Antimicrobial Selection and Length of Hospital Stay in Patients with Community-Acquired Pneumonia

• Objective: To evaluate the relationship between antimicrobial selection and length of hospital stay (LOS) in patients with community-acquired pneumonia (CAP). • Design: Retrospective, observational, multicenter study. • Setting and participants: 376 hospitalized inpatients with mild to moderately severe CAP treated in 9 U.S. hospitals between January 1999 and April 2000. • Measurements: Demographic information, data to calculate a pneumonia severity index (PSI) score, type of antimicrobial therapy, LOS, and treatment outcome (successful or unsuccessful therapy or death) were retrieved from patients’ medical records. An analysis of covariance model with PSI score as a continuous variable and with categorical factors for antimicrobial regimen and treatment site was used to compare LOS for various antimicrobial regimens. • Results: Significant predictors of LOS included PSI score (P < 0.001), treatment site (P = 0.015), and initial regimen (P = 0.017). Initial regimens that included atypical pathogen coverage were associated with a mean LOS that was 1.8 days shorter than the LOS associated with regimens that did not provide such coverage (P = 0.008). A subgroup analysis of patients who remained on the same regimen throughout their hospitalization (excluding intravenous-to-oral switch) showed that those who continued to receive a regimen that provided atypical coverage had a mean LOS that was 3.3 days shorter than that of patients who continued to receive a regimen that did not provide atypical coverage (P < 0.001). • Conclusions: This study supports the association between a decreased LOS and initial therapy that provides atypical coverage. This association remains significant for patients who continue their initial regimen for the duration of hospitalization. Empiric therapy that includes coverage for atypical pathogens may be important in reducing the cost of care for CAP. Community-acquired pneumonia (CAP) is associated with significant morbidity and mortality. In the United States, 3.3 million to 4 million people develop CAP annually, and approximately 20% of these are admitted to hospital [1]. Among outpatients, the mortality rate is low, ranging from less than 1% to 5% [2]; however, mortality averages 12% among patients requiring hospital admission and is almost 40% in those admitted to the intensive care unit (ICU) [2]. Although the majority of patients with CAP are treated in the community, the cost of treating those admitted to hospital is approximately 90% of the total cost of care budget for CAP [3], which is estimated to be approximately $9 billion annually [4]. Due to these costs, an emphasis is placed on reducing the length of hospital stay (LOS) of patients while maintaining the quality of care wherever possible. It is often difficult to determine the etiology of CAP due to problems with obtaining valid specimens and a marginal probability of recovering the causative organism(s) from the specimens. Thus, CAP treatment is often empiric, based on knowledge of the most probable organisms and the local hospital microbiology summary data. Generally, the organisms most commonly associated with CAP are Streptococcus pneumoniae, Haemophilus influenzae, Mycoplasma pneumoniae, Chlamydia pneumoniae, and Legionella pneumophila [5–7]. The selection of empiric antibiotic therapy is frequently based on