A198 THE LONG-TERM IMPACT OF ENVIRONMENTAL EXPOSURES ON HOST HEALTH AND THE RISK FACTORS OF CROHN'S DISEASE

Background

Several environmental factors are associated with Crohn’s disease (CD) in large case-control studies; however, it is not clear how these factors maybe be influenced by age of exposure and if they are related to alterations in pre-disease biological markers of CD risk.

Purpose

To investigate the association between environmental factors in different age groups with future risk of CD onset and assess their relation to other pre-disease biomarkers.

Method

We used an environmental risk assessment questionnaire (ERA) to collect information from healthy first-degree relatives(FDR) of CD enrolled in the CCC-GEM project. ERA was a multi-item questionnaire querying 69 questions under 7 section headings: background, cultural/ethnic, smoking history, medical history, family history, environmental history and pet history. For the environmental and pet sections, current and historical (<1, 2-4, 5-15 years old) data was captured at the time of recruitment. We used Cox proportional hazard models to identify exposures associated with future CD onset. Next, we used regression models to identify the relationship of exposures with biological factors associated with CD risk previously identified by our group i.e.: i) intestinal permeability using urinary fractional excretion of lactulose to mannitol ratio (LMR) with LMR≥0.025 defined as abnormal; ii) subclinical inflammation using fecal calprotectin (FCP) with FCP≥100µg/g; and iii) fecal microbiome composition and diversity using 16S rDNA sequencing. Two-sided p<0.05 (or false discovery rate corrected p<0.05) were considered significant.

Result(s)

A total of 4289 FDRs were recruited, 47% were male, median recruitment age was 17.0 years[6-35]. After a median follow-up of 5.6-years (IQR=3.42-8.67), 86 FDRs developed CD. Living with a dog between age 5-15 (Hazard Ratio (HR)=0.61; 95% confidence interval (CI)=0.39-0.95), and a large family size (>3) in the first year of life (HR=0.41; 95% CI=0.22-0.89) were protective against CD onset. Conversely, having a bird at time of survey (HR=2.84; CI=1.37-5.90), and having a sibling with CD (HR=2.07; 95% CI=1.18-3.63) were risk factors for CD onset. We found that owning a dog between age of 5-15 (Odd Ratio(OR)=0.77, 95% CI=0.65-0.90) was significantly associated with LMR, nine taxa bacterial and higher chao1 diversity index. Having a bird at time of survey was significantly associated with FCP (OR=2.04, 95% CI=1.33-3.11). There was no association between large family size and having a CD sibling with gut microbiome, FCP or LMR.

Conclusion(s)

The study identified four environmental factors associated with future development of CD. Among them, exposure to dogs during early life was protective against CD onset and might be explained by its association with normal gut permeability and microbiome. We also identified that having a bird at recruitment increased risk of CD onset which might be mediated by an increase in subclinical inflammation.

Submitted on behalf of the CCC-GEM consortium

Disclosure of Interest

None Declared

A218 ASSOCIATIONS BETWEEN ADHERENCE TO LITERATURE-DERIVED DIETARY INDICES AND PRE-DISEASE BIOMARKERS: IMPLICATIONS FOR CROHN’S DISEASE PREVENTION

Background

The incidence of Crohn’s disease (CD) is increasing globally, indicating a significant environmental influence such as diet. A plethora of dietary adherence (DA) patterns exist in the literature: the Mediterranean Diet (MD), Empirical Dietary Inflammatory Pattern (EDIP), Specific Carbohydrate Diet (SCD), and low Fermentable Oligosaccharide, Disaccharide, Monosaccharide, or Polyol diet (FODMAP) are all potential candidates to maintain a reduced level of inflammation, improving gastrointestinal function. Contrary, the Westernized diet (WD) is generally reported as a diet promoting inflammation in humans.

Purpose

To determine if DA to literature-derived dietary indices in a cohort of first-degree relatives (FDRs) of CD patients can modulate pre-disease biomarkers.

