Background: Early life factors have been postulated to play a role in development of immune tolerance and intestinal microbiome, which in turn may influence the risk of inflammatory bowel disease.
Methods: We conducted a prospective cohort study of 60,186 U.S. women enrolled since 1976 in the Nurses Health Study (NHS) I and 86,495 women enrolled since 1989 in the NHS II with no history of ulcerative colitis (UC) or Crohn’s disease (CD). Information about breastfeeding, birth weight, and preterm birth were collected in 1992 in NHS I and 1991 in NHS II. Diagnoses of CD and UC were confirmed through review of medical records. We used Cox proportional hazards models to calculate hazard ratios and 95% confidence intervals.
Results: Among 146,681 women over 3,373,726 person-years of follow-up, we documented 248 cases of CD and 304 cases of UC through 2007 in NHS II and 2008 in NHS I. The median age of diagnosis was 51 for CD and 49 for UC. Compared with women who were not breastfed, women who were breastfed had multivariate-adjusted hazard ratios of 0.99 (95% confidence interval, 0.76–1.30) for CD and 1.03 (95% confidence interval, 0.81–1.32) for UC. Similarly, low or high birth weight and preterm birth were not significantly associated with risk of UC or CD.
Conclusions: In 2 large prospective cohorts of U.S. women, we did not observe a significant association between early life factors including having been breastfed, birth weight, preterm birth, and risk of adult-onset UC and CD.
Article first published online 6 February 2013
*Division of Gastroenterology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts
†Division of Gastroenterology and Nutrition, Children's Hospital Boston and Harvard Medical School, Boston, Massachusetts
‡Department of Adult Oncology, Dana-Farber Cancer Institute, Boston, Massachusetts
§Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts.
Reprints: Andrew T. Chan, MD, MPH, Gastrointestinal Unit, Massachusetts General Hospital, 55 Fruit Street, GRJ-728A, Boston, MA 02114 (e-mail: email@example.com).
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.ibdjournal.org).
Funded by R01 CA137178, R01 CA050385, P01 CA87969, P30 DK043351, K08 DK064256, and the Broad Medical Research Program of the Broad Foundation. A. T. Chan is a Damon Runyon Cancer Research Foundation Clinical Investigator. H. Khalili is supported by a career development award from the Crohn’s and Colitis Foundation of American and IBD Working Group. L. M. Higuchi is supported by the National Institute of Diabetes and Digestive and Kidney Diseases (K08 DK064256). A. N. Ananthakrishnan is supported by a Research Scholars Award from the American Gastroenterological Association. A. T. Chan has served as a consultant for Bayer Healthcare, Millennium Pharmaceuticals, and Pfizer Inc.
Received September 17, 2012
Accepted September 22, 2012
The key pathogenic mechanism underlying inflammatory bowel disease (IBD) is an inappropriate immune response to the intestinal microbiome or loss of tolerance in a genetically susceptible host.1 Previous epidemiological studies have shown that a number of early life factors posited to influence development of immune tolerance, including breastfeeding and birth weight, may be potential risk factors for development of adult-onset autoimmune diseases.2–5 In addition, recent studies have illustrated the impact of early life factors in defining the composition of intestinal microbiome.6–8 Although the exact role of the intestinal microbiome in the pathogenesis of IBD remains poorly defined, its composition in genetically predisposed patients may influence the risk of development of IBD.
Limited epidemiological data suggest a link between having been breastfed and risk of developing IBD.9 However, many of these studies lack information on duration of having been breastfed or failed to confirm diagnoses of IBD through rigorous medical record review.10–12 In addition, a number of these studies were underpowered leading to inconclusive results.13,14 One previous study found an association between preterm birth and incident IBD but neither the exposure information nor the diagnoses were validated.15 We therefore sought to examine the association between early life factors, including having been breastfed, preterm birth, and birth weight, and risk of incident ulcerative colitis (UC) and Crohn’s disease (CD) during adulthood in 2 large prospective cohorts of U.S. women, the Nurses Health Study (NHS) I and NHS II, in which validated information about early life factors was collected. These cohorts offered us a unique opportunity to examine early life factors in the context of important lifestyle risk factors that may either confound or modify its association with UC and CD.
