After the recent reports of an oncology patient's death related to a "dirty shower", I thought that I would talk about Legionella pneumophila.


Isolation of the microbe from sputum or bronchoalveolar lavage fluid on selective media is the most reliable means of diagnosis. Buffered charcoal-yeast extract (BCYE) agars with added antibiotics to suppress commensal flora are available commercially, but these media often have decreased sensitivity for isolation of non-pneumophila strains; cefamandole is especially inhibitory. Legionella species lacking β-lactamase, such as L. micdadei and L. bozemanii, will not grow on BCYE formulations containing cephalosporins. A more sensitive medium consists of BCYE with added vancomycin, anisomycin, and polymyxin B. The non-pneumophila strains are easily missed in clinical and environmental specimens if dye-containing media are not used. Colonies of L. micdadei and L. maceachernii are blue on culture media containing bromocresol purple and bromthymol blue dyes, whereas the colonies of other species are yellow-green to apple green; the dyes color the organism, making detection easier. Direct fluorescent antibody (DFA) stains for the visualization of Legionella species in clinical specimens are commercially available for a limited number of species. The sensitivity and specificity of DFA staining for species other than L. pneumophila is not precisely known. A Legionella DNA probe can detect the presence of multiple Legionella spp. Legionella urinary antigen detects only L. pneumophila serogroup 1; it is not useful for other Legionella species. Detection of Legionella spp. in clinical specimens by DNA amplification is a promising technique that has been applied in a limited number of cases. species, but does not differentiate among species. The DNA probe appears to have fewer false-positive reactions than does DFA staining; it is no longer commercially available. Antibody seroconversion in diagnosing infection caused by non-pneumophila species is of uncertain specificity. Reports of infection based on seroconversion alone should be viewed with skepticism.
There are no randomized trials of therapy for Legionella infection; the majority of reported clinical experience concerns infection with L. pneumophila. In vitro susceptibility data and more limited clinical experience indicate that response to the therapy of infection with other species should be similar. Legionella species are susceptible in vitro to erythromycin, macrolides, tetracyclines, trimethoprim-sulfamethoxazole, rifampacin, and fluoroquinolones. The newer macrolide agents are more active than erythromycin both in vitro and intracellularly against the non-pneumophila species. They offer a number of other clinical advantages over erythromycin, including better penetration into tissue and alveolar macrophages, and improved pharmacokinetics permitting once daily dosing. The fluoroquinolones are considerably more active than erythromycin. Based on these factors, the newer macrolides or quinolones are the therapy of choice for infection caused by Legionella species. Patients who are immunocompromised or who are hospitalized with potentially life-threatening infection should receive intravenous therapy with either a macrolide or a fluoroquinolone. Quinolones are preferable when treating transplant patients receiving cyclosporin, because macrolides interfere with the metabolism of these antirejection agents. Erythromycin has been the historical drug of choice based on the observation of clinical response in the majority of patients. However, there are a number of case reports of erythromycin failure in highly immunocompromised patients. Failure of erythromycin may be due to the fact that erythromycin is bacteriostatic rather than bactericidal against intracellular Legionella. The optimal duration of therapy with these agents is uncertain. Data from clinical trials of community-acquired pneumonia suggest that, in immunocompetent patients, 5 to 10 days of therapy with macrolide therapy or 10 to 14 days of therapy with a fluoroquinolone constitutes adequate therapy. Immunocompromised patients should receive longer courses of therapy (14 to 21 days) in order to prevent relapse. Oral therapy may be used as initial treatment in immunocompetent patients who are not seriously ill. Patients receiving initial parenteral therapy may be switched to oral therapy once a clinical response is apparent.
