Introduction

Psoriatic arthritis (PsA) is a chronic, immune-mediated disease characterized by musculoskeletal and extra-musculoskeletal manifestations, including peripheral arthritis, enthesitis, dactylitis, axial involvement, and skin and nail psoriasis, as well as associated conditions such as uveitis and inflammatory bowel disease1,2. Management of PsA is further complicated by frequent comorbidities, such as metabolic syndrome, cardiovascular diseases, anxiety, depression and fibromyalgia, all of which contribute to an increased disease burden and therapeutic challenges3,4,5. This heterogeneity hinders accurate disease monitoring, individualized care and treat-to-target strategies in routine practice.

Over the past two decades, the PsA treatment landscape has evolved substantially with the approval of numerous biologic and targeted synthetic DMARDs (b/tsDMARDs), which inhibit different molecular targets and inflammatory pathways6. These include inhibitors of tumor necrosis factor (TNF), interleukin-12 (IL-12)–IL-23, IL-17, IL-23, phosphodiesterase-4 (PDE4), cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) and Janus kinases (JAKs)6. In response to these developments, treatment recommendations from the Group for Research and Assessment of Psoriasis and Psoriatic Arthritis (GRAPPA) and European Alliance of Associations for Rheumatology (EULAR) have evolved accordingly, adopting a domain-based, patient-centred approach7,8.

These recommendations generally support the use of conventional synthetic DMARDs (csDMARDs) as initial therapy in selected domains, such as peripheral arthritis, with escalation to b/tsDMARDs in cases of inadequate response. In cases of axial involvement or severe enthesitis, early initiation of b/tsDMARDs is often warranted when available. Importantly, the GRAPPA recommendations avoid rigid treatment hierarchies and a one-size-fits-all approach, instead emphasizing clinical flexibility, shared decision-making, regular assessment of treatment response and domain-specific treatment adjustments. Nonetheless, they do not define how many prior treatment failures constitute therapeutic refractoriness, underscoring the need for standardized criteria.

Despite these therapeutic advancements, a 2020 systematic review and meta-analysis reported that fewer than 30% of people with PsA achieve remission when assessed by composite measures, and up to 50% report residual symptoms despite treatment9. These findings highlight the unmet needs in PsA care and have prompted increasing attention to the concept of patients who are ‘difficult to treat’ (D2T). In the absence of standardized terminology, these patients are variably referred to as ‘refractory’, ‘treatment-resistant’, ‘D2T’ or ‘complex’, depending on context and clinical perception. This lack of uniformity limits the comparability of studies and impairs the design of clinical trials, ultimately contributing to therapeutic inertia.

Estimates of the prevalence of treatment-refractory (TR) PsA vary considerably, depending on the specific criteria used. When applying the EULAR definition of D2T rheumatoid arthritis (D2T-RA) — which requires failure of at least two b/tsDMARDs with distinct mechanisms of action (MoAs), evidence of active and/or progressive disease, and problematic management as perceived by the clinician and/or patient10 — prevalence in PsA cohorts is ~16–30%11,12,13. However, the transferability of these criteria to PsA is questionable. Unlike RA, PsA is a multidomain disease with substantial clinical and pathophysiological heterogeneity, distinct treatment pathways, and a complex interplay of skin and joint involvement, mechanical stress, obesity and metabolic syndrome, mental health, and lifestyle factors14. The application of RA-derived criteria can overlook or misclassify important subgroups within the PsA spectrum — for example, patients with a predominant enthesitis phenotype — thereby reinforcing the need for disease-specific definitions that better reflect disease complexity and patient experience15. Thus, the heterogeneity of PsA mandates a framework that captures both biologic treatment failure and the broader contextual complexity of disease management.

Recognizing these unmet needs16,17, the GRAPPA D2T-PsA task force initiated a structured, multi-phase initiative to define more precisely the spectrum of patients whose disease is particularly complex-to-manage (C2M-PsA) or TR-PsA18. This project integrates perspectives from clinicians, health care professionals (HCPs), and — critically — patients, reflecting a shared-decision model of care that is central to GRAPPA’s philosophy.

The primary aim of this project was to develop consensus-based definitions for C2M-PsA and TR-PsA. To ensure that these definitions were grounded in empirical evidence and informed by diverse stakeholder perspectives, the project incorporated three key components to guide the consensus process: a review of existing terminology and conceptual frameworks related to TR-PsA and difficult-to-treat-PsA; a survey of HCPs to identify real-world challenges in PsA management; and an analysis of international patient-reported data to capture emotional, psychosocial and contextual factors contributing to treatment complexity or failure. This Consensus Statement presents the resulting consensus definitions for C2M-PsA and TR-PsA, which are intended to support shared clinical decision making and to enable more effective patient stratification in both observational studies and clinical trials.

