What is autoimmune-associated epilepsy, and how is it different from acute seizures with an autoimmune cause? How can clinicians best diagnose and treat these conditions? Sharp Waves reporter Dr. Maryam Nabavi-Nouri spoke with Dr. Claude Steriade.
The following is the transcript from an episode of ILAE’s podcast, Sharp Waves.
I’m Maryam Nabavi-Nouri; I’m a pediatric neurologist and epileptologist. I practice at the University of Western Ontario, in London, Ontario, Canada. Part of my research and clinical focus revolves around autoimmune mediated epilepsy. I’m interviewing Dr. Claude Steriade, from the New York University School of Medicine, about autoimmune mediated epilepsy and autoimmune encephalitis.
Thank you for joining us, Claude, for a friendly chat and interview on this episode of Sharp Waves. I wanted to start by asking you how you became interested in this field, and what inspired you to focus on autoimmunity and epilepsy.
Dr. Steriade: Sure. I think my interest in autoimmunity and neurology probably started in residency. I had a patient with autoimmune encephalitis, and I was a PGY2 (second-year resident) on the inpatient ward, and you know, that can sometimes be a little bit draining. You see a lot of patients who are very sick and not all of them get better. I had this patient who was very ill and responded beautifully to immunotherapy and it was like a breath of fresh air. This was a disease for which the pathophysiology was fascinating, and it also was extremely rewarding clinically in terms of therapeutic response. That’s probably where my clinical interest in that started.
As I went through residency, I was really drawn to epilepsy for various reasons—probably also in part the really drastic response to treatment that many patients have – you can really turn people’s lives around, you know, in epilepsy as well. So when I went into fellowship in epilepsy, I decided to try to marry those two interests. In addition to my epilepsy fellowship, I joined in some of the autoimmune neurology clinics, at the Cleveland Clinic, and tried to gain some clinical experience there. And I noticed clinically that there were quite a few situations where there was relevant overlap between the two fields.
A lot of patients with autoimmune encephalitis develop seizures, so whenever I was on the cEEG rotation, readding critically ill patients’ EEGs, there were patients in whom that overlap was relevant, so I started wondering if that relevance was there in other settings as well. That’s where my interest in autoimmune associated epilepsy came about, and I developed an interest in trying to hone down the phenotypic signatures of patients who may have an autoimmune mechanism for their epilepsy.
When I came to NYU to start my first attending job, my only attending job, that’s what I continued doing. It’s a field that sort of has it all for me. Clinically it’s such an interesting area, these patients are fascinating to listen to, and you listen to their seizure history, their seizure semiology, when you listen to the other neurologic symptoms they have – they’re interesting to treat, they’re rewarding to treat for the most part, though obviously a lot of patients with autoimmune associated epilepsies can be therapy resistant and they can be very sick, but it can be a very rewarding process. I also continue to be fascinated by their biology, which is why I continue to focus my research effort on that patient population.
Dr. Nabavi-Nouri: Can you very broadly tell us about our current knowledge of immune-mediated epilepsies in the field?
Dr. Steriade: Yeah, and I think we’ve known for a long time that epilepsy in and of itself is a neuroinflammatory disease. There’s a broad literature on neuroinflammation and seizures and the bidirectional relationships between both. In patients whose epilepsy is structural in nature, for example in focal cortical dysplasias, or in any epilepsies, what’s been more, uh, new is that over the past decade or so there’s this entity of epilepsies in which neuroinflammation isn’t just necessarily an epiphenomenon which may feed back into furthering seizures, but it’s the originating or causative factor that is initiating the epilepsy in the first place.
So here I’m not talking about patients who have structural lesions and they have seizures and there’s neuroinflammation. I’m talking about patients who are healthy, have no reason to develop seizures and all of the sudden develop hard to treat, drug-resistant seizures, often temporal lobe in onset, with associated psychiatric and cognitive comorbidities, in whom the original causative etiological problem is autoimmune in nature. And the reason we’ve been able to identify this is the discovery of various antibodies that target neuronal proteins, some of which target intracellular antigens like GAD65 and some of the onconeural antibodies, and some of which target cell surface antigens like LGI-1, CASPR-2, NMDA, GABA-B, and they each present in different ways, as we know, and can respond to immunotherapy, particularly the cell surface antibodies, much better than they might respond to conventional epilepsy treatments like antiseizure medications, epilepsy surgery and so on.
