Association of a data-driven nutrient risk score with incident rheumatoid arthritis in UK Biobank adults: a prospective cohort study.

BACKGROUND

Although specific nutrients have been linked to rheumatoid arthritis (RA), the impact of overall nutrient intake patterns remains unclear. Here, we aimed to explore the association between the nutrient risk score (NRS) and incident RA, as well as the modifying effect of genetic susceptibility and the potential mediating role of obesity.

METHODS

This prospective study included 161,725 participants from the UK Biobank. Nutrients were selected using a LASSO-Cox model. We used Cox proportional hazards models to investigate the association between NRS and the risk of RA. Generalized propensity score (GPS) and inverse probability weighting (IPW) methods were employed to control for potential confounding factors. Stratified analyses were conducted to assess effect heterogeneity across subgroups. Mediation analysis was used to estimate the mediating effect of obesity.

RESULTS

During a median follow-up of 13.4 years, 1,756 participants developed RA. Each unit increase in NRS was associated with a 96% higher risk of RA (HR: 1.96, 95% CI: 1.64-2.33) after adjusting for covariates. The association remained robust in analyses using GPS and IPW weighting. Stratified analyses showed a more pronounced association among individuals reporting short sleep duration (<7 hours) and low levels of physical activity. Furthermore, a gradient association between NRS and RA risk was observed across genetic risk strata. Specifically, compared with participants at low genetic risk with the lowest NRS (Q1), those at high genetic risk with the highest NRS (Q4) had nearly a threefold higher risk of incident RA (HR: 3.28, 95% CI: 2.49-4.32). In addition, several obesity indicators showed potential mediating effects, with the proportion mediated ranging from 3.81% to 13.57%.

CONCLUSION

Higher NRS was associated with an increased risk of incident RA, especially among individuals with high genetic susceptibility. Obesity partially mediated the association, highlighting NRS as a potential tool for RA risk stratification and intervention.