Methods

Steering committee and working group

The development of these definitions began in March 2023 with the approval of the project from the GRAPPA Research Committee and the subsequent establishment of a dedicated steering committee and working group, together forming the GRAPPA D2T-PsA task force. The steering committee included two project co-leads (F.P. and P.M.), two Young-GRAPPiAns (A.L.R. and S.S.), a rheumatologist (V.C.), a dermatologist (W.L.) and a patient research partner (PRP; C.L.). The broader working group consisted of an additional 12 individuals, including 10 rheumatologists and 2 dermatologists (Supplementary Fig. 1).

The project started with a scoping literature review to explore existing definitions and terminologies related to refractory PsA19. This review was followed by two complementary surveys, one capturing the perspectives of HCPs20 and the other focusing on insights from patients21. These findings informed the steering committee in the development of preliminary definitions and proposed terminology for C2M-PsA and TR-PsA. These initial drafts were presented at the first task-force meeting, where discussions centred on refining the definitions around three core components: treatment failure history, characterization of active and/or symptomatic disease, and disease perception.

Scoping literature review

A scoping literature review was conducted in July 2023 to identify existing terminologies, definitions and criteria applied to the concept of D2T-PsA and TR-PsA. The search was carried out in the PubMed, Embase, and Web of Science databases for articles published between January 2000 and September 2022, using combinations of terms including “psoriatic arthritis”, “difficult to treat”, “refractory”, “resistant” and related descriptors. To broaden the scope, both investigators (A.L.R. and S.S.) independently screened all titles and abstracts from the most recent major international conferences: ACR Convergence 2022 and EULAR Congress 2023. Discrepancies were resolved by a third investigator (F.P.). The full findings of this review have been published separately19.

Group for Research and Assessment of Psoriasis and Psoriatic Arthritis health care professionals survey

An international, web-based survey was developed by the working group and disseminated to all GRAPPA members between June and September 2023. Eligible participants were HCPs involved in PsA care. The survey included demographic questions, structured multiple-choice items on proposed criteria, and open-ended questions to capture qualitative insights. Survey development involved iterative rounds of review and input from both clinician-investigators and PRPs. Responses were analysed using descriptive statistics and thematic qualitative analysis. A detailed account of the methodology and results has been published separately20.

Group for Research and Assessment of Psoriasis and Psoriatic Arthritis patient survey

A parallel international survey targeting people with lived experience of PsA was conducted between October 2023 and January 2024, with the aim of identifying patient-perceived features related to treatment ineffectiveness. The questionnaire was developed in English and translated into nine languages (Chinese, Dutch, French, German, Italian, Japanese, Portuguese, Spanish and Turkish). It included both structured items and free-text responses. The survey was distributed electronically with support from patient organizations and GRAPPA PRPs. A medical anthropologist contributed to the analysis, which employed a mixed-methods approach combining quantitative summaries with qualitative content analysis. The full dataset and analysis were published in a separate manuscript21.

Task force meetings and Delphi consensus round

Between March 2023 and December 2024, the GRAPPA D2T task force conducted a structured, multi-phase consensus process to develop and refine definitions for C2M-PsA and TR-PsA. The process began with a dedicated kick-off meeting in March 2023, followed by feedback discussions at the GRAPPA Annual Meeting (July 2023) and a roundtable dialogue at ACR Convergence (November 2023). Findings from the scoping literature review, expert survey and patient survey were synthesized and discussed within the task force. Between March and November 2024, a series of structured virtual task force meetings took place, complemented by broader stakeholder engagement during the EULAR 2024 Congress, where the draft framework was discussed. Additional input was obtained during the GRAPPA Collaborative Research Network (CRN) meeting with industry representatives and during a dedicated GRAPPA Membership session at the 2024 Annual Meeting. From August to November 2024, the Steering Committee convened several virtual meetings to further refine the definitions and incorporate multidisciplinary feedback. Two formal Delphi rounds were conducted between August and November 2024 to evaluate the proposed definitions, overarching statements and clarifying statements. These items were initially drafted by the steering committee based on evidence from the literature review and survey findings, and were subsequently refined through structured discussions with the full task force. In each Delphi round, task force members were asked to indicate their agreement with each item using a binary response format (“agree” or “not agree”). Items that did not reach the predefined consensus threshold of ≥ 75% agreement in round 1 were revised by the steering committee, drawing on qualitative feedback from free-text comments submitted during the online voting and follow-up virtual task force discussions. The revised items were then redistributed for a second round of voting using the same response format. The final task force proposals were subsequently presented and discussed at the GRAPPA-Adjacent-to-ACR meeting (November 2024) and submitted for full GRAPPA membership voting in December 2024.

