Performance of cryptogenic new onset refractory status epilepticus score in a Brazilian cohort after testing for antineuronal antibodies with tissue‐based and cell‐based assays

Abstract Objective This study was undertaken to describe a case series of Brazilian new onset refractory status epilepticus (NORSE) patients and evaluate the sensitivity and specificity of a clinical score to predict cryptogenic etiology (c‐NORSE score) after autoimmune encephalitis (AE) testing. Methods Thirty‐seven patients with NORSE from the Brazilian Autoimmune Encephalitis Network were investigated with brain magnetic resonance imaging (MRI), cerebrospinal fluid (CSF) analysis, and complementary testing with tissue‐based assays and cell‐based assays for antineuronal and antiglial antibodies (abs) in serum/CSF. Final diagnoses were compiled after chart review. Patients were allocated into two groups according to c‐NORSE score: (≥5) high score (HS) or (<5) low score (LS), and clinical variables were compared. c‐NORSE score sensitivity and specificity were calculated. Results We found that 23 (62%) of the NORSE patients were children, 49% were female, 22% had AE, and 51% were classified as c‐NORSE. Eleven patients (30%) were allocated to the HS group and 26 (70%) to LS. Bilateral and symmetric MRI findings were more frequent in the HS group (HS 100% vs. LS 43%, p  = .018), whereas memory or behavioral symptoms were less common (HS 27% vs. LS 89%, p  = .001). All HS patients had c‐NORSE, whereas 69% of the LS patients had other diagnoses, such as AE ( n  = 8), herpes simplex virus encephalitis ( n  = 4), paraneoplastic encephalomyelitis ( n  = 1), acute disseminated encephalomyelitis ( n  = 1), and other causes ( n  = 4). The sensitivity of the c‐NORSE score for predicting cryptogenic cases was 57.9% (95% confidence interval [CI] = 36%–80%), and the specificity was 100% (95% CI = 84%–100%). Significance In this cohort, AE was the most identified cause of NORSE, and 51% were cryptogenic. c‐NORSE score ≥ 5 had a specificity of 100% (95% CI = 84%–100%) for identifying cryptogenic cases, whereas a score < 5 indicates additional investigation should be ordered. Although the sensitivity of the c‐NORSE score was lower than previously reported, suggesting it may vary depending on complementary investigation, it is a useful tool for bedside NORSE evaluation in low‐income countries, where access to antineuronal abs is limited.