Real-world performance of EULAR and GRAPPA definitions of difficult-to-treat psoriatic arthritis within the ORCHESTRA cohort: an observational cohort study

Abstract

Background:

Psoriatic arthritis (PsA) is a complex inflammatory disease with significant musculoskeletal, dermatological, and comorbidity burden. Despite therapeutic advances, many patients remain difficult to treat (D2T). Conventional definitions often overlook distinctions between inflammatory and non-inflammatory drivers, risking inappropriate treatment escalation. Two consensus frameworks—the European Alliance of Associations for Rheumatology (EULAR) and the Group for Research and Assessment of Psoriasis and Psoriatic Arthritis (GRAPPA)—have recently been proposed to better capture this complexity based on disease drivers.

Objectives:

To evaluate the applicability, overlap, and clinical utility of EULAR and GRAPPA definitions for difficult/complex-to-manage (D2M/C2M) and treatment-refractory PsA (TR-PsA) in a real-world cohort.

Methods:

We conducted a retrospective analysis within the prospective Ottawa Rheumatology CompreHEnSive TReatment and Assessment (ORCHESTRA) cohort, including consecutive PsA patients initiating or switching advanced therapy between April 2022 and September 2025. Patients were classified according to EULAR and GRAPPA frameworks into the broader D2M-PsA (EULAR) and C2M-PsA (GRAPPA) groups, and narrower inflammation-restricted TR-PsA subsets. Prevalence, overlap, and agreement were assessed using descriptive statistics and kappa analysis. Baseline clinical, comorbidity, and ultrasound characteristics were compared across groups.

Results:

Seventy-six PsA patients were included. GRAPPA identified more patients as C2M-PsA (35/76; 46%) than EULAR’s D2M-PsA (16/76; 21%). All EULAR D2M-PsA patients were included in C2M-PsA, whereas 19 additional patients were uniquely identified by GRAPPA due to less stringent treatment failure thresholds. Agreement between EULAR D2M-PsA and GRAPPA C2M-PsA was moderate (κ = 0.48). TR-PsA classification showed perfect agreement (κ = 1.000), with identical patients identified by both frameworks. D2M/C2M patients had longer disease duration and greater prior therapy exposure than non-D2T patients, while objective inflammatory measures were largely comparable.

Conclusion:

EULAR and GRAPPA D2T PsA frameworks share conceptual intent but differ in real-world clinical capture. EULAR preferentially enriches for multidomain involvement and inflammatory refractoriness after multiple therapy failures, whereas GRAPPA identifies a broader population reflecting earlier disease complexity. While both frameworks identified the same patients using inflammation-restricted criteria, differences in operationalization limit assumptions of interchangeability.

Plain language summary

Understanding how different frameworks define “difficult-to-treat” psoriatic arthritis in real-world patients

Psoriatic arthritis (PsA) is a chronic inflammatory condition that can affect joints, skin, and many other aspects of health. Although several effective treatments are available, some patients remain difficult to manage. In these patients, ongoing symptoms may be caused by active inflammation, but also by other factors such as pain sensitization, comorbid conditions, or treatment intolerance. Distinguishing between these causes is important to avoid unnecessary escalation of immune-suppressing therapy. Recently, two expert groups—GRAPPA and EULAR—proposed new frameworks to better define “difficult-to-treat” PsA by considering disease drivers rather than treatment failure alone. However, how these frameworks perform in real-world clinical practice has not been well studied. In this study, we applied both frameworks to a real-world cohort of PsA patients from the ORCHESTRA registry in Canada who were starting or changing advanced therapies. We compared how often each framework identified patients as difficult-to-manage, how much they overlapped, and whether they identified similar clinical profiles. We found that the GRAPPA framework identified more patients as having complex or difficult-to-manage PsA, capturing patients earlier in their disease course with fewer treatment failures. In contrast, the EULAR framework identified a smaller group with more advanced disease and greater treatment exposure. When focusing only on patients with true inflammatory treatment refractoriness, both frameworks identified exactly the same patients. Overall, our findings suggest that while GRAPPA and EULAR share similar goals, they identify different patient populations in practice. GRAPPA may be more useful for recognizing early disease complexity, while EULAR may better identify patients with established inflammatory refractoriness. These differences should be considered when using these frameworks in clinical care and research.

