Time related change of etiologi and outcome in two large cohorts of pleural infection

We wished to describe and analyze etiology and outcome in two comprehensive cohorts of patients with pleural empyema. Methods: The first cohort from 3 Danish hospitals from the period 1996-2007 included 158 patients with known etiology was. The second cohort from 9 (6 additional) hospitals in the period 2008-2011 included 437 patients. In both cohorts, pleural effusion samples were analyzed with routine methods for microbiological etiology. In the second cohort, additional 16S PCR, 23S PCR, Triple-primer 16S RNA PCR, and cloning was done on a minority of samples. Patients were categorized into 6 etiology groups. Outcomes included death, the need for surgical treatment, and length of hospital stay (LOS). Non-parametric statistical analyses were used when appropriate; p Results: In the two cohorts, a known etiology was identified in 158 and in 217 patients, respectively. The six etiology groups in the two cohorts amounted to streptococci (34% vs 48%, p In the two cohorts, the length of hospital stay (29 vs 18 days), proportion of referral for surgery (34% vs 15%), 90 day mortality (27% vs 14%) were all lower in the second cohort, p Conclusion: In the more recent second cohort, fewer patients were in the groups of S. aureus or enterobactericae, and larger patient proportion in the streptococci, or other etiology groups. Patients in the more recent second cohort had lower mortality, lower need for referral to surgery, and lower median LOS.