Method

We used food frequency questionnaire (FFQ) data from 2,696 healthy FDR subjects of the Crohn’s Colitis Canada- Genes, Environment, Microbial (CCC-GEM) project. We rederived each of the following scores using our FFQ data, utilizing originally described methods for the MD, EDIP, SCD, low FODMAP, and WD to obtain DA. Each diet was correlated pairwise via Kendall’s Tau.

We fit multivariable regression models to identify the association of DA (top quintile vs remaining) and: i) intestinal permeability using urinary fractional excretion of lactulose to mannitol ratio (LMR), LMR≥0.03 defined abnormal; ii) subclinical inflammation using fecal calprotectin (FCP) measured with BÜHLMANN fCAL® ELISA, FCP≥250µg/g defined inflammation; and iii) fecal microbiome richness and composition using 16S rRNA sequencing. Two-sided p<0.05 for primary and q<0.05 for secondary analysis defined significance.

Result(s)

There were positive correlations between the MD, SCD, and low FODMAP, these diets negatively correlated with the WD. The EDIP negatively correlated with the SCD and low FODMAP, did not correlate with the MD, and positively correlated with the WD. No diet was associated with abnormal LMR or FCP. Only the SCD was associated with increased microbial richness (q=0.03). All diets were associated with microbial genera: the MD (n=18 taxa, (2.0^-7<q-values< 0.04), EDIP (n=9, [2.8^-4-0.05]), SCD (n=13, [3.7^-11-0.05]), low FODMAP (n=14, [1.3^-7-0.05]), and WD (n=1, [0.03]).

Conclusion(s)

This study shows that literature-derived dietary indices correlate generally with each other, yet none were not associated with abnormal LMR or FCP. However, we found that diet can impact microbiome richness and composition. Thus, it is tempting to speculate that diet is a possible intervention capable of maintain microbiome homeostasis to reduce future risk of CD.

Submitted on behalf of the CCC-GEM consortium.

Funding

Crohn’s and Colitis Canada Genetics Environment Microbial (CCC-GEM) III

The Leona M. and Harry B. Helmsley Charitable Trust

Kenneth Croitoru is the recipient of the Canada Research Chair in Inflammatory Bowel Disease

Disclosure of Interest

None Declared

A238 BILE ACID COMPOSITION AND DIETARY FAT: IMPLICATIONS FOR CROHN’S DISEASE IN A COHORT OF HEALTHY FIRST-DEGREE RELATIVES

Background

Crohn’s disease (CD) is a chronic relapsing inflammatory disease of the gastrointestinal tract. The etiology of CD may arise from complex interactions including host genetics, diet, and the intestinal microbiome. Increased consumption of saturated fats, characteristic of the Western diet, is a known risk factor for CD. Dietary fat (DF) is absorbed by the host through the release of primary bile acids (PBAs) and bio-transformed by the microbiome into secondary bile acids (SBAs). Altogether, bile acids (BAs) can act as signaling molecules involved in host immune regulation and potentially in CD onset.

Purpose

To investigate the relationship between CD risk, BAs, and DF, and evaluate the predictive performance of CD onset of these factors by developing machine learning models.

Method

We used samples healthy first-degree relatives (FDRs) recruited as part of the Crohn’s Colitis Canada- Genes, Environment, Microbial (GEM) project. Those who developed CD (n=87) were matched 1:4 by age, sex, follow-up time, and geographic location with control FDRs remaining healthy (n=347).

Serum, urine, and stool BA were measured using ultrahigh Performance Liquid Chromatography-Tandem Mass Spectroscopy. DF types were derived from food frequency questionnaire data. We used conditional logistic regressions to identify associations between CD onset, BAs (n=93), and DFs (n=9). We further explored the relationships of significant CD-related BAs and DF via Generalized Estimation Equations. Finally, we used a tree-based machine-learning algorithm (XGBoost) with 5-fold cross-validation to assess the prediction performance of CD onset using BA from all sources as well as DF. Two-sided p<0.05 was considered significant.

