Abstract
OBJECTIVES: Early presentation and workup for acute infectious (IE) and autoimmune encephalitis (AE) are similar. This study aims to identify routine laboratory markers at presentation that are associated with IE or AE. METHODS: This was a multi-center retrospective study at three tertiary care hospitals in New York City analyzing demographic and clinical data from patients diagnosed with definitive encephalitis based on a confirmed pathogen and/or autoantibody and established criteria for clinical syndromes. RESULTS: Three hundred and thirty-three individuals with confirmed acute meningoencephalitis were included. An infectious-nonbacterial (NB) pathogen was identified in 151/333 (45.40%), bacterial pathogen in 95/333 (28.50%), and autoantibody in 87/333 (26.10%). NB encephalitis was differentiated from AE by the presence of fever (NB 62.25%, AE 24.10%; p < 0.001), higher CSF white blood cell (WBC) (median 78 cells/μL, 8.00 cells/μL; p < 0.001), higher CSF protein (76.50 mg/dL, 40.90 mg/dL; p < 0.001), lower CSF glucose (58.00 mg/dL, 69.00 mg/dL; p < 0.001), lower serum WBC (7.80 cells/μL, 9.72 cells/μL; p < 0.050), higher erythrocyte sedimentation rate (19.50 mm/HR, 13.00 mm/HR; p < 0.05), higher C-reactive protein (6.40 mg/L, 1.25 mg/L; p = 0.005), and lack of antinuclear antibody titers (>1:40; NB 11.54%, AE 32.73%; p < 0.001). CSF-to-serum WBC ratio was significantly higher in NB compared to AE (NB 11.3, AE 0.99; p < 0.001). From these findings, the association of presenting with fever, CSF WBC ≥50 cells/μL, and CSF protein ≥75 mg/dL was explored in ruling-out AE. When all three criteria are present, an AE was found to be highly unlikely (sensitivity 92%, specificity 75%, negative predictive value 95%, and positive predictive value 64%). INTERPRETATIONS: Specific paraclinical data at initial presentation may risk stratify which patients have an IE versus AE.