Objective: To investigate the relationship between metabolomic profiles, genome-wide polygenic risk scores (PRSs) and risk of rheumatoid arthritis (RA).
Methods: 143 nuclear magnetic resonance-based plasma metabolic biomarkers were measured among 93 800 participants in the UK Biobank. The Cox regression model was used to assess the associations between these metabolic biomarkers and RA risk, and genetic correlation and Mendelian randomisation analyses were performed to reveal their causal relationships. Subsequently, a metabolic risk score (MRS) comprised of the weighted sum of 17 clinically validated metabolic markers was constructed. A PRS was derived by assigning weights to genetic variants that exhibited significant associations with RA at a genome-wide level.
Results: A total of 620 incident RA cases were recorded during a median follow-up time of 8.2 years. We determined that 30 metabolic biomarkers were potentially associated with RA, while no further significant causal associations were found. Individuals in the top decile of MRS had an increased risk of RA (HR 3.52, 95% CI: 2.80 to 4.43) compared with those below the median of MRS. Further, significant gradient associations between MRS and RA risk were observed across genetic risk strata. Specifically, compared with the low genetic risk and favourable MRS group, the risk of incident RA in the high genetic risk and unfavourable MRS group has almost elevated by fivefold (HR 6.10, 95% CI: 4.06 to 9.14).
Conclusion: Our findings suggested the metabolic profiles comprising multiple metabolic biomarkers contribute to capturing an elevated risk of RA, and the integration of genome-wide PRSs further improved risk stratification.
Keywords: Arthritis, Rheumatoid; Autoimmune Diseases; Epidemiology.
© Author(s) (or their employer(s)) 2023. Re-use permitted under CC BY. Published by BMJ.