SCCM Pod-523 CCM: Clinical Predictors of Seizures in ICU Patients

---

Host Elizabeth H. Mack, MD, MS, FCCM, is joined by Samuel Snider, MD, and Michael Fong, MD, to discuss a retrospective cohort study that examined factors such as cardiac arrest, brain neoplasms, and EEG patterns to determine their association with status epilepticus and isolated seizures in critically ill patients, aiming to improve monitoring and treatment strategies for high-risk patients (Snider SB, et al. Crit Care Med. 2023 Aug;51:1001-1011). Samuel Snider, MD, is a board-certified neurologist at Brigham and Women's Hospital and an instructor of neurology at Harvard Medical School in Boston, Massachusetts. Michael Fong, MD, is an assistant professor adjunct at the Yale School of Medicine in New Haven, Connecticut.

*If you are unable to play the podcast please click here to download the file.

Transcript:

Dr. Mack: Hello and welcome to the Society of Critical Care Medicine Podcast. I’m your host, Dr. Elizabeth Mack. Today, I’ll be speaking with Dr. Samuel Snider and Dr. Michael Fong. We’ll be talking about the article, “Clinical and Electroencephalographic Predictors of Seizures in Status Epilepticus in 12,450 Critically Ill Adults: A Retrospective Cohort Study,” published in the August 2023 issue of Critical Care Medicine. To access the full article, visit ccmjournal.org.

Dr. Sam Snyder is a neurointensivist at the Brigham and Women’s Hospital in Boston, Massachusetts, and instructor of neurology at Harvard Medical School. Dr. Fong is a neurologist and subspecialty epileptologist at the Westmead Comprehensive Epilepsy Center in Sydney, Australia. Dr. Fong is also an assistant professor adjunct at the Yale School of Medicine, and he is a new coauthor of Hirsch and Brenner’s Atlas of EEG in Critical Care and second author of the ACNS Critical Care EEG Terminology 2021. Welcome, Drs. Snyder and Fong. Before we start, do you have any disclosures to report?

Dr. Snider: No, I have no disclosures, and thank you so much for having us. Happy to be here.

Dr. Fong: Dr. Mack, thanks so much for the kind introduction. I have no relevant disclosures.

Dr. Mack: Wonderful. I really appreciate your contribution to the literature, a very important one. I really enjoyed digging into your article. I’m curious, can you tell us a little bit about the Critical Care EEG Monitoring Research Consortium and a little bit about how the collaborative was formed, how this idea for the study developed? Where did that come from?

Dr. Fong: Yeah, that sounds great. Maybe because I’m a bit more involved in the consortium, I’ll start things off. The Critical Care EEG Monitoring Research Consortium, or CCEMRC, bit of a mouthful, was essentially built over two decades. It was formed by a number of clinicians and neurophysiologists who were interested in better understanding EEG patterns and seizures in the critical care setting.

Over the two decades, and again, I take no credit for its creation, but it started in 2007 and was formed out of the American Clinical Neurophysiology Society ICU EEG Monitoring Consortium, which was then nine academic centers in North America, led by Dr. Larry Hirsch at Yale, my good friend and colleague, and now it’s grown to this significant entity where it has more than 30 centers across North America and some representation in other regions as well.

The collaboration was formed, interestingly, because the old CCEMRC database was retired in 2021. Many listening to this podcast will appreciate that the ACNS standardized critical care EEG terminology was updated in 2021, so many of the terms changed and the database itself was getting a bit outdated. At the retirement of the old database, it was thought that we should provide essentially a final analysis of all the data. That data was originally looked at by Andres Rodriguez Ruiz from Emory in 2016 with 4700 patients. But final analysis that runs over the entire course between 2013 and 2021 now has 12,450 patients in it, so it’s the largest number of prospectively collected critical care EEG data.

Dr. Mack: Well, thank you for sharing that. I really enjoy hearing the history of it and all that’s gone into it. It sounds like quite a bit of work, so thank you for that. Can you remind us a little bit about why this matters? Really, what’s the impact of status epilepticus, of seizures, on mortality and morbidity in adults? I know there’s been some age-old debate on this, but does the impact depend on whether the seizures are convulsive or nonconvulsive?

