Distinct lung microbiota community states are associated with pulmonary nontuberculous mycobacterial disease prognosis

The incidence of nontuberculous mycobacterial pulmonary disease (PNTM) is rising, but the available treatments have limitations. Currently, the understanding of the ecology of the airway microbiota in PNTM is limited, especially regarding community structure, dynamics, and their relationship with clinical outcomes. We used metagenomic sequencing to characterize the lung microbiota in bronchoalveolar lavage fluid (BALF). We evaluated the prognosis of patients with PNTM through respiratory specimen cultures and chest CT scans. PNTM exhibit distinct airway microbiota characteristics compared to controls, however, no significant differences were observed in NTM species. A Dirichlet multinomial mixture model was used to identify two distinct community types (pneumotypes) and investigate their association with host immunity and prognosis. At the 13-month median follow-up, pneumotype 1 (including Mycobacterium, opportunistic pathogens, and anaerobes) presented a lower probability of sustained culture conversion (hazard ratio = 0.29; 95% confidence interval = 0.12–0.73; P = 0.009) than pneumotype 2, indicating a worse prognosis. Based on microbial community abundance and interactions, Ralstonia (NetMoss score = 1.0; log2FoldChange = 3.6) and Dolosigranulum (NetMoss score = 0.6; log2FoldChange = 1.4) emerged as prominent taxa associated with the shift from pneumotype 1 to pneumotype 2, which correlated with differences in clinical outcomes. Our research indicates that distinct baseline microbial characteristics in PNTM patients are associated with prognosis. Furthermore, we identified candidate microbes driving changes in the PNTM microbial community state, serving as potential therapeutic targets.