Background While the relative efficacy of treatments can be demonstrated in relatively small studies with limited follow-up, most safety concerns are infrequent, requiring longer follow-up and larger populations. This is recognized by the regulatory framework, where data from pivotal randomized controlled trials are usually considered sufficient for demonstrating efficacy and non-toxicity, but post-approval safety studies are required for many years to fully evaluate drug-associated risks. Though such regulatory safety-studies often focus on one drug (vs. all others), clinical decision-making requires data across all available treatment options. Long-standing longitudinal clinical registries, like the Anti-Rheumatic Therapies in Sweden (ARTIS) database, thus have a key role in assessing the relative safety of b/tsDMARDs, allowing simultaneous comparison of all drugs used in clinical practice, with consistent definitions of treatment cohorts, follow-up, and outcomes.

Objectives To assess incidence rates of critical safety endpoints for individual b/tsDMARDs used to treat RA, updating previously published reports and including more recently introduced treatments.

Methods Nationwide register-based cohort study including all RA patients in Sweden registered as starting any b/tsDMARD between Jan 1^st 2010 and Dec 31^st 2019, and followed until Dec 31^st 2020. The incidence rates of selected outcomes, identified through national healthcare registers, were compared between individual b/tsDMARDs while adjusting for a range of potential confounders (covering demographics, RA-related characteristics and disease activity, and comorbidity) using Inverse Probability of Treatment Weighting. Probabilities were predicted by multinomial logistic regression, regressing all covariates on treatment status. Exposure time was counted from treatment start until stop (+90 days’ lag time), censored at emigration and death.

Results There were clear differences between patients starting individual b/tsDMARDs, in particular with TNF inhibitors more often used as a first line b/tsDMARD; sarilumab, baricitinib, and tofacitinib predominantly used later in the treatment course; rituximab used more often for older patients, and non-TNFi generally used more frequently for patients with higher disease activity or comorbidity. Expectedly, these differences translated into differences in the crude rate of safety endpoints.

Several differences remained after confounder-adjustment (Table 1), including a higher rate of treatment discontinuation due to adverse events on baricitinib, tofacitinib, and sarilumab. Rituximab was associated with higher rates of several outcomes, but the confounder-adjustment markedly reduced risks and residual confounding likely explain part of the remaining increase. Baricitinib and tofacitinib were associated with higher rates of hospitalised herpes zoster, but not with similarly elevated rates of other serious infections. There were no clear differences in the rate of cardiovascular events or severe depression. Low number of events limit the comparison, in particular for sarilumab and tofacitinib.

Conclusion We found large differences in the rate of treatment discontinuations due to adverse events across b/tsDMARDs, which were not generally mirrored by corresponding differences in the rates for specific serious adverse events.

References N/A

Acknowledgements ARTIS has been or is currently supported by agreements with Abbvie, BMS, Eli Lilly, MSD, Pfizer, Roche, Samsung Bioepis, and Sanofi.

Disclosure of Interests Thomas Frisell: None declared, Hannah Bower: None declared, Eva Baecklund: None declared, Daniela Di Giuseppe: None declared, Bénédicte Delcoigne: None declared, Nils Feltelius Employee of: NF is employed by the Medical Products Agency (MPA), which is a governmental body. The views in this abstract may not represent the views of the MPA, Helena Forsblad-d’Elia: None declared, Elisabet Lindqvist: None declared, Ulf Lindström: None declared, Johan Askling Grant/research support from: Karolinska Institutet has entered into agreements with the following companies, with JA as PI: Abbvie,

BMS, Eli Lilly, Galapagos, Janssen, Pfizer, Roche, Samsung Bioepis and Sanofi.