And so that’s where really the advance has been and now there’s an interest in finding out well, are there patients out there who are walking, talking, with seizure disorders, showing up to our epilepsy clinics, some of them may be not responding to antiseizure medications or epilepsy surgery, who may have an autoimmune cause for their seizures that we’re not identifying because we’re not looking for it in some of these patients.
The question is how many patients are there—not with phenotypically clear autoimmune encephalitis who are intubated in the ICU – how many patients are there in our epilepsy clinics with focal epilepsy of unknown cause who actually have an autoimmune problem? There’s been a number of studies in this area, they’re rigorous studies that tried to exclude patients who phenotypically very clearly have encephalitis. The studies that look at this question – how many patients with focal epilepsy have an autoimmune cause for their seizures, defined as the presence of an autoantibody, it’s about 5%. Which is pretty significant if you think about the number of patients who walk into clinics with focal epilepsy of unknown cause – 1 in 20 people have an autoimmune cause for their seizures and can potentially be diagnosed and treated with immunotherapy and potentially respond to it.
Dr. Nabavi-Nouri: I would say it was a very revealing moment when that paper came out from the ILAE task force, and I would say you have contributed significantly to our understanding of this field. It’s definitely something that most physicians were struggling with, and once the distinction and the definitions came out, for a lot of us working in epilepsy it was like a breath of fresh air because you could coin it differently. What led to the definitions and the task force and the work that you guys published as a result?
Dr. Steriade: Sure. So the history behind that paper was that this task force was created, probably while I was finishing fellowship, so 2018, by the ILAE, to try to improve the diagnosis and treatment of autoimmune-related seizure disorders. That task force is kind of unique because it’s a mix of epileptologists and immunologists. I was lucky enough to be selected – I was very junior when I was selected. I was in this room with the bigs of this field – Andrew McKeon, Jeff Britton, Sarosh Irani, and all these big names.
And one of the main themes that came about in some of these initial meetings was well, what are we even talking about when we say autoimmune epilepsy? That term is all over the place in our clinics, in our electronic health records, the term autoimmune epilepsy comes up a lot, in the literature, the term autoimmune epilepsy if you do a PubMed search, comes up a lot. But what is it that we actually talking about when we say autoimmune epilepsy and if we asked different people in the room, we were getting different answers, and that’s a problem, if the experts don’t agree on that.
In the end, the experts did agree that we’re talking about two problems and lumping them into one, and those two problems are biologically different, they behave differently clinically, and they should be tackled therapeutically differently. And they’re really sort of diseases that need to be addressed with a different mindset in terms of research. So they can’t be lumped in together. Those two entities should be distinguished.
There’s the entity of acute symptomatic seizure secondary to autoimmune encephalitis. Those are just like acute symptomatic seizures from a traumatic brain injury – they are seizures that happen acutely because of an acute brain injury – in this case it’s an inflammatory brain injury – related to autoimmune encephalitis. For example, your patient with NMDA receptor encephalitis who’s admitted – she’s got psychiatric changes, catatonia, and then she has a seizure on the psych ward and gets transferred to neurology. Does that patient have epilepsy? No, she has acute symptomatic seizures from an acute brain problem that needs to be treated, and when you treat it with immunotherapy, almost all the time, not always but almost all the time, the seizures will resolve.
Epilepsy is not that. Epilepsy is an enduring predisposition to recurrent unprovoked seizures. It’s a chronic condition, that’s why it’s so disabling.
The other condition we outlined in the paper was autoimmune-associated epilepsy. That’s the patient showing up to your clinic with five-year history of bitemporal drug resistant epilepsy who’s also got a lot of memory impairment, and you know has comorbid depression, anxiety, she’s also got comorbid type 1 diabetes, very frequent seizures whenever you do an EEG, and then you do a GAD65 and she’s got titers through the roof. MRI shows bitemporal flare changes. That patient has GAD65 autoantibody associated epilepsy.
The term “associated” was chosen very carefully. There was a thought okay, this is actual epilepsy, this is autoimmune epilepsy but then we paused and said well, autoimmune epilepsy implies this is a pure autoimmune problem, and therefore as you treat the autoimmune problem the epilepsy should go away. But is that the case? Not really, so that patient with five years history of bitemporal epilepsy and her MRI shows signal changes in the temporal lobes, that patient very often doesn’t respond to immunotherapy very well. Our GAD65 patients are very refractory. Is the issue that we’re not good at choosing the type of immunotherapy? Maybe, but maybe the other issue is that it’s not purely an autoimmune problem.
