Brazil ICU Encephalitis Surveillance

Our proposal is to develop a sentinel syndromic surveillance strategy to identify encephalitis cases possibly related to emerging pathogens admitted to Intensive Care Units (ICUs) in Brazil. "Sentinel" to allow a diagnostic intensive approach on a smaller number of cases, "syndromic" to guarantee a sensitive criterion to include new or unexpected pathogens, and in ICUs to prioritize potentially severe threats. In a resource-limited setting it won't be possible to monitor and investigate all cases of encephalitis, so a cost-effective algorithm for early identification of the cases that are most likely to be caused by unusual, unexpected or emerging pathogens must be developed. As universal surveillance of encephalitis is not recommended in Brazil, data on incidence, causes and prognosis is not available, leaving a gap in the understanding of the epidemiology of this central nervous system disease in the country. This study will review cases of encephalitis admitted in the last five years to ICUs in a large metropolitan area. Its results will help understand the epidemiology of encephalitis in Brazil and will provide data to build a strategy for early identification of outbreaks and of emerging infectious diseases.

We will use data of a sentinel sample of 19 ICUs of nine of the largest hospitals in the metropolitan area of Rio de Janeiro - Brazil (12 million inhabitants in 2018), from 2015 to 2018, covered with the Epimed Monitor database, which routinely registers all admissions in ICUs. The Epimed Monitor system is present in more than 400 hospitals throughout Brazil, with 750 ICUs and 11,000 monitored beds, reaching over 1,000,000 patients within its base (http://www.utisbrasileiras.com.br/en/epimed/)(Zampieri et al. 2017). Based on previous studies (Boucher et al. 2017; Soares et al. 2011), we expect to identify between 100 and 250 cases of encephalitis.

Inclusion criteria

Identification of suspected case of central nervous system infections in the Epimed Database of non-surgical and non-neonatal ICU admissions: risk greater or equal to 10% - sensitivity of 88.69%, a specificity of 85.57%, area under the Receiver Operating Characteristics (ROC) curve of 0.892 (95% confidence interval 0.864 - 0.921, P<0.0001) - by a diagnostic prediction tool for central nervous system infection (CNSI) developed by the authors for this study. Suspected cases had their medical records reviewed. Those who fulfilled the diagnostic criteria of CNSI as follows were included as a case:

Brain abscess or suppurative intracranial infections: diagnosis of brain abscess requires at least one of the following criteria:

• Organism identified from brain tissue
• Abscess on gross anatomic or histopathologic exam
• Two of the following: headache, dizziness, fever, localizing neurological signs, changing level of consciousness or confusion; and one of the following: organism on microscopic examination of brain or abscess tissue, neuroimaging evidence of infection, or diagnostic serology.

-Encephalitis: Involvement of the brain parenchyma by the infectious agent. Diagnosis of encephalitis by Venkatesan et al19 requires:

• Major criteria: Presence of decreased or altered level of consciousness, lethargy or personality change lasting at least 24 hours and exclusion of encephalopathy due to other non-infectious aetiologies
• Two of the following for possible encephalitis, three of the following for probable or confirmed encephalitis: fever, seizures, new onset focal neurological findings, cerebrospinal fluid (CSF) white blood cell (WBC) count >10x106/L, abnormal neuroimaging and electroencephalography.

-Meningitis: Patients without criteria for encephalitis, and at least one of the following criteria:

• Organism identified from CSF
• The presence of two of the following: fever or headache, meningeal signs or cranial nerve signs (and hypothermia, apnoea, bradycardia or irritability for those under <1 year of age); and one of the following: increased WBC count (≥5 cells/mm3), increased protein and/or decreased glucose in CSF, an organism on Gram stain or culture, or diagnostic serology.

Exclusion criteria

Participants without CNSI case definition, readmissions during the period of data collection (only the first admission was evaluated), evidence of nosocomial or postoperative neuroinfection, diseases of inflammatory or autoimmune etiology, missing important data for analysis or diagnosis.