Abstract
BACKGROUND: The frequency of acute exacerbations (AECOPD) is a critical predictor of disease progression in chronic obstructive pulmonary disease (COPD). However, the dynamics of the lower respiratory microbiome across a spectrum of exacerbation frequency remain poorly characterized, limiting insights into microbial drivers of susceptibility. METHODS: We conducted a cross-sectional study of 39 hospitalized AECOPD patients, stratified into non-frequent (NFE, ≤ 1 event/year, n = 11), moderate (ME, 2 events/year, n = 13), and frequent exacerbators (FE, ≥3 events/year, n = 15). Metagenomic next-generation sequencing (mNGS) was performed on bronchoalveolar lavage fluid (BALF) to profile the airway microbiome. RESULTS: Microbial alpha diversity exhibited a significant, graded decline from NFE to FE groups (e.g., Shannon index: NFE 3.68 ± 0.34, ME 3.02 ± 1.02, FE 0.84 ± 0.54; p < 0.05). Beta diversity analysis revealed distinct community clustering by exacerbation phenotype (PERMANOVA R(2) = 0.19, p = 0.001). The FE group was characterized by a striking dominance of Pseudomonadota (relative abundance: 72.25%), which correlated positively with exacerbation frequency (r = 0.536, p < 0.001). In contrast, commensal taxa including Streptococcus (r = -0.814, p < 0.0001) and others within the Bacillota and Bacteroidota phyla were depleted in FE and were negatively associated with exacerbation frequency. Twelve exacerbation-resilient taxa (83.3% belonging to Bacillota/Bacteroidota) were positively correlated with FEV(1)% predicted (r = 0.322-0.483, p < 0.05). Alpha diversity indices showed a strong inverse association with exacerbation frequency (r = -0.84 to -0.86, p < 0.001) but not spirometric measures. CONCLUSION: Our findings delineate a gradient of airway microbial dysbiosis along the exacerbation frequency spectrum in COPD. The exacerbation-prone phenotype is defined by a loss of microbial diversity, expansion of Pseudomonadota, and depletion of potentially protective commensals. These microbiome features represent promising biomarkers for identifying high-risk patients and may inform future microbiome-targeted therapeutic strategies.