Article Text

Download PDFPDF

Daily bowel movements and Escherichia coli O157 infection
  1. H Kitajima1,
  2. S Ida2,
  3. M Fujimura1
  1. 1Department of Neonatal Medicine, Osaka Medical Centre and Research Institute for Maternal and Child Health, 840 Murodo-cho, Izumi, Osaka, 594-1101, Japan
  2. 2Department of First Paediatric Medicine, Osaka Medical Centre and Research Institute for Maternal and Child Health
  1. Correspondence to:
    Dr H Kitajima, Department of Neonatal Medicine, Osaka Medical Centre and Research Institute for Maternal and Child Health, Izumi, 594-1101, Osaka, Japan;


The usual bowel habits of 259 school age children before the large outbreak of E coli O157 infection in Sakai City in the summer of 1996 were investigated. A daily morning movement and/or frequent bowel movements may be a protective factor against infection, while being constipated contributed to a worsening of symptoms.

  • E coli O157
  • bloody diarrhoea
  • daily bowel movements
  • ETEC, enterotoxigenic E coli
  • HUS, haemolytic-uraemic syndrome
View Full Text

Statistics from

Since 1982, infections caused by Escherichia coli O157:H7 have emerged as a major public health concern in developed countries.1 The condition is often associated with haemolytic uraemic syndrome and may lead to death. The reasons for such serious complications remain unknown. One report of the 1996 Sakai outbreak of E coli O157 infection involved 8938 schoolchildren; 82% had diarrhoea, 72% abdominal cramps, and 27% bloody stools. E coli O157 was isolated from the stools of 2311 (25.9%) children, and 614 (6.9%) schoolchildren were hospitalised. A total of 107 (1.2%) inpatients were diagnosed with haemolytic-uraemic syndrome (HUS); three girls died.2

A telephone survey of 14 children of friends who lived in Sakai just after the outbreak showed that half had only mild symptoms. These children had two or more bowel movements a day when well. By contrast, three of five children admitted to our hospital with haemolytic-uraemic syndrome had a history of constipation. The most severely affected child, who later died, had not had a bowel movement for three or four days. To test the hypothesis that the bacteria could survive and grow in the intestines of children when retained for longer periods, we examined their daily bowel habits before the onset of infection as an indicator of the time the bacteria could be retained.


In March 1998, we conducted a mailed questionnaire survey of 587 school age children who presented to our centre. The questions were: (1) How many times a day did you usually have bowel movements before the outbreak in summer, 1996? (2) When during the day do you empty your bowels? (3) Do you empty your bowels just after meals? (4) How long does it take you to empty your bowels?

The social class and family size of the subjects were similar, as the two main districts studied were new suburban areas for young to middle aged workers of central Osaka.

Statistical analyses were performed by χ2 tests and logistic regression analyses using the SAS program.


There were 259 responses (44%) from 139 boys, 118 girls, and two of unknown gender. The answers were provided by their mothers. E coli O157 was cultured from the stools of 65 (25.1%) of the children. The isolation rate among the study subjects was similar to that (25.9%) of the total number infected.2 The severity of illness could be classified into three groups: A, bloody diarrhoea and abdominal pain; B, abdominal pain and diarrhoea without bloody stool; and C, no symptoms. Groups A/B/C constituted 89/109/61 of the total population, and 25/24/16 among children with a positive stool culture. The age distribution of the subjects was: 6 years, n = 30; 7 years, n = 56; 8 years, n = 39; 9 years, n = 46; 10 years, n = 38; 11 years, n = 48; unknown, n = 2.

The ages, distribution among schools, and severity of symptoms for the respondents were not significantly different from the non-respondents. No family took special measures to alter bowel habits.

Table 1 shows the daily bowel movements and the severity of illness. However, two severely affected children did not respond. Those with one or more daily bowel movements had less severe symptoms compared with those who had less than one movement a day (p = 0.0334 for the total population; and p = 0.0038 (odds ratio 0.124, 95% CI 0.021 to 0.733, p = 0.0214) for children with a positive culture).

Table 1

Daily bowel movements and the severity of illness

Children reported to empty their bowels at a regular time each day (group A/B/C: 50/77/46) clearly experienced less severe illness than those who had irregular habits (group A/B/C: 39/32/15) (p = 0.0261). A daily morning bowel movement and/or more frequent bowel movements within the day (group A/B/C: 23/51/29) were associated with a reduced severity of illness compared with those who had bowel movements only at other times (group A/B/C: 66/58/32) (p = 0.0041). In conclusion, a daily morning bowel movement and/or frequent movements were shown to be a protective factor against symptoms in E coli O157 infection of school age children.


An investigation of the bowel habits of patients before an E coli O157 infection is unprecedented. As medical records are usually made at the time of presentation with an illness and a history of bowel habits prior to infection is rarely recorded, a study showing the relation between the frequency of movements and severity of illness could not be carried out using standard medical records.

Adhesion of enterotoxigenic E coli (ETEC) to the intestinal mucosal cell surface, followed by production of verotoxins that cause haemorrhagic colitis and HUS, may be an important factor in the pathogenesis.3 More frequent bowel habits would protect not only against intestinal E coli O157 infections, but also other ETEC strains, by reducing the transit time. The risk factors for HUS or thrombocytopenic purpura include old age4 and mental retardation.5 In addition, antimotility agents are associated with the progression of enteritis to HUS6 as a result of the same condition. We believe the hypothesis raised by our study deserves further examination in future studies.


We thank Dr AM Weindling for his helpful suggestions, Dr K Morinaga and Miss S Murakami for help with the statistical analysis, and Miss Y Hige for data collection and technical assistance.


View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.