You are here

Inflammatory Bowel Disease Environmental Risk Factors: A Population-Based Case–Control Study of Middle Eastern Migration to Australia

Clinical Gastroenterology and Hepatology, Volume 13, Issue 8, August 2015, Pages 1453 - 1463.e1

Background & Aims

The incidences of the inflammatory bowel diseases (IBDs) Crohn’s disease (CD) and ulcerative colitis (UC) are increasing, indicating gene–environment interactions. Migrants from low-IBD-prevalence countries to a high-prevalence country may help identify the relative contribution of environmental risk factors compared with native Caucasians.

Methods

This prospective case–control study evaluated IBD environmental risk factors of Middle Eastern migrants (MEM) in Australia compared with matched Caucasian IBD subjects, MEM controls, Caucasian controls, and controls in the Middle East using adjusted odds ratios (aOR).

Results

A total of 795 subjects were recruited: 154 MEM cases (75 CD; 79 UC), 153 MEM controls, 162 Caucasian cases (85 CD; 77 UC), 173 Caucasian controls, and 153 controls in Lebanon. Smoking increased CD risk in MEM and Caucasians and reduced UC risk in Caucasians (aOR, 0.77; 95% CI, 0.41–0.98) but not MEM (aOR, 1.45; 95% CI, 0.80–2.62). Antibiotic use reduced the risk of MEM CD (aOR, 0.27; 95% CI, 0.11–0.67) and UC (aOR, 0.38; 95% CI, 0.18–0.80), but increased the risk in Caucasians (CD: aOR, 5.24; 95% CI, 2.13–12.90; and UC: aOR, 6.82; 95% CI, 2.67–17.38). Most hygiene markers (rural dwelling, pet ownership, pet feeding, and farm animal contact) reduced CD and UC risk in MEM (P < .05). In contrast, in Caucasians these hygiene markers lacked significance. Other significant risk factors include IBD family history, appendectomy, tonsillectomy, and breastfeeding.

Conclusions

Differential IBD environmental risk factors exist between migrants and native Caucasians, indicating a dynamic interplay between environmental factors and IBD risk for immigrants that is distinct to those factors most relevant in native Caucasians.

Keywords: Inflammatory Bowel Disease, Risk Factors, Migrants, Environmental, Hygiene Hypothesis.

Abbreviations used in this paper: aOR - adjusted odds ratio, AUD - Australian dollars, CD - Crohn’s disease, CI - confidence interval, IBD - inflammatory bowel disease, MEM - Middle Eastern Migrants, OCP - oral contraceptive pill, OR - odds ratio, SEIFA - socioeconomic indexes for areas, SES - socioeconomic status, UC - ulcerative colitis.

The pathogenesis of the inflammatory bowel diseases (IBDs) Crohn’s disease (CD) and ulcerative colitis (UC) involves an interaction between genetic susceptibility, environmental risk factors, host intestinal microbiota, and immunologic dysregulation. 1 The importance of environmental IBD risk factors is supported by high dizygotic twin concordance rates, 2 the North–South gradient of disease in North America, 3 and the rapidly increasing incidence of IBD in developing countries. 4 Despite the identification of ever-increasing numbers of susceptibility gene loci associated with IBD through genome-wide associated studies, the attributable risk of disease heritability remains at most 23%, stressing the importance of IBD environmental risk factors.4 and 5

The study of IBD in migrant groups transitioning from a developing country deemed protective against IBD, to a developed country with a high prevalence of IBD may provide insight into the interplay between environmental hygiene and immunologic risk factors. South Asian adult migrants to the United Kingdom and Canada have higher IBD incidence rates,5 and 6 but possible risk factors associated with migration have not been evaluated. Speculative hygiene-themed risks include dietary changes, improved sanitation, and dysbiosis from antibiotics,4, 6, and 7 while smoking, breastfeeding, vaccination, and appendicitis may represent immunologically driven risk factors. 8 The relative contribution of these IBD environmental risk factors may differ according to the population studied. A comparison of these relative contributions between native vs migratory populations may identify differential risk factors. Australia is a developed country with one of the highest incidence rates of IBD in the world, 9 in addition to a high migration rate, contributing to 53% of the nation’s annual population growth. 10 Middle Eastern migrants (MEM) accounts for 4.4% of the Australian population clustered within a well-defined area of Sydney, anecdotally with high prevalence rates of IBD. Studies have associated Jewish ancestry with increased genetic susceptibility of IBD. 11 Genetic studies have shown that Jewish and Middle Eastern non-Jewish populations share a common Semitic ancestry, 12 suggesting that non-Jewish Middle Eastern populations also may have a higher susceptibility to the development of IBD in a permissive environment. We hypothesize that non-Jewish Middle Eastern immigrants to Australia may be exposed to differential risks of IBD compared with native Caucasians. This study, therefore, compares known and speculative IBD risk factors in a migrant population compared with native Caucasians. We also aimed to characterize key differences between first- and second-generation migrants.

Methods

Cases

A prospective case–control face-to-face questionnaire study was performed on consecutive ambulatory IBD cases based in the Sydney and South Western Sydney Local Health Districts of New South Wales in Australia. 13 The catchment population of the area is 582,100, with a current smoking status of 14.9% and a median annual household income of Australian dollars $86,944, which is slightly higher than that of the state of New South Wales. Medical records and registry data were reviewed to confirm the diagnosis of UC and CD using conventional clinical criteria. Cases with an inconclusive diagnosis, those with IBD unclassified, and other non-IBD diagnoses were excluded.

Controls

Age- and sex-matched non-IBD controls living in the same catchment (MEM controls and Caucasian controls) were recruited randomly in the community. A further age- and sex-matched group of Middle Eastern controls were recruited randomly from the community in Lebanon (Lebanese controls) to compare the background environmental changes between Australia and the Middle East. Relatives of IBD subjects were excluded.

