Abstract
Context
Determining the microbial etiology in children with empyema thoracis can guide optimal treatment of individual children, and rational antimicrobial use in institutions. Culture methods and molecular diagnostic technologies are available, but there is no evidence-based guideline addressing the optimal diagnostic technology.Objective
This systematic review was undertaken to compare molecular diagnostic methods with culture methods for determining microbial etiology in children with empyema.Evidence acquisition
Literature search was undertaken in seven databases and four clinical trial registries, to identify diagnostic test accuracy (DTA) studies or diagnostic randomized controlled trials, comparing molecular methods with culture, for establishing the microbiological etiology of empyema. The outcomes were diagnostic accuracy parameters. Risk-of-bias and certainty of evidence were evaluated using the GRADE approach.Results
Nine DTA studies were included, however data from eight could be pooled. The pooled sensitivity of polymerase chain reaction (PCR) in pleural fluid specimens was 0.98 (95% CrI 0.91-0.99), and specificity 0.49 (95% CrI 0.19-0.75). The likelihood ratio of a positive result (LR+) was 1.90 (95% CrI 1.20-3.88), whereas LR- was 0.05 (95% CrI 0.01-0.23). Positive predictive value ranged from 0.11 to 1.0, and negative predictive value from 0.90 to 1.0. The diagnostic odds ratio was 38.23 (95% CrI 5.66-231.81). The overall risk of bias and overall applicability concerns were low in 3/8 studies and unclear in 5/8. The certainty of evidence was very low.Conclusion
Pleural fluid PCR has high sensitivity compared to culture, but moderate specificity. The optimal choice needs considerations beyond the comparative diagnostic accuracy.