Quantitative sputum Candida load and microbiome signatures stratify invasive pulmonary candidiasis severity

Original abstract

UNLABELLED: The clinical significance of Candida species isolation from sputum remains ambiguous due to the colonization-infection dilemma. Integrating quantitative culture with microbiome profiling may clarify its role, yet systematic evidence is scarce. In this exploratory study (June 2023-June 2026), we enrolled 140 hospitalized patients suspected of invasive pulmonary candidiasis (IPC) and stratified them into possible IPC (n = 36), probable IPC (n = 71), and negative control (n = 33) groups based on clinical criteria. All underwent quantitative fungal culture of sputum and metagenomic next-generation sequencing (mNGS). Patients were stratified by sputum Candida species load into low (<105 CFU/mL), high (≥105 CFU/mL), and negative control (0 CFU/mL) groups. We investigated associations between Candida species burden, respiratory microbiota, local inflammation (IL-8), and clinical outcomes. Higher Candida species load was significantly associated with prolonged hospitalization, elevated sputum IL-8, an increased microbial dysbiosis index, and depletion of key commensals (e.g., Neisseria, Prevotella, Rothia, Streptococcus, and Veillonella). Counterintuitively, it correlated negatively with several classic bacterial pathogens (such as Acinetobacter baumannii, Escherichia coli, Klebsiella pneumoniae, Mycobacterium tuberculosis, and Pseudomonas aeruginosa). A random forest model identified a nine-species microbiome signature that effectively discriminated probable from possible IPC (AUC, 0.877; 95% CI, 0.662-0.944). In patients with suspected IPC, higher sputum Candida species load is associated with exacerbated respiratory dysbiosis and airway inflammation. Quantitative Candida species assessment combined with microbiome signatures may serve as a complementary tool for stratifying disease severity and ecological disruption, meriting further validation in larger, multicenter cohorts. IMPORTANCE: Respiratory specimens dominate clinical fungal cultures, yet interpreting the role of Candida spp. in sputum remains a diagnostic dilemma. Here, by integrating quantitative culture with metagenomic sequencing, we show that a high sputum Candida species load is a marker of profound ecological disruption. High Candida species burden was associated with significant depletion of key commensals, an elevated dysbiosis index, and heightened local inflammation. Counterintuitively, it correlated negatively with several classic bacterial pathogens, revealing a distinct dysbiotic signature. Moreover, a microbiome-based model accurately discriminated patients with probable invasive pulmonary candidiasis. Our findings suggest that quantitative sputum Candida species load, interpreted within the context of the microbial community, may be a practical biomarker for identifying patients with significant respiratory dysbiosis and inflammation, offering a path toward more refined, ecology-informed therapeutic strategies.