To assess the association between EEG findings and final clinical outcomes in patients diagnosed with encephalitis 

Encephalitis is associated with high rates of mortality and morbidity; causing death in 5.6-20% of patients admitted to the hospital. Forty percent of patients with encephalitis manifest with seizures (sz). Sz have have an impact on final prognosis, and as such, EEG acquisition has been recommended in all patients with encephalitis. 

Patients from university hospitals affiliated with The University of Texas Health Science Center at Houston with a final diagnosis of encephalitis between 2005-2016 were included. Clinical data was retrospectively collected. Clinical outcomes were defined by the Glasgow Outcome Scale (GOS): GOS ≥13 was deemed favorable, and GOS<13 was unfavorable. Timing, number, and EEG reports were abstracted for findings. Statistical analysis included Chi-square and ANOVA tests. 

Of the 340 patients, 42.9% women, with a mean age 48.6 years old. Breakdown of etiology included autoimmune with positive antibodies in 10% and infectious in 36.4%. 143 patients had EEG performed, 26 were monitored with continuous video EEG (cvEEG). 86% had abnormal EEG in first recording, which was associated with a worse outcome, particularly in presence of slow background (p<0.05). Rhythmic patterns and seizure activity on EEG did not correlate with poor outcome (p>0.05). Sz were only captured in 17 patients and were noted on first EEG in 52%. In 70%, sz were captured during cvEEG.  

We demonstrated that abnormal background activity in the first EEG is associated with poor outcomes. Unlike prior studies, our cohort did not demonstrate worse outcomes with sz, which may be due to low number of patients with an EEG-captured sz. Given that cvEEG was utilized in a limited number of patients, with a majority of seizures reported during cvEEG, we suggest lower threshold for long-term monitoring in patients with suspected encephalitis.