In many of these patients there’s a structural substrate for their epilepsy. Many have findings consistent with hippocampal sclerosis on MRI, if you end up doing epilepsy surgery, many of them do have hippocampal sclerosis on histopathology. So it’s associated with an autoimmune problem, and in some patients the autoimmune part of their epilepsy may be the predominant part and if you treat that well, they may respond really well. But in some other patients, maybe it started out as mostly an autoimmune problem, and as the years have gone on it’s become mostly a structural problem. That’s why if you give them steroids, IVIG, Cytoxan and all the rest, they’re not going to get better because the hippocampus is fried.
So that term “associated” was chosen carefully. Actually, Jackie French was the one who suggested that. I came to her and said you know there’s a problem, I feel like we’re going to mislead people with that term and she’s the one who suggested associated and it stuck.
And once you sort of put that on paper, and I’m glad it was well received, I think most people were like okay, clearly there’s two different categories of patients. Now operationally, I feel like there are patients who will go from one category to another, like I have LGI-1 patients, they’ve had LGI-1 for weeks to months, I treat them and not all of them respond. I used to think it was me who didn’t know what I was doing and then the Mayo group came out with their long-term data and even Mayo has some patients at 2 years who still seize – about 20%. So how does this LGI-1 patient who had encephalitis, you’ve been treating him with immunotherapy and they’re continuing to have seizures and it’s been 2 years – does that patient now have epilepsy? Do they still have encephalitis?
I think it comes down, the issue of distinguishing in some of these patients who are transitioning from the acute phase to the chronic phase, the issue of how we’re going to define them as acute symptomatic versus autoimmune-associated epilepsy comes down to a lack of really good biomarkers for active autoimmunity and active inflammation.
Antibody titers aren’t great. In NMDA, antibody titers tend to follow clinical course, but not always. There are patients that have NMDA antibodies when they’re doing really well after years and years. Same with LGI-1. We don’t have great ways of trying to track whether we’re morphing from an inflammatory problem to a structural one. That means it’s a problem therapeutically in terms of clinical decision making.
Why does it matter to have two different categories? What’s the impact of that on your clinical decision making, and on your counseling of patients?
Well the impact of that is: A, patients with acute symptomatic seizures secondary to autoimmune encephalitis tend to do better, have a better prognosis, than patients with autoimmune associated epilepsy, so it helps in your counseling patients. And B, it helps you in terms of how hard to push immunotherapy.
I generally always do a trial of immunotherapy when I make a new diagnosis of, whether it’s autoimmune encephalitis or autoimmune associated epilepsy, I mean you’re going to treat both but your expectations for response are going to be different. Not every autoimmune-associated epilepsy patient behaves the same, and not every autoimmune encephalitis patient behaves the same. There are a number of factors that come into play – the type of antibody, how long they’ve been sick, presence of an underlying malignancy and so on.
Having these two categories are the starting point for expectations in counseling, expected response to immunotherapy and decisions on escalation of immunotherapy. You’re for sure going to escalate to second-line therapy in an autoimmune encephalitis case that’s not responding to first line. In a patient that has autoimmune-associated epilepsy and has zero response to first-line immunotherapy, we don’t know what to do. These patients are usually very sick, and I do tend to try many different therapies, being very open with them that there might not be a response. But we don’t know that that’s the right thing to do, to give someone with GAD65 antibody associated epilepsy Cytoxan when we have clear hippocampal sclerosis on the MRI if they haven’t responded to steroids/IVIG – there are many physicians who would not do that, and they would not be wrong.
I think trying to operationalize the distinction will be important to sortof make it easier for clinicians to do these things that I just talked about – accurately prognosticate and make clinical decisions. I think it’s going to come down to having better biomarkers of active neuroinflammation, active brain autoimmunity in these patients, to inform us whether there’s untreated brain autoimmunity that would benefit from aggressive immunotherapy.
Dr. Nabavi-Nouri: And to that point, we don’t have any set time points. Of course, these are patients that are pretty complex so there’s never, I don’t think there’s been a consensus in terms of whether we use a timeline to decide patients fall on this side of the spectrum versus the other.
Dr. Steriade: That’s a really important comment. There was a discussion about that in the task force – do we say that if you’re still having seizures at one year, do you then have epilepsy? And the decision was made to not have that timeline because it was felt that A, there’s so much heterogeneity in the disease between patients and B, there are some patients who come to us late. They haven’t been diagnosed. I’ve seen an LGI-1 patient who came to me a year after she first got diagnosed and she responded beautifully to steroids, thankfully. Just because she hadn’t been identified as having LGI-1 does that mean she didn’t have encephalitis anymore at the magic 365-day mark? No!