The NHS I is a prospective cohort that began in 1976 when 121,700 U.S. female registered nurses, aged 30 to 55 years, completed a mailed health questionnaire. Follow-up questionnaires are mailed every 2 years to update health information. In 1989, a parallel cohort, the NHS II, enrolled 116,686 U.S. female nurses between the ages of 25 to 42. These women have been followed with similar biennial questionnaires. Follow-up for the current study exceeded 90%. The institutional review board at the Brigham and Women’s Hospital approved this study.
Assessment of Early Life Factors
In the 1991 NHS II and 1992 NHS I questionnaires, participants were asked whether they had been breastfed (response categories: no, yes, not sure) and for how long (response categories: no, <3, 4–8, >9 months, not sure). In a subsample of NHS II participants, Troy et al16 validated the self-report of being breastfed against mother’s report with 82% sensitivity and 86% specificity. The correlation between mother’s and daughter’s reports of breastfeeding duration was 0.74.
In the 1991 NHS II and 1992 NHS I questionnaires, participants were asked to report whether they were born 2 or more weeks premature. When self-reports of preterm birth were compared with similar data collected from a sample of mothers of the participants, 90% agreement was found.17
In the 1991 NHS II and 1992 NHS I questionnaires, participants were asked to report their own birth weight in categories (<2.26, 2.3–2.49, 2.5–3.175, 3.2–3.85, 3.9–4.54, >4.54 kg). In a validation study in NHS II, birth weight was obtained from state birth certificates for 220 randomly chosen female nurses from the NHS II and divided into 5 categories. Actual birth weight was then compared with self-reported birth weight. Seventy percent of nurses reported the same birth weight category as documented on their birth certificate. Overall, the spearman correlation coefficient was 0.74, demonstrating substantial correlation.16
We limited our analysis to individuals who provided information on early life factors on the 1991 NHS II and 1992 NHS I questionnaires. The baseline characteristics of individuals who did not provide such information were similar to participants who provided information on early life factors (for NHS I, mean age: 41.9 versus 43.6, mean body mass index (BMI): 23.4 versus 23.6, ever-smoking: 55% versus 56%; for NHS II, mean age: 34.7 versus 35.0, mean BMI: 24.1 versus 24.1, ever-smoking: 34% versus 36%). Similarly, the incidence of CD and UC during follow-up was similar in both groups.
Race and ethnicity was assessed in 1992 in the NHS I and in 1989 in the NHS II and categorized as white, African, or Hispanic origin. Information on cigarette smoking, weight, and use of oral contraceptives were collected every 2 years. BMI was computed using weight in kilograms divided by height in square meters as reported at baseline. Participants’ self-report of body weight, height, and use of oral contraceptives has been previously validated.18,19 Data on maternal history of smoking during pregnancy were also collected in both cohorts. We did not specifically collect information about gestational diabetes; however, we did use data on history of diabetes in the participant’s mother in our multivariate models.
We have previously detailed our methods for confirming cases of CD and UC.20 In brief, since 1976 in NHS I and 1989 in NHS II, participants have reported diagnoses of UC or CD through an open-ended response on biennial surveys. In NHS I, we have specifically queried participants about diagnoses of UC since 1982 and CD since 1992. In NHS II, we have specifically queried participants about diagnoses of both UC and CD since 1993. When a diagnosis was reported on any biennial questionnaire, a supplementary questionnaire and related medical records were requested and reviewed by 2 gastroenterologists blinded to exposure information. We excluded participants who subsequently denied the diagnosis on the supplementary questionnaire or permission to review their records. Data were extracted on diagnostic tests, histopathology, anatomical location of disease, and disease behavior. Using standardized criteria,21–24 UC diagnosis was based on a typical clinical presentation ≥4 weeks and endoscopic or surgical pathologic specimen consistent with UC (e.g., evidence of chronicity). CD diagnosis was based on a typical clinical history for ≥4 weeks and endoscopy or radiologic evaluation demonstrating small bowel findings or surgical findings consistent with CD combined with pathology suggesting transmural inflammation or granuloma contributed to a diagnosis of CD. Disagreements were resolved through consensus. Among those women from whom we received adequate medical records, the case confirmation rate for IBD was 78% in NHS I and 74% in NHS II.20
Person-time for each participant was calculated from the date of return of baseline questionnaire to the date of the diagnosis of UC or CD, date of last questionnaire, death from any cause, or June 1, 2008 for NHS I and June 1, 2007 for NHS II, whichever came first. We used Cox proportional hazards modeling with time-varying covariates to adjust for other known or suspected risk factors before each 2-year interval to calculate adjusted hazard ratios (HR) and 95% confidence interval (CIs). Because weight may be influenced by preclinical disease, we adjusted for BMI using the baseline value, consistent with previous analyses.25 We observed no heterogeneity in the association of any of the specified early life factors with CD or UC in separate analyses of NHS I and NHS II (all P for heterogeneity >0.50 for both UC and CD). Thus, we pooled individual-level data from NHS I and NHS II and adjusted for cohort membership in all analyses. We also examined the association between early life factors and risk of UC or CD according to strata of history of diabetes in the mother and maternal smoking during pregnancy and evaluated for potential interaction using cross-classified categories of these risk factors and early life factors. We tested the significance of interactions using the log likelihood ratio test comparing the model with cross-classified categories with a model that included these factors as independent variables. We used SAS version 9.1.3 (Cary, NC) for these analyses. All P values were 2 sided, and <0.05 was considered statistically significant.