Keywords

difficult-to-treat psoriatic arthritis EULAR GRAPPA psoriatic arthritis real-world evidence treatment refractoriness

Introduction

Psoriatic arthritis (PsA) is a chronic, multifaceted inflammatory disorder characterized by involvement of the joints, entheses, digits, axial skeleton, skin, and nails.¹ Beyond its musculoskeletal and dermatological manifestations, PsA imposes a substantial burden through a range of comorbidities, including fibromyalgia, depression, obesity, and cardiovascular disease, all of which contribute to significant health and psychosocial consequences.^2–5 Despite notable advances in therapeutic options, a substantial proportion of patients remains difficult to treat (D2T).⁶ The complexity of these D2T cases extends well beyond true inflammatory refractoriness, as non-inflammatory comorbidities frequently impact disease activity measures. In addition, treatment obstacles may arise from medication side effects, contraindications, or limited access. The existing literature on D2T PsA presents considerable heterogeneity in its definition, having variable nomenclature and threshold criteria, often centered on the number of failed therapies⁷ and largely adapted from the European Alliance of Associations for Rheumatology (EULAR) D2T rheumatoid arthritis (RA) framework.⁸ Consequently, most definitions of D2T PsA do not stratify patients according to their specific disease state, those with persistent inflammatory activity and those whose disease burden is predominantly driven by non-inflammatory factors; rather, they encompass all cases within a single D2T category.^6,9–11 This conflation risks obscuring the underlying mechanisms of non-response, hindering the identification of distinct patient subgroups, and may lead to the inappropriate escalation of immunosuppressive therapy when inflammation is not the primary driver of disease activity

To address these challenges, recent initiatives from the EULAR and the Group for Research and Assessment of Psoriasis and Psoriatic Arthritis (GRAPPA) have focused on developing clearer and standardized definitions, which were recently published.^12,13 Both frameworks introduced definitions that capture a broad category of patients, irrespective of the underlying driver of failure, encompassing not only biological non-response but also comorbidities and treatment-related adverse events, as well as a narrower category reflecting true inflammatory refractoriness. Although the specific criteria differ slightly, both frameworks share a similar conceptual structure. One key distinction, however, lies in their intended application: the EULAR criteria were primarily developed to facilitate the identification of more homogeneous patient populations for research, whereas the GRAPPA framework was designed to support clinical shared decision-making and application in observational studies.

Within the broad disease-state category, EULAR proposed difficult-to-manage PsA (D2M-PsA), defined by failure of ⩾2 advanced therapies with different mechanisms of action, together with patient- or physician-perceived inadequate control and unmet treatment targets. In parallel, GRAPPA introduced the broad category of complex-to-manage PsA (C2M-PsA), defined as persistent disease impact despite at least one adequate trial of advanced therapy

Within each framework, a narrower subset of true treatment-refractory cases requires objective evidence of active inflammation (clinical, laboratory, or imaging). EULAR and GRAPPA define this group as treatment-refractory PsA (TR-PsA). Although the criteria within these definitions differ slightly, the concepts are closely aligned. These consensus-based definitions represent an important advance, but their applicability and limitations in routine clinical practice remain untested. The primary objective of this study was to determine the applicability, prevalence, and agreement of TR-PsA and D2M/C2M-PsA EULAR and GRAPPA consensus definitions within a real-world cohort who are initiating a new advanced therapy. Secondary objectives included characterizing and comparing the demographic, clinical, and comorbidity profiles of patients who fulfilled versus did not fulfill these definitions.