The final endorsed definitions were presented at the EULAR 2025 Congress22, concluding the consensus process and establishing a framework co-developed with patients, clinicians, researchers and industry stakeholders. Supplementary Fig. 2 summarizes the process.

Development of the definitions of complex-to-manage and treatment-refractory psoriatic arthritis

Scoping literature review

The scoping review19 identified 565 articles, including 15 publications that directly or indirectly addressed the concept of refractory PsA. Most studies did not apply standardized definitions; instead, terminology varied substantially across publications. Commonly used descriptors included “refractory”, “multi-resistant”, “complex” and “non-responding”, often applied without predefined criteria. Key features reported in these studies included exposure to multiple b/tsDMARDs, persistent disease activity across domains, presence of comorbidities and low treatment satisfaction. The review underscored the conceptual heterogeneity and the absence of disease-specific criteria, emphasizing the need for standardized, PsA-specific definitions.

Health care professionals survey

A total of 223 responses from 47 countries were included in the HCP survey20. Respondents were mostly rheumatologists (80.2%) and dermatologists (17.9%) from Europe (45.7%), Latin America (18.8%) and North America (18.3%). Most respondents supported distinguishing between TR-PsA (or D2T-PsA) and C2M-PsA (82.9%), associating the former with ongoing inflammation and treatment failure and the latter with contextual challenges despite controlled inflammation. A large majority (90.6%) also emphasized the need for objective markers of inflammation in defining TR-PsA. However, there was no consensus on the number of treatment failures required, with suggestions ranging from one to four or more b/tsDMARDs.

Importantly, the shift towards the terminology ‘complex-to-manage’ was driven by early and meaningful involvement of PRPs. During preparatory meetings for the survey design, the initial term ‘difficult-to-manage’ was initially selected. However, particularly when discussing the project with the PRPs, they expressed concern that the label ‘difficult-to-manage’ carried unintended negative connotations. They felt that this phrasing risked assigning blame or stigma to patients, potentially reinforcing perceptions of non-adherence or psychological resistance. In response, the steering committee formally adopted the alternative term ‘complex-to-manage’, which was viewed as more respectful, neutral and aligned with a biopsychosocial understanding of the disease.

Patient survey

A total of 570 patients with PsA completed the multilingual survey21. Fatigue, pain and functional limitation were among the most commonly cited reasons for treatment dissatisfaction. Ranking analysis also showed the negative impact of peripheral arthritis, axial disease, chronic pain, comorbidities and fatigue. Thematic analysis of open responses confirmed a high emotional burden and frequent reference to frustration with ineffective treatment changes, poor communication and lack of individualized care. These insights emphasized the patient-perceived mismatch between clinical assessment and real-world disease experience. Additionally, there were meaningful language-specific differences in patient narratives, emphasizing the need for culturally sensitive communication and definitions. The prominence of these findings aligns with the multidimensional burden captured in the C2M-PsA definition.

Task-force meetings

Findings from the literature review and surveys were presented and discussed during task-force meetings. Members reached initial agreement on key conceptual elements, including the need for a multidimensional, domain-inclusive definition that balances treatment history, persistent disease activity and contextual factors. Feedback from breakout groups revealed tensions regarding how to weigh subjective symptoms versus objective findings and the relevance of comorbidities. As a result, an initial framework was proposed that grouped the construct into three core domains: objective inflammation, patient-reported disease impact and complicating contextual features.

Drafting and endorsement of the definitions and overarching statements

To capture the broader challenges encountered in clinical management beyond objective inflammation, and in acceptance of feedback from PRPs, the steering committee adopted the term ‘complex-to-manage psoriatic arthritis’ (C2M-PsA). This terminology aligns conceptually with similar initiatives, such as the Assessment of SpondyloArthritis International Society (ASAS) definition of D2T axial spondyloarthritis, and reflects the multifactorial nature of disease burden in spondyloarthritis and in PsA23,24. C2M-PsA includes patients with persistent symptoms that are not solely attributable to active inflammation, but also to other contributing mechanisms such as comorbidities, psychosocial factors and high subjective symptom burden. To specifically address the subgroup of patients with true biological non-response and ongoing objective inflammation, the term ‘treatment-refractory PsA’ (TR-PsA) was endorsed to characterize those with persistent disease activity despite multiple therapeutic interventions.

The final terminology, definitions, conceptual framework and overarching statements were voted on by 184 GRAPPA members, including 140 full members, 37 early-career members, 6 PRPs and 1 advanced practice provider. In this final phase, 95.1% of participants (n = 175) endorsed the proposal, confirming strong community support and leading to the formal adoption of the C2M-PsA and TR-PsA definitions along with the overarching principles (Table 1).

