Evaluating EEG's Role in Medication Tapering Decisions

Epilepsy classification systems have evolved over the years, with improved categorization of seizure types and adoption of more widely accepted terminologies. A systematic approach to the classification of seizures and epilepsy is essential for the selection of appropriate diagnostic tests and treatment strategies.

In this episode, Aaron Berkowitz, MD, FAAN, speaks with Roohi Katyal, MD, author of the article “Classification and Diagnosis of Epilepsy,” in the Continuum February 2025 Epilepsy issue.

Dr. Berkowitz is a Continuum® Audio interviewer and a clinical professor of neurology at the University of California, San Francisco.

Dr. Katyal is an assistant professor of neurology and codirector of adult epilepsy at Louisiana State University Health Shreveport in Shreveport, Louisiana.

Additional Resources

Read the article: Classification and Diagnosis of Epilepsy

Subscribe to Continuum: shop.lww.com/Continuum

Earn CME (available only to AAN members): continpub.com/AudioCME

Continuum® Aloud (verbatim audio-book style recordings of articles available only to Continuum® subscribers): continpub.com/Aloud

More about the American Academy of Neurology: aan.com

Social Media

facebook.com/continuumcme

@ContinuumAAN

Host: @AaronLBerkowitz

Guest: @RoohiKatyal

Full episode transcript available here

Dr Jones: This is Dr Lyell Jones, Editor-in-Chief of Continuum, the premier topic-based neurology clinical review and CME journal from the American Academy of Neurology. Thank you for joining us on Continuum Audio, which features conversations with Continuum's guest editors and authors who are the leading experts in their fields. Subscribers to the Continuum Journal can read the full article or listen to verbatim recordings of the article and have access to exclusive interviews not featured on the podcast. Please visit the link in the episode notes for more information on the article, subscribing to the journal, and how to get CME.

Dr Berkowitz: This is Dr Aaron Berkowitz, and today I'm interviewing Dr Roohi Katyal about her article on classification and diagnosis of epilepsy, which appears in the February 2025 Continuum issue on epilepsy. Welcome to the podcast, Dr Katyal, and could you please introduce yourself to our audience?

Dr Katyal: Thank you for having me. I'm very excited to be here. I'm Dr Roohi Katyal. I currently work as Assistant Professor of Neurology at LSU Health Shreveport. Here I also direct our adult epilepsy division at LSU Health along with my colleague, Dr Hotait. 

Dr Berkowitz: Fantastic. Well, happy to have you here. Your article is comprehensive, it's practical, and it focused on explaining the most recent International League Against Epilepsy (ILAE) classification of epilepsy and importantly, how to apply it to provide patients with a precise diagnosis of epilepsy and the particular subtype of epilepsy to guide the patient's treatment. There are so many helpful tables and figures that demonstrate all of the concepts and how to apply them at the bedside. So, I encourage our listeners to have a look at your article, even consider maybe screenshotting some of these helpful tables onto their phone or printing them out for handy reference at the bedside and when teaching residents. Your article begins with the current definition of epilepsy. So, I want to ask you about that definition and make sure we're on the same page and understand what it is and what it means, and then talk through a sort of hypothetical patient scenario with you to see how we might apply these in clinical practice. You talked about, in your article, how the new definition of epilepsy from the ILAE allows for the diagnosis of epilepsy in three different scenarios. So, could you tell us what these scenarios are?

Dr Katyal: So, epilepsy in general is a chronic condition where there is a recurrent predisposition to having seizures. As you mentioned, epilepsy can be diagnosed in one of three situations. One situation would be where an individual has had two or more unprovoked seizures separated by more than 24 hours. The second situation would be where somebody has had one unprovoked seizure and their risk of having recurrent seizures is high. And the third situation would be where somebody had---where the clinical features could be diagnosis of an epilepsy syndrome. An example of that would be a young child presenting with absence seizures and their EEG showing 3 Hz characteristic generalized spike in with discharges. So that child could be diagnosed with childhood absence epilepsy. 

Dr Berkowitz: Perfect. Okay, so we have these three scenarios, and in two of those scenarios, we heard the word unprovoked. Just to make sure everyone's on the same page, let's unpack this word “unprovoked” a little bit. What does it mean for a seizure to be unprovoked versus provoked? 

Dr Katyal: So unprovoked would be where we don't have any underlying provoking features. So underlying provoking features are usually reversible causes of epilepsy. These would be underlying electrolyte abnormality, such as hyperglycemia being a common one which can be reversed. And these individuals usually do not need long-term treatment with anti-seizure medications.