Methods

Study design and population

This single-center observational study is a retrospective analysis of data from the prospective longitudinal cohort, the Ottawa Rheumatology CompreHEnSiveness TReatment and Assessment Clinic (ORCHESTRA), established in April 2022 at the Ottawa University Hospital. The reporting of this study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.¹⁴

ORCHESTRA operates within a shared-care model with referring primary rheumatologists, and it includes patients with inflammatory arthritis: RA, PsA, and ankylosing spondylitis patients who are initiating or switching to a new biologic disease-modifying antirheumatic drug (bDMARD) or targeted synthetic disease-modifying antirheumatic drug (tsDMARD) therapies at the Ottawa Hospital Rheumatology Division. Patients are assessed at baseline (prior to therapy initiation), at 3- and 6-months following treatment initiation, and subsequently every 3 months until individualized treatment targets, defined as low disease activity or remission, are achieved. At each visit, a standardized comprehensive assessment is performed, including demographic characteristics, clinical and ultrasound disease activity measures, laboratory markers of inflammation (erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP)), and comorbidities.

For the present analysis, all consecutive PsA patients enrolled in the ORCHESTRA cohort between April 2022 and September 2025 who met the eligibility criteria were included. As this was a retrospective, exploratory analysis of an ongoing registry, no formal sample size calculation was performed.

Inclusion criteria

Age ⩾18 years at the time of ORCHESTRA enrollment.

Confirmed diagnosis of PsA, as determined by the treating rheumatologist.

Enrollment in the ORCHESTRA cohort, which captures patients referred for initiation or switching of advanced therapy.

Availability of baseline clinical data required to apply EULAR and GRAPPA definitions.

Exclusion criteria

Diagnosis of inflammatory arthritis other than PsA

Age <18 years

Absence of sufficient baseline data to operationalize EULAR or GRAPPA definitions

The study was approved by the Ottawa Health Science Network Research Ethics Board (OHSN-REB 20220187-01H) and conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Written informed consent was obtained from all participants.

Operational definitions

Baseline was defined as the initial ORCHESTRA visit at which patients were assessed to determine whether they fulfilled the proposed definition criteria for both EULAR and GRAPPA based on historical treatment exposure, current clinical assessment, disease activity measures, and patient- and physician-reported perceptions of disease impact (Table 1).^12,13 We adhered closely to the original definitions and their suggested approaches to operationalization, applying minor pragmatic adaptations only when explicit thresholds or measures were not specified in either framework.

Table 1.

Summary of proposed definitions for D2M/C2M-PsA and TR-PsA according to EULAR and GRAPPA frameworks.