Table 1 Overarching principles and definitions for complex-to-manage and treatment-refractory psoriatic arthritis
Full size table

Overarching statements, definitions and clarifying statements

Overarching principles

The five overarching statements provide a conceptual framework to distinguish between complex contextual challenges and true therapeutic resistance in PsA (Fig. 1), serving as a clinical guide for applying the C2M-PsA and TR-PsA definitions in practice (Fig. 2).

Fig. 1: Interplay between complex-to-manage and treatment-refractory psoriatic arthritis.
figure 1

The Group for Research and Assessment of Psoriasis and Psoriatic Arthritis (GRAPPA) consensus definitions define two distinct states of psoriatic arthritis (PsA) in the context of inadequate response to treatment: complex-to-manage PsA (C2M-PsA) and treatment-refractory PsA (TR-PsA). b/tsDMARD), biologic or targeted synthetic DMARD.

Full size image
Fig. 2: Classification of treatment failure in psoriatic arthritis and definitions of complex-to-manage and treatment-refractory psoriatic arthritis.
figure 2

This algorithm illustrates the classification of treatment failure in psoriatic arthritis (PsA), with emphasis on the definitions of complex-to-manage PsA (C2M-PsA) and treatment-refractory PsA (TR-PsA). Patients with a confirmed diagnosis of PsA who remain symptomatic despite advanced therapy are assessed for treatment response. C2M-PsA is defined as a state involving persistent symptoms, considered problematic by the rheumatologist and/or the patient, following inadequate response to at least one biologic or targeted synthetic DMARD (b/tsDMARD), and can be driven by multiple contributing factors. TR-PsA is a more specific subset of C2M-PsA, characterized by failure of three or more therapies with different mechanisms of action (MoAs) (including at least two b/tsDMARDs), persistent symptoms acknowledged by both clinician and patient as problematic, and objective evidence of ongoing inflammation with exclusion of alternative explanations.

Full size image

Overarching statement 1: Psoriatic arthritis is a heterogeneous disease that requires multidisciplinary care; treatment decisions should involve shared decision making, addressing the individual patient holistically

PsA is a clinically heterogeneous condition that affects multiple domains — including skin and nails, musculoskeletal system, eyes, gastrointestinal tract and mental health. Optimal management of PsA requires a multidisciplinary team approach and treatment decisions guided by shared decision-making. Care should be individualized, addressing both the physical and psychological dimensions of disease burden. Socioeconomic and systemic barriers – such as limited access to care, insurance coverage, diagnostic delays and misinformation regarding the disease and/or its treatment – further complicate management and underscore the need for patient-centred and equitable models of care25.

Overarching statement 2: Management of psoriatic arthritis should consider all potentially affected disease domains and be guided by shared decision-making and predefined treatment targets; composite disease activity scores may serve as supportive tools, but it is essential to recognize that some psoriatic arthritis manifestations may not be fully captured by these scores

Effective management of PsA requires systematic assessment of all potentially affected domains. Although treatment should follow predefined targets, there is no universally accepted end point, and treatment goals should be established collaboratively with the patient26.

Composite disease activity scores can aid clinical decision-making, but none fully capture the multidimensional nature of PsA or encompass the entire GRAPPA core domain set27,28. They should therefore complement, rather than replace, individualized assessment of each domain. Moreover, these scores can occasionally reflect factors unrelated to PsA activity, potentially leading to misinterpretation. Selection and interpretation of these assessment tools should be guided by the patient’s specific clinical features and priorities.

Overarching statement 3: A lack of, or inadequate response to, recommended psoriatic arthritis treatments at adequate dosing and over an appropriate length of time, should prompt a re-evaluation of the diagnosis and consideration of the presence of comorbidities or other factors contributing to treatment non-response

For this statement, ‘recommended PsA treatments’ are according to the current GRAPPA management recommendations. This statement underscores the importance of reassessing both PsA diagnosis and potential treatment barriers in cases of inadequate response (examples are described in Supplementary Table 1). When patients do not improve despite appropriate therapy, clinicians should investigate other contributors such as comorbidities, overlapping conditions, poor adherence or alternative diagnoses29,30. This structured re-evaluation aligns with the GRAPPA guidelines and supports a more personalized approach to treatment-resistant cases7.

Overarching statement 4: Reasons for inadequate treatment responses can be classified into complex-to-manage psoriatic arthritis, a broader and more heterogenous group that includes factors such as comorbid conditions, persistent symptoms despite effective control of inflammation, and significant impacts on the quality of life; within this larger group, treatment-refractory psoriatic arthritis constitutes a smaller, specific subgroup characterized by inadequate treatment response and objective evidence of ongoing inflammation despite multiple attempted therapeutic strategies

For this statement, the factors related to C2M-PsA (Supplementary Table 1) include, but are not limited to, mal-adherence, chronic pain, comorbidities and fibromyalgia.

