to rule out seizures. And then, of course, unexplained altered mental status or fluctuating altered mental status in the absence of any brain injury. Other indications, if there is need for paralysis and the patient is at higher risk for seizures, so think severe brain injury and ARDS and you're trying to control their respiratory status through sedation and paralysis, consider monitoring those patients. Patients that require high-dose sedation to monitor depth of sedation, so our status epilepticus patients or refractory intracranial pressure patients. And then, of course, if there's paroxysmal events similar to our patient that are concerning for seizures and we want to better characterize them. And then, finally, some of the non-seizure indications include prognostication. Most commonly, EEG is used in prognosticating after cardiac arrest in patients with anoxic brain injury where absence of reactivity, suppressed background, burst suppression are all known to be poor prognostic indicators. A less common indication not used across all intensive care units is ischemia detection. There is growing recognition of the use of EEG monitoring, particularly in subarachnoid hemorrhage patients that are at risk for delayed cerebral ischemia. And some centers do use a continuous EEG for earlier prediction of delayed cerebral ischemia. A little bit about sort of the logistics, how long to monitor when we do start monitoring, so, you know, we know that for patients that have the most common indication which is post-generalized convulsive status epilepticus, about half of patients with non-convulsive seizures will be picked up in that first 30 to 60 minutes of monitoring. And if you monitor these patients for at least 24 hours, almost all 95, up to 95% of patients who have non-convulsive seizures are identified. EEG, of course, is resource intensive, and it is not, you know, depending on which center you are, what facility you're at, whether you have continuous EEG or not, this may not be a possibility in every scenario. And so there is some data coming out now, particularly this is a newer score that has been available for a few years where we can really use the initial EEG recording. So even the first half hour to an hour, if there is a high burden of epileptiform discharges, if there are high-frequency discharges, or if they've had recent seizures, this score called the two helps to be scored and help predict what the seizure risk is over the next 72 hours. And the score sort of, it's available now on Medscape and freely available online. If you plug in the values, it will give you a predicted seizure risk, and also recommendation on how long you should consider monitoring this patient. So we are starting to incorporate this into our clinical practice as well. Now when we do EEG monitoring, what do we find? And you know, of course, we're looking for electrographic seizures or status epilepticus. And we see this in reports range from 3% to upwards of 30%, depending on the underlying etiology. I'm not going to go for both the sake of time and because we're focusing on non-seizure patterns, just very quickly our approach to if we do see seizures or electrographic status, the treatment approach tends to be very similar for what we have for convulsive status. So starting by using benzodiazepines if they're already intubated and on sedation, going directly to the second line agents, which is phosphenytoin, valproic acid, or levotiracetam. And if seizures are refractory, additional anti-seizure medications are considering anesthetics for seizure or birth suppression. But what we're going to focus more on is the non-seizure findings, which is periodic and rhythmic patterns. So if you have EEG in your ICUs or get EEG, you will often see reports of periodic discharges or lateralized rhythmic delta activity or generalized rhythmic delta activity. And with increasing use of EEG, what we're finding is that up to half the patients that undergo monitoring don't really have seizures but have what we call these periodic and rhythmic patterns. They don't quite look like seizures, but they also don't really meet that benign interictal activity. And so often we call these ictal-interictal continuum patterns where they look seizure-like but don't quite meet the defining criteria for seizures. And so, you know, these are a few examples where they can be lateralized over one hemisphere. We're seeing these sharp, periodic, lateralized periodic discharges. Or this is, or they can be generalized. So we see these sharp waves that you're seeing are what we call generalized periodic discharges. And so the question now that we're seeing these so often in our population is the question is, what is their significance? Are they ictal? Are they interictal? What is their relationship to seizures? Do they result in secondary brain injury? Do they impact outcomes? And of course, should we be treating them? And if so, how aggressively should we be treating them? And so what I'll review is just a little bit of, you know, the growing evidence or body of evidence related to these patterns. So this was one of the largest studies done by the Critical Care EEG Monitoring Consortium where they found that some of these patterns, particularly the lateralized patterns, so lateralized periodic discharges and lateralized rhythmic delta activity and generalized periodic discharges have, are significantly associated with seizures. And this risk tends to be frequency and burden dependent. So the higher the frequency of these discharges, that means the more often they're occurring per second and the more the prevalence. So the longer that they're ongoing for, so the longer your EEG is occupied by these patterns, the greater the risk for having seizures. And so that's one thing to keep in mind. If I connect EEG and I'm seeing high frequency, high burden patterns, I might have a lower threshold to treat these patients. Then what is the evidence for secondary brain injury? And to start with just a quick overview of what we know from convulsive seizures. So convulsive seizures in the acute phase, there is this compensatory phase where there's a sympathetic overdrive. Your cardiac output, blood glucose, blood pressure, sort of all auto-regulating. There's cerebral auto-regulation and homeostasis is sort of maintained. But there comes a breaking point where these auto-regulatory mechanisms can no longer keep up and we start to see a systemic collapse as well as collapse of our cerebral auto-regulation. And so the question is, we know this happens in convulsive seizures, is there any evidence or data to suggest we see similar cerebral dysregulation or changes in cerebral metabolism with these non-seizure patterns or these ictal, interictal, periodic and rhythmic discharges? And in fact, there are now many small focus cohort studies on different patient populations that do demonstrate evidence for some associated secondary brain injury. So this was a study that looked