Difficult-to-Manage (D2M-PsA)/Complex-to-Manage (C2M-PsA)
EULAR (D2M-PsA)	Treatment failure history	Treatment in accordance with EULAR recommendations and failure of ⩾2 advanced therapies with different mechanisms of action, regardless of the reason for failure.^∆
	Evidence of persistence disease activity in any ⩾1 of the following	(A) Failure to achieve or maintain low disease activity	DAPSA > 14 or MDA < 5
		(B) Active EMM	Nail disease, Skin BSA ⩾3%, Uveitis, Active IBD flare^‡‡
		(C) Objective evidence of inflammation: (Clinical, Laboratory, or image)	- SJC ⩾ 1- Dactylitis ⩾ 1 digit- BASDAI ⩾ 4.- SPARCC ⩾ 1- ESR ⩾ 30, CRP ⩾ 10- Image evidence of inflammation in peripheral or axial MSK (US or MRI)
	Clinical perception	PtGA or PhGA ⩾ 4
GRAPPA (C2M-PsA)	Pillar (1): Treatment exposure	Failure of ⩾1 advanced therapy (b/tsDMARD) after at least 3 months of therapy per GRAPPA guidance. (unless discontinued earlier due to adverse events/intolerance then 3 months is not required).^∆
	Pillar (2): Persistent symptoms	Persistent symptoms in ⩾1 domain: (Any of the following)-DAPSA > 14 or MDA < 5-Arthritis (TJC ⩾ 3 or SJC ⩾ 1)-Enthesitis (SPARCC ⩾ 1)-Dactylitis ⩾ 1 digit-Nail disease-Skin involvement (BSA ⩾ 1%)-Axial involvement (BASDAI ⩾ 4)
	Pillar (3): Perceived as problematic	Patient OR physician global assessment (PtGA or PhGA) ⩾ 4
Treatment-Refractory-PsA (TR-PsA) in EULAR and GRAPPA
EULAR (TR-PsA)	Treatment Failure history	Failure of ⩾2 advanced therapies with different modes of action
	Evidence of persistent disease activity by ⩾2 of the following, with [C] being mandatory	A) Failure to achieve or maintain low disease activity	DAPSA > 14 or MDA < 5
		B) Active EMM	Nail disease Skin BSA ⩾ 3%, Uveitis, Active IBD flare
		C) Objective evidence of inflammation: (Clinical, Laboratory, or image)	- SJC ⩾ 1- Dactylitis ⩾ 1 digit- BASDAI ⩾ 4.- SPARCC ⩾ 1- ESR ⩾ 30, CRP ⩾ 10- Image evidence of inflammation in peripheral or axial MSK (US or MRI)
Treatment-Refractory-PsA (TR-PsA) in EULAR and GRAPPA
	Clinical perception	PtGA or PhGA ⩾ 4
	Exclusion	TR-PsA refers to primary and secondary failure, but not discontinuation due to side effects/intolerability/contraindication, and treatment failure is not better explained by other mechanistic or confounding causes^∆
GRAPPA (TR-PsA)	Pillar (1): Treatment exposure	Prior exposure to ⩾3 therapies with different modes of action (including ⩾2 biologic/targeted synthetic DMARDs) with treatment duration of at least 3 months per GRAPPA guidance
	Pillar (2): Persistent symptoms	Persistent symptoms in ⩾1 domain: (Any of the following)-DAPSA > 14 or MDA < 5-Arthritis (TJC ⩾ 3 or SJC ⩾ 1)-Enthesitis (SPARCC ⩾ 1)-Dactylitis ⩾ 1 digit-Nail disease-Skin involvement (BSA ⩾ 1%)-Axial involvement (BASDAI ⩾ 4)
	Pillar (3): Perceived as problematic	PtGA AND PhGA ⩾ 4
	Pillar (4): Objective inflammation (Clinical, Laboratory, or image)	Symptomatic disease with objective evidence of active inflammation in any ⩾1 of these domains:-SJC ⩾ 1Dactylitis ⩾ 1 digit-Nail disease-Skin (BSA ⩾ 1%),-Uveitis^‡‡ -IBD^‡‡ -ESR ⩾30, CRP ⩾ 10^‡ -Synovial WBC > 2000^‡ -Ultrasound evidence of inflammation (i.e., synovitis, tenosynovitis)-Enthesitis: SPARCC > 1 plus ultrasound signs of inflammation-Axial disease (BASDAI ⩾ 4 with MRI evidence of inflammation (sacroiliitis or spinal osteitis).
	Pillar (5): Exclusion of alternative	Exclusion of alternative causes that better explain the cause of treatment failure (e.g. Comorbidities, psychosocial) or treatment-related challenges^€

∆

This is multifactorial drivers beyond objective inflammation and include multiple drivers that are leading to the C2M state (e.g., fibromyalgia, mood/sleep disorders, obesity, OA/CPPD, access/adherence treatment side effects).

‡

Providing that no other causes likely explain these laboratory findings.

‡‡

Uveitis (ophthalmologic confirmation) or inflammatory bowel disease (clinical/ endoscopic/ histologic or fecal calprotectin evidence).

€

Exclusion criteria: Comorbidities or other conditions may contribute to symptoms or biomarker elevations, such as infection, obesity-related CRP, metabolic syndrome, dyslipidemia, fibromyalgia/central sensitization, generalized hypermobility, or mechanical/degenerative pain, gout or CPPD, anxiety, and depression. These factors should be recognized as potential contributors, but do not preclude TR-PsA classification if PsA activity is the primary driver of disease manifestations. Exclusion should occur when these conditions better explain the clinical picture.

