Background:Biological disease modifying anti-rheumatic drugs (bDMARDs) and Janus Kinase inhibitors (JAKi) are both recommended post conventional synthetic disease modifying anti-rheumatic drug (csDMARD) therapy failure in active rheumatoid arthritis (RA), but the data on long-term durability are limited.Objectives:The objective of this study is to analyze a database of patients at the Western University, Rheumatology Center who initiated a bDMARD or JAKi and compare the proportion and characteristics of patients associated with retention of a drug class.Methods:This was a single-center study of 215 adult RA patients (82.76 % females, age 59.8 ± 12.0 years, disease duration 15.5 ± 10.0 years; table 1) failing multiple csDMARDs prior to initiating either bDMARDs (TNF inhibitors, abatacept, rituximab, tocilizumab) or JAKi, between June 2014 (when tofacitinib was approved in Canada) and April 2020. All patients enrolled had failed traditional DMARDs, including methotrexate, hydroxychloroquine, sulfasalazine and/or leflunomide. Durability and predictors of discontinuation were analyzed by Kaplan-Meier and Cox regression analyses for all treatment trials, and for patients receiving bDMARDs/JAKi as a first line after csDMARDs failure.Results:In 215 patients, there were 320 treatment events (148 bDMARDs, 172 JAKi) and 142 discontinuations (53.5% bDMARDs, 46.5% JAKi). Figure 1 represents the Kaplan-Meier survival curve for time to therapy discontinuation in 215 patients receiving bDMARDs vs JAKi. The Cox proportional hazards model was significant with better retention for JAKi, with a hazard ratio (HR) for treatment discontinuation of JAKi compared with bDMARDs of 0.676 (95% CI 0.47-0.97, p=0.034), adjusted for gender, age, disease duration, and line of therapy (Table 1). Moreover, the analysis revealed better retention for both groups as first line advanced therapy compared to later lines of therapy; 57.6% of JAKi and 31.1% of bDMARDs were used as first line advanced therapy. HR for treatment discontinuation for first line vs later lines of therapy was 0.593 (95% CI 0.40-0.88, p=0.01), adjusted for drug class, gender, age, and disease duration (Table 1). The most common reasons for discontinuations were inefficacy (60%), side effects (22%), or other reasons (18%). Inefficacy (58% vs 62%, p=0.8) and side-effects (16% vs 27%, p=0.4) were equally common for bDMARDs and JAKi. Sex, age at treatment onset, and RA duration did not predict discontinuation by Cox regression analyses, and after sub-grouping into bDMARDs and JAKi.Conclusion:EULAR guidelines have placed bDMARDs equal to JAKi as post csDMARD failure therapy in active RA. However, this study demonstrates that JAKi has a greater durability than biologics regardless of gender, age, disease duration, and line of therapy. Therefore, JAKi may be considered as a preferable method of treatment post csDMARD failure in active RA.Figure 1.Kaplan-Meier survival curves for (A) time to discontinuation of therapy in all RA patients receiving bDMARDs versus JAKi; P-value represents Cox regression adjusted for gender, age, disease duration, and line of therapy (B) time to discontinuation of therapy in patients using bDMARDs/JAKi as first line of advanced therapy; P-value represents Cox regression adjusted for drug class, gender, age, and disease durationTable 1.Patient demographics and hazard ratios for discontinuation of bDMARDs versus JAKi by Cox regression modelCharacteristicJAKi (N=172)bDMARD (N=148)MeanAge (years)60.958.559.8Sex (% F)77.888.582.8Disease duration (years)15.315.815.5Line of advanced therapy (% first line)57.631.145.3Drug used (%)Tofacitinib: 93.5Rituximab: 26.4Etanercept: 19.6Adalimumab: 17.6Predictors of Drug DiscontinuationHR (95% CIs)P valuesCrude ModelJAKi vs bDMARDs0.60 (0.43, 0.84)0.003Adjusted modelJAKi vs bDMARDs0.68 (0.47, 0.97)0.034Male vs female0.77 (0.46, 1.31)0.342Age1.01 (0.99, 1.03)0.123RA duration0.99 (0.97, 1.01)0.500Treatment line 1 vs >10.59 (0.40, 0.88)0.010Disclosure of Interests:Karla Machlab: None declared, Samir M. Iskandar: None declared, Tatiana Nevskaya: None declared, Louise Vanderhoek: None declared, Jillian Bylsma: None declared, Sara Hewitt: None declared, Janet Pope Speakers bureau: AbbVie, Amgen, BMS, BI, Gilead, Galapagos, Janssen, Lilly, Medexus, Merck, Novartis, Pfizer, Sanofi, Sandoz, Consultant of: AbbVie, Amgen, BMS, BI, Celltrion, Gilead, Galapagos, Janssen, Lilly, Medexus, Merck, Novartis, Pfizer, Roche, Samsung, Sanofi, Sandoz, Teva, UCB, Grant/research support from: AbbVie, BMS, Merck, Pfizer, Roche, Seattle Genetics