Result(s)

In total, 10 of 93 BAs, and two of nine DFs were significantly associated with increased odds of CD onset (p<0.05). Additionally, five BAs were significantly associated with DF (p<0.05). Serum-derived BAs had the best predictive performance for CD, with a mean AUC of 0.70 [95% CI: 0.63;0.76], followed by stool derived BAs with a mean AUC= 0.65 [0.55;0.75], and followed by urine derived Bas with a mean AUC= 0.57 [0.48;0.66]. Lastly DF was not a predictive marker of CD onset with a mean AUC= 0.50 [0.41;0.60].

Conclusion(s)

This study suggests that BAs are associated with the pathogenesis of CD and the effects may be influenced by DF. Serum-derived BAs may be able to better predict the risk of CD than other stool or urine derived BA, while DF is not directly implicated in CD risk.

Submitted on behalf of the CCC-GEM consortium.

Funding

Crohn’s and Colitis Canada Genetics Environment Microbial (CCC-GEM) III

The Leona M. and Harry B. Helmsley Charitable Trust

Kenneth Croitoru is the recipient of the Canada Research Chair in Inflammatory Bowel Diseases

The International Organization for the Study of Inflammatory Bowel Diseases (IOIBD)

Jingcheng Shao is the recipient of a Data Science Institute Summer Undergraduate Data Science award

Disclosure of Interest

None Declared

A157 DEFINITIONS OF MEDITERRANEAN DIET INCONSISTENTLY ASSOCIATE WITH MARKERS OF GUT BARRIER FUNCTION OR SUBCLINICAL INFLAMMATION IN A POPULATION-BASED COHORT

Background

The Mediterranean Diet (MD) is proposed to reduce the risk of Crohn’s disease (CD) onset in cohort studies, with inconsistent results. This inconsistency may be due to heterogeneity in defining MD scores. Additionally, relationships between MD compliance and intestinal permeability or sub-clinical inflammation are not defined.

Aims

We examined correlations between different MD scores, and determined associations between MD compliance and intestinal permeability or subclinical inflammation in a cohort of first degree relatives of CD patients.

Methods

We used food frequency questionnaire data from 2,112 subjects of the Crohn’s Colitis Canada- Genes, Environment, Microbial (CCC-GEM) project. We obtained 12 MD definitions from the literature and calculated daily percent compliance, we further compared MD scores via pairwise correlations (Kendall’s Tau). We measured intestinal permeability via urinary fractional excretion ratio of lactulose to mannitol (LMR) (LMR≥0.03 defined abnormal), and subclinical inflammation via fecal calprotectin (FCP) measured with BÜHLMANN fCAL® ELISA (FCP≥250 defined abnormal). We fit multivariable regression models between MD compliance and abnormal LMR and FCP, respectively. Two-sided p&lt;0.05 defined significance.

Results

There was large variation in cross-correlations among MD scores, from nil (t=0.0, p=0.54) to highly significant (t=0.97, p&lt;2.2e-16). Associations of MD compliance and abnormal LMR or FCP were in both directions of effect, largely non-significant. Of the 12 MD scores, none associated with abnormal LMR, while 4 associated with abnormal FCP-Odds Ratios =1.22, 1.23, 1.24, and 1.30; p=0.02, 0.02, 0.01, and 0.009, and 95% Confidence Intervals = [1.03,1.45], [1.04,1.45], [1.05,1.47], and [1.07,1.59] respectively. No diet remained significant after correcting for multiple testing.

Conclusions

Currently MD definitions vary widely. Despite discrepancies, we expected consistent directions of effect for MD compliance on LMR or FCP. The largely non-significant associations between MDs suggest limitations in definition, interpretation, and relation to biological outcomes.

Submitted on behalf of the CCC-GEM consortium.

Funding Agencies

CIHRCrohn’s and Colitis Canada Genetics Environment Microbial (CCC-GEM) III;The Leona M. and Harry B. Helmsley Charitable Trust; Justine Timpano is a recipient of a fellowship award from Mount Sinai Hospital; Kenneth Croitoru is the recipient of the Canada Research Chair in Inflammatory Bowel Diseases