Dr Berkowitz: Fantastic. Tell me if I have this right, but when I'm teaching residents, I… did it provoked and unprovoked---there's a little confusing, right? Because we use those terms differently in common language than in this context. But a provoked seizure, the provoking factor has to be two things: acute and reversible. Because some people might say, well, the patient has a brain tumor. Didn't the brain tumor provoke the seizure? The brain tumor isn't acute and the brain tumor isn't reversible, so it would be an unprovoked seizure. I always found that confusing when I was learning it, so I try to remind learners I work with that provoked means acute and reversible, and unprovoked means it's not acute and not reversible. Do I have that right? Am I teaching that correctly?

Dr Katyal: That's correct.

Dr Berkowitz: Great. And then the other important point here. So, I think we were all familiar prior to this new guideline in 2017 that two unprovoked seizures more than twenty-four hours apart, that's epilepsy. That's pretty straightforward. But now, just like we can diagnose MS at the time of the first clinical attack with the right criteria predicting that patient is likely to have relapse, we can say the patient’s had a single seizure and already at that time we think they have epilepsy if we think there's a high risk of recurrence, greater than or equal to sixty percent in this guideline, or an epilepsy syndrome. You told us what an epilepsy syndrome is; many of these are pediatric syndromes that we've studied for our boards. What hertz, spike, and wave goes with each one or what types of seizures. But what about this new idea that a person can have epilepsy after a single unprovoked seizure if the recurrence rate is greater than sixty percent? How would we know that the recurrence rate is going to be greater than sixty percent?

Dr Katyal: Absolutely. So, the recurrence rate over sixty percent is projected to be over a ten year period. So, more than sixty percent frequency rate in the next ten years. And in general, we usually assess that with a comprehensive analysis and test. So, one part of the comprehensive analysis would be, a very important part would be a careful history taking from the patient. So, a careful history should usually include all the features leading up to the episodes of all the prodromal symptoms and warning signs. And ideally you also want to get an account from a witness who saw the episode as to what the episode itself looked like. And in terms of risk assessment and comprehensive analysis, this should be further supplemented with tests such as EEG, which is really a supportive test, as well as neuroimaging. If you have an individual with a prior history of, let's say, left hemispheric ischemic stroke and now they're presenting with new onset focal aware seizures with right arm clonic activity, this would be a good example to state that their risk of having future seizures is going to be high.

Dr Berkowitz: Perfect. Yeah. So, if someone has a single seizure and has a lesion, as you said, most common in high-income countries would be a prior stroke or prior cerebrovascular event, prior head trauma, then we can presume that the risk is going to be high enough that we could call that epilepsy after the first unprovoked seizure. What if it's the first unprovoked seizure and the imaging is unremarkable? There's no explanatory lesion. How would we get to a diagnosis of epilepsy? How would we get to a risk of greater than sixty percent in a nonlesional unprovoked seizure? I should say, no lesion we can see on MRI.

Dr Katyal: You know, in those situations an EEG can be very helpful. An EEG may not always show abnormalities, but when it does show abnormalities, it can help us distinguish between focal and generalized epilepsy types, it can help us make the diagnosis of epilepsy in certain cases, and it can also help us diagnose epilepsy syndromes in certain cases. 

Dr Berkowitz: Perfect. The teaching I remember from a resident that I'm passing on to my residents, so please let me know if it's correct, is that a routine EEG, a 20-minute EEG after a single unprovoked seizure, this sensitivity is not great, is that right? Around fifty percent is what I was told with a single EEG, is that right?

Dr Katyal: Yeah, the sensitivity is not that great. Again, you know, it may not show abnormality in all the situations. It's truly just helpful when we do see abnormalities. And that's what I always tell my patients as well when I see them in clinic. It may be abnormal or it may be normal. But if it does show up normal, that does not rule out the diagnosis of epilepsy. Really have to put all the pieces together and come to that finally diagnosis.

Dr Berkowitz: Perfect. Well, in that spirit of putting all the pieces together, let's walk through together a hypothetical case scenario of a 19-year-old patient who presents after a first event that is considered a possible seizure. First, how do you approach the history and exam in this scenario to try to determine if you think this was indeed an epileptic seizure? 

