Identification and Treatment of Autoimmune Epilepsy

Among the one-third of adults with epilepsy of unknown etiology, an autoimmune cause appears to be the potential culprit in at least 20% of cases, according to new research published by Divyanshu Dubey, MD, and colleagues in JAMA Neurology.¹ To aid clinicians in identifying individuals with autoimmune epilepsy, which does not respond to standard antiepileptic therapy, the study investigators devised an antibody prevalence in epilepsy (APE) score useful for predicting positive serologic findings. Moreover, the investigators showed that seropositive patients exhibited favorable responses to immunomodulatory therapy.

“Traditionally, a lot of people would question whether autoimmune epilepsy really existed,” commented Sean J. Pittock, MD, the director of the Center for Multiple Sclerosis and Autoimmune Neurology and of the Neuroimmunology Laboratory at the Mayo Clinic in Minnesota. “It’s not up for discussion anymore. It’s now a recognized and accepted fact that a proportion of patients with new-onset epilepsy can have an autoimmune etiology.”

Autoimmune neurology is a relatively new but burgeoning field. Since the initial recognition that antibodies specific for neural antigens can lead to a host of central nervous system (CNS) disorders, it has become evident that CNS autoimmunity can cause focal seizures in isolation or a syndrome where seizures take center stage.

Patients with autoimmune-induced seizures tend to have recalcitrant disease that does not respond to antiepileptic drugs. Accurately diagnosing these individuals not only helps avoid unnecessary therapy, but also offers the opportunity to treat patients with immunomodulatory therapies that have been shown to offer benefit, such as corticosteroids and intravenous immune globulin.

The primary aim of the Dubey et al. study was to determine the prevalence of neurologic autoantibodies among adult patients with new-onset or established epilepsy of unknown etiology. Of 112 consecutive patients who fit these criteria, 35% had serum antibodies suggesting that autoimmunity may be to blame for the seizures. However, because some of the antibodies tested can be found in association with other neurological conditions (eg, Alzheimer’s disease), the investigators narrowed the antibody search to include only those targeting the N-methyl-D-aspartate receptor (NMDAR), the voltage-gated potassium channel complex (VGCc), leucine-rich glioma-inactivated protein 1 (LGI1), high-titer glutamic acid decarboxylase 65 (GAD65), and antineuronal nuclear antibody type 1 (ANNA-1 or anti-Hu). Presence of these antibodies, all of which are highly suggestive of an autoimmune cause of epilepsy, was evident in 20.5% of patients.

But there are potentially more important findings from the Dubey et al. study that went well beyond determining the prevalence of neurological autoantibodies. The investigators sought to identify clues in the patient history that might signal the presence of immune-mediated epilepsy. New-onset seizures, neuropsychiatric changes, autonomic dysfunction, a viral prodrome, a brain MRI indicative of limbic encephalitis, and faciobrachial dystonic seizures were all significantly more common in antibody-positive cases than in antibody-negative cases.

“There are published papers going back to 2008 where people have suggested there are clinical features associated with autoimmune epilepsy,” commented Sarosh R. Irani, MD, an Honorary Consultant Neurologist and Senior Clinical Fellow at the University of Oxford. For example, Dr. Irani and others have found that patients with faciobrachial dystonic seizures always have anti-LGI1 antibodies, suggesting a strong one-to-one relationship.² “The Dubey group has now proven some of these clinical features on a prospective level,” Dr. Irani noted.

As part of their study, Dubey et al. created a 9-item, 15-point index—the APE index—that encompasses all of the clinical factors described above along with others, such as cerebrospinal fluid findings indicative of inflammation, seizures refractory to at least 2 antiseizure medications, and the presence of underlying malignancy. The team prospectively assigned a score to each patient in the study to estimate the probability of an autoimmune etiology prior to knowing the antibody test results. They found that applying an APE score threshold of 4 or higher proved to be quite effective at distinguishing those patients with positive serology from those with negative serology (83% vs 19%; P < .001). If using this APE score threshold to select patients for antibody testing in the clinic, the results translate into 82% specificity and 83% sensitivity.

Dr. Irani notes that the APE score largely captures patients with autoimmune encephalitis. “I think it’s really important to work out whether the APE score is rebranding patients who have autoimmune encephalitis or whether autoimmune epilepsy should be given its own designation by name,” he says. “My feeling is that it’s probably somewhere in between. Patients who have limited other features should be designated as having autoimmune epilepsy, because it will enhance their diagnosis and clinicians won’t be waiting for them to be neuropsychiatrically impaired or have an MRI change.”

This latter point is particularly important, because it has significant implications with regard to treatment. Of the 23 patients in the study who were seropositive, 15 (65%) received some form of immunomodulatory therapy. Good seizure outcomes (ie, a >50% reduction in seizure frequency at the first follow-up visit) were strongly tied to the use of immunomodulatory therapy (OR 19.5; 95% CI, 2.2-173.5; P = .008), particularly intravenous methylprednisolone (OR 8.25; 95% CI, 1.2-59.0; P = .04) or plasmapheresis (OR 32.5; 95% CI, 1.6-673.8; P = .02).

“There are many examples in the literature where patients with autoimmune epilepsy preferentially respond to immunotherapy, [but these examples have largely been observed retrospectively]. Assessing this issue prospectively is really important, and that’s where this paper really has a unique position,” says Dr. Irani.

Dr. Pittock offers another take on the findings. “The concept that some of these patients that have medically intractable epilepsy actually have a treatable and potentially reversible condition is very important,” emphasizes Dr. Pittock. “If you diagnose autoimmune epilepsy in these patients early and treat them, you can actually cure them,” he says, drawing from his own clinical experience and research.^3,4

Based on this work, the next step in Dr. Pittock’s mind is to devise an autoimmune-antibody–based score that will predict which patients will respond well to immune therapy. “At the end of the day, a physician is really interested in knowing: Can I make this patient better? That’s the main question.” As it turns out, Dr. Dubey is currently carrying out a fellowship at the Mayo Clinic where he is working on just such a predictive tool.

Published: July 13, 2018