Ethnicity

Cases and controls identified their ethnic origins and only subjects with definite Middle Eastern or Caucasian ethnicities were included. MEM heritage countries were according to the Standard Australian Classification of Countries. 14 Subjects with mixed ethnicity, that is, with only 1 parent originating from countries listed as Middle Eastern or as European in the Standard Australian Classification of Countries, were excluded. This Australian Census data from 2006 showed that there were 193,633 people born in the Middle East who resided in Australia. Almost 40% from the Middle East were born in Lebanon, 16.8% were born in Iraq, 15.7% were born in Turkey, 11.6% were born in Iran, 4.0% were born in Israel, 3.6% were born in Syria and 2% were born in 9 other countries, each contributing less than 4000 persons. The clustering of MEM was in a specific, well-defined region of South Western Sydney, which is the population under investigation.

Data

A validated questionnaire tested for language, and cultural acceptability was used to determine the presence, absence, and timing of exposure to 102 environmental associations. 15 The questionnaire consists of non–open-ended and open-ended questions on exposure to a variety of variables during infancy (age, 0–5 y), childhood (age, 6–10 y), and adolescence (age, 11–16 y). Categoric answers of yes, no, and unsure were used whenever possible, with the unsure option included to reduce recall bias. Subjects were encouraged to consult with family members when answering the questions to increase the accuracy of their responses. Socioeconomic status (SES) at birth was determined by using the postcode of residence at birth to calculate the SES using the Socio-Economic Indexes for Areas (SEIFA) index of disadvantage from the Australia Bureau of Statistics, 16 and was divided into 4 quartiles (<965.78, 965.79–1010.76, 1010.77–1057.72, and >1057.73), with the highest quartile representative of the most advantaged SES. The mean SEIFA index score in Sydney is 1010. An extensive dietary assessment using the questionnaire tool was not feasible, but recall of takeaway-food consumption and vegetarianism was deemed to be most accurate. We categorized the frequency of consumption into daily, twice weekly, weekly, monthly, and almost never.

Sample Size Calculation and Statistical Methods

To detect an odds ratio of at least 1.75 with a power of 80%, at a significance level of 5%, a sample size of at least 316 cases and 316 controls was required. Univariate odds ratios (ORs) were calculated with 95% confidence intervals (CIs) for each potential risk factor in both Middle Eastern and Caucasian groups. A multivariate model was produced using logistic regression, in which a base model first was created using sociodemographic factors (age and sex) and factors consistently associated with IBD (cigarette smoking and family history). SES was not used in our base model because this variable lacked significance. Our base model was tested against each environmental risk factor for both CD and UC in Middle Eastern and Caucasian groups to obtain adjusted odds ratios (aORs) with 95% CIs. The Mann–Whitney U test was used for nonparametric continuous variables. Two-sided P values less than .05 were deemed statistically significant. Statistical analysis was performed using SPSS for Windows, version 20.0 (Chicago, IL).

Ethics

All patients provided consent for this study. Incomplete questionnaires were deemed invalid and therefore there were no missing data. This study was approved by the Human Research Ethics Committees of Sydney South West Local Health and Sydney Local Health Districts (Ethics committee approval code HREC/11/CRGH245).

Results

Demographic Data

A total of 795 subjects comprising 154 MEM IBD cases (75 CD and 79 UC), 162 Caucasian IBD cases (85 CD and 77 UC), 153 MEM controls, 153 Lebanese controls, and 173 Caucasian controls were recruited. Middle Eastern and Caucasian cases and controls were controlled for age and sex (all P > .05) ( Table 1 ). More than 90% of Middle Eastern cases and controls identified themselves as Lebanese ( Supplementary Figure 1 ), 10 consistent with the migratory pattern to Australia. The most common phenotype of UC in both Middle Eastern and Caucasian subjects was extensive disease. The most common location and behavior of disease in both Middle Eastern and Caucasian CD patients was ileocolonic, and nonstricturing, nonpenetrating disease ( Supplementary Table 1 ). MEM and Caucasians compared with controls showed similar characteristics but with some differences according to SEIFA and tertiary education ( Table 1 ).

Table 1 Demography of Middle Eastern and Caucasian Cases and Controls

Variable Middle Eastern Caucasian
Migrant controls, % Lebanese controls, % CD cases, % P, CD vs migrant controls P, CD vs Lebanese controls UC cases, % P, UC vs migrant controls P, UC vs Lebanese controls Controls, % CD cases, % P UC cases, % P
Mean age, y 37.1 35.6 37.6 .45 .31 42.5 .08 .05 44.9 43.7 .58 49.1 .08
Sex, female 50.3 64.1 53.3 .70 .12 55.6 .44 .22 52.0 49.4 .69 48.0 .56
Positive IBD family history                          
 First degree 2.0 1.3 12.0 <.001 <.001 17.7 <.001 <.001 1.7 20.0 <.001 13.0 <.001
 Second degree 5.2 3.3 25.3 <.001 <.001 12.7 .04 <.001 2.9 18.8 <.001 11.7 .01
Education                          
 Primary or less 16.3 21.7 6.7 <.001 <.001 8.9 <.001 .04 9.8 2.4 .03 2.6 .04
 Secondary 51.6 22.9 54.7 .06 <.001 58.2 .07 .03 66.5 67.1 .40 50.6 .15
 Tertiary 32.0 55.6 37.3 .05 .04 31.6 .06 <.001 23.7 30.6 .05 46.8 .03
SEIFA index a                          
 <966 26.1 b 26.7 .57 b 25.3 .39 b 26.1 22.4 .09 22.1 .09
 967–1011 13.1 b 20.0 .04 b 13.9 .32 b 13.1 10.6 .31 7.8 .02
 1012–1058 9.8 b 16.0 .02 b 19.0 <.001 b 9.8 9.4 .48 6.5 .03

a See text. Higher categories denote higher socioeconomic status.

b SEIFA index was not applicable for Lebanese controls.