This statement introduces a critical differentiation within the spectrum of challenging PsA cases (Fig. 1). Inadequate treatment response can be used to categorize patients into the broader and more heterogeneous group of C2M-PsA, which includes factors such as comorbid conditions, persistent symptoms despite effective control of inflammation and substantial impacts on quality of life. Within this broader group, a more specific subset, TR-PsA, is characterized by objective evidence of ongoing inflammation despite multiple attempted therapeutic strategies. This framework enables clinicians to better stratify patients and tailor interventions to both biological and contextual complexities.

Overarching statement 5: For the definition of treatment-refractory psoriatic arthritis, the presence of objective evidence of ongoing inflammation in affected domains is required: arthritis, dactylitis, enthesitis, or skin/nail disease (by inflammation-sensitive imaging modalities for axial disease, or through laboratory assessmentse such as acute-phase reactants or synovial-fluid analysis) must be present in patients with persistent symptoms and inadequate response to treatment

For the definition of TR-PsA, if clinical evaluation of arthritis, dactylitis, enthesitis or skin and/or nail disease is inconclusive, confirmation with inflammation-sensitive imaging (musculoskeletal ultrasonography (MSUS) or MRI) is recommended. If enthesitis is the only sign of active disease, confirmation with inflammation-sensitive imaging is recommended. With regard to laboratory assessments, other causes of elevated acute phase reactants or a high cell-count in synovial fluid must be excluded.

The final overarching statement establishes objective criteria for defining TR-PsA. It reinforces that confirmation of TR-PsA must go beyond symptoms and include verifiable signs of inflammation in at least one domain. Because both peripheral and axial PsA manifestations can be difficult to assess clinically31,32, inflammation-sensitive imaging modalities, such as MSUS or MRI, have a key role in evaluating disease activity. Laboratory parameters such as C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) can also support the TR-PsA definition, although they are elevated in only about half of PsA cases33,34. The task force emphasizes the importance of a comprehensive assessment, accounting for the limitations of isolated biomarkers and ensuring the exclusion of non-inflammatory causes of elevated inflammation (for example, infection or adiposity) or abnormal MSUS35,36,37.

Definitions of complex-to-manage and treatment-refractory psoriatic arthritis

Definition of complex-to-manage refractory psoriatic arthritis: ‘Complex-to-manage psoriatic arthritis’ is a disease state characterized by persistent symptoms despite at least one adequate trial of a biologic DMARD or targeted synthetic DMARD recommended for psoriatic arthritis treatment; this category extends beyond pure biological non-response to also include factors such as comorbidities, overlapping conditions and treatment-related challenges

That a patient with PsA is in a C2M state can be acknowledged by the treating clinician and/or the patient in the case of treatment inefficacy and signs and/or symptoms regarded as problematic.

For this definition, ‘recommended PsA treatments’ are treatments that are recommended according to the current GRAPPA management recommendations. The C2M-PsA category includes factors including, but not limited to, comorbidities, overlapping conditions, chronic pain, non-adherence, treatment side effects, socioeconomic or psychological barriers (Supplementary Table 1).

The task force acknowledges that C2M-PsA is a dynamic state that can improve with appropriate interventions. Many factors could lead to this complex presentation, including mental-health issues, chronic-pain syndromes, or logistical barriers to accessing care (see Supplementary Table 1). Recognizing this variability, the inclusion criterion of failure of at least one approved treatment is aimed at capturing both ends of the spectrum: from patients with longstanding complexity to those who might surprise clinicians with an unexpectedly positive response to initial therapy.

This category underscores the multifactorial nature of PsA management challenges and the need for flexible, adaptive care strategies38,39. C2M-PsA can be identified from either the clinician’s or the patient’s perspective, supporting a holistic approach that maximizes the potential for meaningful improvement (Supplementary Fig. 3).

Definition of treatment-refractory psoriatic arthritis: ‘Treatment-refractory psoriatic arthritis’ is a disease state defined as the failure to respond to ≥ 3 previous treatments for PsA with different modes of action (including ≥ 2 biologic/targeted synthetic DMARDs), AND persistent symptoms perceived as problematic by both the treating clinician and the patient, AND the presence of objective evidence of ongoing inflammation

That a PsA patient is in a TR state must be recognized by both the clinician and the patient in the case of treatment inefficacy and signs and/or symptoms regarded as problematic. ‘Previous treatments for PsA’ refers to csDMARDs or b/tsDMARDs according to the current GRAPPA management recommendations.

