Outcomes associated with empiric azithromycin use among patients hospitalized with non-severe community-acquired pneumonia: emulation of a target trial

Abstract

Background: There remains equipoise regarding the benefit of empiric azithromycin with beta-lactam therapy in patients hospitalized with non-severe CAP.

Methods: We performed a target trial emulation using a propensity-weighted cohort from 68 hospitals in Michigan. Adult patients hospitalized with non-severe CAP receiving beta-lactam antimicrobial therapy with or without azithromycin were included. CAP was defined by ICD-10 discharge diagnosis code of pneumonia and >2 signs/symptoms of pneumonia with radiographic findings. Patients with severe CAP, risk factors for multi-drug-resistant organisms, a macrolide allergy, or who received non-standard CAP treatment, doxycycline, an alternative macrolide, or a fluroquinolone were excluded. Time zero was encounter start. The primary outcome was time (days) to clinical stability. Secondary outcomes included composite 30-day mortality and rehospitalization, ICU transfer, and antibiotic duration.

Results: Of 66,657 patients hospitalized for CAP between September 2015 and July 2024, 28.5% (19,010) met inclusion criteria, of whom 93.8% (17,822/19,010) received empiric azithromycin. After IPTW, there was no difference in time to clinical stability between those receiving and not receiving empiric azithromycin (3 [IQR 3-4] vs 3 days [IQR 3-4], aHR 1.00 [0.96-1.05], p=.91). Empiric azithromycin was associated with lower composite 30-day mortality and rehospitalization (10.8% vs. 15.1%, aHR 0.73 [0.62-0.87], p=0.0004). There were no differences in ICU transfer (0.9% vs 1.4%; aHR 0.85 [0.48-1.49], p=.57), or total antibiotic duration (6 [IQR 5-8] vs. 7 [IQR 5-9] days, p=.23).

Conclusion: Adding azithromycin to beta-lactam therapy in patients hospitalized with non-severe CAP was not associated with time to clinical stability but was associated with lower composite 30-day mortality and rehospitalization.