Empiric antibiotic therapy and mortality among medicare pneumonia inpatients in 10 western states: 1993, 1995, and 1997

Study objectives: To examine the association of empiric inpatient antibiotic treatment of community-acquired pneumonia (CAP) with mortality, and whether this association varies from year to year. Design: Population-based, retrospective study adjusting for demographics, comorbidities, and clinical characteristics. Setting: Acute-care hospitals in 10 western states. Patients: A group of 10,069 Medicare beneficiaries aged ≥ 65 years who were hospitalized with CAP during fiscal years 1993, 1995, and 1997. Measurements and results: We examined the risk for mortality during the 30 days after admission to the hospital. The impact of specific antibiotic regimens varied greatly from year to year. In 1993, therapy with a macrolide plus a β-lactam was associated with significantly lower mortality than therapy with either a β-lactam alone (adjusted odds ratio [AOR], 0.42; 95% confidence interval [CI], 0.25 to 0.69) or other regimens that did not include a macrolide, β-lactam, or fluoroquinolone (AOR, 0.35; 95% CI, 0.20 to 0.62). Those associations were not observed in 1995 or 1997. Lower mortality was associated with fluoroquinolone monotherapy compared with β-lactam monotherapy in 1997 (AOR, 0.27; 95% CI, 0.07 to 0.96) and with macrolide monotherapy compared with other regimens in 1995 (AOR, 0.24; 95% CI, 0.06 to 0.93), but the number of patients who received these regimens was small. Conclusions: The inclusion of a macrolide or a fluoroquinolone in initial empiric CAP treatment was associated with improved survival, but this association varied from year to year, perhaps as a result of a temporal variation in the incidence of atypical pathogen pneumonia. Improved testing and surveillance for atypical pathogen pneumonia are needed to guide empiric therapy.