Clinical Features, Risk Factors, and Outcomes of Pediatric Acute Encephalitis: A Retrospective Cohort Study From 2018 to 2024.

Background

Early diagnosis of infectious versus autoimmune encephalitis in children is critical for optimizing patient care.

Methods

We performed a retrospective cohort study of children one to <18 years of age who had magnetic resonance imaging (MRI) reports with key words "encephalitis" or "meningoencephalitis" at Children's Healthcare of Atlanta from 2018 to 2024 and met clinical diagnostic criteria for acute encephalitis. We extracted data from the medical record and stratified by infectious, autoimmune, or unknown/other etiologies. We performed multivariable logistic regression to form a predictive score to differentiate infectious versus autoimmune encephalitis, using two-thirds of the cohort for training and one-third for validation.

Results

A total of 121 children were included, of which 54 (45%) had infectious, 47 (39%) autoimmune, and 20 (17%) other/unknown etiologies. Compared to infectious encephalitis, children with autoimmune encephalitis were more commonly female (70.2% vs 48.1%, P = 0.02), had an abnormal brain MRI (85.1% vs 68.5%, P = 0.0001), and abnormal neurological exam (70.2% vs 31.5%, P = 0.001). Children with infectious encephalitis presented more commonly with fever (79.6% vs 31.9%, P < 0.0001) and a higher cerebrospinal fluid leukocyte count (median 40 vs 12 cells/μL, P = 0.02). A weighted predictive score utilizing clinical findings (sex, fever, parenchymal enhancement on MRI, and season of admission) distinguished between autoimmune and infectious encephalitis with sensitivity 81.5%, specificity 72.3%, and an internal validation of P = 0.01.

Conclusions

Infectious and autoimmune encephalitis were associated with differences in clinical, laboratory, and imaging features in children and could be distinguished with moderate sensitivity and specificity using a clinical score.