Assessment of nasopharyngeal Streptococcus pneumoniae colonization does not permit discrimination between Canadian children with viral and bacterial respiratory infection: a matched-cohort cross-sectional study.

BACKGROUND: Readily-available diagnostics do not reliably discriminate between viral and bacterial pediatric uncomplicated pneumonia, both of which are common. Some have suggested that assessment of pneumococcal carriage could be used to identify those children with bacterial pneumonia. The objective of this study was to determine if nasopharyngeal pneumococcal colonization patterns differed between children with definite viral disease, definite bacterial disease, and respiratory disease of indeterminate etiology.
METHODS: Three groups of subjects were recruited: children with critical respiratory illness, previously healthy children with respiratory illness admitted to the ward, and previously healthy children diagnosed in the emergency department with non-severe pneumonia. Subjects were categorized as follows: a) viral infection syndrome (eg. bronchiolitis), b) bacterial infection syndrome (ie. pneumonia complicated by effusion/empyema), or c) 'indeterminate' pneumonia. Subjects' nasopharyngeal swabs underwent quantitative PCR testing for S. pneumoniae. Associations between categorical variables were determined with Fisher's exact, chi-square, or logistic regression, as appropriate. Associations between quantitative genomic load and categorical variables was determined by linear regression.
RESULTS: There were 206 children in Group 1, 122 children in Group 2, and 179 children in Group 3. Only a minority (227/507, 45%) had detectable pneumococcal carriage; in those subjects, there was no association of quantitative genomic load with age, recruitment group, or disease category. In multivariate logistic regression, pneumococcal colonization > 3 log copies/mL was associated with younger age and recruitment group, but not with disease category.
CONCLUSIONS: The nasopharyngeal S. pneumoniae colonization patterns of subjects with definite viral infection were very similar to colonization patterns of those with definite bacterial infection or indeterminate pneumonia. Assessment and quantification of nasopharyngeal pneumococcal colonization does not therefore appear useful to discriminate between acute viral and bacterial respiratory disease; consequently, this diagnostic testing is unlikely to reliably determine which children with indeterminate pneumonia have a bacterial etiology and/or require antibiotic treatment.