Reference Section
Infection, Antibiotic Use, and the Development of Crohnýs Disease
B USINESS BRIEFING: US GASTROENTEROLOGY REVIEW 2005
1
a report by
Michael F Picco, MD, PhD
Assistant Professor of Medicine, Mayo Medical School and Director, Gastroenterology and Hepatology
Fellowship Program, Mayo Clinic
Introduction
Crohnýs disease (CD) is a heterogeneous disorder of
unclear etiology. Genetic predisposition is the most
important risk factor but environmental influences
clearly play a major role.1 This is evident from studies of
identical twins where concordance rates (both twins
have the disease) for CD are only about 50%. Specific
environmental risk factors that have been linked to CD
include cigarette smoking, urban residence, and
socioeconomic status.2 The majority of patients with
CD are diagnosed by the third decade of life, suggesting
that childhood exposures may play a role. Common
exposures that have been linked to the development of
CD are early childhood infections and the use of
antibiotics to treat these infections.


Epidemiology
Several investigators have proposed a relationship
between infection and the development of CD. Higher
rates of CD have been observed in the first half of the
year (winter months) in Sweden,suggesting a role of viral
infection.3 Risk of inflammatory bowel disease (IBD)
was increased among those who had infectious events
during childhood especially in the first six months of
life.4,5 Lower socioeconomic status, possibly due to
overcrowded conditions, was thought to play a role.


These findings are in complete contrast to other
observations, which suggest a lack of childhood
infection predisposes to the development of CD. Some
have proposed that access in childhood to a separate
bathroom and hot water tap, as measures of
socioeconomic status, were associated with a higher
likelihood of CD.6 Others contend that lack of
exposure in childhood to intestinal helminthes results
in later development of immunological diseases such as
CD.7 Bacterial and parasitic interactions with the
developing host immune system may be important in
future reactions to environmental antigens resulting in
later chronic disease.


Obviously, this area is very controversial with widely
disparate opinions. Further fueling this controversy is
the contention that childhood use of antibiotics also
affects CD risk. Early studies suggested that antibiotic
use was higher among CD patients.8,9 However, these
studies were based on prevalent cases and relied on
patient recall. Case ascertainment and recall bias could
have had a role in affecting their conclusions. Also,
antibiotics may be used to treat diarrhea that can be the
first manifestation of CD, thus making it appear that
antibiotics were associated with CD, although the
majority of infections reported in these studies were
upper respiratory.


In a recent study, the impact of antibiotic usage on the
development of CD was determined through the
matching of 587 incident CD cases and 1,460 controls
through a population database from the UK.10 They
adjusted for age, gender, and smoking history as possible
confounding factors and enrolled only incident cases to
minimize recall bias. Diagnoses thought to suggest
possible CD were gastrointestinal infections, perianal
sepsis, anemia, irritable bowel syndrome (IBS),
abdominal pain, or diarrhea.These were more common
in the two years prior to diagnosis, suggesting that they
may have represented early CD, and any antibiotic
usage in this period may have been to treat early disease
manifestations. These events were uncommon and
numbers were similar to the control population in the
two to five years prior to diagnosis.They restricted their
analysis in this two- to five-year period and included
only those patients who had no gastrointestinal
symptoms and had not used gastrointestinal drugs.


Antibiotic exposure here was thought to be less likely
due to gastrointestinal events, reducing the bias that
antibiotics were used to treat early Crohnýs disease
(ýreverse causalityý bias).Their logistic regression analysis
found a modest association of the development of CD
with prior antibiotic usage (odds ratio, 1 in 53)
particularly tetracyclines (odds ratio, 1 in 33). The
authors concluded that there is a significant association
between antibiotic use and the diagnosis of CD two to
five years later.


It is not clear from this study whether the antibiotics
may predispose to CD or if the infections that are
treated by them are the predisposing factor. The
indications for the antibiotics were not discussed other
than if they were for gastrointestinal infections. This
Michael F Picco, MD, PhD, is a
Consultant at the Mayo Clinic and
Assistant Professor of Medicine at
the Mayo Medical School. He is also
the Director of the Gastroenterology
and Hepatology Fellowship Program
at the Mayo Clinic in Jacksonville,
and a member of the Ad Hoc
Committee on Training for the
American College of
Gastroenterology. He is a member
of the American Gastroenterological
Association, American College of
Gastroenterology, and the Crohnýs
and Colitis Foundation of America.


