A real-world retrospective cohort study reveals the clinical utility of metagenomic next-generation sequencing in lower respiratory tract infections

BACKGROUND: Lower respiratory tract infections (LRTIs) are complicated by diverse pathogens, posing challenges to traditional diagnostics. However, robust evidence on LRTI pathogen spectra and metagenomic next-generation sequencing (mNGS) clinical utility remains limited. METHODS: A retrospective analysis was conducted among 815 patients with suspected LRTIs who underwent mNGS and conventional microbiological testing(CMT) of bronchoalveolar lavage fluid. We evaluated the pathogen spectrum, the diagnostic value of mNGS across different infection categories, and its utility in guiding antibiotic therapy. RESULT: Following exclusions, 754 patients demonstrated 84.5% mNGS positivity. mNGS detected DNA viruses (33.85%, EBV predominating), bacteria (30.83%), fungi (23.30%), mycobacteria (9.43%), and special pathogens (2.59%). Confirmed pathogens included Mycobacterium tuberculosis (n = 124), Candida albicans (n = 118), Pseudomonas aeruginosa (n = 87), Pneumocystis jirovecii (n = 65), Haemophilus influenzae (n = 50) and Aspergillus fumigatus (n = 48). mNGS showed higher positivity than CMT (84.5% vs 53.6%, P < 0.05), with sensitivities of 90.6% (LRTIs), 73.3% (bacterial), 74.7% (fungal), and 81.9% (tuberculosis); specificities were 22.6%, 56.5%, 73.5%, and 96.4%. Its high sensitivity but modest specificity necessitates cautious interpretation. mNGS guided treatment adjustments in 48.4% of patients, with higher rates in critically ill patients (60.1% vs 45.7%, P < 0.05), though clinical improvement was lower in this group (54.7% vs 79.2%, P < 0.05). CONCLUSIONS: mNGS comprehensively detects pathogens in LRTIs, including bacteria, fungi, mycobacteria, DNA viruses, and special pathogens. While its broad diagnostic value and treatment guidance utility are significant, integration with clinical context is essential to distinguish true pathogens from colonization.