ABSTRACT BACKGROUND Differentiating the etiology of acute febrile respiratory illness in children is a challenge in low-income, malaria-endemic settings because the main pathogens responsible (viruses, bacteria, and malaria parasites) overlap in clinical presentation and frequently occur together as mixed infections. The critical task is to rapidly identify bacterial pneumonia to enable appropriate antibiotic treatment, ideally at point of care. Current diagnostic tests are insufficient and there is a need for the discovery and development of new tools. Here we report the identification of a unique biomarker signature that can be identified in blood samples. METHODS Blood samples from 195 pediatric Mozambican patients with clinical pneumonia were analyzed with an aptamer-based high dynamic range assay to quantify ∼1200 proteins. For discovery of new biomarkers, we identified a training set of patient samples in which the underlying etiology of the pneumonia was established as bacterial, viral or malaria. Proteins whose abundances varied significantly between patients with verified etiologies (FDR&lt;0.01) formed the basis for predictive diagnostic models that were created using machine learning techniques (Random Forest, Elastic Net). These models were validated on a dedicated test set of samples. RESULTS 219 proteins had significantly different abundances between bacterial and viral infections, and 151 differed between bacterial infections and a mixed pool of viral and malaria infections. Predictive diagnostic models achieved &gt;90% sensitivity and &gt;80% specificity, regardless of whether one or two pathogen classes were present. Bacterial pneumonia was strongly associated with markers of neutrophil activity, in particular neutrophil degranulation. Degranulation markers included HP, LCN2, LTF, MPO, MMP8, PGLYRP1, RETN, SERPINA1, S100A9, and SLPI. CONCLUSION Blood protein signatures highly associated with neutrophil biology reliably differentiated bacterial pneumonia from other causes. With appropriate technology, these markers could provide the basis for a rapid diagnostic for field-based triage for antibiotic treatment of pediatric pneumonia.

Biomarkers to distinguish bacterial from viral pediatric clinical pneumonia in a malaria endemic setting