Dr. Fong: Well, what great questions and incredibly topical questions. Maybe I’ll start off, being the EEG-er among the two of us, and then Dr. Snyder, being the neurointensivist, might follow up. It’s always important to remember that convulsive status epilepticus is absolutely harmful. The original three-month mortality data was between 19% to 24%, so it’s highly associated with mortality. That’s convulsive status epilepticus.

Part of the reason for this work was, it is, as a concept, gaining more information that the burden of seizures matter, even if they are nonconvulsive. There is emerging data that, for patients who have a low burden of seizures, they generally do a little bit better compared to patients who have a high burden of seizures, and that seems to be an independent risk factor. It seems to be perhaps not necessarily impacting mortality as much, but it certainly is impacting functional outcomes, where patients with high seizure burdens have worse functional outcomes in the longer term.

That’s the sort of premise as to why we think certainly treating convulsive status epilepticus is important but also to define the states associated with nonconvulsive seizures, especially a higher seizure burden.

Dr. Snider: Yeah, I think that’s a really good answer. I don’t have a ton to add to that, other than there’s been a lot published on predictors of seizures in the hospital, and there’s been relatively less published about predictors of status in the hospital. Whether you want to call it status or just general patients with very high burden of clinical seizures, we know those things are associated with adverse outcomes, whereas patients with isolated seizures and the association with adverse outcome is much less. So, just thinking about a clinically important end point, status, and very high seizure burdens are going to be really important things to understand what causes them.

Dr. Fong: Maybe just a quick plug in addition to that. When we talk about high seizure burden versus low seizure burden, the ACNS has taken on the literature that seizure burdens of greater than 20% of the record, so any 12 minutes in any 60-minute recording, has been associated with harms. They have now used that as the threshold for the updated definitions of status epilepticus.

Dr. Mack: Great. Thank you so much. I appreciate perspectives from you both. Take us to school. Tell us a little bit about RPPs and their frequency and meaning in critically ill adults.

Dr. Snider: Sure. Maybe I’ll lead with this one. RPPs refer to rhythmic and periodic patterns. With the wider and wider availability of continuous EEG, so more monitoring data that’s been accrued, we’ve noticed that there are things that show up on the record that are not seizures but seem to be associated with seizures, and they fall into the category of rhythmic and periodic patterns.

There’s been a fair amount written about this already, but you can think about them in terms of periodic discharges and rhythmic delta activity. You can think about them in terms of whether they’re generalized or lateralized, so you have generalized periodic discharges and lateralized periodic discharges. Then you have generalized rhythmic delta activity and lateralized rhythmic delta activity.

There was sort of a seminal paper in 2017 using this dataset, for the patients who were recruited up to that time and measured the associations with seizures for these different RPPs. What we knew going into the study was that lateralized periodic discharges have the strongest association with seizures, generalized periodic discharges being slightly less, and then lateralized rhythmic delta activity, LRDA, that seems also to be associated with seizures, but the generalized rhythmic delta activity, GRDA, it does not have a clear association with seizures in the literature. It’s a little bit debated, but it’s not clearly known to be associated with seizures. It’s thought to maybe represent more just encephalopathy or a brain that’s not functioning normally.

Dr. Fong: Just to add on to that for people looking at the terminology and then all of the subsequent studies, there is another pattern within rhythmic and periodic patterns, and that is spiking wave or sharpened wave. But the reason that that doesn’t make it into the literature so often is because it’s a rare pattern. It’s one of those that is hard to study but tends to be in line with the lateralized or generalized periodic discharges.

Dr. Mack: Great. Thank you for that. If you don’t mind, tell us a bit about the study design. You were retrospectively looking at more than 12,000 patients from three centers from 2013 to 2021. Give us a sense of what that looked like. That’s quite a bit of an undertaking.

Dr. Snider: Yeah, sure. This amazing database that Michael was telling us about, we have this huge sample of patients who had been monitored in the hospital and we knew whether or not they had seizures and we knew this information about what sorts of clinical diagnoses, what sorts of rhythmic and periodic patterns they had that had already been abstracted out of the EEG. It was in this nice database form for us to analyze.

