The lung microbiome is increasingly implicated in chronic obstructive pulmonary disease (COPD) pathogenesis. However, its long-term dynamics and interactions with key clinical features—including inhaled corticosteroid (ICS) use, smoking, and lung function—remain poorly defined. We conducted a prospective two-year study of 43 Korean male patients with COPD who provided sputum samples annually (n = 129). Bacterial communities were profiled using 16S rRNA gene sequencing. Associations between microbial composition and clinical characteristics—including inhaled corticosteroid (ICS) use, smoking status, lung function (FEV₁), and recent exacerbations—were evaluated using negative binomial mixed models (NBMMs) with and without time interaction terms, adjusting for potential confounders. At baseline, overall microbial diversity did not significantly differ between patients with COPD and ex-smoker controls without airflow limitation; however, several low-abundance genera showed differential abundance. Cross-sectional NBMMs revealed that ICS use, current smoking, and reduced FEV₁ % predicted were associated with distinct taxonomic profiles. ICS use was associated with reduced relative abundances of Veillonella, Catonella, and Saccharimonas. Persistent smoking was linked to increased abundances of Actinomyces and Bulleidia and decreased Lautropia. Patients with FEV₁ % predicted < 50% exhibited lower Alloprevotella levels. In longitudinal models, ICS use was associated with increasing temporal trends in Megasphaera and Alloprevotella. Persistent smokers showed attenuated changes in Butyrivibrio and Pseudomonas, while those with severe airflow limitation exhibited increased Bacteroides and decreased Atopobium, Gemella, Kingella, and Tannerella over time. Acute exacerbations were not significantly associated with microbial composition at baseline or during follow-up. Clinical factors in COPD are associated with distinct temporal shifts in the airway microbiome of patients with COPD. Longitudinal profiling combined with NBMMs with time-interaction terms revealed subtle microbial shifts with potential clinical implications that were not evident in cross-sectional analyses. These findings underscore the potential utility of temporal microbiome signatures in stratifying COPD patients and guiding future therapeutic strategies.
Moon et al. (Wed,) studied this question.