ABSTRACT To explore the diagnostic efficiency, clinical concordance, and precision treatment value of metagenomic next-generation sequencing (mNGS) for severe pulmonary infections in children in the pediatric intensive care unit (PICU), and to provide evidence for improving microbiological diagnosis and optimizing anti-infective strategies. A retrospective cohort study included 89 children with severe pneumonia in the PICU in 2024. All underwent routine microbiological testing and mNGS of bronchoalveolar lavage fluid (BALF). Detection rates, pathogen composition, co-infection identification, diagnostic concordance, and treatment impact were analyzed. Metagenomic next-generation sequencing demonstrated high diagnostic sensitivity in the PICU setting, achieving a positive detection rate of 90.0% (80/89) and identifying a diverse spectrum of 103 pathogens, including 50.5% viruses, 43.7% bacteria, 38.8% co-infections (vs 11.6%), and 86.3% diagnostic concordance (vs 55.8%, P < 0.01). Among 46 patients included in the therapeutic outcome analysis (22 in the mNGS-guided group), 21 patients in the mNGS-guided group improved. Multivariate logistic regression analysis, adjusting for confounding factors (age, underlying diseases, PaO 2 /FiO 2 ratio, PRISM III score, and preoperative antibiotic use duration), confirmed that mNGS-guided therapy was an independent protective factor for achieving the primary outcome (OR = 5.23, 95% CI: 1.87–14.61, P = 0.002) and secondary outcomes (C-reactive protein reduction ≥50%: OR = 4.89, 95% CI: 1.72–13.93, P = 0.003; oxygenation improvement: OR = 5.67, 95% CI: 1.98–16.21, P = 0.001). Metagenomic next-generation sequencing demonstrated high diagnostic sensitivity in the PICU setting, guiding precision therapy, and improving prognosis. IMPORTANCE It supports metagenomic next-generation sequencing (mNGS) as a supplementary tool for pediatric intensive care unit (PICU) refractory infections, guides anti-infective adjustments, and informs tiered diagnostic pathways for resource-limited settings to optimize cost-effectiveness.
Xu et al. (Mon,) studied this question.