Background: Although large genetic studies have mapped the variants driving rheumatoid arthritis (RA) susceptibility, the extent to which these same factors dictate clinical phenotype is not yet understood. Alongside HLA polymorphisms, genome-wide association studies (GWAS) have identified over 100 non-HLA variants contributing to disease risk. It remains to be established whether a high cumulative burden of these susceptibility variants simply increases the risk of developing RA or whether it actively influences disease severity.

Objectives: This study aimed to evaluate whether a high burden of non-HLA RA susceptibility variants was associated with an earlier disease onset and a more aggressive disease course, measured by radiographic damage.

Methods: Analysis was performed using the Norfolk Arthritis Register (NOAR). Participants in this inception cohort were followed prospectively for up to 10 years, with a total of 2,198 longitudinal radiographs of hands and feet, and were evaluated for presence of erosions. Patients who met either 2010 or 1987 ACR criteria were included (n=1540). A polygenic risk score (PRS) was constructed as a weighted sum of risk alleles from 143 validated non-HLA susceptibility loci, utilizing effect sizes derived from large-scale RA GWAS [1,2]. The PRS was standardized to a mean of 0 and a standard deviation (SD) of 1; consequently, all risk estimates are reported per SD increase in the genetic score. The association between the PRS and age of onset was tested using Cox Proportional Hazards modelling, treating age-at-onset as a continuous variable to calculate Hazard Ratios (HR), controlling for anti-CCP status and the shared epitope (SE). Association between PRS and development of erosions was modelled longitudinally using Generalized Estimating Equations (GEE). This approach accounted for repeated measures over time and adjusted for key clinical covariates, including disease duration, age of onset, and treatment.

Results: Survival analysis demonstrated that the PRS is associated with age of onset. Each standard deviation increase in the PRS was associated with a significantly increased hazard of disease onset (HR 1.06; 95% CI: 1.003–1.12; p = 0.04). This confirms that a higher polygenic burden is associated with earlier development of RA, shifting the age of diagnosis toward younger age groups. This effect was independent of anti-CCP status and the SE. Further analysis demonstrated the PRS was also significantly associated with radiographic erosions (OR 1.14 per SD; 95% CI 1.02–1.27; P=0.022). In multivariable analysis adjusting for anti-CCP and shared epitope status, the association between PRS and erosions remained stable in magnitude (OR 1.14 per SD) but was attenuated to borderline significance (P=0.056).

Conclusions: We observed that a non-HLA susceptibility PRS is associated with earlier RA onset and modest joint damage accumulation. However, the modest effect size limits its utility as a standalone predictor. Future work will include updating PRS to include new found susceptibility variants, as well as on integrating PRS with clinical variables to better inform patient prognosis.

REFERENCES: [1] Ishigaki et al, 2021, Nature Genetics. [2] Padyukov, 2022, Seminars in Immunopathology

Acknowledgments: NIL.

Disclosure of Interests: None declared.