The association of antibiotic treatment regimen and hospital mortality in patients hospitalized with legionella pneumonia

Abstract

Background: Guidelines recommend azithromycin or a quinolone antibiotic for treatment of Legionella pneumonia. No clinical study has compared these strategies. Methods: We performed a retrospective cohort analysis of adults hospitalized in the United States with a diagnosis of Legionella pneumonia in the Premier Perspectives database (1 July 2008-30 June 2013). Our primary outcome was hospital mortality; we additionally evaluated hospital length of stay, development of Clostridium difficile colitis, and total hospital cost.We used propensity-based matching to compare patients treated with azithromycin vs a quinolone. All analyses were repeated on a subgroup of more severely ill patients, defined as requiring intensive care unit admission or mechanical ventilation or having a predicted probability of hospital mortality in the top quartile for all patients. Results: Legionella pneumonia was diagnosed in 3152 adults across 437 hospitals. Quinolones alone were used in 28.8%, azithromycin alone was used in 34.0%, and 1.8% received both. Crude hospital mortality was similar: 6.6% (95% confidence interval [CI], 5.0%-8.2%) for quinolones vs 6.4% (95% CI, 5.0%-7.9%) for azithromycin (P = .87); after propensity matching (n = 813 in each group), mortality remained similar (6.3% [95% CI, 4.6%-7.9%] vs 6.5% [95% CI, 4.8%-8.2%], P = .84 for the whole cohort, and 14.9% [95%CI, 10.0%-19.8%] vs 18.3%[95% CI, 13.0%-23.6%], P = .36 for the more severely ill). There was no difference in hospital length of stay, development of C. difficile, or total hospital cost. Conclusions: Use of azithromycin alone or a quinolone alone for treatment of Legionella pneumonia was associated with similar hospital mortality. Few patients receive combination therapy.