[Largest outbreak of legionellosis associated with spa baths: comparison of diagnostic tests].

In July 2002, a large outbreak of legionellosis occurred in a bathhouse with spa facilities in Miyazaki Prefecture. Among the visitors, 295, including suspected cases had pneumonia and/or symptoms of fever, coughing, etc. Of these, 37% were hospitalized and 7 died. Clinical samples from 95 mainly inpatients were collected and microbiologically tested at our laboratory. Legionella pneumophila serogroup (SG) 1 was isolated from 3 of 24 in sputum culture, and none of the 3 had been treated effectively with antibiotics at sputa collection. L. pneumophila antigen in urine was detected by using enzyme immunoassay and/or immunochromatographic kits in 23 of 75 patients. Serum antibodies to L. pneumophila SG1 and Legionella dumoffii were detected in 5 each of 66 patients--9 cases including a case at mixed infection-by microplate agglutination test and/or indirect immunofluorescence assay. At our laboratory, 32 were diagnosed with legionellosis. In this outbreak, 14 were diagnosed at other laboratories, resulting in 46 confirmed cases. Urine antigen was detected more frequently by Binax NOW immunochromatographic assay than by Biotest EIA-31% versus 16% of cases tested. Both assays detected urine antigen only in samples collected within 4 weeks after onset. Antigen concentration in urine enhanced sensitivity-58% and 51%-and extended the period of antigen detection beyond 5 weeks. Both antibody titers to L. pneumophila SG1 and L. dumoffii in more than 90% of sera collected within 3 weeks after onset were < 1:16. The rate of serum antibody titer to > or = 1:128 within 3 weeks was 1.6%, during 4 to 6 weeks less than 10%, and after 7 weeks or more 8 to 25%. After an administrative report was published, L. pneumophila DNA in sputa was detected in 5 of 17 patients by nested PCR, resulting in extra 3 cases. Altogether, urinary antigen detection and PCR were more effective in laboratory diagnostic tests than culture and serology. Culture combined with molecular epidemiology is critical, however, for confirming the source of infection.