Biomarkers to support paediatric LRTI antibiotic decision making by clinicians

<bold>Introduction:</bold> Antibiotics are overused in paediatric LRTI. This is due to the limitations of current diagnostics and the clinical overlap of the 2 aetiologies. There is a need for new diagnostics to support judicious antibiotic decision making by clinicians. Biomarkers, identified from proteomics, measured in the easily accesible respiratory tract offer a potential solution <bold>Aim:</bold> Determine the diagnostic performance of four novel biomarkers, haptoglobin (HP), angiotensinogen (AGT), PIGR and Annexin A2 (AX2), measured in the upper respiratory tract (URT) of children (clinical study - UK & Bangladesh) with LRTI compared to reference standard Method Biomarker levels were measured in URT samples (n=119) by ELISA. An expert review panel assigned cases to aetiological groupings (bacterial vs viral) and antibiotic recommended vs antibiotic not recommended groups (reference standard). The diagnostic performance of biomarkers in isolation was determined and logistic regression was used to develop diagnostic models. Statistics performed in R. <bold>Results:</bold> Biomarkers had limited diagnostic performance in isolation (Area under the curve, AUC 0.50-0.59). Combined biomarker models performed better for both aetiological (PIGR & AX2, AUC 0.68) and antibiotic recommendation (HP, AGT, PIGR & AX2; AUC 0.74) grouping. Combined biomarker & clinical diagnostic models demonstrated excellent performance to predict both aetiological (AGT & Coryza, Symptom Length, Temperature; AUC 0.90) & antibiotic recommendation grouping (AGT, PIGT, HAP & Fever, Coryza, Age; AUC 0.97). <bold>Conclusion:</bold> Airway based biomarkers, particularly in conjunction with clinical characteristics, demonstrate the potential to aid antibiotic decisions in paediatric LRTI