Risk Stratification of Dengue Cases Requiring Hospitalization.

Area of Science:

• Immunology
• Computational Biology
• Infectious Disease Epidemiology

Background:

• Dengue pathogenesis is characterized by immune-driven inflammation, often leading to severe disease.
• Accurate early identification of severe dengue cases is crucial for timely intervention and resource allocation.

Purpose of the Study:

• To develop a machine learning model for predicting dengue severity using a minimal set of inflammatory biomarkers.
• To support clinical decision-making and patient triage in dengue-endemic regions.

Main Methods:

• Quantified 48 inflammatory mediators from plasma samples of dengue patients (DF and DWS/SD) at admission.
• Applied a random forest approach to identify predictive biomarkers for disease severity.
• Correlated biomarker levels with clinical parameters like lymphocyte and platelet counts, and liver enzymes.

Main Results:

• 43 of 48 immune mediators were differentially expressed in dengue patients versus controls; 26 differed between DF and DWS/SD.
• Key severity-associated markers included Hepatocyte Growth Factor (HGF), Tumor Necrosis Factor-beta (TNF-beta), Macrophage Inflammatory Protein-1 beta (MIP-1-beta), and Stem Cell Growth Factor beta (SCGF-beta).
• A model using these markers and fever duration achieved nearly 80% accuracy in distinguishing DWS/SD from DF.

Conclusions:

• Targeted cytokine profiling shows promise for early identification of severe dengue.
• This approach can aid in clinical decision-making for hospitalization and potentially alleviate healthcare burdens in endemic areas.