Discontinuation of tofacitinib and TNF inhibitors in patients with rheumatoid arthritis: analysis of pooled data from two registries in Canada

• Published in: BMJ open Vol. 13; no. 3; p. e063198
• Main Authors: Movahedi, Mohammad, Choquette, Denis, Coupal, Louis, Cesta, Angela, Li, Xiuying, Keystone, Edward C, Bombardier, Claire, Investigators, OBRI
• Format: Journal Article
• Language: English
• Published: England British Medical Journal Publishing Group 06.03.2023; BMJ Publishing Group LTD; BMJ Publishing Group
• Series: Original research
• Subjects: Arthritis, Rheumatoid - drug therapy; Bias; Cardiovascular disease; Comorbidity; Decision making; Drug dosages; Humans; Inflammatory diseases; Missing data; Observational studies; Ontario; Patients; Questionnaires; Registries; Reimbursement; Retrospective Studies; Rheumatoid arthritis; Rheumatology; Survival analysis; THERAPEUTICS; TNF inhibitors; Tumor Necrosis Factor Inhibitors - therapeutic use; Tumor necrosis factor-TNF; Ontario; Canada; THERAPEUTICS; RHEUMATOLOGY
• ISSN: 2044-6055; 2044-6055
• DOI: 10.1136/bmjopen-2022-063198

ObjectivesThe similarity in retention of tumour necrosis factor inhibitors (TNFi) and tofacitinib (TOFA) was previously reported separately by the Ontario Best Practices Research Initiative and the Quebec cohort Rhumadata. However, because of small sample sizes in each registry, we aimed to confirm the findings by repeating the analysis of discontinuation of TNFi compared with TOFA, using pooled data from both these registries.DesignRetrospective cohort study.SettingPooled data from two rheumatoid arthritis (RA) registries in Canada.ParticipantsPatients with RA starting TOFA or TNFi between June 2014 and December 2019 were included. A total of 1318 patients were included TNFi (n=825) or TOFA (n=493).Outcome measuresTime to discontinuation was assessed using Kaplan-Meier survival and Cox proportional hazards regression analysis. Propensity score (PS) stratification (deciles) and PS weighting were used to estimate treatment effects.ResultsThe mean disease duration in the TNFi group was shorter (8.9 years vs 13 years, p<0.001). Prior biological use (33.9% vs 66.9%, p<0.001) and clinical disease activity index (20.0 vs 22.1, p=0.02) were lower in the TNFi group.Discontinuation was reported in 309 (37.5%) and 181 (36.7%) TNFi and TOFA patients, respectively. After covariate adjustment using PS, there was no statistically significant difference between the two groups in discontinuation due to any reason HR=0.96 (95% CI 0.78 to 1.19, p=0.74)) as well as discontinuation due to ineffectiveness only HR=1.08 (95% CI 0.81 to 1.43, p=0.61)).TNFi users were less likely to discontinue due to adverse events (AEs) (adjusted HRs: 0.46, 95% CI 0.29 to 0.74; p=0.001). Results remained consistent for firstline users.ConclusionsIn this pooled real-world data study, the discontinuation rates overall were similar. However, discontinuation due to AEs was higher in TOFA compared with TNFi users.