Dr Katyal: So, if I'm seeing them in the clinic or in the outpatient setting and they're hopefully presenting with somebody who's already seen the seizure itself, my first question usually is if they had any warning signs or any triggers leading up to the episode. A lot of times, you know, patients may not remember what happened during the episode, but they may remember if they felt anything different just before or the day prior, something different may have happened around that time. Yeah, so they may report that. Then a very important aspect of that would be talking to somebody who has seen the episode, a witness of the episode; and ideally somebody who has seen the onset of the episode as well, because that can give us very important clues as to how the event or the episode started and how it progressed. And then another very important question would be, for the individual who has experienced it, is how they felt after the episode ended. So, you can get some clues as to if they had a clear postictal state. Other important questions would be if they had any tongue biting or if they lost control of their bladder or all those during the episode. This, all those pieces can guide us as to if the seizure was epileptic, or the episode was epileptic or not.

Dr Berkowitz: Fantastic. That's very helpful guidance. All right. So, let's say that based on the history, you're relatively convinced that this patient had a generalized tonic clonic seizure and after recovering from the event, you do a detailed neurologic exam. That's completely normal. What's your approach at this point to determining if you think the seizure was provoked or unprovoked, since that's, as you said, a key component of defining whether this patient simply had a seizure, or had a seizure and has epilepsy?

Dr Katyal: The important findings would be from the laboratory test that may have been done at the time when the patient first presented with the seizure. So, we want to rule out features like hypoglycemia or other electrolyte abnormalities such as changes in sodium levels or big, big fluctuations there. We also want to rule out any other metabolic causes or other reasons such as alcohol withdrawal, which can be a provoking factor. Because these would be very important to rule out is if we find a provoking reason, then this individual may not need to be on long term anti-seizure medication. So very important to rule that out first. 

Dr Berkowitz: Great. So, let's say you get all of your labs and history and toxicology screen and no provoking factors there. We would obtain neuroimaging to see if there's either an acute provoking factor or some type of lesion as we discussed earlier. Let's say in this theoretical case, the labs are normal, the neuroimaging normal. There is no apparent provoking factor, there's no lesion. So, this patient has simply had a single unprovoked seizure. How do we go about now deciding if this patient has epilepsy? How do we try to get ourselves to either an above sixty percent risk and tell this patient they have epilepsy and probably need to be on a medicine, or they have a less than sixty percent risk and that becomes a little more tricky? And we'll talk about that more as well. 

Dr Katyal: For in a young patient, especially in a young patient as a nineteen year old as you present, one very important aspect if I get this history would be to ask them about absolutely prior history of similar episodes, which a lot of times they may not have had similar episodes.

But then with this age group, you also want to ask about episodes of brief lapses in awareness or episodes of sudden jerking or myoclonic jerking episodes. Because if you have brief lapses of awareness, that could signify an absence seizure in this particular age group. And brief, sudden episodes of myoclonic jerking could be brief myoclonic seizures in this age group. And if we put together, just based on the clinical history, you could diagnose this patient with a very specific epileptic syndrome, which could be juvenile myoclonic epilepsy in the best case. Let's say if you ask about episodes of staring or relapses of awareness, that's not the case, and there's no history of myoclonic jerking episodes or myoclonic seizures, then the next step would be proceeding to more of our supplemental tests, which would be an EEG and neuroimaging. In all cases of new-onset seizure especially should have comprehensive assessment with EEG and neuroimaging to begin with, and we can supplement that with additional tests wherever we need, such as genetic testing and some other more advanced testing. 

Dr Berkowitz: That's very helpful. OK, so let's say this particular patient, you talk to them, you talk to their family, no prior history of any types of events like this. No concerns for spells that could---unlike absence, no concern for movements that could sound like myoclonus. So, as you said, we would be looking for those and we could get to part one of the definition. There is more than one spell, even though we're being consulted for one particular event. But let's say this was the only event, we think it's unprovoked, the neuroimaging is normal. So, you said we proceed to an EEG and as you mentioned earlier, if the EEG is abnormal, that's going to tell us if the risk is probably this more than sixty percent and the patient should probably be on a medicine.

But common scenario, right, that the patient has an event, they have a full work up, we don't find anything. We're convinced it was a seizure. We get our routine EEG as we said, very good, an affirmative test, but not a perfectly sensitive test. And let's say this person's routine EEG turns out to be normal. So how would you discuss with the patient their risk of a future seizure and the considerations around whether to start an anti-seizure medicine if their work-up has been normal, they've had a single unprovoked seizure, and their EEG is unrevealing?

