Predictors of Frequent Acute Respiratory Infections in Children With Tracheostomies.

BACKGROUND: Children with tracheostomies experience frequent and recurrent acute respiratory infections (ARIs). While cultured respiratory pathogens can inform ARI diagnosis, it is unknown if their presence in the airway affects future ARI risk. OBJECTIVE: To identify predictors of frequent (3+) ARIs within 36 months of tracheostomy. METHODS: We conducted a single-center, retrospective cohort study of children with tracheostomies placed between 2010 and 2016. Medical records were reviewed for each encounter in which a respiratory culture was obtained over the 3 years post-tracheostomy. ARIs were defined using encounter ICD-9/10 codes. Logistic and Poisson regression were used to model the association between clinical and microbiologic predictor variables with having frequent (3+) ARIs and the total number of ARIs per child. Mediation analysis using stepwise regression models further evaluated the role of Pseudomonas aeruginosa. RESULTS: Among 436 children, 631 diagnosed ARIs occurred within 36 months of tracheostomy; 20.2% of children had 3+ ARIs. P. aeruginosa was isolated in 25% of children and was more common among those with 3+ ARIs compared with 0-2 ARIs (56.8% vs 20.7%, p < 0.001). Those with early P. aeruginosa isolation were more likely to have 3+ ARIs (aOR 3.38, 95% CI 1.97-5.81), and this relationship persisted when analyzing ARIs and P. aeruginosa counts. Identification of P. aeruginosa partially mediated the relationship of ventilator dependence with ARI frequency. CONCLUSION: Isolation of P. aeruginosa, particularly early and repeated isolation, is associated with frequent ARIs in the 3 years after tracheostomy and is an important partial mediator. Findings may inform risk stratification and targeted treatment strategies.