Relapses, Comorbidities, and Predictors of Outcome in Anti-GABAA Receptor Encephalitis

Objectives: To characterize the magnetic resonance imaging (MRI) lesion dynamics, comorbidities, predictors of relapse, and outcomes in anti-γ-aminobutyric acid type A receptor (GABAAR) encephalitis, and assess the utility of LIM-domain-only-protein 5 (LMO5) antibodies as tumor markers. Methods: GABAAR antibodies were confirmed by 2 techniques in serum or cerebrospinal fluid. Long-term outcomes were defined as good (modified Rankin scale, mRS = 0–1) or poor (mRS 2–5) at ≥12 months. LMO5 antibodies were assessed by cell-based assays and Western blot. Results: Thirty-three patients were identified (4 children, 29 adults; median age, 5.5 and 60 years; 61% male). Ten patients (10/32, 31%) had concurrent systemic autoimmunity. Adults presented with seizures and cognitive/behavioral symptoms, often with thymoma, gastrointestinal, or other tumors (18/33, 55%), whereas children frequently had seizures and ataxia with cerebellar MRI lesions. Multifocal T2/fluid-attenuated inversion recovery hyperintensities were present at onset in 23 of 31 (74%) or developed later in those with absent or single lesions. Lesions showed dynamic changes, suggesting ongoing inflammation even without clinical correlate. Relapses occurred in 17 of 31 (55%, all adults) and were associated with older age (p = 0.02) and lack of second-line immunotherapy (p = 0.02). Four patients (4/33, 12%) died. After a 32.5-month median follow-up, 9 of 20 (45%) had persistent cognitive deficits, and 6 of 20 (30%) had a poor outcome, which was associated with relapses (p = 0.04). LMO5 antibodies were absent in patients and controls. Interpretation: Anti-GABAAR encephalitis shows age-dependent presentations, most commonly seizures. MRI reveals dynamic changes consistent with an ongoing “clinically silent” inflammation. Relapses and cognitive sequelae are common and associate with not receiving second-line immunotherapy. LMO5 antibodies lack tumor-predictive value. ANN NEUROL 2026.