TR-PsA represents a more stringent subset within the broader construct of C2M-PsA, and specifically identifies patients with therapeutic resistance due to persistent active inflammation (Fig. 2, Supplementary Fig. 4). The first criterion addresses persistent PsA-related symptoms. The second criterion refers to treatment failure across multiple agents with different MoAs. The third criterion emphasizes a high symptom burden acknowledged by both clinician and patient. The fourth criterion requires objective confirmation of active inflammation, which can be done via clinical examination (for example, swollen joints or active skin disease), imaging (by MSUS or MRI, especially for isolated axial disease or enthesitis), or by laboratory markers (for example, increased CRP or synovial-fluid analysis). This confirmation is necessary to distinguish inflammatory disease from mimickers such as fibromyalgia, hypermobility syndrome or mechanical pain40,41. Finally, alternative explanations for the four criteria must be carefully evaluated and excluded before fulfilling the definition of TR-PsA.

The requirement of inadequate response to at least three therapies, with at least two being b/tsDMARDs, was strongly supported during the Delphi consensus. Although clinicians usually initiate treatment with a csDMARD, often leading to a good response, in other situations a b/tsDMARD might be started directly, particularly for those presenting to dermatologists with severe skin disease, or for those with axial involvement or associated conditions, which typically do not respond well to treatment with csDMARDs. Therefore, prior exposure to csDMARDs is not mandatory for meeting the definition of TR-PsA. Instead, the emphasis is on an inadequate response to treatments with different MoAs, ensuring that the classification captures true therapeutic resistance regardless of treatment sequence (see the clinical vignettes provided in Supplementary Box 1).

Importantly, treatment intolerance or discontinuation due to adverse events should not be counted towards the definition of TR-PsA, as these scenarios do not reflect resistance to therapy. This distinction reinforces the notion that TR-PsA is defined by biological non-response with active inflammation despite adequate therapeutic trials. Moreover, given the expanding landscape of therapeutics targeting distinct immunopathogenic pathways across PsA domains, a threshold of at least three targeted treatments was considered both clinically appropriate, to enhance specificity, and sensitive, to identify patients in whom multiple treatment options have failed without achieving adequate disease control.

Finally, recognizing the importance of defining a specific time frame to ascertain TR-PsA status, the task force concluded that additional evidence is needed to support a consensus owing to the complexity of evaluating treatment responses across diverse therapies. This gap highlights an important area for future research, as well as the need to differentiate between primary and secondary treatment failure (Table 2).

Table 2 Research agenda for complex-to-manage refractory psoriatic arthritis and treatment-refractory psoriatic arthritis
Full size table

Clarifying statements

Following the formal endorsement of the terminology and proposed framework, a series of clarifying statements were drafted and voted upon by the task force to further delineate the boundaries and operationalization of the C2M-PsA and TR-PsA definitions. These statements were designed to resolve areas of ambiguity and guide their consistent application in clinical and research settings.

Clarifying statement 1: complex-to-manage psoriatic arthritis

That a patient with PsA is in a C2M state can be acknowledged by the treating clinician and/or the patient in the case of treatment inefficacy and signs and/or symptoms regarded as problematic. This statement reached 89% agreement within the task force.

Clarifying statement 2: treatment-refractory psoriatic arthritis

That a patient with PsA is in a TR state must be recognized by both the clinician and the patient in the case of treatment inefficacy and signs and/or symptoms regarded as problematic. This statement reached 94% agreement within the task force.

Clarifying statement 3: comorbid conditions

Active disease of associated conditions (such as acute anterior uveitis or inflammatory bowel disease) will also contribute to TR-PsA and should ideally be assessed by the appropriate specialist (e.g. ophthalmologist, gastroenterologist) as part of a multidisciplinary approach that includes the rheumatologist. This statement reached 89% agreement within the task force.

Clarifying statement 4: minimum duration of therapy

The minimal treatment duration of treatment with a b/tsDMARD before assessing treatment failure should be 12–24 weeks according to the MoA. This statement reached 94% agreement within the task force. For C2M-PsA only, this minimum treatment duration does not apply when therapy is discontinued owing to intolerance or adverse effects.

Clarifying statement 5: exclusion criteria for treatment-refractory psoriatic arthritis

Reasons such as contraindications, intolerance, lack of access to treatment, and non-adherence will be incorporated only into the C2M-PsA definition, in order to ensure that the TR-PsA group is as pure as possible and that the TR-PsA definition truly reflects ‘treatment-refractory’ disease with persistent inflammation. This statement reached 100% agreement within the task force.

Discussion

This project represents the first formal effort to develop consensus definitions for C2M-PsA and TR-PsA. These definitions are aimed at standardizing terminology, improving patient stratification and providing a foundation for research and clinical care in PsA cases marked by therapeutic complexity or resistance.

The need for this initiative reflects the fact that, despite therapeutic advances, a substantial proportion of patients continue to report suboptimal outcomes, including residual pain, functional limitations and impaired quality of life42,43. Until now, the absence of standardized criteria to characterize such patients has hindered research comparability and may have contributed to therapeutic inertia in clinical practice. At the same time, growing recognition of disease heterogeneity and variable treatment responses in PsA underscores the importance of a more nuanced classification that takes into consideration all patient characteristics44,45,46.