CD, Crohn's disease; UC, ulcerative colitis; SEIFA, Socio-Economic Indexes for Areas.

Family History

MEM and Caucasian IBD cases were associated significantly with a first-degree relative with IBD compared with controls (MEM CD, 12.0%; MEM UC, 17.7%; MEM controls, 2.0%; Caucasian CD, 20.0%; Caucasian UC, 13.0%; Caucasian controls, 1.7%; P < .01). Second-degree relatives with IBD also were increased compared with controls (MEM CD, 25.3%; MEM UC, 12.7%; controls, 5.2%; Caucasian CD, 18.8%; Caucasian UC, 11.7%; Caucasian controls, 2.9%; P < .01). MEM IBD cases had a higher first-degree family history of IBD in UC (aOR, 11.83; 95% CI, 3.12–44.85) than Caucasians (aOR, 5.83; 95% CI, 1.45–23.41), and a lower first-degree family history of CD (aOR, 6.60; 95% CI, 1.57–27.82) than Caucasians (aOR, 11.23; 95% CI, 3.01–41.88) (Table 2 and Table 3).

Table 2 Environmental Risk Factors for MEM and Caucasian Crohn’s Disease Compared With MEM and Caucasian Controls

Variable a MEM CD cases vs MEM controls Caucasian CD cases vs Caucasian controls
Unadjusted OR 95% CIs Adjusted OR 95% CI Unadjusted OR 95% CIs Adjusted OR 95% CIs
Family history                
 First degree 6.82 1.79–26.00 6.60 1.57–27.82 14.08 4.00–49.61 11.23 3.01–41.88
 Second degree 6.15 2.55–14.85 5.28 2.07–13.47 7.79 2.78–22.10 7.75 2.45–24.53
Smoking, ever                
 Yes 1.92 1.10–3.36 2.01 1.08–3.76 1.95 1.15–3.31 2.14 1.18–3.87
Medical conditions                
 OCP use 1.01 0.45–2.28 0.55 0.20–1.53 1.97 1.05–3.68 2.41 1.04–5.63
 Appendectomy 3.22 0.85–12.17 5.15 1.76–15.09 1.51 0.75–3.04 0.73 0.24–2.19
 Tonsillectomy 2.87 1.11–7.41 4.04 1.82–8.96 1.77 0.87–3.63 1.08 0.53–2.62
 Use of antibiotics at ages 0–16 y 0.21 0.06–0.77 0.27 0.11–0.67 5.14 2.31–11.30 5.24 2.13–12.90
Rural dweller                
 Infancy 0.22 0.09–0.54 0.19 0.41–0.41 1.34 0.77–2.31 1.11 0.58–2.15
 Childhood 0.42 0.19–0.94 0.35 0.18–0.68 1.03 0.59–1.78 0.71 0.23–2.15
 Adolescence 0.42 0.17–1.04 0.27 0.12–0.60 0.95 0.49–1.83 0.74 0.32–1.72
Pet ownership                
 Infancy 0.79 0.33–1.91 0.28 0.13–0.64 1.20 0.70–2.06 1.36 0.68–2.74
 Childhood 0.87 0.38–1.97 0.24 0.11–0.52 2.87 1.58–5.21 3.27 1.59–6.71
 Adolescence 0.61 0.26–1.45 0.27 0.12–0.57 2.73 1.49–5.01 2.91 1.44–5.88
Farm animal contact                
 Infancy 0.45 0.12–1.71 0.12 0.03–0.43 0.95 0.48–1.89 0.08 0.38–1.88
 Childhood 0.42 0.11–1.60 0.16 0.05–0.50 0.62 0.31–1.28 0.57 0.24–1.34
 Adolescence 0.53 0.11–2.70 0.08 0.02–0.34 0.72 0.36–1.44 0.79 0.35–1.80
Sharing bedroom                
 Infancy 0.97 0.30–3.13 0.47 0.21–1.04 0.58 0.34–1.01 0.43 0.23–0.80
 Childhood 1.21 0.35–4.21 0.36 0.16–0.80 0.60 0.35–1.02 0.54 0.29–1.01
 Adolescence 0.71 0.30–1.70 0.39 0.19–0.77 0.80 0.47–1.36 0.80 0.43–1.49
Pet feeding                
 Infancy 0.20 0.04–0.91 0.12 0.04–0.37 1.05 0.59–1.87 1.22 0.62–2.41
 Childhood 0.34 0.11–1.08 0.10 0.03–0.30 1.18 0.69–2.02 1.23 0.64–2.34
 Adolescence 0.16 0.04–0.73 0.13 0.05–0.36 1.30 0.77–2.21 1.39 0.75–2.59
Access to hot water                
 Infancy 1.46 0.79–3.03 1.71 0.68–4.30 0.92 0.43–1.96 0.57 0.21–1.62
 Childhood 1.42 0.69–2.96 1.56 0.63–3.87 1.01 0.45–2.28 0.54 0.19–1.57
 Adolescence 1.75 0.81–3.78 1.97 0.77–5.00 0.88 0.35–2.18 0.49 0.15–1.63
Takeaway consumption, daily vs almost never                
 Infancy 1.28 0.92–3.45 1.97 0.03–2.88 0.54 0.24–1.23 1.84 0.30–2.31
 Childhood 3.37 1.15–12.11 2.65 0.08–0.86 0.31 0.15–0.65 4.23 0.06–6.32
 Adolescence 1.10 0.41–2.97 0.96 0.31–3.01 0.47 0.23–0.97 0.69 0.03–14.94
 Vegetarian 0.44 0.09–2.08 0.33 0.06–1.69 1.48 0.46–4.82 0.73 0.16–3.23

a Other variables assessed and not shown owing to nonsignificant associations: vaccination for tuberculosis, measles, and mumps; sibship number; grommet insertion; prematurity; exposure to intrauterine smoking; maternal age; public swimming pool use; number of bedrooms; and number within household.

