Background We recently reported a glucocorticoid (GC)-sparing effect of Abatacept (ABA) in patients with RA in real life practice (reference 1).

Objectives To analyze the association between ABA GC-sparing effect and both disease characteristics, such as seropositivity, and abatacept effectiveness in RA patients.

Methods This is a pooled analysis of nine European RA registries (ARTIS, ATTRA, BIOBADASER, DANBIO, GISEA, NOR-DMARD, ORA, SCQM, Reuma.pt) and one North-American (RHEUMDATA). Inclusion criteria were a diagnosis of RA, initiation of ABA treatment, and at least one subsequent follow-up visit with data on GC use. The basal dosage of GC was categorized as none (dose =0 mg/d), low (>0 and ≤5 mg/d), moderate (>5 and ≤10.0 mg/d) or high dose (>10.0 mg/d). We assessed the predictors of reaching both a EULAR good/moderate responder status and stopping/tapering GC in a multivariable model including age, sex, disease duration, number of previous biologics, DAS28 at baseline (BL), RF/ACPA positivity, and concomitant treatment with methotrexate.

Results Of the 1719 patients with information on GC dose, 634 (36.9%) patients were on a low, 745 (43.3%) were on moderate, and 340 (19.8%) patients were on high dose GCs at BL. Patients with moderate/high dose of GCs had significantly higher BL disease activity (mean DAS28 5.3 vs 4.9, p<0.001) and were more frequently RF or ACPA-positive (73.6% vs 65.4%, p=0.02). Among the 373 patients who stopped or reduced GC, 276 (74.0%) were EULAR good/moderate responders at 12 months, compared to only 63.4% of those who did not reduce GC (p<0.001). 78.4% of the patients who could stop or reduce GC within 12 months were RF positive, compared to 70.8% among those who did not reduce GC. Similarly, 74.3% of patients stopping or reducing GC were ACPA-positive, compared to 62.2% ACPA-positive patients among those who did not reduce GC. In the multivariable model, shorter disease duration (OR 0.95, 95%CI: [0.91; 0.99], p=0.01), lower number of previous biologics (OR 0.76, 95%CI: [0.59; 0.99], p=0.04), lower initial DAS28 (OR 0.56, 95%CI: [0.41; 0.77], p=0.0003), and seropositivity (OR 1.86, 95%CI: [1.09; 3.29], p=0.03) were significantly associated with stopping and reducing GC and being a EULAR good or moderate responder.

Conclusions These results add further evidence to the association between seropositivity and ABA effectiveness by demonstrating that a higher proportion of patients with seropositive RA on ABA therapy are able to reduce or stop concomitant GC compared to seronegative patients, despite having higher doses of GC at baseline.

1. Gottenberg J, Mariette X, Hernandez MV, Iannone F, Lie E, Canhão H, Pavelka K, Turesson C, Lund Hetland M, Finckh A. Glucocorticoid-Sparing Effects of Abatacept in Real Life Practice: Data from a Paneuropean Analysis of RA Registries [abstract]. Arthritis Rheumatol. 2015; 67 (suppl 10).

Acknowledgement unrestricted research grant from BMS

Disclosure of Interest J.-E. Gottenberg: None declared, D. Courvoisier: None declared, M. Hetland: None declared, C. Turesson: None declared, H. Canhão: None declared, M. V. Hernandez: None declared, F. Iannone: None declared, E. Lie: None declared, K. Pavelka: None declared, D. Choquette: None declared, X. Mariette: None declared, A. Finckh Grant/research support from: BMS