Detecting legionellosis by unselected culture of respiratory tract secretions and developing links to hospital water strains

Citation
Jr. Kohler et al., Detecting legionellosis by unselected culture of respiratory tract secretions and developing links to hospital water strains, J HOSP INF, 41(4), 1999, pp. 301-311
Citations number
37
Categorie Soggetti
Clinical Immunolgy & Infectious Disease
Journal title
JOURNAL OF HOSPITAL INFECTION
ISSN journal
01956701 → ACNP
Volume
41
Issue
4
Year of publication
1999
Pages
301 - 311
Database
ISI
SICI code
0195-6701(199904)41:4<301:DLBUCO>2.0.ZU;2-O
Abstract
For a 13-month period, all respiratory tract secretions submitted for routi ne bacteriology from a large hospital complex were cultured for legionella, irrespective of clinical diagnosis and laboratory requests. Ten cases of l egionellosis were detected in this manner, three of which met a strict epid emiological definition of hospital-acquired. Therefore, the 16 warm-water s ystems of the hospitals, spread out over two locations, were examined for t he presence of legionella. Legionella pneumophila was found in 15 warm wate r systems, with a distinct pattern of serogroups between the two locations. Legionella of the same serogroups as those isolated from patients were pre sent in each hospital water supply. The isolates were further typed by mono clonal antibodies and by genomic macrorestriction analysis. Similarity betw een clinical and environmental isolates was found in seven cases. In these cases, acquisition from the hospital water supply appears very likely. The strains of the remaining three patients did not match those in hospital wat er, suggesting that community-acquired legionellosis was occurring as well. This study suggests that routinely culturing respiratory tract secretions of pneumonia patients for legionella can help diagnose unsuspected cases of legionellosis. Typing legionella strains beyond the serogroup level with t ools such as macrorestriction analysis is useful to define sources of infec tion, which can then be targeted for control measures.