Because of that we decided to not have a timeline and instead say that whether someone falls into one or the other category comes down to the whole clinical picture, whether there’s ancillary evidence of an active autoimmune encephalitis, whether they have active CSF, an MRI showing signs encephalitis and so on, uh, to make that decision, that distinction, but you know it’s not perfect.
Dr. Nabavi-Nouri: In what subset of patients with drug-resistant epilepsy do you consider an autoimmune cause that you would go the extra mile of investigating them and how far? How extensive do you investigate them?
Dr. Steriade: So there’s a couple of groups that have tried to answer this. There’s the Dutch group, Martin Titulaer’s group, developed a score called the ACES score, the antibodies contributing to epilepsy and seizures score, in which they enrolled patients who did not have a clear encephalitis clinically and tested them for neural autoantibodies – those are all patients with focal epilepsy of unknown cause – and then developed a score which indicates whether you’re likely to find an antibody or not. There were various elements in there which were kind of reminiscent of encephalitis – behavioral changes, autonomic changes and so on.
The Oxford group, Sarosh Irani’s group, also came up with a score, which they didn’t give a catchy name to, with various other clinical features – they also included piloerection semiology, retention scores, absence of epilepsy risk factors. Their score wasn’t validated in an independent cohort. But there are tools out there. Are those tools perfect? No. they’re a starting point.
We’re trying to develop a score that would hopefully be a little bit easier to implement operationally speaking, because what does it mean, “cognitive changes” and “behavioral changes” in an epilepsy patient? That can be kind of hard to define, and different epileptologists might have different interpretations. So there are ways we can improve on the ACES score. We’re doing a multicenter study looking at patients with not only focal epilepsy of unknown cause but also disease control cohorts and really systematically phenotyping them and testing them for antibodies and whether they’re antibody positive in serum, testing CSF as well, and in both cohorts doing serum and CSF in hopes of defining the phenotype better.
I think a core issue in some of the prior studies is that the basic assumption is that positive antibody equals antibody-associated epilepsy, and no antibody equals no antibody-associated epilepsy. But if you examine these statements there are potential issues, right, so. The way that “positive antibody” is defined actually is different between reputable places. Some places will count a positive anti-glycine as a definite positive. Other places will say well, that assay is problematic and there can be false positives. Different places will count low-titer CASPR-2 differently.
The other aspect of the assumption, that if you don’t have an antibody you definitely don’t have autoimmune associated epilepsy, also is potentially problematic. There may be antibodies out there that we haven’t defined. And also maybe not every autoimmune associated epilepsy is antibody mediated.
So we all have these patients that we could have sworn that their panels would come back positive. They walked and talked and smelled autoimmune, you know? They had late-onset bitemporal epilepsy that just exploded, they had an abnormal MRI with bitemporal changes, they had very rapid cognitive decline – they had to be autoimmune, and their panel comes back negative.
So what we’re trying to do in our current study is have a more clean control cohort because in all prior studies, they only looked at focal epilepsy of unknown cause and they divided antibody positive versus negative, but it makes you think like, both of those groups may have some noise in them. So by having a disease control group, which means a patient with a clear cause for their epilepsy, like they have a malformation of cortical development or some other clear cause that’s not autoimmune, then you know for sure that when you’re trying to do your analysis and compare your autoimmune to your non-autoimmune group that your non-autoimmune group is definitely non-autoimmune.
Dr. Nabavi-Nouri: Glad you brought this up. I was actually going to ask you about this influx of multiple different scoring systems that clinically apply to different categories, different disease/epilepsy categories, in particular the APE/APE2 score and the ACES score. What would you see as the next step in unifying our approach and better implementing these various scores in clinical practice?
Dr. Steriade: Yeah. So I have no hard data, like I didn’t do a randomized clinical trial to prove that one score is better than the other. This is just my opinion, but we wrote a systematic review, an opinion piece in JAMA Neurology last year, and what we ended up recommending in that paper was that if you think the patient has acute symptomatic seizures from autoimmune encephalitis, then use the APE2 score. Because really the APE2 score uses many elements that are autoimmune encephalitis related. And if you think your patient may have an autoimmune associated epilepsy, then use the ACES score, because that was developed explicitly in an epilepsy group.
