Abstract
Polygenic risk scores (PRS) predict inflammatory bowel disease (IBD) susceptibility, but their associations with outcomes after diagnosis are uncertain. We studied 5,021 UK Biobank participants with Crohn's disease (CD; n = 1,427) or ulcerative colitis (UC; n = 3,594). Cox-model estimates for disease-specific standardized PRS were pooled across 20 imputed datasets. A clinically extended model included disease duration, log-transformed C-reactive protein, previous intestinal surgery, and baseline aminosalicylate, corticosteroid, and immunomodulator use in addition to demographic and lifestyle covariates. Benjamini-Hochberg false-discovery rate (FDR) correction covered six disease-outcome tests. In CD, each 1-SD higher PRS was associated with intestinal surgery (HR = 1.61, 95% CI 1.31-1.99; P < 0.001; FDR < 0.001); the high versus low tertile estimate was HR = 2.70 (95% CI 1.58-4.62; FDR = 0.004). CD PRS was not associated with gastrointestinal cancer or mortality. In UC, associations with surgery (P = 0.482), gastrointestinal cancer (HR = 1.21, 95% CI 0.99-1.48; P = 0.065; FDR = 0.130), and mortality (HR = 1.10, 95% CI 1.00-1.22; P = 0.058; FDR = 0.130) were nonsignificant. The CD surgery association persisted among participants without previous surgery (HR = 1.62, 95% CI 1.29-2.04; P < 0.001). These genetic epidemiology associations do not establish incremental prediction or clinical utility and require external validation.</p>