Abstract Details

Title A simple bioinformatics approach to disentangle the etiology and prognosis of CNS infections

Topic Infectious Disease

Presentation(s) P4 - Poster Session 4 (5:30 PM-6:30 PM)

Poster/Presentation
Number 4-024

Objective To describe a bioinformatics approach to disentangle the etiology and prognosis of Central Nervous System (CNS) infections.

Background CNS infections cause substantial morbidity and mortality. Rapid etiological determination allows optimization of treatment. We used multiple correspondence analysis (MCA) to determine etiology and prognosis in a cohort of patients with CNS infections.

Design/Methods Adult patients with CNS infections were prospectively enrolled from 5 hospitals in Singapore, between August 2013 and December 2016. Demographics, neuroimaging, neurophysiology and biochemical results were collected. The modified Rankin Scale (mRS) was recorded at enrolment and 2 weeks; 0-2 was classified as good, 3-5 as poor outcome. Etiologies were classified as bacterial, viral, tuberculosis, fungal or autoimmune based on laboratory results. Independent variables were grouped into a smaller number of uncorrelated dimensions (factorial axes) that describe the spread of points. All independent variables used in univariable analysis were used to construct the MCA factorial axes. Interpretation of results was based on the distance between data points and their position along the dimensions. Points similar with respect to the independent variables are closer to each other.

Results Out of 199 patients, 110 (55.2%) had infection, 10 (5.0%) had autoimmune etiology, and 79 (39.7%) were unknown. Age over 65 years, altered mental status, facial focal neurological signs, muscle weakness, neck stiffness, abnormal movements, abnormal MRI and mRS score 3-5 at enrolment were significantly associated with poor outcome at two weeks. Laboratory measurements such as blood and CSF white cell count and differential counts, CSF protein levels and HIV status were indicative of different etiologies. MCA simplified complex datasets by reducing the number of variables to uncover patterns, and represented the results graphically.

Conclusions Using MCA, initial clinical presentation was associated with prognosis, while laboratory parameters informed the etiology. We identified the usefulness of bioinformatics to analyze the wealth of clinical and laboratory data.

Authors/Disclosures
Kevin Tan, MD, FAAN (National Neuroscience Institute) PRESENTER	Dr. Tan has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Eisai. Dr. Tan has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Roche. Dr. Tan has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Merck.
	No disclosure on file
	No disclosure on file
	No disclosure on file
Derek T. Soon, MBBS, PhD (NUHS)	Dr. Soon has received personal compensation in the range of $500-$4,999 for serving as an Expert Witness for Legal Clinic.
	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file
Angela Vincent, MBBS, MSc (John Radcliffe Hospital)	Angela Vincent, MBBS, MSc has received personal compensation in the range of $0-$499 for serving on a Speakers Bureau for Lundbeck Foundation. Angela Vincent, MBBS, MSc has received intellectual property interests from a discovery or technology relating to health care.
	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file

��ɫ��

��ɫ��