ABSTRACT
BACKGROUND: Bacterial cultures from tracheal aspirates (TA) and bronchoalveolar lavage (BAL) specimens can be used to assess patients with artificial airways for lower respiratory tract infections (LRTI). TA collection may be advantageous in situations of limited resources or critical illness. Literature comparing these diagnostic modalities in pediatric populations is scarce. METHODS: Single-center, retrospective analysis of 52 pediatric patients with an artificial airway undergoing evaluation for LRTI. All patients had a TA specimen collected for semiquantitative Gram stain and culture followed by BAL within 48 h. Microbiologic diagnosis of LRTI was defined as a BAL sample with >25% neutrophils and growth of >104 colony-forming units/ml of one or more bacterial species. The test characteristics of TA were compared with these BAL results as the reference standard. Concordance in microorganism identification was also assessed. RESULTS: Overall, 24 patients (47%) met criteria for LRTI using BAL as the diagnostic standard. TA samples positive for an isolated organism had poor sensitivity for acute LRTI when compared with BAL, regardless of semiquantitative white blood cell (WBC) count by Gram stain. Using a TA diagnostic threshold of organism growth and at least "moderate" WBC yielded a specificity of 93%. Positive predictive value was highest when an organism was identified by TA. Negative predictive value was >70% for TA samples with no WBC by semiquantitative analysis, with or without growth of an organism. Complete concordance of cultured species was 58% for all patients, with a higher rate seen among those with endotracheal tubes. CONCLUSIONS: The role of cultures obtained by TA remains limited for the diagnosis of acute LRTI as demonstrated by the poor correlation to BAL results within our cohort. Optimal strategies for diagnosing LRTI across patient populations and airway types remain elusive.
