Comprehensive Association Analyses of Extraintestinal Manifestations in Inflammatory Bowel Disease

Background & Aims: Patients with inflammatory bowel disease (IBD) frequently develop extraintestinal manifestations (EIMs) that contribute substantially to morbidity. We assembled the largest multicohort data set to date to investigate the clinical, serologic, and genetic factors associated with EIM complications in IBD. Methods: Data were available in 12,083 unrelated European ancestry IBD cases with presence or absence of EIMs (eg, ankylosing spondylitis [ankylosing spondylitis and sacroiliitis], primary sclerosing cholangitis [PSC], peripheral arthritis, and skin and ocular manifestations) across 4 cohorts (Cedars-Sinai Medical Center, National Institute for Diabetes and Digestive and Kidney Diseases IBD Genetics Consortium, Sinai Helmsley Alliance for Research Excellence Consortium, and Risk Stratification and Identification of Immunogenetic and Microbial Markers of Rapid Disease Progression in Children with Crohn's Disease cohort). Clinical and serologic parameters were analyzed by means of univariable and multivariable regression analyses using a mixed-effects model. Within-case logistic regression was performed to assess genetic associations. Results: Most EIMs occurred more commonly in female subjects (overall EIM: P = 9.0E-05, odds ratio [OR], 1.2; 95% CI, 1.1–1.4), with CD (especially colonic disease location; P = 9.8E-09, OR, 1.7; 95% CI, 1.4–2.0), and in subjects who required surgery (both CD and UC; P = 3.6E-19, OR, 1.7; 95% CI, 1.5–1.9). Smoking increased risk of EIMs except for PSC, where there was a “protective” effect. Multiple serologic associations were observed, including with PSC (anti-nuclear cytoplasmic antibody; IgG and IgA, anti-Saccharomyces cerevisiae antibodies; and anti-flagellin) and any EIM (anti-nuclear cytoplasmic antibody; IgG and IgA, anti-Saccharomyces cerevisiae antibodies; and anti–Pseudomonas fluorescens-associated sequence). We identified genome-wide significant associations within major histocompatibility complex (ankylosing spondylitis and sacroiliitis, P = 1.4E-15; OR, 2.5; 95% CI, 2.0–3.1; PSC, P = 2.7E-10; OR, 2.8; 95% CI, 2.0–3.8; ocular, P = 2E-08, OR, 3.6; 95% CI, 2.3–5.6; and overall EIM, P = 8.4E-09; OR, 2.2; 95% CI, 1.7–2.9) and CPEB4 (skin, P = 2.7E-08; OR, 1.5; 95% CI, 1.3–1.8). Genetic associations implicated tumor necrosis factor, JAK-STAT, and IL6 as potential targets for EIMs. Contrary to previous reports, only 2% of our subjects had multiple EIMs and most co-occurrences were negatively correlated. Conclusions: We have identified demographic, clinical, and genetic associations with EIMs that revealed underlying mechanisms and implicated novel and existing drug targets—important steps toward a more personalized approach to IBD management.