NOTE. Values in bold indicate statistically significant results (P < .05).

Table 3 Environmental Risk Factors for MEM and Caucasian Ulcerative Colitis Cases Compared With MEM and Caucasian Controls

Variable a MEM UC cases vs MEM controls Caucasian UC cases vs Caucasian controls
Unadjusted OR 95% CIs Adjusted OR 95% CIs Unadjusted OR 95% CIs Adjusted OR 95% CIs
Family history                
 First degree 10.77 2.99–38.75 11.83 3.12–44.85 8.41 2.24–31.50 5.83 1.45–23.41
 Second degree 2.63 0.99–6.95 3.70 1.28–10.65 4.45 1.44–13.75 4.62 1.32–16.16
Smoking                
 Yes 1.37 0.79–2.36 1.45 0.80–2.62 0.65 0.36–1.16 0.77 0.41–0.98
Medical conditions                
 OCP use, ever 1.01 0.46–2.23 0.97 0.40–2.39 2.02 1.04–3.93 1.68 0.71–3.94
 Appendectomy 3.20 1.09–9.33 5.00 1.59–15.70 0.65 0.27–1.59 0.47 0.17–1.29
 Tonsillectomy 1.75 0.86–3.56 2.22 1.02–4.69 2.86 1.45–5.69 2.48 1.15–5.37
 Use of antibiotics at ages 0–16 y 0.43 0.22–0.85 0.38 0.18–0.80 4.88 2.17–10.95 6.82 2.67–17.38
Rural dweller                
 Infancy 0.13 0.60–0.28 0.09 0.04–0.24 0.66 0.36–1.23 0.60 0.30–1.20
 Childhood 0.17 0.08–0.34 0.12 0.05–0.30 0.59 0.32–1.10 0.57 0.28–1.13
 Adolescence 0.12 0.05–0.28 0.10 0.04–0.27 0.64 0.31–1.31 0.58 0.25–1.36
Pet ownership                
 Infancy 0.57 0.29–1.12 0.43 0.20–0.93 1.07 0.62–1.85 0.71 0.38–1.34
 Childhood 0.38 0.21–0.71 0.27 0.13–0.57 2.19 1.21–3.94 2.00 1.04–3.85
 Adolescence 0.57 0.31–1.02 0.41 0.20–0.81 2.08 1.14–3.77 1.64 0.85–3.14
Farm animal contact                
 Infancy 0.26 0.10–0.64 0.13 0.04–0.44 0.44 0.18–1.05 0.42 0.17–1.04
 Childhood 0.24 0.10–0.61 0.12 0.04–0.40 0.69 0.34–1.42 0.65 0.30–1.40
 Adolescence 0.24 0.90–0.65 0.15 0.05–0.50 0.52 0.24–1.14 0.42 0.18–0.98
Sharing bedroom                
 Infancy 0.66 0.34–1.30 0.60 0.28–1.30 0.65 0.37–1.13 0.48 0.26–0.90
 Childhood 0.54 0.26–1.13 0.43 0.19–0.97 0.41 0.24–0.72 0.26 0.14–0.50
 Adolescence 0.86 0.49–1.60 0.73 0.36–1.47 0.63 0.36–1.11 0.38 0.20–0.74
Pet feeding                
 Infancy 0.21 0.08–0.55 0.15 0.05–0.46 1.07 0.62–1.85 0.57 0.28–1.15
 Childhood 0.18 0.08–0.43 0.13 0.05–0.35 2.19 1.21–3.94 0.72 0.38–1.38
 Adolescence 0.30 0.14–0.63 0.24 0.10–0.56 2.08 1.14–3.77 0.69 0.37–1.29
Access to hot water                
 Infancy 1.19 0.60–2.35 1.85 0.78–4.36 3.01 0.97–4.78 3.45 0.35–5.43
 Childhood 1.12 0.56–2.23 1.69 0.71–3.99 2.43 0.80–7.38 3.88 0.13–5.31
 Adolescence 1.32 0.65–2.70 2.16 0.87–5.37 1.65 0.52–5.14 2.74 0.76–9.88
Takeaway consumption, daily vs almost never                
 Infancy 1.02 0.78–2.45 1.22 0.56–2.34 0.40 0.18–0.87 0.33 0.04–2.64
 Childhood 2.25 0.14–36.98 3.96 0.19–8.18 0.21 0.10–0.44 0.75 0.07–7.97
 Adolescence 6.21 1.58–24.35 6.96 1.66–29.18 0.31 0.15–0.64 1.00 0.02–2.45
 Vegetarian 0.87 0.26–2.90 1.19 0.32–4.43 2.75 0.96–7.88 1.91 0.60–6.06

a Other variables assessed and not shown owing to nonsignificant associations: vaccination for tuberculosis, measles, and mumps; sibship number; grommet insertion; prematurity; exposure to intrauterine smoking; maternal age; public swimming pool use; number of bedrooms; and number within household.

NOTE. Values in bold indicate statistically significant results (P < .05).

Smoking

In MEM, smoking was associated significantly with the risk of CD (aOR, 2.01; 95% CI, 1.08–3.76) but not UC (aOR, 1.45; 95% CI, 0.80–2.62). In the Caucasian population, smoking was associated significantly with CD (aOR, 2.14; 95% CI, 1.18–3.87), although it was associated negatively with the development of UC (aOR, 0.77; 95% CI, 0.41–0.98) (Table 2 and Table 3).

Appendectomy and Tonsillectomy

In the MEM population, appendectomy was associated significantly with CD (aOR, 5.15; 95% CI, 1.76–15.09) and UC (aOR, 5.00; 95% CI, 1.59–15.70). Tonsillectomy also was associated with the development of CD (aOR, 4.04; 95% CI, 1.82–8.96) and UC (aOR, 2.22; 95% CI, 1.02–4.69). In the Caucasian population, a previous appendectomy was not associated with either CD or UC. Tonsillectomy, however, was associated with UC (aOR, 2.48; 95% CI, 1.15–5.37) but not CD (Table 2 and Table 3).

