Background: A method for predicting ulcerative colitis (UC) onset has not been established. Serum autoantibodies have been suggested as potential predictive biomarkers for UC onset. We aimed to validate the risks associated with serological and environmental factors and construct a model for predicting UC development.

Methods: Using the population-based cohort studies (n > 83,000), we identified 42 individuals who were diagnosed with UC later in life and compared them with matched healthy controls. We analyzed serum anti-integrin αvβ6 antibody (anti-αvβ6) and anti-endothelial protein C receptor antibody (anti-EPCR) titers, and lifestyle and dietary habits to explore UC onset predictors. The predictive performance of the models was evaluated based on these predictors.

Results: The sensitivity and specificity of anti-EPCR for predicting UC onset were 51.4% and 97.8%, respectively, comparable to those of anti-αvβ6 (52.5% and 97.6%, respectively). The proportion of individuals with insomnia was significantly higher in the preclinical UC group (adjusted odds ratio = 2.14, 95% confidence interval [CI] 1.11-4.04, p = 0.019). The predictive performance of anti-EPCR alone was high with an area under the curve (AUC) of 0.89 (95%CI 0.83-0.96), and that of anti-EPCR combined with anti-αvβ6 was even better with an AUC of 0.92 (95%CI 0.87-0.97); the lifestyle model had lower predictive accuracy (AUC = 0.65, 95%CI 0.55-0.74).

Conclusions: Anti-EPCR and anti-αvβ6 each strongly predict UC onset. The combined anti-EPCR and anti-αvβ6 model had stronger predictive performance than the single models.