Background Cryptogenic new-onset refractory status epilepticus (cNORSE) currently lacks comprehensive knowledge regarding its clinical dynamics, prognostic factors and treatment guidance. Here we present the longitudinal clinical profiles, predictive factors for outcomes and the optimal duration of immunotherapy in patients with cNORSE.

Methods This retrospective secondary endpoint analysis investigated patients with cNORSE identified from a prospective autoimmune encephalitis cohort at a national referral centre in Korea. The main outcomes included longitudinal functional scales, seizure frequency and the number of antiseizure medications. Measures encompassed NORSE-related clinical parameters such as the duration of unconsciousness, immunotherapy profiles, cytokine/chemokine analysis, and serial MRI scans.

Results A total of 74 patients with cNORSE were finally analysed (mean age: 38.0±18.2; 36 (48.6%) male). All patients received first-line immunotherapy, and 91.9% (68/74) received second-line immunotherapy. A total of 83.8% (62/74) regained consciousness within a median duration of 30 days (14–56), and 50% (31/62) achieved good outcome (mRS ≤2) at 2 years. Poor 1-year outcomes (mRS ≥3) were predicted by the presence of mesial temporal lobe (mTL) and extra-mTL lesions at 3-month MRI, and prolonged unconsciousness (≥60 days). Those with mTL atrophy exhibited a higher seizure burden post-NORSE. The optimal duration of immunotherapy appeared to be between 18 weeks and 1-year post-NORSE onset.

Conclusions This study elucidates longitudinal clinical dynamics, functional outcomes, prognostic factors and immunotherapy response in patients with cNORSE. These findings might contribute to a more standardised understanding and clinical decision-making for cNORSE.