Among 146,681 women (60,186 in NHS I and 86,495 in NHS II) who provided information on early life factors, we documented 248 incident cases of CD and 304 incident cases of UC through 2007 in NHS II and 2008 in NHS I over 3,373,726 total person-years of follow-up. The baseline characteristics according to whether they were breastfed in early life are shown in Table 1. Compared with women who were not breastfed, women who were breastfed were older, less likely to be born preterm, had low birth weight, had a mother who smoked during pregnancy, or used oral contraceptives.
We did not observe a significant association between having been breastfed and risk of UC or CD (Table 2). Compared with women who were not breastfed, women who were breastfed had an age-adjusted HR of 0.99 (95% CI, 0.76–1.29) for CD and 1.03 (95% CI, 0.81–1.31) for UC. These estimates were not significantly altered after adjusting for potential confounders including smoking, BMI, birth weight, preterm birth, history of diabetes in the mother, maternal history of smoking during pregnancy, and oral contraceptive use. The multivariate-adjusted HRs were 0.99 (95% CI, 0.76–1.30) for CD and 1.03 (95% CI, 0.81–1.32) for UC. We also evaluated the effect of duration of being breastfed on subsequent risk of UC and CD (Table 3). Compared with women who were not breastfed, the multivariate-adjusted HR of CD was 1.26 (95% CI, 0.83–1.93) for women who were breastfed for ≤3 months, 0.94 (95% CI, 0.66–1.32) for women who were breastfed for 4 to 8 months, and 1.02 (95% CI, 0.63–1.66) for women who were breastfed for ≥9 months. Similarly for UC, the multivariate-adjusted HRs were 1.09 (95% CI, 0.73–1.72) for women who were breastfed for ≤3 months, 0.97 (95% CI, 0.65–1.46) for women who were breastfed for 4 to 8 months, and 0.90 (95% CI, 0.56–1.46) for women who were breastfed for ≥9 months.
Preterm Birth and Birth Weight
Similar to breastfeeding, we did not observe a statistically significant association between having been born preterm and risk of UC or CD (Table 4). The multivariate-adjusted HRs were 0.69 (95% CI, 0.37–1.29) for CD and 0.82 (95% CI, 0.46–1.43) for UC. We also evaluated the association between birth weight and risk of UC and CD (Table 5). Compared with women with normal birth weight, women with low birth weight had a multivariate-adjusted HR of 1.29 (95% CI, 0.79–2.11) for CD and 0.91 (95% CI, 0.55–1.49) for UC. Similarly, women with high birth weight (>4.54 kg) did not have an increased risk of developing UC or CD. The multivariate-adjusted HRs were 1.06 (95% CI, 0.71–1.58) for CD and 1.02 (95% CI, 0.71–1.47) for UC.
