Characteristics and Outcomes of Patients with Cryptogenic Encephalitis in the Neurological Intensive Care Unit
Mohamed Ridha1, Jackson Roberts2, Carla Kim2, Elizabeth Carroll2, Jan Claassen2, Kiran Thakur2
1Department of Stroke and Neurocritical Care, The Ohio State University, 2Department of Neurology, Columbia University Irving Medical Center
Objective:

We sought to characterize the clinical features of cryptogenic encephalitis cases in a tertiary center neurologic intensive care unit (NICU and determine which clinical and radiographic features are associated with poor outcome. 

Background:

Encephalitis is associated with significant morbidity and mortality, and severe cases are frequently managed in the NICU setting. A definite etiology may not be obtained during the index hospitalization despite extensive evaluation, presenting a challenge to clinicians when discussing prognosis with families. Clinical and radiographic predictors of outcomes in cases of severe cryptogenic encephalitis have not been well described. 

Design/Methods:

We performed a single, center retrospective study of clinical predictors of poor outcome among patients with cryptogenic encephalitis admitted to a tertiary NICU. We first compared clinical features between cryptogenic encephalitis cases with a poor outcome, as defined by Glasgow Outcome score less than 4, and good outcome in univariate analyses. We then performed multivariate logistic regression with the significant variables from univariate analyses as predictors and poor versus good outcome as the outcome. 

Results:

Among 238 total encephalitis cases admitted to the NICU, 143 were classified as cryptogenic (60.1% of all NICU encephalitis cases). The mean age was 42.5 years, and the most common presenting symptoms were altered mental status (74.8%), fever (58.0%), and gastrointestinal symptoms (46.8%). A total of 48 patients (33.6%) had a poor outcome. In univariate analyses, older age, NICU length of stay, mechanical ventilation, initial Glasgow Coma Scale, active malignancy, and leptomeningeal enhancement were associated with poor outcome. Using multivariate logistic regression, we identified that only older age (beta = -0.026; p = 0.005) and active cancer (beta = -2.56; p = 0.024) remained associated with poor outcome. 

Conclusions:

We find that active malignancy in patients with severe cryptogenic encephalitis is an independent predictor of outcome; the mechanism of this association warrants further exploration. 

10.1212/WNL.0000000000205857