RESUMO
BACKGROUND: Among children with tracheostomies, little is known about how respiratory culture results differ between states with and without acute respiratory infections (ARI), or the overall test performance of respiratory cultures. OBJECTIVE: To determine the association of respiratory culture organism isolation with diagnosis of ARI in children with tracheostomies, and assess test characteristics of respiratory cultures in the diagnosis of bacterial ARI (bARI). METHODS: This single-center, retrospective cohort study included respiratory cultures of children with tracheostomies obtained between 2010 and 2018. The primary predictor was ARI diagnosis code at the time of culture; the primary outcomes were respiratory culture organism isolation and species identified. Generalized estimating equations were used to assess for association between ARI diagnosis and isolation of any organism while controlling for potential confounders and accounting for within-patient clustering. A multinomial logistic regression equation assessed for association with specific species. Test characteristics were calculated using bARI diagnosis as the reference standard. RESULTS: Among 3578 respiratory cultures from 533 children (median 4 cultures/child, interquartile range (IQR): 1-9), 25.9% were obtained during ARI and 17.2% had ≥1 organism. Children with ARI diagnosis had higher odds of organism identification (adjusted odds ratio 1.29, 95% confidence interval [CI] 1.16-1.44). When controlling for covariates, ARI was associated with isolation of Haemophilus influenzae, Moraxella catarrhalis, Streptococcus pneumoniae, and Streptococcus pyogenes. Test characteristics revealed a 24.3% sensitivity, 85.2% specificity, 36.5% positive predictive value, and 76.3% negative predictive value in screening for bARI. CONCLUSION: The utility of respiratory culture testing to screen for, diagnose, and direct treatment of ARI in children with tracheostomies is limited.