BASDAI, bath ankylosing spondylitis disease activity index; bDMARD, biologic disease-modifying antirheumatic drug; BSA, body surface area; C2M-PsA, complex-to-manage psoriatic arthritis; CRP, C-reactive protein; D2M-PsA, difficult-to-manage psoriatic arthritis; DAPSA, disease activity in psoriatic arthritis score; ESR, erythrocyte sedimentation rate; EULAR, European Alliance of Associations for Rheumatology; GRAPPA, Group for Research and Assessment of Psoriasis and Psoriatic Arthritis; IBD, inflammatory bowel disease; MDA, minimal disease activity; MRI, magnetic resonance imaging; PhGA, physician global assessment; PsA, psoriatic arthritis; PtGA, patient global assessment; SJC, swollen joint count; SPARCC, spondyloarthritis research consortium of Canada enthesitis index; TJC, tender joint count; TR-PsA, treatment-refractory psoriatic arthritis; tsDMARD, targeted synthetic DMARD; WBC, white blood cell count.

Both the EULAR and GRAPPA frameworks are conceptually designed to identify patients with PsA who are challenging to manage, distinguishing between predominantly non-inflammatory complexity (D2M-PsA, EULAR; or C2M-PsA, GRAPPA) and true inflammatory treatment refractoriness (TR-PsA). Accordingly, operationalization incorporated three core domains applicable to all patients, with two additional criteria applied to identify the subset with true inflammatory TR-PsA.

Prior advanced therapy failure

The first requirement was a history of failure of prior advanced therapy after an adequate therapeutic trial, guided by EULAR and GRAPPA PsA treatment recommendations.^15,16 Advanced therapies included tumor necrosis factor inhibitors (TNFi), interleukin-17 inhibitors (IL-17i), interleukin-12/23 inhibitors (IL-12/23i), and Janus kinase inhibitors (JAKi). A minimum treatment duration of approximately 12 weeks was required, in line with treat-to-target consensus principles, which recommend early assessment for improvement at around 3 months and target attainment by 6 months. This minimum duration was used to preserve inclusion, recognizing that in routine clinical practice, treatment switching may occur earlier in cases of clear non-response. Treatment duration varied across the cohort.

Persistent disease activity

The second requirement was evidence of persistent disease activity despite therapy, assessed using clinical, laboratory, and imaging measures, with thresholds applied as specified in the original EULAR and GRAPPA operationalizations. For axial disease, EULAR refers to inflammatory back pain without specifying a quantitative activity measure, whereas GRAPPA suggests a Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) ⩾ 4. Given the retrospective nature of data capture and the absence of a validated axial PsA activity instrument, the BASDAI ⩾ 4 threshold was pragmatically applied to both definitions to ensure consistency.

Perception of problematic disease

The third domain captured the perception of disease as problematic from both the patient and physician perspectives. In accordance with GRAPPA guidance, patient and physician global assessments were used, applying a cutoff of ⩾4, acknowledging that this threshold is pragmatic and not formally validated.

Additional criteria for identification of treatment-refractory PsA (TR-PsA)

Two additional criteria were required to classify patients as having true inflammatory TR-PsA. These included objective evidence of ongoing inflammation and exclusion of non-inflammatory drivers of symptoms. Patients whose treatment changes were primarily due to contraindications or intolerance, or whose disease activity was predominantly attributable to non-inflammatory conditions such as osteoarthritis or fibromyalgia, were not classified as having inflammatory treatment-refractory disease.

Musculoskeletal ultrasound assessment

Musculoskeletal ultrasound (US) examinations were performed at the end of each clinical visit as part of routine clinical care by a rheumatologist experienced in musculoskeletal sonography, who was not involved in D2T subgroup classification, as D2T phenotypes were applied retrospectively. All scans were conducted using a Logiq E10 ultrasound system (GE Healthcare, Milwaukee, WI, USA) equipped with a broadband 6–20 MHz linear probe.

