Composition and dynamics of the respiratory tract microbiome in intubated patients

Objective: Lower respiratory tract infection (LRTI) is a major contributor to respiratory failure requiring intubation and mechanical ventilation. LRTI during mechanical ventilation contributes to morbidity and mortality in intubated patients. We sought to understand the dynamics of respiratory tract microbiota following intubation and the relationship between microbial community structure and infection. Design: Cohort study with longitudinal sampling. Setting: Medical Intensive Care Unit at the Hospital of the University of Pennsylvania. Subjects: 15 subjects with respiratory failure requiring intubation and mechanical ventilation. Interventions: We sampled oropharyngeal (OP) and deep endotracheal (ET) secretions within 24 hours of intubation and every 48-72 hours thereafter. Bacterial community profiling was carried out by purifying DNA, PCR amplification of 16S rRNA gene sequences, deep sequencing, and bioinformatic community analysis. Measurements and Main Results: Healthy subjects typically have diverse upper respiratory tract (URT) and lower respiratory tract (LRT) microbiota, where the LRT community is present in lower abundance and probably derived from the URT by microaspiration. In contrast, critically ill subjects had lower initial diversity at both sites. Diversity further diminished over time on the ventilator. In several subjects, the bacterial community was dominated by a single taxon over multiple time points. The clinical diagnosis of LRTI ascertained by chart review correlated with low community diversity and dominance of a single taxon. Dominant taxa matched clinical bacterial cultures where cultures were obtained and positive. In several cases, dominant taxa included bacteria not detected by culture, including Ureaplasma and Enterococcus. Conclusions: Longitudinal analysis of respiratory tract microbiota in critically ill patients provides insight into the pathogenesis and diagnosis of LRTI. 16S rRNA gene sequencing of endotracheal aspirate samples holds promise for expanded pathogen identification.