Article Text

Download PDFPDF

Detection bias and the association between inflammatory bowel disease and Salmonella and Campylobacter infection
  1. Mark S Riddle,
  2. Chad K Porter
  1. Department of Enteric Disease, Naval Medical Research Center, Silver Spring, Maryland, USA
  1. Correspondence to Dr Mark S Riddle, Department of Enteric Disease, Naval Medical Research Center, 503 Robert Grant Avenue, Silver Spring, MD 20910, USA; mark.riddle{at}

Statistics from

We read with great interest the paper by Jess et al reporting an increased risk of incident inflammatory bowel disease (IBD) following Campylobacter and Salmonella infections.1 Their results were consistent with prior epidemiological studies of large electronic databases yielding data supportive of an association between the onset of IBD and antecedent acute infectious diarrhoea.2 3 However, Jess et al discounted the potential causal association between infectious gastroenteritis and IBD onset in a susceptible population and concluded that the observed association was a function of detection bias due to increased microbial surveillance among those with IBD-like symptoms. In support of their argument, similarities in temporal IBD risk patterns following Campylobacter- or Salmonella-confirmed infection and for pathogen-negative stool cultures were noted.

Clearly, differential detection subsequent to increased microbial screening could bias effect estimates especially given the overlap in IBD symptoms and those associated with acute enteric infection. However, the conclusions drawn by Jess et al seem inconsistent with the reported results and do not fully account for study design limitations.

The authors assume that the symptoms, clinical presentation and differential diagnosis under which bacterial stool cultures were collected were similar for those with and without culture-positive stool results. Importantly, they do not report on the visit-associated clinical signs or symptoms necessary to inform the context of patient clinical care in which stool cultures were obtained. Failure to include these data precludes an assessment of potential differences which might bias the findings and interpretation. For example, based on the reported results, it is unclear whether stool cultures were performed to identify the causative agent of an episodic acute enteric infection (infrequently done) or to rule out potential alternative etiologies for chronic or recurrent diarrhoea. Importantly, the authors report a more than fourfold higher rate of additional diagnostic tests (eg, colonoscopies, x-rays, MRI scans) after negative stool culture (a diagnostic criteria for IBD) compared to those with a positive stool culture, suggesting that initial cultures were unlikely collected to identify the etiologic agent associated with an episode of acute gastroenteritis. Additional bias related to differential outcome surveillance between stool-culture-positive and -negative study groups may also explain the findings and should be considered in context of the results. The estimate of roughly 1200 IBD cases per 100 000 person years within 1 year of a negative stool culture is approximately 40-fold higher than IBD incidence estimates from large, population-based studies and highlights the points above.

Lastly, epidemiological studies should be interpreted within the context of additional emerging data on IBD pathogenesis, which suggests that IBD involves an inappropriate host response to intestinal microbes. Recent literature provides evidence that Campylobacter infection facilitates translocation of commensal, non-invasive Escherichia coli via mechanisms including disruption of epithelial tight junctions and increased paracellular trafficking of bacteria, a lipid raft-dependent endocytic mechanism, and through utilization of M cells to promote transcytosis of non-invasive bacteria across an intact intestinal epithelial barrier.4 Therefore, it is not unreasonable to conclude that human infection with Campylobacter (and other invasive gastrointestinal pathogens such as Salmonella) could facilitate translocation of luminal antigens, including commensal bacteria, across the intestinal epithelial barrier and, in certain individuals, prime an inappropriate mucosal immune response towards commensal flora, resulting in a loss of tolerance to these ‘self’ antigens.

In summary, we feel that given study design limitations and the potential for diagnostic bias, the results of Jess et al may have multiple explanations beyond those posited by the authors. While difficult to perform, additional studies with unbiased and well-characterised clinical and infectious diarrhoea exposures to evaluate the pathogen-specific risks for IBD are needed.


View Abstract


  • Disclaimer The views expressed in this letter are those of the author and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense or the US Government.

  • Competing interests None.

  • Provenance and peer review Not commissioned; internally peer reviewed.

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.