Letters

Increased risk of irritable bowel syndrome after bacterial gastroenteritis: cohort study

BMJ 1999; 318 doi: https://doi.org/10.1136/bmj.318.7183.565 (Published 27 February 1999) Cite this as: BMJ 1999;318:565
  1. Luis A García Rodríguez, director, (lagarcia{at}ceife.es),
  2. Ana Ruigómez, epidemiologist.
  1. Centro Español de Investigación Farmaco- epidemiológica, Madrid 28004, Spain
  1. Correspondence to: Dr Rodríguez
  • Accepted 30 October 1998

Evidence exists of an increased risk of irritable bowel syndrome after an episode of bacterial gastroenteritis. 1 2 In one study, 12out of 38patients presented with bowel dysfunction 1year after salmonella gastroenteritis.1 In another study, 386patients with bacterial gastroenteritis were surveyed by questionnaire 6months after infection, and 27(7%) had developed irritable bowel syndrome.2 Both studies, however, lacked a control group.3

Subjects, methods, and results

Our source population came from the General Practice Research Database, which contains clinicalinformation on patients recorded by general practitioners in the United Kingdom.4

We identified patients aged 25to 74with a bacteriologically confirmed first episode of gastroenteritis, from a recent study that examined the association between acid suppressing drugs and the development of gastroenteritis.5 We excluded all patients with a history of irritable bowel syndrome, cancer, or alcoholism. We sampled a comparison cohort from the source population from which patients with gastroenteritis had been ascertained. We applied the same eligibility criteria as were used for the ascertainment of the gastroenteritis cohort, but the subjects had to be free of gastroenteritis at the start of follow up. We then followed up both cohorts during 1year until the earliest occurrence of one of the following end points: diagnosis of irritable bowel syndrome, cancer, alcohol misuse, or death. To confirm diagnoses, we sent a questionnaire to the doctors of all patients with irritable bowel syndrome in the gastroenteritis cohort and a 10% random sample (n=216) to the doctors of patients with irritable bowel syndrome in the general population. We used Poisson regression to calculate incidence, relative risk, and 95% confidence intervals of irritable bowel syndrome associated with gastroenteritis. The general population cohort was used as the reference group, and the estimates were adjusted for age, sex, and body mass index.

Our study included 584308 subjects in the general population cohort and 318patients in the gastroenteritis cohort (table). During the 1year follow up, 2027(0.3%) subjects in the general population cohort and 14(4.4%) in the gastroenteritis cohort had a diagnosis of irritable bowel syndrome recorded on computer. One hundred and sixty nine subjects (78%) from the random sample of the general population cohort, and 12(86%) subjects in the gastroenteritis cohort, had confirmed irritable bowel syndrome on the basis of the questionnaire. We analysed all patients in the general population with a diagnosis of irritable bowel syndrome recorded on computer and all cases confirmed by the doctor in the gastroenteritis cohort. The adjusted relative risk of irritable bowel syndrome was much higher in the gastroenteritis cohort than in the general population cohort (difference 11.9,95% confidence interval 6.7to 21.0).

Characteristics of study cohorts, and incidence and relative risk of irritable bowel syndrome associated with gastroenteritis. Values are numbers (percentages) unless stated otherwise

View this table:

Comment

We found that, in the year after an episode of gastroenteritis, patients were 10times more likely to have irritable bowel syndrome than were subjects in the general population. All incident cases of irritable bowel syndrome in the gastroenteritis cohort were confirmed by the doctor, but we obtained only a random sample of patients with irritable bowel syndrome in the general population owing to the large number of cases (>2000 patients). When we adjusted for this, the incidence of irritable bowel syndrome was reduced to 2.7per 1000per year in the general population, and the crude relative risk associated with gastroenteritis was increased to 14.4.

Only one other study has analysed the risk of irritable bowel syndrome in an unselected group of patients with gastroenteritis.2 The authors estimated a rate of newly diagnosed irritable bowel syndrome of about 7%, similar to our finding (4%). Our study, however, is the first to include a control group and therefore to provide estimates of relative risk.3 We conclude that bacterial gastroenteritis is a major independent risk factor for irritable bowel syndrome.

Acknowledgments

We thank the staff at the Epidemiology and Pharmacology Information Centre, the doctors participating in the general practice research database, and Saga Johansson and Mari-Ann Wallander for providing the initial idea for the study and for commenting on the paper.

Contributors: LAGR designed the study protocol, participated in data analysis, and wrote the paper; he will act as guarantor for the paper. AR helped with data analysis and contributed to the paper.

Funding: None.

Competing interests: Research funds were provided for the original study from which the gastroenteritis cohort was identified.

References

View Abstract