An integrated biomarker panel for early computational prediction of dengue severity during the acute phase

Serum cytokines and chemokines play a pivotal role in dengue immunopathogenesis and may serve as biomarkers of disease severity. This study aimed to identify applicable blood biomarkers during the acute stage of infection to support early severity evaluation using Principal Component Analysis (PCA)-based feature integration. Ninety-two dengue patients were consecutively enrolled from hospital admissions with laboratory-confirmed dengue during the 2015-2017 epidemic. Only those with NS1-positive or PCR-positive cases were included in this study, spanning asymptomatic, symptomatic, and severe cases, which were assessed for hematological, viral, and cytokine/chemokine parameters. Key variables-platelet count, GPT, creatinine, NS1, NST, IL-18, RANTES, IL-6, IFN-α2, and IL-7-were incorporated into a Prediction Panel. Validation showed that several biomarker combinations achieved up to 90 % accuracy in distinguishing severity groups, while others reached ∼70 %. Principal component analysis further refined the panel, showing accuracy above 90 % when NS1 or NST were included (A-B: 92.1 %, A-C: up to 91.8 %). These findings highlight the potential of an integrated biomarker-based Prediction Panel for early and reliable assessment of dengue severity within the first week of illness.