Associations between empirical antimicrobial therapy at the hospital and mortality in patients with severe community-acquired pneumonia

Introduction: The aim of the study was to examine different antibiotic choices and their relation to outcomes. Methods: We reviewed patients with severe community-acquired pneumonia (SCAP) from two multicenter studies. Empirical antimicrobial regimens were classified as: macrolides alone (group M); macrolides plus betalactams (group MB); macrolides plus betalactam/betalactamase inhibitor (group MBI); every regimen including aminoglycosides (group A); non-pseudomonal third-generation cephalosporins alone (group C); another betalactam alone (first- and second-generation cephalosporins, or betalactam/betalactamase inhibitor) (group B); fluoroquinolones (group F); and other regimens (group Misc). Results: Initial distribution of regimens was: group MB: 261 patients; group A: 65 patients; group C: 31 patients; group B: 23 patients; group M: 18 patients; group MBI: 13 patients; group F: 11 patients; group Misc: 38 patients. The lowest overall mortality was associated with initial treatment with a macrolide plus other agent (or alone). No deaths were documented among the 13 patients receiving amoxicillin/clavulanate plus a macrolide. The excess mortality for initial treatment with group A was significantly higher (14.2%; CI 95% 27.3-1.1) than the overall mortality rate between patients receiving a macrolide plus other agents. No significant differences were documented when mortality was adjusted for intubated patients. Conclusion: Clinicians select the empirical antibiotic regimen after classifying patients according to likely pathogens and prognosis. The inclusion of a macrolide as part of the initial therapeutic regi-men for SCAP appears to be as safe and effective as alternative options. Addition of a macrolide agent to a betalactam/betalactamase inhibitor or using a macrolide alone was a marker for less severe disease.