Antibiotics and Oral Contraceptive Pill Use

Frequent antibiotic use (defined as having taken at least 4 courses of antibiotics in any given year before age 16 years) was associated with a reduced likelihood of CD (aOR, 0.27; 95% CI, 0.11–0.67) and UC (aOR, 0.38; 95% CI, 0.18–0.80) in the MEM population. In the Caucasian population, antibiotic use before age 16 years was associated with a significantly increased likelihood of both CD (aOR, 5.24; 95% CI, 2.13–12.90) and UC (aOR, 6.82; 95% CI, 2.67–17.38). Oral contraceptive use was significant only in Caucasians for CD, increasing the likelihood of CD (aOR, 2.41; 95% CI, 1.04–5.63) (Table 2 and Table 3).

Hygiene

In MEM, proxy variables of microbial exposure of living in a rural area, pet ownership, pet feeding, and farm animal contact from all years before age 16 all were associated with a reduction in CD and UC (all risk factors, P < .05). In Caucasians, opposing hygiene risk factors on the development of IBD were noted compared with MEM. CD risk increased with pet ownership during the childhood years (aOR, 3.27; 95% CI, 1.59–6.71) and adolescence (aOR, 2.91; 95% CI, 1.44–5.88). UC risk increased with pet ownership during the childhood years (aOR, 2.00; 95% CI, 1.04–3.85). On the other hand, bedroom sharing reduced the likelihood of CD during the infancy years (aOR, 0.43; 95% CI, 0.23–0.80) and UC during all years from infancy to 16 years. Rural dwelling, pet feeding, and farm animal contact (except during adolescence, which is associated significantly with reduced UC risk) all lacked significant associations with CD and UC (Table 2 and Table 3).

Breastfeeding

In the Middle Eastern population, breastfeeding was associated significantly with a decreased likelihood of developing IBD if maintained for at least 3 months for CD and for at least 6 months for UC ( Figure 1 ). A longer duration of breastfeeding conferred further protection against IBD. Breastfeeding in the Caucasian population also significantly reduced the likelihood of IBD if maintained for at least 3 months for CD and for at least 6 months for UC. A duration-response effect was evident in all comparisons in which a greater duration of breastfeeding had strong protective associations with the development of CD and UC.

gr1

Figure 1 The effect of duration of breastfeeding in months on the development of (A) Middle Eastern migrant ulcerative colitis and (B) Crohn's disease cases versus Middle Eastern migrant controls; and (C) Caucasian ulcerative colitis and (D) Crohn's disease cases versus Caucasian controls.

Diet

In the Middle Eastern population, takeaway consumption was not associated with the development of an IBD in all age groups except for UC in adolescence. Being a vegetarian was not associated with IBD. Similarly, in Caucasians, the dietary variables of frequency of takeaway consumption and being a vegetarian were not associated significantly with CD and UC (Table 2 and Table 3).

Comparison of Middle Eastern Migrant Controls and Lebanese Controls

Environmental risk factors for MEM controls were compared against Lebanese controls to determine changes in lifestyle after migration ( Table 4 ). Australian MEM migrants were more likely to use the oral contraceptive pill (aOR, 3.23; 95% CI, 1.46–7.60) and antibiotics (aOR, 3.22; 95% CI, 1.81–5.71) compared with Lebanese controls. MEM migrants were less likely to share a bedroom before the age of 16 years (P < .05), have farm animal contact in childhood (aOR, 1.91; 95% CI, 1.01–3.63), or live in a rural area in infancy (aOR, 0.20; 95% CI, 0.08–0.47) compared with Lebanese controls. MEM migrants were more likely to feed pets in childhood only (aOR, 2.07; 95% CI, 1.21–3.53). There were no significant differences in smoking or dietary factors between the 2 groups.

Table 4 Environmental Changes When Comparing Australian Middle Eastern Controls With Controls From Lebanon

Variable Unadjusted OR 95% CI Adjusted OR 95% CI
Family history        
 First degree 1.52 0.25–9.23 1.43 0.23–9.08
 Second degree 1.64 0.53–5.14 1.97 0.61–6.30
Smoking, ever        
 Yes 0.78 0.51–1.25 0.79 0.49–1.28
Medical conditions        
 OCP use 2.60 1.24–5.43 3.23 1.46–7.60
 Appendectomy 0.66 0.23–1.89 0.63 0.21–1.86
 Tonsillectomy 1.26 0.64–2.50 1.17 0.58–2.37
 Use of antibiotics at ages 0–16 y 3.22 1.84–5.65 3.22 1.81–5.71
Rural dweller        
 Infancy 1.23 0.77–1.94 0.20 0.08–0.47
 Childhood 1.17 0.75–1.84 1.25 0.78–2.00
 Adolescence 0.99 0.59–1.68 1.08 0.62–1.90
Pet ownership        
 Infancy 0.76 0.47–1.24 0.77 0.46–1.27
 Childhood 1.29 0.81–2.06 1.32 0.81–2.14
 Adolescence 1.11 0.70–1.77 1.14 0.70–1.85
Farm animal contact        
 Infancy 1.53 0.86–2.71 1.85 0.96–3.58
 Childhood 1.63 0.93–2.87 1.91 1.01–3.63
 Adolescence 1.40 0.79–2.86 1.50 0.79–2.85
Sharing bedroom        
 Infancy 0.43 0.21–0.88 0.44 0.21–0.92
 Childhood 0.34 0.14–0.85 0.34 0.14–0.85
 Adolescence 0.37 0.19–0.69 0.34 0.18–0.65
Pet feeding        
 Infancy 1.62 0.92–2.83 1.73 0.96–3.11
 Childhood 2.01 1.20–3.36 2.07 1.21–3.53
 Adolescence 1.58 0.95–2.62 1.66 0.98–2.82
Access to hot water        
 Infancy 0.65 0.36–1.17 0.55 0.27–1.09
 Childhood 0.76 0.43–1.35 0.70 0.34–1.32
 Adolescence 0.76 0.43–1.35 0.69 0.36–1.34
Takeaway consumption, daily vs almost never        
 Infancy 0.89 0.34–1.40 0.77 0.20–1.63
 Childhood 1.12 0.60–1.66 0.97 0.52–1.54
 Adolescence 1.14 0.60–1.72 1.03 0.57–1.59
 Vegetarian 1.31 0.48–3.62 1.39 0.49–3.95

NOTE. Values in bold indicate statistically significant results (P < .05).

