AUTOIMMUNE EPILEPSY: THE UNDERESTIMATED DIAGNOSIS - FROM RED FLAGS TO IMMUNOTHERAPY

Abstract

Autoimmune-associated epilepsy (AAE) is the epilepsy phenotype most likely to be missed in neurology clinics. Neural autoantibodies are identified in a substantial fraction of adults with epilepsy of unknown aetiology, mostly undiagnosed before testing, and in a sizeable proportion of new-onset refractory status epilepticus (NORSE) and febrile infection-related epilepsy syndrome (FIRES). The barrier to recognition is not the absence of effective therapy, but the absence of systematic testing: autoimmune epilepsy is not rare; it is rarely tested. This review argues for a shift from antibody-driven testing ("test when encephalitis is suspected") to syndrome-driven testing ("epilepsy-plus" phenotypes triggering a reflex panel). We synthesise the 2016-2026 literature across three axes. First, an antibody-by-antibody guide (Table 1) contrasts surface-antigen syndromes (LGI1, NMDAR, CASPR2, GABA-B, GABA-A, AMPA) with the intracellular-antigen GAD65 syndrome. Second, a twelve-domain Red Flag Checklist (Table 2) operationalises the "epilepsy-plus" concept, with a ≥2-flag trigger rule mandating reflex testing of paired serum and cerebrospinal fluid. Third, a 2026 immunotherapy algorithm (Table 3) stratifies first-line, second-line, and refractory therapy on the principle that timing is the active ingredient. We address three ongoing debates - GAD65 low- versus high-titre pathogenicity, seronegative AAE, and immunotherapy duration - proposing a risk-stratified, treat-to-remission framework. The cost of delay is measurable in hippocampi, cognitive function, and lives; the barrier is not cost but habit.