Authors
H. LEE (1), L. HANNES (1), M. VANCAMELBEKE (2), V. BALLET (3), M. FERRANTE (4), S. VERMEIRE (4), I. CLEYNEN (1) / [1] KU Leuven, Leuven, Belgium, Laboratory for Complex Genetics, Department of Human Genetics, [2] KU Leuven, euven, Belgium, Department Chronic Diseases, Metabolism & Ageing (CHROMETA), Translational Research Center for Gastrointestinal Disorders (TARGID), [3] UZ Leuven, Leuven, Belgium, Department of Gastroenterology and Hepatology, [4] KU Leuven, Leuven, Belgium, Department Chronic Diseases, Metabolism & Ageing (CHROMETA), Translational Research Center for Gastrointestinal Disorders (TARGID)
Introduction
Family history of inflammatory bowel disease (IBD) is the strongest risk factor for IBD. There currently, however, is limited understanding of the contribution of genetic risk scores to familial aggregation of IBD.
Aim
We aimed to evaluate the association between the IBD polygenic risk score (PRS) and familial IBD, and determine its contribution to familial IBD.
Methods
We included 54 multiple-affected families (≥ 3 first-degree relatives affected) of European ancestry, including 189 affected IBD patients (156 Crohn’s disease; 33 ulcerative colitis), and 133 unaffected relatives. For all individuals, Immunochip genotypes were available. Weighted PRSs with estimates derived from literature were calculated using PRSice-2.0, including clumping and different p-value thresholds (pT) to select which variants to include in the score. Explained variance (Nagelkerke pseudo-R²) was calculated across different pTs. To account for possible intra-familial correlations, the association between PRS and familial IBD was evaluated in age- and sex-adjusted generalized mixed regression models including family as random effects. Sporadic cases (n=1768) and non-IBD controls (n=868) with Immunochip genotypes were used for comparison.
Results
Using pT=0.05 for PRS calculation, we found that affected relatives had a higher PRS than unaffected relatives (P=1.00x10-02), sporadic cases (P=4.58x10-02), and non-IBD controls (P<2.20x10-16). The risk of disease in families increased by 1.23-fold (95% confidence interval (CI) 1.21–1.24) for every incremental standard deviation in PRS. Individuals in the highest quartile had a 3.45-times higher risk of IBD (95% CI 1.77–6.72) than those in the lowest quartile. However, the proportion of the explained variance between affected and unaffected family members was smaller than that of sporadic IBD and non-IBD controls; and the best pT was different for familial or sporadic IBD. In familial IBD, the best-fit PRS was at pT=6.90x10-03 and explained 5.3% (P=3.07x10-04) of variance, whereas in sporadic IBD, the best-fit PRS was at pT=0.08 and explained 16.7% (P=8.48x10-63). For sporadic IBD, a typical increase in the proportion of variance explained was seen with more liberal p-value thresholds, and levelling off at ~pT=0.1. What was striking however, was that in familial IBD this additive genetic variance was only observed until pT=0.01, after which explained variance dropped dramatically.
Conclusions
Higher IBD polygenic risk increases the risk for familial IBD as it does for sporadic IBD. In sporadic IBD, the increased risk is defined by variants of all p-value levels (until ~pT=0.1). In familial IBD, the difference between those affected or not is in the higher-effect variants (p-value<0.01).
