Abatacept used in combination with non-methotrexate disease-modifying antirheumatic drugs: a descriptive analysis of data from interventional trials and the real-world setting

Background: Methotrexate (MTX) remains the anchor drug in rheumatoid arthritis (RA) treatment, but is poorly tolerated or contraindicated in some patients. There is a wealth of data supporting the use of abatacept in combination with MTX, but data on alternative conventional synthetic disease-modifying antirheumatic drug (csDMARD) combinations with abatacept are scarce. Methods: In this post-hoc exploratory analysis, efficacy and safety data were extracted from abatacept RA studies in which combination with csDMARDs other than MTX was permitted: three interventional trials (ATTAIN, ASSURE, and ARRIVE) and one real-world study (ACTION). Patients with moderate-to-severe RA received abatacept in combination with MTX, hydroxychloroquine, sulfasalazine, azathioprine, or leflunomide for 6 months to 2 years according to the study design. Change from baseline in physical function (Health Assessment Questionnaire-Disability Index (HAQ-DI); all studies) and 28-joint Disease Activity Score (C-reactive protein) (DAS28 (CRP); ATTAIN, ARRIVE, and ACTION), American College of Rheumatology response rates (ATTAIN), and safety were assessed for individual and pooled csDMARD combinations for each trial. A meta-analysis was also performed on pooled data for HAQ-DI and DAS28 (CRP) across interventional trials. Results: Across all four studies, 731 patients received abatacept plus one non-MTX csDMARD (hydroxychloroquine n = 152; sulfasalazine n = 123; azathioprine n = 59; and leflunomide n = 397) and 2382 patients received abatacept plus MTX. Mean changes from baseline in HAQ-DI scores for abatacept plus MTX (all csDMARDs pooled) vs abatacept plus a nonMTX csDMARD were -0.54 vs -0.44 (ATTAIN), -0.43 vs -0.43 (ASSURE), and -0.39 vs -0.36 (ARRIVE). Mean changes from baseline in DAS28 (CRP) and ACR response rates were also similar with abatacept plus MTX or non-MTX csDMARDs. Data for individual non-MTX csDMARDs (pooled across studies) and real-world data were consistent with these findings.