The main focus of this study was, there’s been a fair bit written about predictors of electrographic seizures, but there’s much less written about predictors of electrographic status epilepticus, particularly in a sample of this size, this degree of clinical heterogeneity. What motivated this really was this idea of, are there unique clinical or electrographic predictors of status or is it just the same set of things that predict seizures, right? That’s what we set out to do.

With this in mind, we wanted to look at, what are the independent set of predictors for this outcome status epilepticus, right? So we wanted to build a multivariate model, and the type of regression that we chose allowed us to, for each predictor variable that we looked at, measure the association separately for isolated seizures and status epilepticus. So, for every variable we could say, okay, if you have this, are you more likely to have isolated seizures than to not have seizures at all? And are you more likely to have status epilepticus than isolated seizures? So we got two sets of odds ratios, if you’re following along with what I’m saying. Does that make sense?

Dr. Mack: Yeah, absolutely.

Dr. Snider: Michael, did you want to add anything there?

Dr. Fong: No, that sounds right.

Dr. Snyder: Okay. So what we did is we used an unbiased, fairly standard way of screening through a lot of candidate predictor variables, and what we found in our primary result is that the following clinical diagnoses were associated with status. A diagnosis of cardiac arrest, you’re about odds ratio of 10 for having status as compared to not having status, diagnoses of brain tumors, and then a clinical diagnosis of seizures or status listed as the reason for EEG monitoring.

Those were the three primary clinical predictors of status epilepticus. Again, these are all independent of each other. Then, among the rhythmic and periodic patterns, we found that lateralized periodic discharges, LPDs, the increased odds ratio for status there was about 7 to 8, GPDs, generalized periodic discharges. Then this other type of pattern called brief ictal rhythmic discharges, or BIRDs. That set of variables were what we found to be independently predictive of status epilepticus.

But the nice thing about this regression modeling strategy that we chose was it also gave us the ability to measure the odds ratios for status as compared to isolated seizures. So for the same set of variables, we could see, you know, were they really raising the risk of status or were they just raising the risk of seizures in general? Of those variables that I listed, it was only clinical diagnoses of cardiac arrest, LPDs, lateralized periodic discharges, GPDs, and then a clinical diagnosis of seizures or status.

Those were the only four things that actually increased your odds of status as compared to isolated seizures. Then, interestingly, patients with LRDA, lateralized rhythmic delta activity, while they were at a much higher risk of seizures, they actually appeared to be at a slightly lower risk of status relative to seizures. That was the primary set of findings from the paper, which was that clinical diagnoses of cardiac arrest or clinical seizures or status, GPDs and LPDs, were the four things that really seemed to independently increase your risk for status, even as compared to isolated seizures.

Dr. Fong: Following on from that, for clinicians, it’s important to understand the absolute numbers; the odds ratios are incredibly helpful, but in this large cohort, the probability of seizure for everybody was 9.8% and the probability of status epilepticus for everybody was 3.5%. But if you take someone with lateralized periodic discharges, their probability of seizure was just under 45% and their probability of status epilepticus was over 16%. So it just helps to put into context those odds ratios, it is significant increases in the baseline rate of seizure and status epilepticus.

Dr. Snider: Yeah, that’s a great point, Michael. I mean, if you had a patient where you knew that there was 20% chance of status epilepticus, right, you would almost certainly want to change your management or institute some sort of aggressive prophylactic therapy up front.

Dr. Mack: Absolutely. That really rings true to me as a clinician, and it seems we’ve been looking for predictors of these sorts of outcomes for a long time, particularly outcomes that impact such important factors like mortality. So thank you for that. Anything else from the results that you’d like to share? This has really been some great comment, particularly the predictor information.

Dr. Snider: Sure. We did a number of other analyses looking at if you’re able to substratify these RPPs based on different characteristics, as in how frequently they’re showing up on the record, other sorts of modifying characteristics. And we did not find that considering those modifying characteristics helped you predict the clinical outcome of seizures or status.