Emerging conceptual frameworks have further underscored the importance of differentiating between non-inflammatory persistent symptomatic PsA (NIPsA) and persistent inflammatory PsA (PIPsA)47. NIPsA refers to cases in which patients have persistent symptoms despite well-controlled inflammation, often driven by factors such as structural damage, fibromyalgia, central sensitization or hypermobility syndrome, akin to C2M-PsA. By contrast, PIPsA encompasses cases in which active inflammation persists despite multiple therapies, similar to TR-PsA. Recognizing this distinction is essential, as NIPsA could benefit more from multidisciplinary care and non-pharmacological strategies, whereas PIPsA might require optimization of immunomodulatory therapy29,48.

Informed by a scoping literature review19, international HCP and patient surveys20,21, and structured deliberations by the task force, the GRAPPA definitions adopt a multidimensional, patient-centred perspective. C2M-PsA is a broader category that extends beyond pure biological non-response to include persistent symptomatic burden despite the use of at least one b/tsDMARD, including factors such as comorbidities, overlapping conditions, or psychosocial and cultural barriers49,50,51 (Supplementary Table 1). Importantly, with this category the fluctuating nature of disease complexity is acknowledged and the potential for clinical improvement through targeted, multidisciplinary interventions is highlighted38,39.

TR-PsA is a narrower, more stringent subset within C2M-PsA, requiring the presence of four elements: persistent PsA-related symptoms; failure of at least three treatments with different MoAs (including at least two b/tsDMARDs); symptoms being considered problematic by both clinician and patient; and objective evidence of persistent inflammation. Thus, all patients with TR-PsA also fulfil the criteria for C2M-PsA, but not necessarily vice versa (Table 3 and Supplementary Box 2). These criteria are aimed at identifying patients with refractory inflammatory disease who might benefit from novel therapeutic approaches or enrolment in clinical trials and translational research focused on elucidating underlying disease mechanisms52,53. The inclusion of at least three different PsA treatments, rather than a narrower range, accounts for the diversity of therapeutic strategies across PsA domains. The potential inclusion of one csDMARD among the three prior treatments acknowledges regional differences, as cost constraints or reimbursement policies might mandate the use of a csDMARD prior to initiation of advanced therapies54.

Table 3 Comparison of complex-to-manage psoriatic arthritis and treatment-refractory psoriatic arthritis
Full size table

In contrast to the definition of TR-PsA, that of C2M-PsA is intentionally more inclusive, requiring failure of only one b/tsDMARD. Many cases of PsA, such as those involving obesity, multimorbidity or psychosocial barriers, are perceived as complex even before starting advanced treatments. However, as clinical experience shows, some of these patients can respond well to the first b/tsDMARD. Thus, requiring failure of one b/tsDMARD treatment trial allows for identification of complex cases without prematurely categorizing them as such. This layered framework (Fig. 2) provides conceptual granularity and practical thresholds for both clinical and research use.

The thresholds of at least one b/tsDMARD failure for C2M-PsA and at least prior therapies with different MoAs (including at least two b/tsDMARDs) for TR-PsA are consensus-based choices intended to balance inclusivity (early identification of clinically complex presentations) with specificity (delineating true biological non-response). These thresholds were selected to remain applicable across health care systems in which access and costs influence treatment sequencing and to avoid mandating csDMARD failure. Although developed with broad international input, we recognize that access to targeted therapies, imaging and laboratory infrastructure, and socioeconomic factors vary across regions and might influence classification of C2M-PsA and TR-PsA; in low- and middle-resource settings, constrained access to bDMARDs and JAK inhibitors could limit treatment sequencing, yet the treatment exposure requirement within the TR-PsA definition — namely, failure of at least three treatments with different MoAs, including at least two b/tsDMARDs — remains universal. We acknowledge that these thresholds are pragmatic and somewhat arbitrary and that their performance might differ by disease stage and treatment-access context. We propose prospective validation stratified by disease duration and resource context, which might, in time, support alternative exposure thresholds and region-sensitive adaptations while preserving the core consensus definitions.

To enhance practical implementation of the definitions and strengthen reproducibility by reducing interobserver variability, we also provide explicit operational guidance. The structure for this implementation involves five specific pillars: treatment exposure; persistent symptoms; symptoms perceived as problematic; objective inflammation; and exclusion of alternative explanations for ongoing inflammation. For the ‘persistent symptoms’ pillar, we specify composite thresholds (for example, failure to achieve minimal disease activity and domain-specific criteria55. For the subjective component, ‘perceived as problematic’, we prefer anchoring to the patient-acceptable symptom state56 and introduce a parallel physician item, the physician-acceptable state screen, with numeric fallbacks when patient-acceptable symptom state and/or physician-acceptable state screen are unavailable. For TR-PsA, ‘objective inflammation’ is standardized across clinical, laboratory, imaging, synovial fluid, and active associated conditions, with prespecified thresholds. Full details are provided in Supplementary Box 2.

