Infection with Legionella bacteria is one of the major causes of community-acquired pneumonia (1). In Japan, the major sources of Legionella infection are hot springs and public baths (2). Among Legionella species, L. pneumophila serogroup 1 accounts for most human infections (3). Legionellosis outbreaks caused by a combination of L. pneumophila serogroup 1 or other serogroups have rarely been reported. We report an outbreak of Legionnaires’ disease caused by L. pneumophila strains of serogroup 1 and serogroup 13.

During June 1–17, 2015, the Health Centers in Kanagawa and Shizuoka Prefectures, Japan, were notified of 7 cases of Legionella pneumonia. All patients with pneumonia were admitted to 1 of 5 hospitals. All patients were male (mean age 66.3 years), 4 had diabetes mellitus, and 1 had hepatic cirrhosis and liver cancer. Diagnosis of pneumonia at the hospitals was based on clinical presentation and immunochromatographic detection of L. pneumophila serogroup 1 antigen in urine. All patients recovered and were discharged.

In epidemiologic interviews, all 7 patients stated that they had visited a spa house in Odawara, Kanagawa, Japan, before illness onset. The latent period was not accurately determined because 5 of the 7 patients frequently visited this spa house and some patients had visited it again after illness onset. The spa house had 7 circulating systems, including filtration and heating components, and 9 baths for men.

Sputum samples from 5 urinary antigen–positive patients and environmental samples from the spa house were collected for epidemiologic investigations and cultured for Legionella at the Kanagawa Prefectural Institute of Public Health. L. pneumophila serogroup 1 was detected in sputum from 4 patients (Table), and L. pneumophila serogroups 1 and 13 were detected in sputum from 1 patient (patient 2). Baths 1–5, but not baths 6–9, contained L. pneumophila. Epidemiologic investigation and laboratory results revealed that failure to adequately chlorinate the bath water and the circulating systems resulted in colonization of Legionella at the spa house.

Pulsed-field gel electrophoresis (PFGE) comparison (4) of clinical and environmental isolates revealed that the L. pneumophila serogroup1 strains produced 3 PFGE profiles: A and B, with a 1-band difference, and C (Table). Patient 7 was infected with 2 L. pneumophila serogroup 1 isolates possessing PFGE profiles A and B. The PFGE profile B of L. pneumophila serogroup 13 was identical to that of serogroup 1. We determined the sequence type (ST) of L. pneumophila strains (5,6) and identified 4 new STs: ST2113, ST2114, ST2115, and ST2121. ST2114 differed from ST2121 by only 1 nt in neuA and differed from ST2113 (serogroup 13) by only 2 alleles (mip and neuA), suggesting that these 3 STs are closely related and that 1 of 3 strains (ST2113, ST2114, or ST2121) may be derived from another by homologous recombination. All examined isolates lacked the lag-1 gene, a virulence-associated marker (7).

By using whole-genome sequencing, we applied core-genome multilocus sequence typing (cgMLST) with 50 genes (8) different from the 7 genes in sequence-based typing, thereby confirming the sequence-based typing data (Table). ST2114 and ST2115 isolates were each divided into 2 cgMLST profiles. cgMLST profile b differed from profile a by only 1 nt on lpg1503 and differed from profile d by 5 nt on lpg0812, near the lipopolysaccharide coding region, suggesting that the strain of profile d may be derived from another by homologous recombination. However, the remaining cgMLST profiles (c, e, and f) from strains not isolated from patients differed from profile b by 30, 42, and 43 alleles, respectively.

Multiple L. pneumophila strains with different genetic characteristics exist in the environment and pose infection risks (9). Among the strains studied, dual infections with L. pneumophila serogroup 1 and serogroup 13 strains (patient 2) and L. pneumophila serogroup 1 strains with different genomic subtypes (patient 7) were detected. Results from 3 genetic methods revealed that L. pneumophila serogroup 1 and 13 strains are closely related, although the serogroups differ. Results of this study were consistent with the hypothesis that multiple infections are more likely with less virulent strains and more likely in persons with medical conditions predisposing them to Legionnaires’ disease (10).

Our study of this outbreak suggests that the spa house was colonized by several L. pneumophila strains that were genetically related despite belonging to different serogroups and that 2 strains caused infection. Further analysis of the divergence of outbreak strains in genomes related to Legionella serogroup and sequence types is ongoing. This analysis clarifies the in-depth genetic relations among L. pneumophila strains, such as recombination sites and periods required for divergence. We recommend that the spa house provide high quality management and effective infection control practices according to an infection control manual (e.g., completion of documentation relating to infection control practices and training of employees) and that customers be aware of the sanitary status of spa houses.

Top