Dr Katyal: And I'm assuming neuroimaging is normal as well in this case?

Dr Berkowitz: Correct. Yeah. 

Dr Katyal: We have a normal EEG; we have normal neuroimaging as well. So, in this case, you know, it's more of a discussion with the patient. I tell them of that, you know, the risk of seizure may not be higher than sixty percent in this case with all the tests being normal so far and there's no other prior history of similar episodes. So, we have a discussion with them about the risks that can come with future seizures and decide where the medication should be started or not. 

Dr Berkowitz: And so how do you approach this discussion? The patient will say, Doctor Katyal, I had one seizure, it was very frightening. I got injured. You told me I can't drive for however many months. One cannot drive in that particular state. But I don't really like taking medicines. What is my risk and what do you think? Should I take a medicine? 

Dr Katyal: I'll tell you this because normally I would just have a direct conversation with them, discuss all the facts that we have. We go over the seizure one more time just to make sure we have not missed any similar episodes or any other episodes that may be concerning seizure, which ruled out all the provoking factors, any triggers that may be seen inseizures like this in a young age. And another thing would be to basically have a discussion with them, you know, these are the medication options that we can try. And if there is another seizure, you know, these are the these are the restrictions that would come with it. And it's a very individualized decision, to be honest. That, you know, not everyone may want to start the medication. And you'll also find that some patients who, you know, some individuals are like no,  I want to go back to driving. I don't want to be in this situation again. I would like to try a medication and don't want to ever have a seizure. So, I think it's a very individualized decision and we have a discussion with the patient based on all of these tests. And I would definitely maintain follow-up with them to make sure that, you know, things have not changed and things have---no seizures have recurred in those cases. 

Dr Berkowitz: Yeah, great to hear your approach. And similar experience to you, right, where some patients say, I definitely don't want to take the medication, I'll roll the dice and I hope I don't have another seizure. And we say, we hope so also. As you said, let's keep a close eye. And certainly, if you have another seizure, it's going to be a lifelong seizure medicine at that point. And some patients who, as you said, say, wait, I can't drive for months. And if I don't take a medicine and I have a seizure in the last month, I would have to have another period of no driving. Maybe in that case, they would want to start a medicine. That said, we would present that either of these are reasonable options with risks and benefits and these are the medications we would offer and the possible side effects and risk of those, and make a joint decision with the patient.

Dr Katyal: Absolutely correct. Mentioned it perfectly well that this is a very individualized decision and a joint decision that we make with the patient. 

Dr Berkowitz: Fantastic. Another topic you touch on in your article is the definition of resolved epilepsy. How is that defined in the guidelines? 

Dr Katyal: Yes. So, an epilepsy can be considered resolved if an individual has been seizure-free for at least ten years and has been off of IV seizure medications for at least five of those years. Another situation where epilepsy can be considered resolved would be if they have an age-defined epilepsy syndrome and now they are beyond the relevant age group for the syndrome. 

Dr Berkowitz: That's very helpful. So again, a very clear definition that's helpful in these guidelines. And yet, as I'm sure you experience your practice, as I do in mine, sometimes a little challenging to apply. So, continuing with our made-up hypothetical patient here, let's say at some point in the subsequent years, they have a second unprovoked seizure, still have a normal EEG but they do go on an anti-seizure medicine. And maybe four or five years later, they're seizure-free on a low dose of an anti-seizure medicine. And they say, you know, do I really still need this medicine? I'd really like to come off of it. What do you think? Is that safe? How do you talk about that with the patient? This definition of ten years and five years off medicine seems to be---and maybe unless someone's seeing a lot of children and young adults, a relatively uncommon scenario. It's we've had a first unprovoked seizure. We never figured out why. We don't really know why they had the seizure. We can't really gauge their subsequent risk. They're on medicines, they don't want to be on them and it's only been a few years, let's say three, four, or five years. How do you frame discussion with the patient?

Dr Katyal: Yeah, so that's the definition of being resolved. But in terms of tapering off medications, we can usually consider tapering off medications earlier as well, especially if they've been seizure-free for two or more years. Then again, as we mentioned earlier, it would be a very individualized decision and discussion with the patient, that we could consider tapering off of medication. And we would also want to definitely discuss the risk of breakthrough seizures as we taper off and the risks or the lifestyle modifications that would come with it if they have another breakthrough seizure. So, all those things will go into careful concentration when we decide to taper off, because especially driving restriction may be a big, you know, hard stop for a lot of patients that, you know, now is not a time to taper off medication. So, all of these factors will go into consideration and we could consider tapering off earlier as well. 

