Multivariate Risk-Factor Analysis of Short-Term Poor Prognosis in Children With Autoimmune Encephalitis

Abstract

Background

Autoimmune encephalitis in children is a severe immune-mediated disorder with significant neuropsychiatric manifestations and long-term morbidity. Despite advances in immunotherapy, poor short-term prognosis remains a challenge, necessitating identification of prognostic factors for optimized interventions.

Methods

This retrospective study analysed 96 pediatric autoimmune encephalitis patients from a single center. Demographic, clinical, and laboratory data were compared between good (n = 60) and poor (n = 36) prognosis groups. Statistical analyses included univariate tests, multivariate logistic regression, and receiver operating characteristic curve evaluation.

Results

Univariate analysis identified autonomic symptoms, status epilepticus, impaired consciousness, infection, fever, abnormal MRI, and elevated cerebrospinal fluid cell counts as significant predictors. Multivariate analysis confirmed MRI abnormalities (odds ratio [OR] = 4.39), infection (OR = 3.03), autonomic symptoms (OR = 4.09), disorders of consciousness (OR = 4.29), and fever (OR = 3.69) as risk factors. The combined receiver operating characteristic model achieved an area under the curve of 0.816 (sensitivity 72.22%, specificity 83.33%), outperforming individual predictors.

Conclusion

Multivariate analysis identified abnormal MRI findings, infections, fever, autonomic symptoms, and impaired consciousness as independent predictors of poor short-term prognosis in children with autoimmune encephalitis.