Integrated Host/Microbe (IHM) Metagenomics of the Lower Airway to Diagnose PediatricRespiratory Infections, Identify Etiologic Pathogens, and Predict Outcomes

SUMMARY Lower respiratory tract infection (LRTI) leads to more deaths each year than any other type of infection and disproportionately affects children. The past three years have seen record pediatric hospitalizations due to RSV, influenza, and SARS-CoV-2, highlighting the burden of LRTI in this vulnerable demographic. LRTI remains diagnostically challenging in children due to high rates of viral/bacterial co-infections, noninfectious syndromes that mimic LRTI, frequent incidental pathogen carriage, and the limitations of existing clinical diagnostics. As a result, accurate and timely LRTI diagnosis is difficult to achieve in pediatric critical care, leading to the inappropriate use of empirical antibiotics, the emergence of resistant pathogens, and adverse patient outcomes. Respiratory infections involve a dynamic relationship among three key features: pathogens, the airway microbiome, and the host immune response. However, existing clinical tests rely primarily on pathogen detection, limiting their diagnostic and prognostic utility. Our group has pioneered integrated host/microbe (IHM) metagenomic next generation sequencing (mNGS) methods that enable accurate, culture-independent LRTI diagnosis by simultaneously assessing all three key LRTI features from a single tracheal aspirate sample. Here, we propose a prospective, multicenter cohort study of 400 critically ill children with acute respiratory failure requiring mechanical ventilation that is designed to validate and extend the IHM diagnostic we developed. Aim 1 will independently validate the performance of our existing IHM LRTI diagnostic classifier in distinguishing LRTI from non-infectious acute respiratory conditions and in identifying likely etiologic pathogens. Aim 2 will develop a novel host gene expression classifier specifically for bacterial LRTI rule-out, which would reduce unnecessary antibiotic use in a principled manner. For both Aims 1 and 2, we will additionally develop parsimonious host-based PCR versions of the classifiers for rapid, point-of-care LRTI diagnosis and bacterial LRTI rule-out where mNGS capacity is unavailable. Finally, Aim 3 will develop novel IHM classifiers to predict LRTI outcomes, including prolonged mechanical ventilation and persistent acute respiratory distress syndrome (ARDS), which can facilitate prioritization of resources and intensive interventions to the highest-risk patients. This study promises to address the unmet need for accurate molecular LRTI diagnostics that detect emerging pathogens, enable judicious antimicrobial treatment, and predict outcomes in critically ill children. Our multidisciplinary team of translational scientists with experience in innovative metagenomic approaches is well- positioned to accomplish the study goals. The results of this study will directly inform the design of a future clinical trial evaluating the impact of IHM diagnostics on clinical management and patient outcomes.