Specific associations between fungi and bacteria in broncho-alveolar aspirates from mechanically ventilated intensive care unit patients.

The detection of fungi in the human respiratory tract may represent contamination, colonization or a respiratory infection. To develop effective management strategies, a more accurate and comprehensive understanding of the lung fungal microbiome is required. Therefore, the objective of the present study was to define the "mycobiome" of mechanically ventilated patients admitted to an intensive care unit (ICU) using broncho-alveolar aspirate ("sputum") samples and correlate this with clinical parameters and the bacterial microbiota. To this end, the mycobiome of 33 sputum samples was analyzed by Internal Transcribed Spacer2 (ITS2) amplicon sequencing of the ribosomal operons. The results show that in the investigated sputa of mechanically ventilated patients Candida spp. were most frequently detected, independent of pneumonia or antimicrobial therapy. The presence of Candida excluded in most cases the presence of Malassezia, which was the second most-frequently encountered fungus. Moreover, a hierarchical clustering of the sequence data indicated a patient-specific mycobiome. Fungi detected by culturing (Candida and Aspergillus) were also detected through ITS2 sequencing, but other yeasts and fungi were only detectable by sequencing. While Candida showed no correlations with identified bacterial groups, the presence of Malassezia and Rhodotorula correlated with oral bacteria associated with periodontal disease. Likewise, Cladosporium correlated with other oral bacteria, whereas Saccharomyces correlated more specifically with dental plaque bacteria and Alternaria with the nasal-throat-resident bacteria Neisseria, Haemophilus and Moraxella. In conclusion, ITS2 sequencing of sputum samples uncovered patient-specific lung mycobiomes, which were only partially detectable by culturing, and which could be correlated to specific nasal-oral-pharyngeal niches.