Socioeconomic status and long-term risk of inflammatory bowel disease: a prospective cohort study with mediation analysis

Background: The association of socioeconomic status with incident inflammatory bowel disease and the contribution of lifestyle factors remain unclear. We aimed to comprehensively investigate the association between socioeconomic status (SES) and the risk of inflammatory bowel disease (IBD), and to assess the mediating effect of lifestyle factors in a prospective cohort. Methods: We assessed SES using household income, education level, and employment status, and categorised participants into high-, medium-, and low-SES groups using latent profile analysis. We combined lifestyle factors, including smoking, alcohol consumption, diet, physical activity, body mass index, and sleep duration, to calculate a healthy lifestyle score as a mediator. The primary outcome was incident IBD. We used Cox proportional hazards models and mediation analysis to examine associations between SES and incident IBD and to assess the mediation effect. Results: Participants free of IBD at baseline were included (n = 374,507), with a mean age of 55.84 years, of whom 51.5% were female. Overall, 76,292 (20.4%) were classified as high, 197,130 (52.6%) as medium, and 101,085 (27.0%) as low patterns. Over 14.57 years of follow-up, we identified 2381 incident IBD cases, 1703 ulcerative colitis (UC) and 843 Crohn's disease (CD), with an overall incidence rate of 44.9 per 100,000 person-years. After multivariable adjustment, a high SES pattern was associated with a 45% lower IBD risk compared to low SES (hazard ratio (HR) = 0.55; 95% confidence interval (CI) = 0.48-0.63). Meanwhile, all SES components showed beneficial effects, with higher education level demonstrating the strongest association (HR = 0.58; 95% CI = 0.51-0.65). The combination of high SES and healthy lifestyle conferred the greatest protection (HR = 0.37; 95% CI = 0.29-0.46). Mediation analyses indicated that lifestyle factors explained 11.34-16.50% of the observed association between SES and incident IBD. Similar protective associations were observed for both UC and CD. Conclusions: Healthy lifestyle factors partially explained the association between SES and incident IBD. However, these findings should be interpreted cautiously given the observational design, potential residual confounding, and assumption-dependent mediation estimates.