Diagnostic yield of routine pleural fluid culture and influence of prior antibiotic therapy: A population-based cohort study.

BACKGROUND

Pleural fluid cultures are routinely performed, however, reported diagnostic yields vary widely. We conducted a Danish, retrospective, population-based cohort study to investigate the frequency of positive pleural fluid cultures, and to assess the impact of prior antibiotic exposure. Our particular focus was on patients with suspected or confirmed pulmonary or pleural infection.

METHODS

Suspected infection was defined as a pleural or pulmonary infection considered likely by the treating clinicians prior to pleurocentesis, whereas confirmed infection was based on the final documented discharge diagnosis. Data were retrieved from the regional microbiological database and patients' electronic medical records. Specimens from patients admitted to the intensive care unit were excluded.

RESULTS

Among 840 pleural fluid specimens, 45 were culture-positive (5.4%, 95% CI: 3.8%-6.9%). In patients with suspected pulmonary or pleural infection, 24 of 220 samples were positive (10.9%, 95% CI 6.8%-15.0%), and in those with a confirmed infectious diagnosis, 29 of 206 cultures were positive (14.1%, 95% CI 9.3-18.9%). Multivariable logistic regression showed a significant inverse association between duration of prior antibiotic therapy and pleural fluid culture positivity (p = 0.0037).

CONCLUSIONS

The overall diagnostic yield of pleural fluid cultures in an unselected clinical population was low, but higher among patients with clinically suspected infection. Our findings suggest that pleural fluid cultures should be guided by clinical suspicion rather than performed routinely after pleurocentesis. Whenever possible, pleural fluid should be collected prior to antibiotic initiation to optimise diagnostic yield.