Generational Differences in Migrant Middle Eastern Cases

The mean age at diagnosis for second-generation migrants was significantly lower than that for first-generation migrants for CD (28.3 vs 56.6 y; P = .01) and UC (28.4 vs 53.8 y; P = .01). The mean years of residence in Australia were longer in first-generation than in second-generation CD migrants (39.8 vs 29.4 y; P = .04) and for second-generation UC migrants (41.1 vs 31.3 y; P = .05). The mean years of residence in Australia for MEM controls was 23.9 years, resulting in significant differences between the mean years of residence between MEM cases and controls (P = .01). The median age of MEM controls was significantly older at migration compared with MEM cases (14.3 vs 5.1 y; P = .01). Generational differences in MEM cases are shown in Table 5 . When comparing second-generation with first-generation migrants for CD, rural dwelling was less likely compared with first-generation migrants (aOR, 0.25; 95% CI, 0.66–0.94), whereas tonsillectomy was found to be more likely compared with first-generation migrants (aOR, 6.70; 95% CI, 1.21–37.30). A second-degree family history of CD in second-generation migrants also was more likely than in first-generation migrants (aOR, 4.39; 95% CI, 1.04–18.61). There were no significant differences in the environmental factors between first-generation and second-generation MEM ulcerative cases.

Table 5 Selected Results When Comparing Second- With First-Generation Migrants for Crohn’s Disease and Ulcerative Colitis

Variable Crohn’s disease Ulcerative colitis
Unadjusted OR 95% CI Adjusted OR 95% CI Unadjusted OR 95% CI Adjusted OR 95% CI
Smoking, yes 0.35 0.13–0.95 0.34 0.12–1.09 0.49 0.18–1.32 0.32 0.10–1.09
Family history, first degree 1.24 0.29–5.38 0.33 0.52–2.06 1.64 0.50–5.41 0.84 0.20–3.62
Family history, second degree 2.40 0.77–7.47 4.39 1.04–18.64 3.31 0.64–17.04 2.60 0.42–16.21
Tonsillectomy 7.40 1.57–35.00 6.70 1.21–37.30 1.02 0.34–3.10 0.99 0.26–3.85
Rural dwelling 0.39 0.14–1.03 0.25 0.66–0.94 0.29 0.90–0.93 0.46 0.11–1.98

NOTE. Values in bold indicate statistically significant results (P < .05).

Nonsignificant Findings

Variables that were nonsignificant for both Middle Eastern and Caucasian populations were vaccinations for tuberculosis, measles, and mumps; sibship number; grommets insertion; being born more than 4 weeks premature; exposure to intrauterine smoking; maternal age; frequency of use of a public swimming pool before age 16; number of bedrooms and number of residents in the household from infancy to age 16 years; and access to hot water during childhood.

Discussion

This study used a migrant population to study an extensive number of IBD environmental risk factors compared with a nonmigrant group. The use of a migrant group to investigate a comprehensive range of environmental factors constitutes a novel approach to the epidemiologic study of IBD risk factors. Similarities and disparities of environmental exposure factors were found between Caucasian IBD cases and Middle Eastern IBD cases vs their respective controls. Migrants have a greater sensitivity to immunologically modifying IBD environmental risk factors, particularly cigarette smoking, hygiene-related factors, and antibiotic exposure. Other significant associations include a family history of IBD, highlighting changes to IBD risk as Middle Easterners migrate from developing to developed countries.

Middle Eastern migrants had significantly more risk factors that were proxy measures of the hygiene hypothesis compared with the Caucasian population, suggesting that migrants may be more sensitive to environmental risk factors that influence the intestinal microbiome. Rural dwelling, farm animal contact, bedroom sharing, pet ownership, and pet feeding significantly protected against the development of CD and UC in MEM. Comparatively, bedroom sharing was the only factor that reduced the likelihood of UC in Caucasians. This likely is owing to the sudden exposure to environmental changes from developing to developed nations, causing environmental effects to be more apparent in first- and second-generation migrants. The hygiene hypothesis has been deemed important in explaining the increase of IBD in non-Western countries, which are becoming more Westernized. Improved sanitation with resulting reduced exposure to bacterial infections lead to abnormal immunologic development and decreased tolerance against microbial antigens. 17 However, data from Canada, a developed nation, did not find supportive evidence for the hygiene hypothesis. 18 Migrants from developing countries may be more sensitive in showing gene-to-environmental interactions. In contrast to older studies, access to hot water and higher SES were not associated with the development of IBD in our study. 19 This may indicate that these markers are no longer sensitive or reliable indicators of hygiene in the 21st century, in which access to hot water is ubiquitous.

The lack of association between smoking and IBD is well recognized in developing countries in Asia, the Middle East, and Africa. 17 Our study, however, indicated that once migrants relocate to a developed country, their risk of CD from smoking is not different from that of Caucasians in the developed country. This was seen in the increased risk of CD in MEM smokers (aOR, 2.01), which was similar to that of Caucasians (aOR, 2.14). Thus, smoking as a CD risk factor was only evident in Australia and not in the Middle East.

