The Effectiveness of Beta-lactam Monotherapy, Beta-lactam and Macrolide Combination Therapy or Fluoroquinolone Monotherapy in Patients Hospitalized with Community-Acquired Pneumonia: a Cluster-Randomized Cross-Over Trial
Background: Optimal empirical treatment for community-acquired pneumonia is unknown. We compared empirical treatment strategies of beta-lactam monotherapy (BL), beta-lactam/macrolide combination (BLM) and fluoroquinolone monotherapy (FQL) in hospitalized community-acquired pneumonia (CAP).
Methods: A cluster-randomized cross-over trial was performed in seven Dutch hospitals. CAP patients initially admitted to a non-ICU ward and treated with antibiotics were eligible, independent of antibiotics chosen. Deviations for medical reasons were allowed and were considered per protocol treatment. The primary endpoint was 90-day mortality, with a non-inferiority margin of 3%. Secondary endpoints were time to oral treatment and length of stay. Intention-to-treat (ITT), per-protocol (PP) and on-treatment (OT) analyses were performed. All analyses were adjusted for cluster-period effects and confounders.
Results: Of 3,325 eligible patients, 656, 739 and 888 patients were included in the BL, BLM and FQL periods, respectively. Protocol adherence was 94%, 88%, and 93%, and 71%, 73% and 80% of patients received the preferred antibiotic class in the respective periods. Baseline characteristics did not differ between study periods. Crude 90-day mortality rates were 9.0% (n=59), 11.1% (n=82), and 8.8% (n=78) in the BL, BLM, and FQL periods, respectively. BL was non-inferior to BLM and FQL for day-90 mortality, except for FQL in OT analysis compared to FQL (Figure 1). Median (IQR) length of stay was 6 (4-8), 6 (4-10) and 6 (4-8) for BL, BLM and FQL, respectively. BLM was associated with a longer length of stay (Figure 2). Median (IQR) duration of IV treatment was 4 (3-5), 4 (3-5) and 3 (0-4) in the respective periods. FQL was associated with a shorter duration of IV treatment (Figure 3), however, this did not result in a shorter length of stay.
Conclusion: A strategy of preferred empirical treatment with BL for patients hospitalized with CAP to non-ICU wards was non-inferior to BLM or FQL strategies for 90-day mortality. BLM was associated with a longer length of stay, and FQL with a shorter duration of IV treatment.
Figure 1: absolute risk differences in 90-day mortality
Figure 2: hazard ratios for length of stay
Figure 3: hazard ratios for time to oral treatment
D. F. Postma,
J. J. Oosterheert, None
M. Bonten, None