Utility of Induced Sputum in Assessing Bacterial Etiology for Community-Acquired Pneumonia in Hospitalized Children.

Abstract

Diagnostic testing for bacterial etiology of community-acquired pneumonia (CAP) is insensitive. Induced sputum (IS) is an attractive option for the evaluation of the lower respiratory tract. Children aged 0-18 years with CAP were enrolled in the Etiology of Pneumonia in the Community (EPIC) study between 2010 and 2012. Blood and respiratory specimens were assessed by culture and polymerase chain reaction (PCR). The radiographic CAP was determined by a study radiologist. Sputum was induced with hypertonic saline. IS specimen was high quality (HQ) if Gram stain showed >25 white blood and <10 epithelial cells per low-powered field; all others were low quality (LQ). We compared IS pathogen prevalence between HQ and LQ IS, and by radiographic pneumonia. Pathogen concordance with EPIC etiology was assessed. Length of stay (LOS) was compared by receipt of IS pathogen-concordant antibiotics. Out of 977 children, 916 (94%) children enrolled in Memphis, Tennessee, produced IS; 794 (87%) had radiographic CAP and 174 (19%) were HQ. HQ IS yielded pathogenic bacteria more often than LQ (64% vs 44%; P < .01); however, pathogens were isolated at similar rates in HQ IS in patients with and without radiographic CAP (64% vs. 63%; P = .6). Pathogens from study specimens matched an IS pathogen in only 9/42 (21%) patients with radiographic CAP. Median LOS was similar among patients with radiographic CAP regardless of receipt of IS pathogen-concordant antibiotics (3.1 days), non-pathogen-concordant antibiotics (2.7 days), or no antibiotics (3.2 days; P = .5). Bacterial pathogens were isolated from most IS cultures regardless of radiographic CAP and quality of IS. IS cultures infrequently corresponded with sterile site cultures. Isolation of pathogens from pediatric IS reflects oropharyngeal carriage and is insufficient to determine bacterial etiology of CAP.