Background: Despite an increasing number of treatment options with different mechanisms of action for axial spondyloarthritis (axSpA), including biologic and targeted synthetic disease-modifying drugs (b/tsDMARDs), some patients fail to achieve a satisfactory treatment response. In 2025, the Assessment of SpondyloArthritis international Society (ASAS) published consensus-based, expert definitions of difficult-to-manage (D2M) and treatment-refractory axSpA (the latter being a subgroup of D2M (ref 1)). For D2M axSpA, these include: 1. Treatment according to recommendations and failure of ≥2 b/tsDMARDs (primary/secondary failure, discontinuation due to side effects/intolerability/contraindication) with different modes of action (MOA); 2. Insufficient control of signs/symptoms of axSpA defined as ≥1 of: a) high/very high disease activity, b) signs/symptoms suggestive of active disease, c) rapid radiographic progression, d) axSpA symptoms causing reduction in quality of life; 3. Present signs/symptoms perceived as problematic by rheumatologist and/or patient. These definitions provide a framework for identifying D2M patient groups, and it is therefore relevant to explore how these definitions perform in a real-life setting.

Objectives: Inspired by the ASAS definitions, we aimed to explore how the application of increasingly strict definitions impacts the prevalence of D2M axSpA in a large Nordic observational routine-care patient cohort. A further objective was to describe the characteristics of patients fulfilling the different definitions.

Methods: Observational cohort study, including patients initiating a first-ever b/tsDMARD from 1999 onwards in the Nordic rheumatology registries (Denmark DANBIO-DRQ/Sweden SRQ/Finland ROB-FIN/Norway NOR-DMARD/Iceland ICEBIO) with >3 years of follow-up after treatment start (=baseline). Data-cut varied between the registries (2021-2023). Disease characteristics were retrieved from the rheumatology registries and previous prescriptions and comorbidities through linkage to national administrative registries. Data were pooled for analyses. As a first step, three nested cohorts were defined by discontinuation (irrespective of reason) of ≥2, ≥3, or ≥4 b/tsDMARDs during follow-up (component A). Additional components were then added sequentially in each cohort: B) ≥2 different mechanisms of action; C) discontinuation of the most recent b/tsDMARD due to primary/secondary failure or side effects/intolerability/contraindications. Finally, based on data from the most recent visit during treatment with the most recent b/tsDMARD, two further components were sequentially added: D) insufficient control of signs/symptoms of the disease (defined as Axial Spondyloarthritis Disease Activity Score (ASDAS) ≥2.1 or CRP ≥10 mg/L); E) problematic signs/symptoms (defined as patient global visual analogue scale ≥30 mm).

Baseline characteristics were described for patients fulfilling the D2M definitions when components A, B, and C were applied. No imputation of missing data was performed.

Results: We identified 18,083 patients with axSpA in the registries (41% female; baseline mean age 41 (SD 13) years and median ASDAS 3.2 (IQR 2.6-3.9)) ( Table 1 ), of whom 41% started their first b/tsDMARD in 2015 or later. Proportions of patients discontinuing ≥2, ≥3, or ≥4 b/tsDMARDs during follow-up were 34%, 16%, and 7%, respectively ( Table 1 ).

After adding all components (B-E), the prevalence of patients fulfilling the D2M definitions decreased to 0.6-1.1% ( Figure 1 ).

Baseline characteristics of patients discontinuing ≥2, ≥3, or ≥4 b/tsDMARDs plus meeting components B and C, are shown in Table 1 . With increasing number of discontinued b/tsDMARDs, patients had numerically higher baseline ASDAS and patient global scores, and shorter education. Furthermore, compared with the overall cohort, patients failing b/tsDMARDs tended to be more frequently female and to have more comorbidities including a higher prevalence of depression and use of opioids. The median time from the start of the first b/tsDMARD until discontinuing ≥4 b/tsDMARDs plus meeting components B and C was 5.6 years.

Conclusions: This large registry-based study showed that multiple b/tsDMARD discontinuations were frequent in patients with axSpA. Depending on the stringency of the applied definitions, the proportions of D2M axSpA varied from <1% to 34%. The female overrepresentation and higher prevalence of depression and use of opioids in D2M patients warrant further investigation. Overall, our results provide new evidence on the complexity of D2M axSpA.

REFERENCES: [1] Poddubnyy D, et al. Ann Rheum Dis. 2025. DOI: 10.1016/j.ard.2025.01.035.

Acknowledgments: NIL.

Disclosure of Interests: Daniela Di Giuseppe: None declared, Louise Majormoen Nielung: None declared, Merete Lund Hetland Sandoz, UCB (Institution). Novartis (Personal and institution), Abbvie (Institution), Abbvie, Alfasigma,, Eli Lilly, UCB, Novartis, Sandoz (institution), Johan Askling Abbvie, BMS, Eli Lilly, Galapagos NV/Alfasigma S.p.A, MSD, Pfizer, Roche, Samsung Bioepis, Sanofi, Dan Nordström Pfizer, UCB, MSD, Novartis, Pfizer, UCB, BMS, MSD, UCB, Sella Aarrestad Provan: None declared, Bjorn Gudbjornsson: None declared, Ólafur Pálsson: None declared, Bénédicte Delcoigne: None declared, Lene Wohlfahrt Dreyer MSD, UCB, Eli Lilly, AbbVie and BMS (payment to institution), Brigitte Michelsen Novartis, Novartis, Novartis, UCB and Pfizer (paid to employer), Mirjam de Vries Pfizer, UCB and Abbvie, Johan K Wallman AbbVie, Amgen, Eli Lilly, Novartis, Pfizer, Bente Glintborg AbbVie, Sandoz, EliLilly, AlfaSigma (paid to institution).