Dr Berkowitz: That's very helpful. Yeah, as you said, when we're tapering off medications, if that's the direction the patient wants to go during that period, obviously we wouldn't want them to drive, or be up on a ladder, or swimming alone. You said that some patients might say, actually, I'll keep the medicine, whereas some might say, OK, I'll hold off on all these activities and hope that I can be off this medication. I remember epilepsy colleagues quoting to me at one point that all comers, when a patient's been seizure-free for two years, they estimate the risk of relapse, of having another seizure, somewhere around thirty to forty percent. In your expert opinion, is that about what you would quote to a patient as well,. about a thirty to forty percent, all comers? Obviously not someone who's had a history of status epilepticus and has a lesion or a syndrome, but in the sort of common situation of some unprovoked seizures in an adult, we don't have a clear ideology. Is that thirty to forty percent figure, more or less, you would place the risk when you talk to the patient? Or? 

Dr Katyal: Yeah, absolutely, especially if the neuroimaging is completely normal, all their EE GS have been normal. They have been in this situation---you have a young patient with two seizures separated by so many years. After three or four years of being on the medication and, you know, the patient has been adhering. There are no more seizures. Thirty to forty percent seems reasonable, and this is what I usually tell them that the risk of, as we taper off medications, that risk is not zero but it's low. And around thirty percent is relatively where we would place the risk at.

Dr Berkowitz: We've said in this theoretical case that the EEG is normal. But last question, I've heard some practitioners say that, well, let's say the patient did have an abnormal EEG early on. Not a syndrome, but had maybe a few focal spike wave discharges or sharps and that made you convinced that this patient had epilepsy. But still becomes seizure free for several years. I've heard of some practitioners repeating the EEG before tapering the anti-seizure medicines and I always wonder, would it change anything? It's a brief twenty-minute period. They still have one spike, but I tell them they can't come off. If the spikes are gone, it may be because of the medication, and maybe when I take them off they would have a spike. And how do you use---do you use or how do you use EEG in that decision of whether to taper a medicine?

Dr Katyal: Yeah. In general, I would not always use an EEG for considering tapering off medication. Again, it's very individualized decision. I can give you a hypothetical example, but it's a fairly common one, is that if an individual with let's say focal seizures with impaired awareness, they live alone, they live by themselves. Oftentimes they'll say that, I'm not sure if I'm missing any seizures because nobody has seen them. I may or may not be losing awareness, but I'm not too certain. They have not had any definite seizures for history in the last couple of years and are now considering tapering off medication. So, this may be a situation where I may repeat an EEG, and perhaps even considering the longer EEG for them to understand their seizure burden before we decide to taper off medication. But in most situations, especially if we consider the hypothetical situation you had mentioned for the young patient who had to witness seizures separated by several years and then several years without any seizures, that may be a good example to consider tapering off medication, especially considering all the tests that had been normal before then. 

Dr Berkowitz: That's very helpful to hear. And of course, this is your expert opinion. As you said, no guidelines and different people practice in different ways, but helpful to hear how you approach this common and challenging scenario for practitioners. Well, I want to thank you again, Dr Katyal. This has been a great opportunity to pick your brain on a theoretical case, but one that I think presents a number of scenarios that a lot of us---myself as a general neurologist, as well as you and your colleagues as epileptologists, we all see in general practice patients with unprovoked seizures and a revealing workup, and how to approach this challenging scenario based on the guidelines and on your expert opinion. I learned a lot from your article. Encourage our readers again to take a look. A lot of very helpful tables, figures, and explanations, some of the concepts we've been discussing. So again, today I've been interviewing Dr Roohi Katyal about her article on classification and diagnosis of epilepsy, which appears in the most recent issue of Continuum on Epilepsy. Be sure to check out Continuum audio episodes from this and other issues. And thank you again so much to our listeners for joining us. 

Dr Katyal: Thank you for having me.

Dr Monteith: This is Dr Teshamae Monteith, Associate Editor of Continuum Audio. If you've enjoyed this episode, you'll love the journal, which is full of in-depth and clinically relevant information important for neurology practitioners. Use this link in the episode notes to learn more and subscribe. AAN members, you can get CME for listening to this interview by completing the evaluation at continpub.com/audioCME. Thank you for listening to Continuum Audio.