Escherichia coli and community-acquired gastroenteritis, Melbourne, Australia

Emerging Infectious Diseases, Oct, 2004 by Roy M. Robins-Browne, Anne-Marie Bordun, Marija Tauschek, Vicki R. Bennett-Wood, Jacinta Russell, Frances Oppedisano, Nicole A. Lister, Karl A. Bettelheim, Christopher K. Fairley, Martha I. Sinclair, Margaret E. Hellard

As part of a study to determine the effects of water filtration on the incidence of community-acquired gastroenteritis in Melbourne, Australia, we examined fecal samples from patients with gastroenteritis and asymptomatic persons for diarrheagenic strains of Escherichia coli. Atypical strains of enteropathogenic E. coli (EPEC) were the most frequently identified pathogens of all bacterial, viral, and parasitic agents in patients with gastroenteritis. Moreover, atypical EPEC were more common in patients with gastroenteritis (89 [12.8%] of 696) than in asymptomatic persons (11 [2.3%] of 489, p < 0.0001). Twenty-two random isolates of atypical EPEC that were characterized further showed marked heterogeneity in terms of serotype, genetic subtype, and carriage of virulence-associated determinants. Apart from the surface protein, intimin, no virulence determinant or phenotype was uniformly present in atypical EPEC strains. This study shows that atypical EPEC are an important cause of gastroenteritis in Melbourne.

**********

Strains of Escherichia coli that cause diarrhea are classified into pathotypes (or virotypes) according to their specific virulence determinants (1). These virulence determinants give each pathotype the capacity to cause a clinical syndrome with distinctive epidemiologic and pathologic characteristics. For example, enterohemorrhagic E. coli (EHEC) may cause hemorrhagic colitis and the hemolytic uremic syndrome because of their production of Shiga toxins, whereas enteroaggregative E. coli (EAEC) are associated with persistent diarrhea in children in less-developed countries (1). Enteropathogenic E. coli (EPEC) share several key virulence determinants with the most common varieties of EHEC, but lack Shiga toxins, and

cause nonspecific diarrhea in infants in less-developed countries (2,3). EPEC also differ from EHEC in that they typically carry an EPEC adherence factor plasmid (EAF). This plasmid encodes both bundle-forming pili (Bfp) that promote bacterial adherence to mammalian cells and are required for virulence (4) and a transcriptional activator, known as Per, that upregulates genes, such as eae, within a pathogenicity island termed the locus for enterocyte effacement (LEE) (5). LEE is required to produce attaching-effacing lesions, which are characteristic of EPECinduced pathology. A subset of EPEC, known as atypical EPEC, does not carry EAF and hence does not produce Bfp (3). The role of EPEC in disease is uncertain.

The principal reservoir of EHEC is food animals, in particular, cattle, which harbor these bacteria in the distal intestinal tract and from which bacteria can spread to humans through fecally contaminated food or water (1). Although the other pathotypes of diarrheagenic E. coli generally do not originate in animals, they may also spread to humans through food or water contaminated with excrement. Recently, we conducted a study to determine if the water supply of Melbourne, Australia's second largest city with >3 million inhabitants, is a source of intestinal pathogens that are responsible for community-acquired gastroenteritis. Among the pathogens that were sought were diarrheagenic E. coli, including atypical EPEC, which emerged as the predominant cause of gastroenteritis in this community.

Materials and Methods

The design of the Water Quality Study (WQS), which was conducted from September 1997 to February 1999, has been reported previously (6). Briefly, 600 Melbourne families, with at least two children 1-15 years of age, were enrolled in the study. Each family was allocated at random to receive a real or sham water treatment unit, which was installed in the kitchen of their home and supplied water through a separate faucet. Family members, comprising 2,811 persons, were followed for 15 months (68 weeks). Each participating household had a nominated member who completed a weekly questionnaire regarding the presence, duration, and severity of gastrointestinal symptoms. The primary endpoint of the study was highly credible gastroenteritis, which was defined as exhibiting any of the following symptoms in a 24-hour period: two or more loose stools, two or more episodes of vomiting, one loose stool together with abdominal pain or nausea or vomiting, or one episode of vomiting with abdominal pain or nausea. Cases of highly credible gastroenteritis were deemed to be distinct if the participant was symptom-free for at least 6 days.

Sample Collection and Processing

Participants in the study were asked to collect fecal specimens during episodes of gastroenteritis. A total of 795 specimens collected during 2,669 reported episodes of gastroenteritis were examined for rotavirus, adenovirus, Norwalk-like viruses, Giardia spp., and Cryptosporidium spp. and were cultured for Salmonella spp., Shigella spp., Campylobacter spp., Vibrio spp., Yersinia spp., Aeromonas spp., Plesiomonas spp., and Clostridium difficile, as described previously (6,7). Baseline frequencies of these pathogens in the study population were determined during the 4-month period, May through August 1997, immediately preceding the WQS. Frequencies were examined by investigating 1,091 fecal specimens from a convenience sample of participants. Participants who provided a baseline specimen were similar to those who did not provide a specimen in age, sex, and family background.

Examination of Feces for E. coli

Sufficient funds were available to investigate 1,250 samples for diarrheagenic E. coli. Of these samples, 500 were randomly selected from 1,091 fecal samples obtained from healthy persons in the baseline study, and 750 samples were randomly selected from the 795 samples obtained from participants with highly credible gastroenteritis in the WQS.