Epidemiological evidence of legionellosis transmission through domestic hot water supply systems and possibilities of control.

The use of epidemiological markers for a survey is essential because of the ubiquity of legionellae, particularly L. pneumophila serogroup (SG) 1. The two settings under study were an administrative building associated with a fatal Legionnaires' Disease (LD) case due to L. pneumophila SG 1, and a hospital with 11 cases due to L. pneumophila SG 1 and 3 cases due to L. anisa. Monoclonal antibody serotyping allowed us to establish a link between the outbreaks of LD and the contamination of the hot water supply systems. Two subtypes of L. pneumophila SG 1 and L. anisa were detected in the hospital water system. However the finding of only one subtype of L. pneumophila SG 1 in the patients raised the problem of the difference in virulence of each Legionella strain. By means of aerosol tests on normal and cyclophosphamide-immunosuppressed guinea pigs, we demonstrated the potential pathogenicity of hospital water in the immunosuppressed animals. Two decontamination methods were applied: chlorination and a rise in hot water temperature. Time-limited control measures in the hospital were inadequate, resulting in only temporary eradication, followed by a rapid recolonization of legionellae and the appearance of new nosocomial cases. Decontamination of the administrative building was found to be effective when a constant concentration of 5 ppm free chlorine was obtained at tap outlets, and the water temperature was maintained at 55 C.