Several aspects of this work deserve emphasis. First, the development process was grounded in stakeholder engagement, including early and sustained involvement of PRPs, who played a critical role in shaping terminology, especially the replacement of ‘difficult-to-manage’ with ‘complex-to-manage’. Unlike other initiatives that rely solely on input from a small number of patients, we aimed to capture a wider range of patient perspectives by developing a multilingual patient survey that yielded nearly 600 responses from people with PsA worldwide. Furthermore, the framework builds on established principles from prior consensus efforts, including ASAS and EULAR definitions of D2T disease10,23, while adapting them to the specific pathophysiology and clinical features of PsA.

The requirement of objective evidence of inflammation in the TR-PsA definition was widely supported by GRAPPA membership, aligning with the need to differentiate inflammatory disease from chronic pain, which can be difficult with physical examination alone57,58. However, the limitations of currently available biomarkers and imaging modalities were acknowledged, underscoring the importance of expert clinical judgment. In practice, whichever imaging modality is available should be used; MSUS is particularly valuable in peripheral disease, as it can detect synovitis, erosions, tenosynovitis, paratenonitis and dactylitis, whereas MRI is especially relevant for axial disease, enabling assessment of bone-marrow oedema, sacroiliitis and deep soft-tissue inflammation. The modalities are complementary, and feasibility and availability should guide prioritization59. Regarding the timing of therapeutic-failure adjudication, we considered a minimum exposure of 12–24 weeks before determining failure, reflecting the heterogeneous onset kinetics of PsA therapies. This window mirrors pharmacodynamic response profiles and aligns with treat-to-target guidance that recommends reassessment at approximately 12 weeks and attainment of the treatment target by ~6 months60. Consistent with this approach, the 2023 EULAR recommendations for PsA management advises reassessing csDMARD efficacy at ~12 weeks, aiming for ≥ 50% improvement by 3 months, and achievement of the treatment target by 6 months8, and the GRAPPA 2021 recommendations support reassessment and potential escalation every 12–24 weeks7. We acknowledge that optimal timing might vary by MoA and disease domain, underscoring the need for prospective validation (Table 2).

A key limitation of this initiative lies in the fact that the proposed definitions were developed through a structured consensus process rather than being derived from prospective cohort data. As such, their predictive value remains to be established. However, the definitions were grounded in extensive international stakeholder input and explicitly designed to generate hypotheses and guide future studies. The diverse views within the task force also reflect ongoing uncertainties in the field — such as optimal thresholds for treatment failure or the precise role of contextual factors — but these uncertainties were addressed through transparent voting and discussion. Importantly, a major strength of this effort is its emphasis on inclusivity and clinical applicability, which was achieved by integrating the views of HCPs and patients across multiple continents and specialties. This patient-informed approach ensured that the definitions account for both biological resistance and the broader realities of PsA management.

Another limitation of our drug exposure criterion for TR-PsA is that it counts MoAs, but prior therapies are not weighted by relative efficacy or domain coverage. Therefore, prospective validations should stratify outcomes according to the therapeutic target of the treatment class (for example, TNF, IL-17, IL-23, JAK or PDE4) and also evaluate potency- and domain-weighted exposure metrics.

Despite the achievement of a strong consensus on the definitions, further validation in clinical and research settings is essential. Future studies should evaluate the performance of the definitions in prospective cohorts, registries and pragmatic trials, assess their prognostic utility and explore their role in guiding treatment decisions. To facilitate structured uptake and ensure scientific rigour, five quality indicators have been proposed: external validation in independent cohorts; inclusion in clinical recommendations; citation frequency; use in clinical trials; and application in translational research (Supplementary Table 2). These efforts will ensure that the definitions are both clinically meaningful and methodologically robust, ultimately strengthening their role in real-world application and policy development. Finally, to ensure their continued relevance over time, a formal re-evaluation by the GRAPPA task force is planned 3 years post-publication (Supplementary Fig. 2), allowing for incorporation of new data and clinical experience to determine whether refinement is warranted.

Conclusions

These consensus-derived GRAPPA definitions provide a shared framework for standardizing terminology, supporting individualized care and improving patient stratification in both research and practice. They offer practical tools for clinicians and researchers, serving as a foundation for future observational studies and clinical trials, and supporting the earlier recognition of patients with C2M or TR disease. Prospective validation across diverse settings and phenotypes is needed to establish their reliability, responsiveness and clinical utility.