Consideration for ICU Admission in New-onset Encephalitis Patients: A Clinical Decision Model Development and Validation Study Based on Two Multicenter Cohorts
Ralph Habis1, Ashley N Heck2, Paris Bean2, John Probasco1, Romergryko G. Geocadin1, Rodrigo Hasbun2, Arun Venkatesan1
1Johns Hopkins Encephalitis Center, Department of Neurology, Johns Hopkins University School of Medicine, Maryland, USA, 2Department of Medicine, Section of Infectious Disease, McGovern Medical School, UTHealth Science Center, Houston, Texas, USA
Objective:

To Develop a risk score for predicting intensive care unit (ICU) admission in patients with new-onset encephalitis.

 

Background:

Encephalitis patients can experience severe neurological or systemic complications. ICU care, when appropriate and timely, may be associated with better prognosis. Therefore, it is essential to identify encephalitis patients promptly who require ICU for optimal management and allocation of resources, particularly when ICU bed capacity is limited. 

Design/Methods:

This retrospective study was conducted on 634 adult patients diagnosed with encephalitis using the International Encephalitis Consortium guidelines. The study included 17 hospitals in Houston, Texas (cohort-1) and 2 hospitals in Baltimore, Maryland (cohort-2), and included 372 and 262 encephalitis patients, respectively. The study had two parts: 1) description of clinical characteristics, complications, and outcomes of ICU-admitted patients compared to non-ICU-admitted patients, and 2)creation and validation of a risk score to predict ICU admission.

Results:

Of 634 encephalitis patients, 255 (40%) were admitted to ICU ; of these, 19% developed status epilepticus, 42% required vasopressors, and 75% required mechanical ventilation. Patients admitted to ICU had higher mortality (18.4% vs. 2.9%, p<0.001) and worse outcomes. On logistic regression of cohort-1 focal neurological signs, seizure, leukocytosis, FOUR score <14 at presentation, and medical history of CKD, were associated with ICU (p<0.05). Using these variables, a risk model was created with a total of 9 points and an area under receiver operating characteristics curve (AUROC) of 0.77 (95% CI 0.72-0.82), p<0.001. Patients were classified into 3 subgroups for possible consideration of ICU care: low risk [12/87 (14%)], intermediate risk [120 /246 (49%], and high risk [13/14 (93%)]. This model was successfully externally validated in cohort-2 with an AUROC of 0.757 (95% CI 0.687-0.826, p<0.001).

Conclusions:

This risk score may aid physicians as they decide to provide intensity of care  for adult patients presenting with new onset encephalitis.

10.1212/WNL.0000000000206001