To report current status of diagnosis and treatment in patients with cryptogenic new-onset refractory status epileptics (C-NORSE) in Japan.

Approximately half of the patients with NORSE is cryptogenic, but it is often difficult to differentiate C-NORSE from secondary NORSE in an emergency situation because it is not easy to get antibody test results quickly. We therefore developed C-NORSE score to identify C-NORSE patients quickly in 2017.

Among 721 patients with suspected autoimmune encephalitis (AE) tested for neuronal surface antibodies between Jan. 2007 and Dec. 2025, 63 (admitted to 34 hospitals) were diagnosed with C-NORSE.

Among the 721 patients, C-NORSE was the second most common disorder (n=63, 9%) after anti-NMDAR encephalitis (n=107, 15%). High C-NORSE score (≥5/6) was seen in 58/61 (95%) and exclusively in C-NORSE patients. 34 (54%) were female. Median age at onset was 29 years (range, 5-74 years). CSF analysis revealed pleocytosis in 67%, but not oligoclonal bands. Brain MRI was initially normal in 31/61 (51%) but abnormal at follow-up in 55/60 (92%); autoimmune-limbic encephalitis (ALE) MRI pattern was less frequently seen than ALE-Plus pattern (8% vs. 42%). First-line immunotherapy was used in 62 (98%) a median of 2 days, while second/third-line immunotherapy in 29 (46%) a median of 20 days. Tocilizumab has increasingly been used since 2022. The last follow-up outcome at median 16.8 months (IQR 6-45) was poor (mRS ≥3) in 40/63 (63%), with a mortality rate of 11% (n=7). Poor outcome was associated with initial brain MRI abnormalities, but not with ALE or ALE-Plus pattern, or the use of second/third-line immunotherapy. 59 (94%) developed drug-resistant post-NORSE epilepsy.

C-NORSE is the second most common disorder in patients with suspected AE. C-NORSE score is useful for identification of C-NORSE patients. Second/third-line immunotherapy didn't improve functional outcome or prevent development of post-NORSE epilepsy, but further study is needed.