Burden of Disease and Drug Response for Patients With Rheumatoid Arthritis by Shared Epitope and Anticitrullinated Protein Antibody Status

Objective To characterize disease burden among adults with rheumatoid arthritis (RA) by both shared epitope (SE) and anticitrullinated protein antibody (ACPA) status, and to determine how their responses to abatacept, tumor necrosis inhibitors, and Janus kinase inhibitors may differ.

Methods Using data from 2 observational cohorts (FORWARD, Veterans Affairs RA [VARA]), individuals with RA were classified by SE/ACPA status. Outcomes included disease activity (Patient Activity Scale II [PAS-II], Routine Assessment of Patient Index Data 3 [RAPID3], Disease Activity Score in 28 joints), Rheumatic Disease Comorbidity Index (RDCI), lifetime disease-modifying antirheumatic drug (DMARD) exposure, and healthcare usage. Differences by SE/ACPA classification were determined with multiple linear regression. Response to DMARD initiation was assessed with linear regression for continuous measures of disease activity and logistic regression for achieving a change as large as the minimum clinically important difference.

Results A total of 3243 individuals were included (FORWARD, n = 917; VARA n = 2326). RDCI among ACPA-negative individuals was lower than in ACPA-positive individuals in both cohorts (β [95% CI]: FORWARD −0.35 [−0.65 to −0.05], P = 0.02; VARA −0.35 [−0.63 to −0.08], P = 0.01). In FORWARD, there were significant differences in disease burden, including lower disease activity (PAS-II −0.77 [−1.10 to −0.44], P < 0.001), lower healthcare usage (rheumatology visits −0.18 [−0.35 to 0.0], P = 0.046), and higher DMARD counts (0.43 [0.02 to 0.85], P = 0.04) among SE+/ACPA+ individuals. ACPA+ abatacept initiators were more likely to experience clinically important improvements in PAS-II and DAS28, but RAPID3 was not significantly associated with abatacept response.

Conclusion Our results highlight important differences in disease burden by SE/ACPA status and suggest that ACPA status, rather than correlative SE status, may be the stronger predictor of abatacept response among individuals with RA.