Final Progress Report

Proposal No. IBD-0204R2
Principal Investigator:  Scott M. Montgomery, Ph.D.
Applicant Organization:  Karolinska Institutet (Stockholm, Sweden)
Project Title:  Markers of perinatal bowel colonization and pediatric Crohn's disease risk
Period of Award:  September 1, 2007 – February 28, 2009

Summary of project aims:
 
• To create register-based datasets for epidemiological research into early life exposures relevant to patterns of bowel flora colonisation (using epidemiological markers) and therefore Crohn’s disease risk. These datasets will also represent a resource for further research.

• To specifically investigate if birth by caesarean section is a risk for Crohn’s disease as perinatal bowel colonisation may represent an important period for establishment of homeostasis. Other aspects of delivery mode are also of interest.

• To antibiotic therapy in the mother (during pregnancy) or offspring (in infancy and childhood) increases Crohn’s disease risk. Antibiotic therapy may disrupt bowel colonisation, impair homeostasis and therefore increase Crohn’s disease risk. Antibiotic therapy is identified by diagnoses requiring this treatment.

• To assess if factors relevant to maternal immune function in pregnancy are associated with Crohn’s disease risk in offspring, including maternal diseases, smoking and maternal age. 


Accomplishments towards meeting the project aims

• A number of Sweden’s administrative databases were used to create two datasets for the project. One dataset contained details of pregnancy and perinatal characteristics for cases and controls. The second contained information on health-related exposures and diagnoses for cases and controls that occurred in infancy and childhood. All children in Sweden with a diagnosis of pediatric Crohn’s disease during the study period were identified and their data were linked with that for their mothers, with information on all inpatient care (for mother and child), details of the pregnancy and delivery, as well as vital status at all time points in these longitudinal datasets. The same information was collected for the matched control population. Detailed disease characteristics from medical records for a subset of patients (Stockholm County) were included and revealed that the Inpatient Register successfully identified the vast majority of relevant patients. More detailed information about the datasets is provided by the published paper submitted with this report and the Seventh Annual BMRP Investigator Meeting presentation (posted on the BMRP website).

In addition to being used for the specific aims of the project, these data represent a useful resource for future research to test other hypotheses or extend this work. Researchers at Karolinska Institutet are currently working with these data; so further research outputs will be produced.

• The association of birth by caesarean and other delivery characteristics with early-onset Crohn’s disease was robustly tested in this epidemiological study, with careful control for potential confounding factors, such as period, delivery unit, socioeconomic position of the family and maternal IBD.

• Maternal infections in pregnancy and Crohn’s disease risk in offspring was investigated using the newly created dataset. Existing material was augmented to create a second dataset that could be used to robustly test whether infections in infancy and childhood associated with antibiotic use may increase Crohn’s disease risk.

• The dataset with details of pregnancy and delivery also contained other maternal characteristics and diagnoses, allowing epidemiological investigation of whether such factors were relevant to Crohn’s disease risk among offspring. 


Significant results

• Overall, birth by caesarean was not a risk for Crohn’s disease. However, there was a modest and statistically significant risk increase among boys but not girls. This is consistent with the sex differences in the incidence of pediatric Crohn’s disease and may indicate sex differences in immune responses to infecting or colonising micro-organisms.

• Other aspects of delivery and pregnancy potentially associated with pattern of bowel colonisation were not associated with Crohn’s disease risk. However, maternal age did appear to modify the association of birth by caesarean section and Crohn’s disease risk. This suggests that other characteristics in early life associated with caesarean section can alter the outcome; either by altering exposures during delivery or signalling other associated characteristics operating at times other than delivery. As expected, a diagnosis of maternal IBD is a risk for Crohn’s disease in offspring, but this did not explain the modest association with caesarean section, as we were able to adjust for maternal IBD, as well as repeat the analyses after the exclusion of subjects with a mother who had IBD.

• Hospital admissions associated with antibiotic therapy in the first five years of life were associated with a notably increased and statistically significant risk of Crohn’s disease. Markers of antibiotic therapy after age five years were not associated with subsequent Crohn’s disease. This is consistent with the hypothesis that disruption of bowel flora by antibiotic therapy in early infancy may result in permanently disrupted homeostasis and increased Crohn’s disease risk. 


Lay summary of the progress report

This study successfully obtained and combined data from several Swedish administrative registers (computerised national data compilations). We were able to conduct an epidemiological study of conditions in pregnancy, at the time of birth and in infancy that may influence Crohn’s disease risk. It would not be feasible to conduct interview or questionnaire-based studies about these events many years after they occurred due to memory problems. This study was able to capitalise on the information collected at the time by Swedish national registers. The datasets created for this project represent a research resource for this and future studies

An inappropriate immune response against normal microbes in the gut may be responsible for Crohn’s disease. This project examined the hypothesis that preventing normal microbial colonisation, particularly around the time of birth, would result in failure to recognise these microbes, increasing the risk of an inappropriate immune response against them. Birth by caesarean section could disrupt the first bowel colonisation, but this was not identified as a substantial a risk for Crohn’s disease: caesarean section cannot explain why the disease has become so much more common in recent years. A modestly raised Crohn’s disease risk was only observed among boys delivered by caesarean section but not girls: this adds to the existing evidence that there are sex differences in risks for childhood Crohn’s disease. While not representing a major risk in terms of public health, the mechanisms underlying the association with birth by caesarean section are worthy of investigation to help understand the causes of Crohn’s disease, and particularly in the context of further exposures and susceptibility factors. The association of caesarean birth with Crohn’s risk varied by maternal age: this unexpected finding suggests that some other factors linked with age and delivery mode may be risks themselves or modify the effect of caesarean delivery.

As an individual’s immune system continues to mature rapidly in the first years after birth, we postulated that disruption of the bowel bacteria in infancy before age five years could impair the relationship between the immune system and the normal bacteria in the gut, resulting in an inappropriate immune response mounted against the normal bacteria. A sustained course of antibiotic therapy is one such exposure that could disrupt colonisation. Inpatient treatment for diagnoses requiring antibiotic therapy was used as an indicator of this exposure. We found that this marker of antibiotic exposure before, but not after, five years of age was robustly associated with later Crohn’s disease risk. This window of susceptibility in the first years of life may help to explain why earlier findings on antibiotic therapy have been inconsistent. Most studies have considered antibiotic use throughout life rather than examining the relatively brief period in infancy when the developing immune system – as it develops to recognise the normal bowel bacteria - may be most affected. 

This project has provided epidemiological evidence suggesting that disruption of bowel colonisation in the first years of life is relevant to Crohn’s disease risk. Birth by caesarean section is not a major risk in itself, but likely to be one of several factors in early life that have become more common in recent years contributing to the increased risk of Crohn’s disease.