But the one thing that we did note is, for LPDs, which, as Michael mentioned, were associated with a 15% risk of status, if the sharpness of the LPD was such that it was defined as spiky, less than 70 milliseconds from onset to peak, that was associated with a very big bump in the risk of both isolated seizures and status. These spiky LPDs, nearly 30% of those patients had status epilepticus and over 60% had seizures. So there’s going to be further studies looking at that specific variable as an additional predictor.

Dr. Mack: Wonderful. You’re making me imagine a day where all of our ICU patients have pulse ox leads on their chest and leads on their heads as well. I wonder if that day will come.

Dr. Fong: That day is soon.

Dr. Mack: Yeah, especially in your world. Based on the findings, any ideas about where the consortium may go from here, what you all might be looking at?

Dr. Fong: Excellent question. I don’t speak for the entire consortium, but just in terms of where this work is going, this work and the consortium work over the past 10, 20 years has really set the foundation for improving outcomes in critically ill patients. The work so far has been about defining the patterns that are associated with seizures and, in this case, status epilepticus. But the next steps are really looking toward which patterns and what types of status epilepticus are associated with harm and, as an extension of that, just design well-controlled treatment trials with the aim of improving clinical outcomes. And not only the consortium, but many centers around the world are working toward that same goal.

Dr. Mack: Great. Thank you so much. Anything else you’d like to share?

Dr. Snider: No, I think Michael hit a lot of the highlights there, I mean, I’m sort of excited about using these data as a benchmark to think about designing some treatment trials. One of the things with these RPPs is that the management of these is very, very heterogeneous and I’m sure there’s massive variability in how they’re treated center to center, even doctor to doctor. And having some standard numbers for the incidence of seizure and status, with certain kinds of RPPs, maybe it can be a starting place to think about, might you design a randomized treatment trial? This could be a way to help think about how to power your trial.

Dr. Fong: I have nothing else to add about the manuscript, but I’ll mostly just use the opportunity to say thank you incredibly for having us on to talk about the work and, as an extension of that, it’s important to mention that this is work on behalf of the consortium, so I put out a huge thanks to all of the consortium members and all the clinicians and neurophysiologists and, when I was a fellow, adding in countless hours of time to the database. So for all of those who have helped over the years, I certainly say thank you. And I certainly thank you for the opportunity to speak on the topic.

Dr. Snider: That’s fantastic. I’m glad you said that. Yeah, the attendings and fellows who were entering this data into the database really made this possible. We should also call out our co-senior authors, Dr. Hirsch at Yale and Dr. Lee here at Brigham, for supervising the study.

Dr. Mack: Thank you all for taking the time to share with us your findings and really your vision and how this all came together. I think it’s very inspirational for other investigators. This concludes another episode of the Society of Critical Care Medicine Podcast. If you’re listening on your favorite podcast app and you like what you heard, consider rating and leaving a review. For the Society of Critical Care Medicine Podcast, I’m Dr. Elizabeth Mack. Thank you.

Announcer: Elizabeth H. Mack, MD, MS, FCCM, is a professor of pediatrics and chief of pediatric critical care at Medical University of South Carolina Children’s Health in Charleston, South Carolina, USA.

Join or renew your membership with SCCM, the only multiprofessional society dedicated exclusively to the advancement of critical care. Contact a customer service representative at +1 847 827-6888 or visit sccm.org/membership for more information.

The SCCM Podcast is the copyrighted material of the Society of Critical Care Medicine and all rights are reserved. Find more episodes at sccm.org/podcast. This podcast is for educational purposes only. The material presented is intended to represent an approach, view, statement, or opinion of the presenter that may be helpful to others. The views and opinions expressed herein are those of the presenters and do not necessarily reflect the opinions or views of SCCM. SCCM does not recommend or endorse any specific test, physician, product, procedure, opinion, or other information that may be mentioned.

Some episodes of the SCCM Podcast include a transcript of the episode’s audio. Although the transcription is largely accurate, in some cases it is incomplete or inaccurate due to inaudible passages or transcription errors and should not be treated as an authoritative record.

Disclaimer