We considered the possibility that history of diabetes in the mother or smoking during pregnancy may modify the effect of these early life factors on risk of UC and CD in stratified analyses according to these potential effect modifiers. The effect of breastfeeding, preterm birth, or birth weight on risk of UC or CD did not seem to be significantly different within any strata according to history of diabetes in the mother or smoking during pregnancy (all Pinteraction > 0.10). We also assessed whether these potential effect modifiers are independently associated with risk of UC or CD. Although maternal smoking during pregnancy was not associated with risk of UC or CD, history of diabetes in the mother was associated with an increased risk of CD (multivariate-adjusted HR = 1.64, 95% CI, 1.15–2.33) but not UC (see Table, Supplemental Digital Content 1, http://links.lww.com/IBD/A75).
Because a diagnosis of UC or CD could influence recall of early life factors, we performed sensitivity analyses limiting our follow-up to the time period after the first exposure ascertainment (in 1991 in NHS II and in 1992 in NHS I). Compared with women who were not breastfed, women who were breastfed had a multivariate-adjusted HRs of 0.93 (95% CI, 0.69–1.27) for CD and 1.04 (95% CI, 0.79–1.36) for UC. Similarly, in these sensitivity analyses, there was no association between preterm birth or birth weight and risk of UC or CD (data not shown).
The composition of an infant’s intestinal flora is determined through exposure to bacteria from the mother and the environment. Although the exact role of intestinal microbiome in the pathogenesis of IBD remains poorly defined, its specific make up and/or alteration in its composition in genetically predisposed patients may influence the risk of development of IBD. Studies have shown distinct patterns of microbial consortium in gastrointestinal tracts of preterm birth,8 extremely low birth weight,7 and infants who have been breastfed.26 Taken together, these findings may suggest a plausible biological mechanism linking events in early life to development of later-onset IBD. However, in 2 large prospective cohorts of U.S. women, we did not observe an association between having been breastfed, preterm birth, birth weight, and risk of UC and CD. We did observe an association between history of diabetes in the mother and risk of CD. However, it is unclear whether gestational diabetes is an early life factor that also may be associated with CD; future studies are needed to investigate the association between gestational diabetes and risk of CD or UC in their offspring.
A systematic review and meta-analysis of previous studies observed that breastfeeding is associated with a lower risk of both UC and CD.9 Our results may contrast with these findings because we examined risk of incident CD and UC in adulthood (after age 39), whereas the majority of the studies included in previous meta-analyses were composed of pediatric or childhood IBD cases.12,13,27–31 Genome-wide studies32,33 have shown that genetic susceptibility loci for IBD are largely consistent according to age of diagnosis. However, cases of IBD diagnosed in childhood or adolescence have distinct phenotypic features and natural history compared with cases of IBD diagnosed in adulthood, suggesting the likelihood of divergent environmental risk factors that may influence risk of early-onset IBD compared with adult-onset IBD. Alternatively, it is possible that the impact of early life factors on disease pathogenesis in a genetically susceptible individual may become apparent only within childhood or adolescence, which would dilute the ability to detect associations with incident disease diagnosed in adulthood.
Our study has several notable strengths. First, our prospective study design avoids the potential recall and selection biases of retrospective case–control studies, which collect data on diet and lifestyle after diagnosis of CD or UC. Second, we confirmed all cases of CD and UC through medical record review, a significant advantage over studies that rely on self-report or discharge codes, which may not accurately reflect true diagnoses. Third, the availability of detailed and validated information on BMI, previous oral contraceptive use, smoking, and other important early life factors allowed us to control for a number of potential confounding factors that may have influenced our observed associations. Fourth, we used validated data on early life factors.
We acknowledge several limitations. First, our cohorts, which consist of female health professionals, may not be representative of the overall U.S. population. However, as we have previously reported,20 our age-specific incidence of CD and UC are largely similar to rates from other U.S. populations. In addition, previous studies have shown that the prevalence of risk factors, such as smoking and BMI, in our cohort is consistent with those of the broader population of U.S. women.34,35 Second, although validation studies have demonstrated that data on early life factors reported by these health professionals are highly accurate, we cannot exclude the possibility of exposure misclassification. Inaccurate reporting of breastfeeding duration or birth weight could have biased our study against finding significant associations. Third, our definition of preterm birth in our questionnaires was delivery 2 or more weeks preterm. However, others define preterm birth as delivery 3 or more weeks preterm.36 Thus, it is possible that alternative definitions of preterm birth, representing more significant prematurity, may be associated with risk. Fourth, the number of cases in some of our exposure categories was relatively small, limiting the precision of our risk estimates. Moreover, we are unable to exclude the possibility of a more modest association between these early life factors and risk of adult-onset IBD. Finally, we do not have detailed information about other potential early life factors, including mode of delivery and early exposure to antibiotics, which may also influence risk of IBD.
