The effect of non–shared epitope HLA–DRB1 alleles on rheumatoid arthritis (RA) is poorly understood. This study was undertaken to investigate the effects of several HLA–DRB1 alleles, independent of the shared epitope, on the risk of developing anti–citrullinated protein antibody (ACPA)–positive or ACPA‐negative RA in a large case–control study.

HLA typing for the DRB1 gene was performed in 1,352 patients with RA and 922 controls from the Swedish Epidemiological Investigation of Rheumatoid Arthritis study. Relative risks (RRs) and 95% confidence intervals (95% CIs) were calculated.

DRB1*13 was found to protect against ACPA‐positive RA when stratifying for the shared epitope and using a dominant genetic model (RR 0.41 [95% CI 0.26–0.64]). Furthermore, DRB1*13 neutralized the effect of the shared epitope in ACPA‐positive RA (RR 3.91 [95% CI 3.04–5.02] in patients who had the shared epitope but not DRB1*13, and RR 1.22 [95% CI 0.81–1.83] in patients with both the shared epitope and DRB1*13, as compared with patients negative for both the shared epitope and DRB1*13). However, we did not replicate the previous published risk of ACPA‐negative RA conferred by DRB1*03 when a dominant genetic model was used (RR 1.29 [95% CI 0.91–1.82]). Similarly, no significant effect of DRB1*03 on RR for ACPA‐negative RA was seen using the recessive genetic model (RR 1.18 [95% CI 0.6–2.4]). In contrast, the combination of DRB1*03 and DRB1*13 was significantly associated with increased risk of developing ACPA‐negative RA (RR 2.07 [95% CI 1.17–3.67]).

Our findings indicate that the DRB1*13 allele plays a dual role in the development of RA, by protecting against ACPA‐positive RA but, in combination with DRB1*03, increasing the risk of ACPA‐negative RA.