Quinolones versus macrolides in the treatment of legionellosis: a systematic review and meta-analysis.
J Antimicrob Chemother. 2014 Sep;69(9):2354-60.
Burdet C1, Lepeule R2, Duval X3, Caseris M4, Rioux C4, Lucet JC5, Yazdanpanah Y6.
1APHP, Bichat Hospital, Infectious Diseases Department, Paris, France APHP, Bichat Hospital, Biostatistics Department, Paris, France INSERM, IAME, UMR 1137 and Université Paris-Diderot, Paris, France firstname.lastname@example.org.
2APHP, Beaujon Hospital, Internal Medicine Department, Clichy, France.
3INSERM, IAME, UMR 1137 and Université Paris-Diderot, Paris, France INSERM CIC 1425, APHP, Bichat Hospital, Paris, France.
4APHP, Bichat Hospital, Infectious Diseases Department, Paris, France.
5INSERM, IAME, UMR 1137 and Université Paris-Diderot, Paris, France APHP, Bichat Hospital, Infection Control Unit, Paris, France.
6APHP, Bichat Hospital, Infectious Diseases Department, Paris, France INSERM, IAME, UMR 1137 and Université Paris-Diderot, Paris, France.
Legionellosis is a life-threatening disease. The clinical superiority of quinolones or macrolides for treating patients with legionellosis has not been established.
We performed a systematic review and meta-analysis of studies reporting data that allowed the comparison of quinolones versus macrolides in the treatment of proven legionellosis published from 1 January 1985 to 31 January 2013. We collected baseline aggregate patient characteristics. Studied outcomes included mortality, clinical cure, time to apyrexia, length of hospital stay and occurrence of complications in each treatment group. Treatment effect was assessed using a Mantel-Haenszel random effects model.
Among 1005 abstracts reviewed, 12 studies were selected (n=879 patients). No randomized controlled trial was performed directly comparing quinolone and macrolide efficacy in legionellosis. Mean age was 58.3 years, 27.7% were women and Fine score was ≥ 4 in 35.8%. Among 253 patients who received quinolone monotherapy, 10 died (4.0%). Among 211 patients who received macrolide monotherapy, 23 died (10.9%). The pooled OR of death for treatment with a quinolone versus a macrolide was 0.5 (95% CI 0.2-1.3, n=8 studies, 464 patients). Length of stay was significantly shorter in the quinolone monotherapy group. The difference was 3.0 days (95% CI 0.7-5.3, P=0.001, n=3 studies, 263 patients). Neither of two tests for heterogeneity was significant (I (2)=0% for both, P=1). Other studied outcomes were not significantly different among treatment groups.
Few clinical data on legionellosis treatment are available. This first meta-analysis showed a trend toward a lower mortality rate and a significant decrease in length of hospital stay among patients receiving quinolones. These results must be confirmed by a randomized controlled trial.