What the workup should be – definitely serum and CSF in all suspected autoimmune encephalitis – I think that’s really standard. The question is that if you have a patient with suspected autoimmune-associated epilepsy, does every patient need to be tapped? I tend to tap most. There’s no clear evidence right now that that’s necessary – most of the antibodies you’re going to find are antibodies that have pretty good sensitivity in serum like CASPR-2, GAD-65. I’ve seen some chronic LGI-1s that show up in clinic. The antibodies that have high sensitivity in CSF like NMDA, GABA-B, those don’t tend to show up in an epilepsy clinic, so it may not be the right thing to do to tap everybody. In the JAMA Neurology paper we sort of said “consider CSF” to leave it up to people what they do in the autoimmune epilepsy suspected cases.
Dr. Nabavi-Nouri: One of the questions that lies heavy with me is how long you treat your patients for who have a diagnosis or who you’re convinced have a diagnosis of autoimmune-associated epilepsy.
Dr. Steriade: I think the issue of how long to treat is an important one and a difficult one. We don’t have great guidelines. There is now a consensus publication on long-term management of autoimmune encephalitis, that was published in JNNP (the Journal of Neurology, Neurosurgery & Psychiatry) this year or last. And even then, it’s not – different experts will treat for different lengths of time after an autoimmune encephalitis to prevent relapses. And then the autoimmune associated epilepsy groups, it’s even less sort of clear what to do.
I think generally, if someone responds, I keep them on for at least a year before thinking about tapering off. Sometimes I do trials of response, and 6 to 12 weeks is usually a good duration of treatment to see if someone’s steroid responsive or immunotherapy responsive to decide what to do next. I think it’s really important in these patients to have a clear idea in your mind, like before you do an immunotherapy trial – like, what are going to be my parameters in terms of deciding if I was successful or not? Because you know, many patients feel better on steroids, or some patients feel worse after their IVIG.
There needs to be some objective marker of response, whether that’s a seizure diary, cognitive testing,having objective markers of response and in a very clear sort of pre-treatment plan, of what you’re going to consider a success or a failure, can be helpful because it’s kind of easy in these patients to end up in this sort of loop of endless treatments and then realize well, am I making this person better, or am I exposing them to really toxic therapies?
Dr. Nouri: I wanted to conclude by talking about your, personally about your successful career as a young clinician investigator and what advice you would have to our listeners, especially as it pertains to forming networks and securing grants?
Dr. Steriade: I think the most important thing is to choose to do something that you love. Because it’s a lot of hard work, and if there’s a grant deadline coming up and there are late nights, and if you don’t have joy for what you’re writing about, it’s going to feel like very, very hard work. If there’s joy, then it’s not going to feel like work.
So I think like for me, finding something that I felt passionate about and still feel passionate about is really why I think I’ve been successful, is because I still feel excited even when the circumstances are objectively a little bit difficult and the hours are long, you know. Finding something that you love, sticking with people who support you and believe in you. I’ve been so lucky – every institution I’ve been in, I’ve had incredible mentors. And maybe I’m biased because I’m a young woman, but they’ve been women who get it and who have been through the same sort of challenges that let’s face it, sometimes young women are faced with different types of challenges than our male counterparts. I’m not saying our male counterparts don’t have challenges but they’re different.
But when I was in fellowship my mentor there, Lara Jehi, was just incredible – she is the reason I really believed in myself and ended up really pursuing this seriously. She looked at me and she was like, “You are going to end up being an autoimmune epilepsy expert.”
Having somebody who believes in you – it’s true when you’re a kid and it’s true when you’re an adult too – someone needs to believe in you and give you a belief in yourself. And then I came here and I had Jackie French, who’s like my work mom and she believed in me, and she makes me believe in myself. And obviously having mentors who have the intellectual ability to help you, the network ability to help you, has been huge. I wouldn’t be where I am now if it hadn’t been for those two and many more people who were able to read my grants and read my papers and link me with the people who would help me do the work.
Dr. Nouri: Having trained with you from early days, I truly admire what you’ve accomplished and your energy and the knowledge you brought to the field. Because of the love you have for the field you’ve been able to contribute so much to it.
Dr. Steriade: Thanks, Maryam. It’s always so much fun talking about this stuff with you.
This episode was reported by Dr. Maryam Nabavi-Nouri, and edited and produced by Nancy Volkers.
Sharp Waves content is meant for informational purposes only and not as medical or clinical advice.
The International League Against Epilepsy is the world’s preeminent association of health professionals and scientists, working toward a world where no person’s life is limited by epilepsy.
Contact us with feedback or episode ideas at [email protected]