Background: Our aims were to describe the aetiologies of culture-positive pleural infections and to evaluate the choice of empiric antimicrobial treatment regimens according to antimicrobial sensitivity, and to evaluate the possible influence of this on outcome.
Methods: All cases over a 9-y period were identified from 3 hospitals using the laboratory databases of the clinical microbiology departments, and were verified by evaluating the medical records.
Results: We identified 291 isolates in pleural fluid cultures from 158 patients. These included viridans streptococci (25%), Staphylococcus aureus (18%), anaerobic bacteria (17%), Enterobacteriaceae (12%), Staphylococcus epidermidis (10%), and Streptococcus pneumoniae (7%), with differences between nosocomial and community-acquired infections. The mortality (overall 27%) was highest among the patients with Enterobacteriaceae (50%) and S. aureus (36%) infections, and in patients with mixed infections (34%). The actual empiric treatment or the recommended penicillin plus metronidazole had low antimicrobial coverage (49%) compared to the proposed cefuroxime plus metronidazole (78%). Thoracentesis was often delayed (median 2 days). The adequacy of empiric antimicrobial therapy was independently correlated with mortality (odds ratio 0.43, 95% confidence interval 0.30-0.62).
Conclusions: The early diagnosis of pleural infection could be optimized. In this North-European patient population, we suggest that the recommended empiric antimicrobial treatment be changed to cefuroxime plus metronidazole for community-acquired and nosocomial infections.