Background Since the introduction of biological disease-modifying antirheumatic drugs (bDMARDs), treatment of rheumatoid arthritis (RA) has changed. Remission is a feasible treatment target for many patients. As a result, there is an increased interest in appropriate treatment strategies after achieving disease control, such as bDMARD discontinuation while in remission. However, its adoption in typical clinical practice has not been widely studied.
Objectives We conducted a cross-national longitudinal observational study of bDMARD to describe the proportions of RA patients in remission who discontinued these agents, and to assess the potential predictors of discontinuation.
Methods We utilized data from US CORRONA and Japanese NinJa registries. Patients who used bDMARDs on at least two consecutive visits and experienced remission defined by the Clinical Disease Activity Index (CDAI) ≤2.8 at least once were included. The outcome of interest was bDMARD discontinuation while in remission. Discontinuation was described with the Kaplan-Meier method as well as the cumulative incidence function method to account for competing risk events, i.e., loss of remission and switch to another bDMARD. The baseline predictors of discontinuation were assessed in the Cox regression models.
Results The numbers of eligible patients were 4,382 in the CORRONA (34.4% of 12,721 bDMARD users among 22,684 total patients) and 413 in the NinJa (18.4% of 2,243 bDMARD users among 11,205 total patients). In the Kaplan-Meier analysis, discontinuation at 5 years among patient who remained in remission was observed in 28.1% in CORRONA and 22.7% in NinJa (p=0.002, log-rank test; Figure Panels A and B). In the cumulative incidence function analysis, the proportions of discontinuation at 5 years among patients who were initially in remission were 11.3% in CORRONA and 12.2% in NinJa (Figure Panels C and D). It was not significantly different (p=0.782, Gray's test) due to greater attrition to loss of remission in CORRONA (83.9%) compared to NinJa (77.2%). In Cox regression, reduced CDAI (hazard ratio for 1 point decrease: 1.49 [1.37,1.61] in CORRONA and 1.79 [1.33,2.38] in NinJa) and use of glucocorticoids (HR: 1.59 [1.19,2.12] in CORRONA and 2.11 [1.03, 4.31] in NinJa) predicted higher rates of bDMARD discontinuation in both cohorts. The overall R2 for the full models including sex, race, age, RA duration, biologic class, methotrexate use, duration of bDMARD use, and index year, was 0.074 in CORRONA and 0.168 in NinJa.
Conclusions We found a slightly higher rate of bDMARD discontinuation in CORRONA than NinJa. Possible contributors to the discrepancy may include health care system differences (e.g. out of pocket cost) as well as cultural differences (e.g. patient preference vs doctor recommendation). While several variables were associated with more frequent discontinuation, they did not provide robust explanation of the variability in practice.
Acknowledgements We would like to thank all CORRONA and NinJa contributors (patients, clinicians, and investigators) for maintaining such valuable sources of information.
The CORRONA RA registry has been supported through contracted subscriptions in the last two years by Abbvie, Amgen, AstraZeneca, Genentech, Horizon Pharma, Eli Lilly, Novartis, Pfizer, Vertex and UCB. All non-pharmaceutically-initiated CORRONA research activities are conducted entirely independent of any pharmaceutical input. NinJa was supported in part by a Health and Labor Sciences Research Grant from the Ministry of Health, Labor, and Welfare of Japan.
Disclosure of Interest : K. Yoshida Paid instructor for: Abbott Japan, H. Radner: None declared, M. Mjaavatten: None declared, J. Greenberg Shareholder of: CORRONA, Consultant for: AstraZeneca, Pfizer, Employee of: CORRONA, A. Kavanaugh Grant/research support: Abbott, Janssen, Amgen, BMS, Roche, UCBM, M. Kishimoto Speakers bureau: Santen, Mitsubishi-Tanabe, Pfizer, Abbott Japan, K. Matsui: None declared, M. Okada Speakers bureau: Santen, Mitsubishi-Tanabe, Pfizer, Abbott Japan, G. Reed Employee of: CORRONA, Y. Saeki Grant/research support: Tanabe-Mitsubishi, Chugai, Janssen, Nippon-Kayaku, Nichi-Iko, S. Tohma Grant/research support: Pfizer Japan, Eisai, Chugai, J. Kremer Shareholder of: CORRONA, Grant/research support: Abbott, Amgen, BMS, Genetech, Novo- Nordisk, Pfizer, Novartis, Eli Lilly, Employee of: CORRONA, D. Solomon Grant/research support: Eli Lilly, Amgen, CORRONA, Consultant for: Pfizer, Novartis, Eli Lilly