Retrospective application of proposed pediatric autoimmune encephalitis (AE) criteria by Cellucci et al and characterization of high-dose corticosteroid (HD-CS) use in children admitted with encephalitis between 2000-2024.

Prompt determination of etiology remains challenging as viral and immune-mediated presentations may appear similar. AE benefits from early initiation of HD-CS, and while limited evidence exists for corticosteroid use in viral encephalitis diagnostic, diagnostic uncertainty may delay treatment initiation.

Retrospective review of electronic medical records was performed at a tertiary children’s hospital. Patients aged 3 months to 18 years with discharge diagnosis of viral or AE, meningoencephalitis, encephalomyelitis, or acute disseminated encephalomyelitis (ADEM) from 2000-2024 were included. Statistical analysis was conducted using R. Pediatric AE criteria was retrospectively applied as defined by Cellucci et al. Logistic regression analyzed presentation variables to determine likely diagnosis. Associations between HD-CS use, hospital length of stay (LOS) and change in modified Rankin scale (mRS) from admission to discharge were calculated.

Among 68 cases, 29.4% had a discharge diagnosis of ADEM, 19.1% antibody-positive AE, 17.6% antibody-negative AE, 16.2% viral encephalitis, 11.8% PCR-confirmed viral encephalitis, and 5.9% unclear etiology. Antibody-positive AE patients included 90.0% ADEM, and 83.3% antibody-negative patients who received HD-CS, compared to 0% viral, 18.2% presumed viral, and 50.0% of unknown etiology. Upon retrospective Cellucci criteria application, all antibody-negative, 75.0% unknown-etiology and 72.7% presumed viral cases met possible AE criteria at presentation; 66.7% antibody-negative, 54.5% presumed viral, and 50% unknown-etiology patients met probable antibody-negative AE criteria. Time from presentation to HD-CS was associated with LOS (r=0.49, p=0.0015) but not a change in mRS (r=-0.06, p=0.6737).

Although many PCR-negative presumed viral encephalitis patients met Cellucci criteria for probable antibody-negative AE, only 33.3% of those  received HD-CS therapy. Time from presentation to HD-CS was robustly associated with LOS but needs interpretation in light of treatment heterogeneity.