In conclusion, in 2 large prospective cohorts of U.S. women, we did not find a significant association between early life factors including having been breastfed, preterm birth, birth weight, and risk of UC and CD in adulthood.
Author contribution: H. Khalili: study concept and design; acquisition of data; analysis and interpretation of data; drafting of the manuscript; and statistical analysis. A. N. Ananthakrishnan: acquisition of data and critical revision of the manuscript. L. M. Higuchi: acquisition of data; analysis and interpretation of data; and critical revision of the manuscript for important intellectual content. J. M. Richter: study concept and design; acquisition of data; analysis and interpretation of data; and critical revision of the manuscript. C. S. Fuchs: acquisition of data; and critical revision of the manuscript. A. T. Chan: study concept and design; analysis and interpretation of data; drafting of the manuscript; and critical revision of the manuscript.
1. Abraham C, Cho JH. Inflammatory bowel disease. N Engl J Med. 2009;361:2066–2078.
2. Jacobsson LT, Jacobsson ME, Askling J, et al.. Perinatal characteristics and risk of rheumatoid arthritis. BMJ. 2003;326:1068–1069.
3. Mandl LA, Costenbader KH, Simard JF, et al.. Is birthweight associated with risk of rheumatoid arthritis? Data from a large cohort study. Ann Rheum Dis. 2009;68:514–518.
4. Simard JF, Costenbader KH, Hernan MA, et al.. Early life factors and adult-onset rheumatoid arthritis. J Rheumatol. 2010;37:32–37.
5. Young KA, Parrish LA, Zerbe GO, et al.. Perinatal and early childhood risk factors associated with rheumatoid factor positivity in a healthy paediatric population. Ann Rheum Dis. 2007;66:179–183.
6. Vael C, Desager K. The importance of the development of the intestinal microbiota in infancy. Curr Opin Pediatr. 2009;21:794–800.
7. LaTuga MS, Ellis JC, Cotton CM, et al.. Beyond bacteria: a study of the enteric microbial consortium in extremely low birth weight infants. PLoS One. 2011;6:e27858.
8. Mai V, Young CM, Ukhanova M, et al.. Fecal microbiota in premature infants prior to necrotizing enterocolitis. PLoS One. 2011;6:e20647.
9. Klement E, Cohen RV, Boxman J, et al.. Breastfeeding and risk of inflammatory bowel disease: a systematic review with meta-analysis. Am J Clin Nutr. 2004;80:1342–1352.
10. Bergstrand O, Hellers G. Breast-feeding during infancy in patients who later develop Crohn's disease. Scand J Gastroenterol. 1983;18:903–906.
11. Corrao G, Tragnone A, Caprilli R, et al.. Risk of inflammatory bowel disease attributable to smoking, oral contraception and breastfeeding in Italy: a nationwide case-control study. Cooperative Investigators of the Italian Group for the Study of the Colon and the Rectum (GISC). Int J Epidemiol. 1998;27:397–404.
12. Ekbom A, Adami HO, Helmick CG, et al.. Perinatal risk factors for inflammatory bowel disease: a case-control study. Am J Epidemiol. 1990;132:1111–1119.
13. Gilat T, Hacohen D, Lilos P, et al.. Childhood factors in ulcerative colitis and Crohn's disease. An international cooperative study. Scand J Gastroenterol. 1987;22:1009–1024.
14. Gruber M, Marshall JR, Zielezny M, et al.. A case-control study to examine the influence of maternal perinatal behaviors on the incidence of Crohn's disease. Gastroenterol Nurs. 1996;19:53–59.
15. Sonntag B, Stolze B, Heinecke A, et al.. Preterm birth but not mode of delivery is associated with an increased risk of developing inflammatory bowel disease later in life. Inflamm Bowel Dis. 2007;13:1385–1390.
16. Troy LM, Michels KB, Hunter DJ, et al.. Self-reported birthweight and history of having been breastfed among younger women: an assessment of validity. Int J Epidemiol. 1996;25:122–127.
