Patterns of initial antibiotic therapy for community-acquired pneumonia in U.S. hospitals, 2000 to 2009.

BACKGROUND: Although clinical guidelines for management of community-acquired pneumonia (CAP) in non-intensive care unit ("non-ICU") hospitalized patients have changed substantially over the last decade, it is unknown how treatment of this disease has evolved over this period.
METHODS: Using data from >100 U.S. hospitals, we identified all adults (aged ≥18 years) hospitalized for CAP between January 1, 2000, and June 30, 2009 ("study period"). We excluded patients admitted to ICU <24 hours of admission, those not starting antibiotics <24 hours of admission, those not receiving antibiotics for ≥48 hours (if alive), and those with probable healthcare-associated pneumonia. We defined "initial therapy" as all parenteral antibiotics received ≤24 hours of admission, and we examined changes in such therapy over the study period. The statistical significance of changes in initial therapy was ascertained using 2-tailed χ tests.
RESULTS: We identified 40,392 patients who met all selection criteria. In 2000, the most frequently used initial regimens were levofloxacin (24.0% of all such admissions), ceftriaxone (9.0%), cefotaxime (7.3%), ceftriaxone plus levofloxacin (3.2%) and azithromycin plus cefotaxime (3.0%); in 2009, they were ceftriaxone plus azithromycin (18.5%), levofloxacin (12.7%), ceftriaxone (6.6%), moxifloxacin (4.7%) and ceftriaxone + levofloxacin (3.2%). Use of single-agent regimens declined between 2000 and 2009 (from 48.2%-30.0%); use of vancomycin almost doubled (13.1%-23.3%). All findings were statistically significant (P < 0.01).
CONCLUSIONS: Initial antibiotic therapy for non-ICU CAP has changed substantially in the United States over the past decade, in line with evidence of widespread antibiotic resistance, evolving treatment guidelines and, most recently, quality improvement initiatives that tie hospital payments to guideline-based care.