Antibiotic use may affect IBD risk by affecting the intestinal microbiome and immunity. 20 Our study found that antibiotic use in childhood increased the risk of CD and UC in Caucasians, but reduced the risk in Middle Easterners. The decreased risk of antibiotics similarly was found recently in an inception population-based cohort study of Asia–Pacific countries that currently are seeing an increase in IBD incidence. 21 The opposing effects of antibiotics might be related to antibiotic-prescribing patterns and indications. The increased risk in Caucasians may be owing to the decrease in intestinal microbial diversity and detection bias around the time of the first IBD diagnosis. 22 The reduced risk in Middle Easterners may be owing to antibiotics being a surrogate marker of recurrent gastroenteritis that occurs at a rate 8 times higher in the Middle East than in Australia. 23 Increased bacteria exposure in developing countries may increase intestinal microbiome diversity before migration to Australia as a result of the development of immunologic tolerance. 24

For Middle Eastern migrants, a positive family history of IBD was associated very strongly with the development of both CD and UC when compared with Middle Eastern migrants in Australia. However, adjusted odds ratios were even higher when compared with Lebanese controls in Lebanon with adjusted odds ratios of 8.55 and 17.89 for CD and UC, respectively. These risks are among the highest reported familial IBD risk in any population, and suggest a high genetic predisposition in this population for developing IBD given appropriate epigenetic and environmental risk factors. Migrants offer the advantage of possessing a shorter time between disease onset and risk factor exposure when they transition from developing to developed countries.6 and 25 The MEM group shows that features such as IBD prevalence traditionally is low in the Middle East for non-Jewish residents, 26 and immigrants to Western nations have shown increased IBD incidence.27 and 28 The practice of consanguinity in Middle Easterners, which has prevalence rates of up to 36%, 29 may compound the risk of developing IBD further through amplification of recessive genetic risk factors.

Overall, our study also showed similar results to previous studies for many risk factors. Notably, having been breastfed for at least 6 months’ duration was associated with a reduced risk of subsequent development of CD and UC.15 and 21 This protective factor was shown in both Caucasians and the MEM populations.

There were some limitations to our study. To minimize the effects of recall bias we encouraged subjects to discuss early life influences with family members. We also included the option of “unsure” to minimize guesses by participants. Although the Middle Eastern population is heterogeneous with potential differences in genetic susceptibility to IBD between different groups and countries, this risk was minimized because more than 90% of our cases were Lebanese and we used a Lebanese control group. In addition, we used a previously validated questionnaire that was extensive but not exhaustive and so some proposed environmental triggers such as vitamin D exposure were not assessed. However, these limitations would apply to both Caucasians and Middle Easterners. Finally, the effects may be the result of chance alone. This was a large study of migrants and a sample size calculation suggested adequacy to show a reasonable effect size. There also was internal consistency as shown by duration response effects of breastfeeding and the generation effects in which second-generation migrants with a longer duration of environmental exposure had stronger risk factor odds ratios. The significant values shown also are consistent with previous publications.

Conclusions

In conclusion, this study showed that a migrant population has differential IBD risk factors compared with Caucasians in a developed country. The risk factors that were greater in recent immigrants than in Caucasians were related to the hygiene hypothesis, smoking, and the effects of family history. In Caucasians, however, the use of antibiotics was associated strongly with the development of both CD and UC, but was associated with a lower rate of developing IBD in migrants. Differential risk exposures between different populations may help explain the discrepancies of IBD risk factor data in the literature.

Acknowledgments

The authors thank the other members of the IBD Sydney Organisation for patient recruitment and contributing to this and other research projects.

Supplementary Material

 

fx1

Supplementary Figure 1

Supplementary Table 1 IBD phenotype of Middle Eastern migrant and Caucasian ulcerative colitis and Crohn's disease patients

  Middle Eastern (n = 135) Caucasian (n = 151)
Ulcerative colitis    
 E1: Ulcerative proctitis 14.0 10.9
 E2: Left-sided (distal) 40.4 40.6
 E3: Extensive UC 45.6 48.4
Crohn's disease    
 L1: ileal 32.1 19.5
 L2: colonic 20.5 36.8
 L3: ileocolonic 47.4 43.7
 p: perianal 28.2 26.4
 B1: non-stricturing non-penetrating 53.8 51.7
 B2: stricturing 23.1 26.4
 B3: penetrating 23.1 22.9

