Determining the proportion of patients with epilepsy whose seizures are actually influenced by an immune response is an important goal in epilepsy research. As our ability to identify relevant neuroinflammatory processes improves, we are likely to continue finding neuroinflammation as an important aspect of both epileptogenesis and ictogenesis. It is fascinating to consider whether, one day, we might more widely treat epilepsy with some form of immune therapy. This emerging question relates to how the immune system impacts seizure occurrence in what are considered nonautoimmune epilepsies. Regardless of these considerations of the future, the autoimmune epilepsies we already recognize clearly benefit from immunosuppressant treatments. 

The growth in our recognition of autoimmune epilepsy has changed the world of epilepsy treatment. Specific autoantibodies have been characterized that cause conditions that may be considered as forms of epilepsy. Whether these neuroinflammatory illnesses should be regarded as epilepsies or autoimmune encephalitides is sometimes debated, but I consider them to be forms of epilepsy because epilepsy refers to any condition that produces an intrinsic predisposition to seizures. This definition holds true when a cause is not identified or when the cause is a tumor, a stroke, or autoimmune inflammation. To elaborate a little, a person whose seizures relate to a brain tumor has epilepsy; the tumor is certainly relevant to the epilepsy, but it does not eliminate the epilepsy diagnosis even if the treatment may include treatment for the tumor. However, there is still a division among neurologists about how to label the condition when there is immune involvement in addition to how broadly we should test for it.  

Some of these autoimmune epilepsies are associated with autoantibodies that target neuronal surface proteins (eg, LGI1, NMDAR, GABABR, AMPAR). Regardless of the antibody, the autoimmune epilepsies are often initially identified by certain classic presentations. This includes having frequent seizures at the time of developing epilepsy, the lack of initial response to antiseizure medications, and the development of cognitive or psychiatric symptoms within weeks or months of the first seizures. Having a viral prodrome in association with the onset of epilepsy increases the suspicion further. Individuals with this history should be considered for testing for an autoimmune cause because identifying autoantibodies would lead to an important change in treatment. Early immunotherapy is indicated for autoimmune epilepsies and can treat both the seizures and the cognitive or psychiatric problems. However, uncertainty remains about how widely patients with epilepsy should be screened for autoantibodies and how to best treat when abnormality is identified. 

When acute treatment (eg, solumedrol, plasmapheresis, intravenous immunoglobulin) is not fully successful, I find that collaborative care with a rheumatologist is helpful when considering or planning chronic immunosuppression, such as with a B-cell–depleting agent. With chronic treatment, some forms of autoimmune epilepsy respond better to immunotherapy than others, which can lead to questions about the duration of treatment when the response is present but incomplete. The situation becomes challenging when deciding whether the patient’s post-immunotherapy epilepsy or neuropsychological state is due to structural brain changes that are now permanent and/or ongoing inflammation.