Electroencephalographic outcomes and predictors of epilepsy in autoimmune encephalitis.

Background: An abnormal EEG is pivotal in diagnosis, exclusion of mimickers and prognosticating epilepsy in Autoimmune Encephalitis (AE). However, little is known about the short and long term electroencephalographic outcomes and predictors of epilepsy in AE. This study aims to describe the seizure characteristics and electrophysiological markers of various AE subtypes and assess the clinical and electrophysiological predictors of autoimmune epilepsy.

Methods: Clinical features and EEGs in 74 patients (acute phase=39 and post-acute phase defined after a minimum eight weeks after acute phase=35) of AE fulfilling the proposed criteria were reviewed in their respective acute phases and at six months follow-up.

Results: The mean age of presentation (N = 74, 45 females) was 21.8 (21.8 ± 17.0) years. 38 (51 %) patients were <18 years. Seizures were present in 55 (74 %) patients with poor response to ASMs (p = 0.039). 39 (52.7 %) EEGs were abnormal in acute phase. Anti-NMDAR AE had most frequently abnormal EEG (63.4 %). Poor background reactivity and theta range slowing were most common abnormalities. FIRDA, EDB and delta range slowing were seen in seropositive AE (P = 0.003). Mutism, psychiatric features and incontinence correlated with abnormal EEG (p = 0.013, p = 0.028 and p = 0.025). Background slowing and epileptiform discharges predicted worse cognitive scores at follow-up (p = 0.012). Eight (11.9 %) patients developed epilepsy. Status epilepticus at presentation (p = 0.009), seronegative status (p = 0.0020), delayed initiation of immunotherapy (p = 0.012), abnormal MRI (p = 0.003) and abnormal EEG (p = 0.004) at onset indicate development of autoimmune epilepsy CONCLUSIONS: FIRDA, EDB and delta range slowing with refractory seizures suggest AE. Epileptiform abnormalities, status epilepticus and seronegativity predict autoimmune epilepsy.