References

  • 1 D. Baumgart. Inflammatory bowel disease: cause and immunobiology. Lancet. 2007;369:1627-1640 Crossref
  • 2 N.P. Thompson, R. Driscoll, R.E. Pounder, et al. Genetics versus environment in inflammatory bowel disease: results of a British twin study. Br Med J. 1996;312:95-96 Crossref
  • 3 S.C. Ng, C.N. Bernstein, M.H. Vatn, et al. Geographical variability and environmental risk factors in inflammatory bowel disease. Gut. 2013;62:630-649 Crossref
  • 4 E.V. Loftus. Clinical epidemiology of inflammatory bowel disease: incidence, prevalence, and environmental influences. Gastroenterology. 2004;126:1504-1517 Crossref
  • 5 K. Fransen, M. Mitrovic, C.C. van Diemen, et al. The quest for genetic risk factors for Crohn's disease in the post-GWAS era. Genome Med. 2011;25:13
  • 6 C.N. Bernstein, F. Shanahan. Disorders of a modern lifestyle: reconciling the epidemiology of inflammatory bowel diseases. Gut. 2008;57:1185-1191 Crossref
  • 7 S. Mahid, K.S. Minor, R.E. Soto, et al. Smoking and inflammatory bowel disease: a meta-analysis. Mayo Clin Proc. 2006;81:1462-1471 Crossref
  • 8 P.C. Lunney, R.W. Leong. Review article: ulcerative colitis, smoking and nicotine therapy. Aliment Pharmacol Ther. 2012;36:997-1008 Crossref
  • 9 J. Wilson, C. Hair, R. Knight, et al. High incidence of inflammatory bowel disease in Australia: a prospective population-based Australian incidence study. Inflamm Bowel Dis. 2010;16:1550-1556 Crossref
  • 10 Australian Bureau of Statistics. WA fastest growing population in the country 2012. Available: http://www.abs.gov.au/ausstats/abs@.nsf/lookup/3101.0Media Release1Jun 2011 . Accessed: November 4, 2012.
  • 11 E.E. Kenny, I. Pe’er, A. Karban, et al. A genome-wide scan of Ashkenazi Jewish Crohn's disease suggests novel susceptibility loci. PLoS Genet. 2012;8:e1002559 Crossref
  • 12 M.F. Hammer, E.T. Wood, M.R. Bonner, et al. Jewish and Middle Eastern non-Jewish populations share a common pool of Y-chromosome biallelic haplotypes. Proc Natl Acad Sci U S A. 2000;97:6769-6774 Crossref
  • 13 Anonymous. A Picture of Health. SLHD Health Profile 2013. NSW Government, 2013, Sydney, NSW, Australia, p6 https://www.slhd.nsw.gov.au/pdfs/SLHD_Health_Profile_2013.pdf . Accessed April 16, 2015.
  • 14 Anonymous. Standard Australian Classification of Countries (SACC), 1998 (Revision 2.03). in Australian Bureau of Statistics, Canberra, ACT, Australia. http://www.abs.gov.au/AUSSTATS/abs@.nsf/0/1837ADE79569F330CA2572680017FE07?opendocument . Accessed April 16, 2015.
  • 15 R.B. Gearry, A.K. Richardson, C.M. Frampton, et al. Population-based cases control study of inflammatory bowel disease risk factors. J Gastroenterol Hepatol. 2010;25:325-333 Crossref
  • 16 Anonymous. Census of Population and Housing: Socio-Economic Indexes for Areas (SEIGA), Australia - Data only. in Australian Bureau of Statistics. Canberra, ACT, Australia. http://www.abs.gov.au/AUSSTATS/abs@.nsf/Lookup/2033.0.55.001Main+Features12006?OpenDocument . Accessed April 16, 2015.
  • 17 N.A. Molodecky, G.G. Kaplan. Environmental risk factors for inflammatory bowel disease. Gastroenterol Hepatol. 2010;6:339-346
  • 18 D.K. Amre, P. Lambrette, L. Law, et al. Investigating the hygiene hypothesis as a risk factor in pediatric onset Crohn's disease: a case-control study. Am J Gastroenterol. 2006;101:1005-1011
  • 19 A.E. Duggan, I. Usmani, K.R. Neal, et al. Appendicectomy, childhood hygiene, Helicobacter pylori status, and risk of inflammatory bowel disease: a case control study. Gut. 1998;43:494-498 Crossref
  • 20 R. Ungaro, C.N. Bernstein, R. Gearry, et al. Antibiotics associated with increased risk of new-onset Crohn's disease but not ulcerative colitis: a meta-analysis. Am J Gastroenterol. 2014;109:1728-1738 Crossref
  • 21 S.C. Ng, W. Tang, R.W. Leong, et al. Environmental risk factors in inflammatory bowel disease: a population-based case-control study in Asia-Pacific. Gut. 2014; Epub ahead of print
  • 22 T. Jess, J. Simonsen, N.M. Nielsen, et al. Enteric Salmonella or Campylobacter infections and the risk of inflammatory bowel disease. Gut. 2011;60:318-324 Crossref
  • 23 Anonymous. World Health Organization. Health statistics and information systems. Mortality and burden of disease estimates for member states in 2004. http://www.who.int/healthinfo/global_burden_disease/estimates_country/en/ . Accessed April 17, 2015.
  • 24 J.R. Korzenik. Past and current theories of etiology of IBD: toothpaste, worms, and refrigerators. J Clin Gastroenterol. 2005;39:S59-S65 Crossref
  • 25 Y. Ko, R. Butcher, R.W. Leong. Epidemiological studies of migration and environmental risk factors in the inflammatory bowel diseases. World J Gastroenterol. 2014;20:1238-1247 Crossref
  • 26 H. Abdul-Baki, I. ElHaji, L.M. El-Zahabi, et al. Clinical epidemiology of inflammatory bowel disease in Lebanon. Inflamm Bowel Dis. 2007;13:475-480 Crossref
  • 27 X. Li, J. Sundquist, K. Hemminki, et al. Risk of inflammatory bowel disease in first- and second-generation immigrants in Sweden: a nationwide follow-up study. Inflamm Bowel Dis. 2011;17:1784-1791 Crossref
  • 28 I. Carr, J.F. Mayberry. The effects of migration on ulcerative colitis: a three-year prospective study among Europeans and first- and second- generation South Asians in Leicester (1991-1994). Am J Gastroenterol. 1999;94:2918-2922
  • 29 B. Barbour, P. Salameh. Consanguinity in Lebanon: prevalence, distribution and determinants. J Biosoc Sci. 2009;41:505-517 Crossref

Footnotes

Gastroenterology and Liver Services, Concord Hospital, Sydney, Australia

Gastroenterology and Liver Services, Bankstown Hospital, Sydney, Australia

§ Faculty of Medicine, UNSW Australia

American University of Beirut Medical Center, Beirut, Lebanon

IBD Sydney Organisation, Sydney, Australia

# Department of Gastroenterology and Hepatology, Liverpool Hospital, Sydney, New South Wales, Australia

∗∗ Department of Gastroenterology and Hepatology, St George Hospital, Kogarah, New South Wales, Australia

‡‡ Faculty of Medicine, University of Otago, Christchurch, New Zealand

Reprint requests Address requests for reprints to: Rupert W. Leong, MD, Concord Hospital, Level 1 West, Hospital Road, Concord, New South Wales 2139 Australia. fax: (61) 2-9767-6767.

Conflicts of interest The authors disclose no conflicts.