17. Gardener H, Munger KL, Chitnis T, et al.. Prenatal and perinatal factors and risk of multiple sclerosis. Epidemiology. 2009;20:611–618.
18. Troy LM, Hunter DJ, Manson JE, et al.. The validity of recalled weight among younger women. Int J Obes Relat Metab Disord. 1995;19:570–572.
19. Hunter DJ, Manson JE, Colditz GA, et al.. Reproducibility of oral contraceptive histories and validity of hormone composition reported in a cohort of US women. Contraception. 1997;56:373–378.
20. Khalili H, Huang ES, Ananthakrishnan AN, et al.. Geographical variation and incidence of inflammatory bowel disease among US women. Gut. 2012;61:1686–1692.
21. Loftus EV Jr, Silverstein MD, Sandborn WJ, et al.. Crohn's disease in Olmsted County, Minnesota, 1940-1993: incidence, prevalence, and survival. Gastroenterology. 1998;114:1161–1168.
22. Loftus EV Jr, Silverstein MD, Sandborn WJ, et al.. Ulcerative colitis in Olmsted County, Minnesota, 1940-1993: incidence, prevalence, and survival. Gut. 2000;46:336–343.
23. Fonager K, Sorensen HT, Rasmussen SN, et al.. Assessment of the diagnoses of Crohn's disease and ulcerative colitis in a Danish hospital information system. Scand J Gastroenterol. 1996;31:154–159.
24. Moum B, Vatn MH, Ekbom A, et al.. Incidence of inflammatory bowel disease in southeastern Norway: evaluation of methods after 1 year of registration. Southeastern Norway IBD Study Group of Gastroenterologists. Digestion. 1995;56:377–381.
25. Hu FB, Willett WC, Li T, et al.. Adiposity as compared with physical activity in predicting mortality among women. New Engl J Med. 2004;351:2694–2703.
26. Palmer C, Bik EM, DiGiulio DB, et al.. Development of the human infant intestinal microbiota. PLoS Biol. 2007;5:e177.
27. Koletzko S, Griffiths A, Corey M, et al.. Infant feeding practices and ulcerative colitis in childhood. BMJ. 1991;302:1580–1581.
28. Rigas A, Rigas B, Glassman M, et al.. Breast-feeding and maternal smoking in the etiology of Crohn's disease and ulcerative colitis in childhood. Ann Epidemiol. 1993;3:387–392.
29. Dietary and other risk factors of ulcerative colitis. A case-control study in Japan. Epidemiology Group of the Research Committee of Inflammatory Bowel Disease in Japan. J Clin Gastroenterol. 1994;19:166–171.
30. Persson PG, Leijonmarck CE, Bernell O, et al.. Risk indicators for inflammatory bowel disease. Int J Epidemiol. 1993;22:268–272.
31. Klein I, Reif S, Farbstein H, et al.. Preillness non dietary factors and habits in inflammatory bowel disease. Ital J Gastroenterol Hepatol. 1998;30:247–251.
32. Kugathasan S, Baldassano RN, Bradfield JP, et al.. Loci on 20q13 and 21q22 are associated with pediatric-onset inflammatory bowel disease. Nat Genet. 2008;40:1211–1215.
33. Imielinski M, Baldassano RN, Griffiths A, et al.. Common variants at five new loci associated with early-onset inflammatory bowel disease. Nat Genet. 2009;41:1335–1340.
34. Sarna L, Bialous SA, Jun HJ, et al.. Smoking trends in the Nurses' Health Study (1976-2003). Nurs Res. 2008;57:374–382.
35. van Dam RM, Li T, Spiegelman D, et al.. Combined impact of lifestyle factors on mortality: prospective cohort study in US women. BMJ. 2008;337:a1440.
36. Beck S, Wojdyla D, Say L, et al.. The worldwide incidence of preterm birth: a systematic review of maternal mortality and morbidity. Bull World Health Organ. 2010;88:31–38.
inflammatory bowel disease; Crohn’s disease; ulcerative colitis; early life factors; breastfeeding; birth weight; prematurity; Nurses Health Study
Supplemental Digital Content
© Crohn's & Colitis Foundation of America, Inc.