His clinical and research interests
are in inflammatory bowel disease
and chronic diarrhea. He completed
his fellowship in Gastroenterology
and Hepatology at the Johns
Hopkins Medical School and
received his PhD in cinical
investigation at the Johns
Hopkins School of Hygiene
and Public Health.


2
B USINESS BRIEFING: US GASTROENTEROLOGY REVIEW 2005
Reference Section
might have suggested that certain infections may lead to
CD rather than specific antibiotics.A family history of
IBD is not likely to have confounded the study results,
but other factors should be considered. Although
smoking history did not seem to affect their findings, its
measurement in the study was crude, and no provision
was made for ex-smokers. Smoking is clearly associated
with both infections and CD. Higher socioeconomic
status and urban residence have both been linked to CD
and are not addressed in this study. Subjects living in
urban areas do seem to have a higher risk of CD and,
because of closer living conditions, may have a greater
risk of infection. Higher socioeconomic status has also
been associated with a higher risk of CD and may result
in more documented infections because of healthcare-
seeking behaviors.These risk factors meet the definition
of potential confounding variables and need to be
carefully considered when determining the existence of
a relationship between prior antibiotic use and CD.


Treatment Observations
Other evidence that intestinal infection and/or
antibiotics play a role in the development of CD comes
from clinical observations on antibiotic efficacy in
treating some forms of this disorder.The fecal stream is
important in the maintenance of inflammation in CD.


If an affected segment is bypassed, inflammation
diminishes. Specific antibiotics, especially
metronidazole,have been found useful in inflammatory
and penetrating CD of the colon and terminal
ileum.11,12 Metronidazole has been found to be
efficacious probably through its effect on bacteria,
particularly anaerobic species in these locations. Both
ciprofloxacin and metronidazole have been effective in
the treatment of perianal disease possibly also due to
effects on gut flora that are potentiating mucosal
inflammation. Antibiotics have some role in the
modification of existing gut flora leading to clinical
improvement. Unfortunately, there is only limited data
on efficacy for these agents and their mechanism of
activity is poorly understood. Their use, although
somewhat controversial, has become accepted
practice.13 They tend to be inexpensive and have
acceptable side effect profiles, which has also led to
their popularity.


Host Interaction with Gut Flora
It is generally accepted that host interactions with gut
bacteria are important in the development and course
of CD. Genetic predisposition to CD may rest in
abnormal responses to gut antigens ý namely bacteria
and bacterial products. Unfortunately, understanding
the precise role of gut flora has been problematic.The
interaction of environmental and genetic influences on
the development of CD has been suggested by recent
information regarding mutations in the CARD15
gene.1 CD is a heterogeneous disease with varied
clinical manifestations or phenotypes resulting in
complex genetics. Hugot et al. discovered linkage of
CD with a specific region of Chromosome 16 (IBD1
locus).14 This led others to find clinically specific
variants of the CARD15 (NOD2) gene, which is
located within this region. CARD15 is important in
regulating nuclear factor kappa B expression in
response to bacterial lipopolysaccharide. These
mutations may cause abnormal host interactions with
intestinal micro-flora resulting in CD.These mutations
may be most important in altering the interaction of
Paneth cells with gut bacteria and explain the
preponderance of ileal disease among patients with
CARD15 mutations.15,16 The highest concentration of
Paneth cells and Peyerýs patches is in the ileum. Patients
with CARD15 mutations are up to 40 times more
likely to develop CD. CARD15 represents a genetic
predisposition to an environmental (bacterial) antigen
in some patients with this disorder. This is the most
compelling evidence yet for the interaction of host
susceptibility and gut flora. The mechanism of this
interaction remains to be determined. Substances that
alter this gut flora such as antibiotics could have varied
effects on the development and clinical course of CD.