at PET correlates across the heterogeneous population and found that periodic discharges, particularly generalized periodic discharges and lateralized discharges have hypermetabolism, the red shows hypermetabolism, similar to what we see with seizures. This was a study in subarachnoid hemorrhage patients that looked at brain tissue oxygenation and found that high-frequency discharges are associated with significant reductions in brain tissue oxygenation. Finally, this was a study in traumatic brain injury patients that found in patients during periods of periodic discharges, there's a drop in brain glucose and a rise in the lactate-pyruvate ratio suggesting evidence for secondary brain injury. So we've demonstrated that, you know, certain patterns, particularly hybrid and high-frequency, are associated with seizures. Similarly, some of these patterns, particularly high-frequency, are associated with secondary brain injury. What about outcomes? And so this also, and this is across disease categories, acute brain injury, hypoxic encephalopathy, acute status, as well as systemic illness in the absence of any structural brain injury, so your sepsis patients and liver failure patients. The increasing burden or increasing prevalence, the x-axis just shows increasing burden. So 0% being these patterns are absent, and 100% means the recording is continuously occupied by these patterns. So as the burden increases, there is a significant increase in the probability of poor discharge outcomes, and this is functional outcomes, adjusted for disease severity and illness severity. So, you know, we've shown there's data that they're associated with seizures, secondary brain injury, potentially associated with worse outcomes. Should we be treating them, and how aggressively should we be treating them? And unfortunately, this is where the data is most lacking, where there's some small focus cohort studies that have shown treatment may not be associated with outcomes, or high-intensity treatment may, in fact, worsen outcomes, but those are all single cohort, observational, retrospective, confounded by indication, and so it's hard to draw any conclusions from those. What does this result in? Practice variation. Seizure medications do end up being prescribed to many of these patients. They're even discharged in medications. There's widespread practice variation, and we know medications themselves are not benign and result in adverse effects. So how do we approach the treatment? The only sort of large, randomized trial was in comatose survivors of cardiac arrest. In this study, they randomized patients with periodic and rhythmic discharges to intense treatment, which they define as increasing seizure medications until these patents are completely suppressed for at least 48 hours, versus standard of care. And what they found was that there's no difference in three-month outcomes, three-month CPC or mortality. However, this is, you know, in a specific cardiac arrest population where we know that there's, you know, poor prognostic profile to begin with, and we can't apply this to all brain injury populations. Additionally, their target of complete suppression for 48 hours, it's not data-driven, and that may not be the target. The target may be give a little bit of seizure medication so that it's, you know, or suppress activity by a small percentage. We don't know that yet, and so it's unclear whether the target itself was the right target to study. And then finally, the control group, while they didn't get anti-seizure medications as part of standard of care, did receive anesthetics or sedation, and so it's hard to draw a definitive conclusion. And so we're still left at a point of what do we do? How do we treat these patients? And so as work is going on, part of my interest is how to treat these patients. The consensus has really arrived on appropriately triaging and identifying who to treat. So the first question is, is there a clinical correlate? If I have a patient who has lateralized periodic discharges and I'm examining them, and they're talking to me and they're fine, I should not chase those discharges. But if I have a patient where there's a clear clinical correlate, I would consider treatment. If there is a significant metabolic derangement, if the patient is extremely uremic, has generalized periodic discharges, I will consider aggressively chasing and treating the metabolic derangement. If there is a history of seizures, particularly immediately prior to the EEG monitoring, I will have a lower threshold of treatment. And then finally, if it's high burden, high frequency, high prevalence, I would consider treatment. And then very quickly, we can sometimes consider ancillary testing. So biomarkers, as we know, some of these patterns are associated with secondary brain injury. MRI scans and seizure patients have shown that seizures can result in cortical diffusion restriction. So you can do an MRI to see if there's a similar pattern of diffusion restriction. There's studies showing that these patterns can have PET correlates. So this is an example of a patient treated for status with no improvement in their exam. EEG shows these periodic discharges. We get a PET scan, and there's corresponding hypermetabolism that prompts increased treatment. There is multimodal monitoring. So we can use, in some severe TBI patients, we do use intracranial electrodes, where again, this is a patient with brain injury, persistently altered. The scalp EEG shows some lateralized rhythmic delta activity, which we may or may not chase, you know, based just on the scalp findings. But we did find this is what the depth electrode shows. So this patient had a deep electrode. And whenever we saw this rhythmic activity on the scalp, there were clear seizures on the depth electrode, which prompted treatment. So very quickly to conclude, if you see a patient that you're monitoring and you see these periodic and rhythmic patterns, the first question is, clinical correlate or not? If yes, consider treatment. If there is no clinical correlate, look at the frequency, the burden, the prevalence, and use that to guide treatment. If it's low frequency, low burden, continue to monitor. Treat metabolic derangements. And then if the patient is not improving, there's worsening EEG, I would consider treatment. And the choice of treatment, again, would depend on patient comorbidities. You can use sedating versus non-sedating, depending on whether they're intubated or not, you know, what their clinical profile is. And if you treat, monitor both for clinical as well as EEG improvement, and know that clinical improvement can sometimes lag behind. And if there's no improvement in either, consider ancillary testing before throwing on multiple medications for these patients. So with that, I'm going to end. I know we've gone a little bit over. Thank you.

continuous EEG monitoring

seizure detection

traumatic brain injury

periodic rhythmic patterns

patient outcomes