Risk Factors for Incident Inflammatory Bowel Disease According to Disease Phenotype

Background & Aims We examined whether relationships between known risk factors for Crohn’s disease (CD) and ulcerative colitis (UC) differ according to disease phenotype, defined by Montreal classification, at the time of diagnosis. Methods We performed a prospective cohort study of 208,070 adults from the Nurses’ Health Study (NHS), NHSII, and Health Professionals Follow-Up Study (HPFS). Dietary, lifestyle, and medical data were obtained at baseline and every 2–4 years. We confirmed cases of inflammatory bowel disease (IBD) and their phenotypes via medical record review. We tested for heterogeneity across CD subtypes using the likelihood ratio test and for linear heterogeneity across UC subtypes using the meta-regression method. Results We ascertained 346 cases of CD and 456 cases of UC over 5,117,021 person-years of follow-up (1986–2016 for NHS and HPFS; 1991–2017 for NHSII). Fiber intake was associated with decreased risk for ileocolonic but not ileal or colonic CD (Pheterogeneity = .04). Physical activity was associated with decreased risk of nonstricturing and nonpenetrating CD but not of penetrating CD (Pheterogeneity = .02). Higher body mass index and current smoking were associated with decreased risk of proctitis and left-sided UC but not of pan-UC (Plinear heterogeneity= .004 and .02, respectively). The associations between other risk factors examined and risk of CD and UC did not differ by disease phenotype (all Pheterogeneity > .06). Conclusions In 3 large prospective cohorts, we observed that dietary and lifestyle risk factors for IBD may differ according to disease phenotype. These findings highlight the need for disease stratification in future epidemiologic studies. We examined whether relationships between known risk factors for Crohn’s disease (CD) and ulcerative colitis (UC) differ according to disease phenotype, defined by Montreal classification, at the time of diagnosis. We performed a prospective cohort study of 208,070 adults from the Nurses’ Health Study (NHS), NHSII, and Health Professionals Follow-Up Study (HPFS). Dietary, lifestyle, and medical data were obtained at baseline and every 2–4 years. We confirmed cases of inflammatory bowel disease (IBD) and their phenotypes via medical record review. We tested for heterogeneity across CD subtypes using the likelihood ratio test and for linear heterogeneity across UC subtypes using the meta-regression method. We ascertained 346 cases of CD and 456 cases of UC over 5,117,021 person-years of follow-up (1986–2016 for NHS and HPFS; 1991–2017 for NHSII). Fiber intake was associated with decreased risk for ileocolonic but not ileal or colonic CD (Pheterogeneity = .04). Physical activity was associated with decreased risk of nonstricturing and nonpenetrating CD but not of penetrating CD (Pheterogeneity = .02). Higher body mass index and current smoking were associated with decreased risk of proctitis and left-sided UC but not of pan-UC (Plinear heterogeneity= .004 and .02, respectively). The associations between other risk factors examined and risk of CD and UC did not differ by disease phenotype (all Pheterogeneity > .06). In 3 large prospective cohorts, we observed that dietary and lifestyle risk factors for IBD may differ according to disease phenotype. These findings highlight the need for disease stratification in future epidemiologic studies.