Conclusion
Descriptive evidence from childhood exposure to
infection and antibiotics, treatment observations, and
recent genetic research all point to the involvement of
host antibiotic flora in the development and maintenance
of inflammation in CD. Conflicting epidemiologic
evidence is likely due to inaccuracies of information and
recall bias regarding childhood exposures.In the complex
interaction of bacteria and bacterial antigens with the
host, it is not clear which species or antigens are the
culprits in this process. Early antibiotic exposure could
conceivably alter the developing immune system
resulting in lack of tolerance to common gut antigens
and subsequent development of CD. This might also
explain the increased risk of CD in higher
socioeconomic groups who may have been more likely
to receive antibiotics in childhood.


The discovery of the CARD15 mutation has added
clear biologic plausibility to a host interaction with
intestinal flora leading to the development of CD.


Although this mechanism may only apply to a subset
of patients with this heterogeneous disorder, it still
remains a monumental advance in the search for the
cause of CD.17 a73
Infection, Antibiotic Use and the Development of Crohnýs disease
B USINESS BRIEFING: US GASTROENTEROLOGY REVIEW 2005
3
References
1. Duerr R H,ýThe genetics of inflammatory bowel diseaseý, Gastroenterol. Clin. N.Am. 31 (2002), pp. 63ý76.


2. Sandler R S Eisen GM, ýEpidemiology of Inflammatory Bowel Diseaseý, Inflammatory Bowel Diseases, 5th ed. Kirsner
(2000), pp. 89ý112.


3. Ekbom A, Helmick C, Zack M,Adami H O ýThe epidemiology of inflammatory bowel disease: a large population-based study
in Swedený, Gastroenterology 100 (2) (1991), pp 350ý358.


4. Ekbom A, Zack M,Adami H O, Helmick C,ýIs there a clustering of inflammatory bowel disease at birth?ý, Am. J. Epidem.


134 (1991), pp. 876ý886.


5. Whorwell P J, Holdstock G,Whorwell G M, ýBottle feeding, early gastroenteritis and inflammatory bowel diseaseý, BMJ,1
(1979), pp. 382ý385.


6. Gent A E, Hellier M D, Grace R H, et al.,ýInflammatory bowel disease and domestic hygiene in infancyý, Lancet 343 (1994),
pp. 766ý770.


7. Weinstock J V Summers, R W, Elliot D E, et al.,ýThe possible link between de-worming and the emergence of immunological
diseaseý, J. Lab. Clin. Med. 139 (2002) pp 334ý338.


8. Wurzelmann J I, Lyles C M, Sandler R S,ýChildhood infections and the risk of inflammatory bowel diseaseý, Dig. Dis. Sci.


39 (1994), pp 555ý560.


9. Gilat T, Hacohen D, Lilos P, Langman M J, ýChildhood factors in ulcerative colitis and Crohný . An international
cooperative studyý, Scand. J. Gastroenterol. 22 (1987) pp.1,009ý1,024.


10. Card T, Logan R F A, Rodrigues L C,Wheeler J G,ýAntibiotic use and the development of Crohný Gut 53 (2004),
pp. 246ý250.


11. Lofberg R,ýReview article:medical treatment of mild to moderate Crohný Aliment. Pharmacol.Ther. 17 (Suppl 2)
(2003), pp. 18ý22.


12. Scribano M, Prantera C,ýReview Article:medical treatment of moderate to severe Crohný Aliment. Pharamacol.Ther.


17 (Supp 2) (2003), pp. 23ý30.


13. Shanahan F, Bernstein C,ýAntibiotics as fist-line therapy for Crohný . Is there a consensus?ý, Inflamm. Bowel Dis. 10
(2004), pp. 324ý326.


14 Hugot J P, Laurent-Puig P, Gower Rousseau C,ýMapping a susceptibility locus for Crohný hromosome 16ý,
Nature 379 (1996), pp. 821ý823.


15. Gasche C, Grundtner P, ýGenotypes and phenotypes in Crohný Do they help in clinical management?ý, Gut 54
(2005):162-7.


16. Mc Grimm and P Pavli. NOD2 mutations and Crohný are Paneth cells and their microbial peptides the link? Gut 53
(2004), pp. 1,558ý1,560
17. Picco M F,ýCrohný Blame it on the bugs?ý, Inflamm. Bowel Dis. 10 (2004), pp. 488ý489.