Background
Currently, the microbial etiology of community-acquired pneumonia in children remains challenging. While Gram stain and sputum culture are commonly used to detect bacterial pathogens, it is unclear whether these approaches can predict single pathogen from bronchoalveolar lavage fluid (BALF) culture.
Methods
A retrospective study involving 287 children hospitalized for pneumonia was conducted. Sputum specimens were collected on admission; and BALF specimens were collected within 24 h after admission. Taking BALF culture as the reference standard, the sensitivity and specificity of Sputum Gram stain (SGS), sputum culture, and BALF Gram stain (BGS) were calculated. The agreement between these approaches and BALF culture was compared using kappa statistics.
Results
For SGS, the specificity was 23%. The overall sensitivity was 70%, including 87% for Gram-positive (G+) cocci, 56% for Gram-negative (G-) cocci, and 50% for G-bacilli. For sputum culture, the specificity was 70%. The overall sensitivity was 64%, including 71% for Streptococcus pneumoniae, 71% for Moraxella catarrhalis, and 64% for Haemophilus influenzae. For BGS, the specificity was 71%. The overall sensitivity was 60%, including 77% for G+cocci, 38% for G-cocci, and 44% for G-bacilli. While SGS had poor agreement with BALF culture, both sputum culture and BGS had moderate agreement with BALF culture.
Conclusions
Both sputum culture and BGS are helpful in predicting single bacterial pathogen from BALF culture among children with community-acquired pneumonia. Sputum cultures and BGS can provide early clues for BALF pathogen when BALF culture results are pending or bronchoscopy is not performed.