An outbreak of gastroenteritis occurred at the end of May 2011 in Kumamoto, Japan, among persons who attended 1 of 2 parties held in a Japanese restaurant on May 29. We reviewed case records for the 94 persons who attended the parties. A total of 48 persons became ill; 43 of them attended the first party (a total of 86 attended), and 5 attended the second party (a total of 8 attended). The ill participants had not eaten any food in common except for the meals served at the restaurant. The main symptoms of the patients were diarrhea (83%), abdominal pain (69%), fever (44%; mean temperature 37.2°C), and nausea (29%). The mean incubation period was 19 h.

A routine protocol to identify bacteria and viruses (Technical Appendix) was used by our laboratory to examine 54 fecal specimens from 44 party participants and 10 members of the restaurant kitchen staff (7 party participants and all of the kitchen staff were asymptomatic). Atypical E. coli (lactose negative; OUT:HNM) strains harboring the eae gene and E. coli OUT:H18 strains harboring the stx2d and astA (but not eae) genes were isolated from 24 and 3 specimen, respectively; 7 specimens yielded both strains (Table 1). The stx2-positive/eae-negative E. coli strains were found to be serotype O183 (a recently described O serotype) by agglutination testing with O183-specific antiserum (S. Iyoda, M. Ohnishi, unpub. data).

Figure

Figure. . XbaI-digested pulsed-field gel electrophoresis profiles of isolates from fecal specimens collected from patients during an outbreak of human gastroenteritis associated with Escherichia albertii, Japan. Extra bands observed in 2 E....

All atypical E. coli strains showed identical or nearly identical XbaI-digested DNA banding patterns by pulsed-field gel electrophoresis, and the 10 E. coli O183:H18 strains also exhibited identical patterns (Figure). The source of the infection was most likely the meals served in the restaurant, but a bacteriological examination of the meal or of the ingredients used to prepare the meal was not possible because none of the food was preserved for analysis.

The lactose-negative/eae-positive features of the OUT:HNM strains suggested that these strains might be E. albertii. We examined additional biochemical properties of these strains and found that they exhibited the E. albertii–specific features described (4,5). These features include nonmotility, inability to ferment xylose and lactose, and inability to produce β-d-glucuronidase. The E. coli O183:H18 strains demonstrated common phenotypic and biochemical properties of E. coli (7).

To determine whether the E. albertii–like OUT:HNM strains were E. albertii, we randomly selected 6 strains and determined their phylogeny by multilocus sequence analysis as described (5) (Technical Appendix Table). Results indicated that although the E. coli O183:H18 strain analyzed in parallel belongs to E. coli sensu stricto, the E. albertii–like OUT:HNM strains belong to the E. albertii lineage; all 6 strains showed identical sequences (Technical Appendix Figure).

We further examined the intimin subtype by sequencing the eae gene, the chromosome integration site of the locus of enterocyte effacement encoding the eae gene and a set of type III secretion system genes, and the presence and subtype of the cdtB gene as described (5). Results showed that the E. albertii strains had intimin σ, which is rarely identified in enteropathogenic E. coli or enterohemorrhagic E. coli; the locus of enterocyte effacement was integrated into the pheU tRNA gene; and the cdtB gene of the II/III/V subtype group was present. These features are consistent with recently described genetic features of E. albertii (5).

We divided the party participants into 4 groups according to strain isolation patterns and statistically assessed the association of strain isolation patterns with incidence of clinical symptoms (Table 2). The results indicated that persons infected with only E. albertii or persons infected with E. albertii and E. coli O183:H18 had diarrhea and abdominal pain more frequently than did uninfected persons (p<0.05) and that the incidence of asymptomatic carriers was lower among persons infected only with E. albertii.

Nucleotide sequences obtained in this study have been deposited in the DNA Data Bank of Japan/European Molecular Biology Laboratory/GenBank database. Accession numbers and other information on sequence analyses are shown in the Technical Appendix.

In this gastroenteritis outbreak, E. albertii or stx2-positive E. coli O183:H18 was isolated from 24 ill patients; both strains were isolated from 7 patients. Thus, although the responsible meal or food was not identified, it was most likely contaminated with these 2 microorganisms. The contribution or involvement of E. coli O183:H18 in this outbreak is unknown because there were 3 patients from whom only E. coli O183:H18 was isolated and because there were no differences in clinical symptoms between persons infected with E. coli O183:H18 and persons not infected (Tables 1, 2). In contrast, E. albertii was isolated from a larger number of patients, and many fecal specimens yielded only E. albertii (Table 1).

The proportion of persons who had clinical symptoms was also higher for E. albertii–positive party participants than for uninfected persons (Table 2). Therefore, it is plausible that E. albertii was the major causative pathogen of this outbreak. This information indicates that E. albertii can cause gastroenteritis outbreaks among humans (5).

More attention should be given to sporadic cases and outbreak cases caused by this emerging pathogen. It may also be informative to revisit past outbreak cases caused by eae-positive atypical E. coli if pathogens were recorded as being nonmotile, unable to ferment lactose and xylose, and unable to produce β-d-glucuronidase.