A standardized scanning protocol was followed, obtaining dorsal and palmar views in a neutral joint position. A total of 46 joints were examined bilaterally: metacarpophalangeal (MCP) joints 1–5, proximal interphalangeal (PIP) joints 1–5, distal interphalangeal (DIP) joints 1–5, wrists, elbows, knees, ankles, and metatarsophalangeal (MTP) joints 2–5. Power Doppler (PD) settings were standardized with a pulse repetition frequency of 500 Hz, low wall filter, and gain adjusted until background signal disappearance. Synovitis was defined according to the OMERACT US Task Force definitions. A semi-quantitative scoring system (0–3) was used for grading both gray-scale (GS) and PD components¹⁷:

GS synovitis: 0 = none; 1 = mild; 2 = moderate; 3 = marked synovial hypertrophy.

PD signal: 0 = no flow; 1 = ⩽ 2 vessel (or signals) in small joints / ⩽ 3 in large joints; 2 = PD signal > grade 1 but < 50% of synovial area; 3 = PD signal occupying > 50% of the GS synovial area.

A combined OMERACT–EULAR synovitis score (0–3) was derived for each joint based on GS and PD findings, and the global OMERACT–EULAR Synovitis Score (GLOESS) was calculated by summing combined synovitis scores across the 46 assessed joints (range 0–138).¹⁸

Enthesitis was assessed by the GS features (hypoechogenicity, thickening, enthesophytes, erosions, and calcifications) and PD signal at enthesis sites. All findings were scored on a scale between 0 and 3. Relevant features were summed to achieve an inflammatory and chronicity score, as described in our group’s previous publications.¹⁹

In this analysis, the criteria component of objective sonographic inflammation was defined as follows:

Ultrasound synovitis: gray-scale (GS) score ⩾1 AND PD score ⩾1

Ultrasound tenosynovitis: GS score ⩾1 and PD score ⩾1

Ultrasound enthesitis: presence of hypoechogenicity or entheseal thickening, and PD signal

Statistical analysis

Descriptive statistics were employed to summarize the demographic and clinical characteristics of the patients. The normality of data distribution was assessed using the Kolmogorov–Smirnov test. Continuous variables were presented as mean ± standard deviation or median with interquartile range (IQR), depending on the distribution. Categorical variables were expressed as frequencies and percentages. Between-group comparisons were performed using the Chi-square test or Fisher’s exact test for categorical variables, and the independent Student’s t-test or Mann–Whitney U test for continuous variables, as appropriate. The agreement between the EULAR and GRAPPA definitions was evaluated using kappa analysis. The strength of agreement was interpreted according to the guidelines proposed by Landis and Koch: 0.01–0.20 slight, 0.21–0.40 fair, 0.41–0.60 moderate, 0.61–0.80 substantial, and 0.81–1.00 almost perfect. A p-value <0.05 was considered statistically significant.²⁰ Given the limited sample size and low number of outcome events within several subgroups, analyses were restricted to descriptive summaries and unadjusted comparisons only. No multivariable regression analyses were performed. These analyses were considered exploratory and hypothesis-generating. All statistical analyses were conducted using SPSS (version 22.0; IBM Corp., Armonk, NY, USA).

Results

Baseline characteristics and ultrasound findings

Seventy-six patients with PsA were enrolled in the study, of whom 42 (55.3%) were female. The mean age was 48.3 ± 12.7 years. The median disease duration was 4.0 years (IQR, 1.0–14.8 years), and the median symptom duration was 8.0 years (IQR, 3.0–18.0 years). The most common clinical phenotype was symmetric polyarthritis, observed in 64 patients (84.2%), while enthesitis and dactylitis were observed in 39 (51.3%) and 42 (55.3%) patients, respectively. Ultrasound assessment revealed a median GLOESS score of 19.0 (IQR, 11.0–28.0), indicating moderate synovial inflammation. The median Enthesitis Inflammation Score was 6.0 (IQR, 2.0–11.0), with low median enthesis Doppler activity of 0.0 (IQR, 0.0–1.0). Additional sociodemographic, clinical, and ultrasound findings across definitions are summarized in Table 2.

Table 2.

Baseline sociodemographic, clinical characteristics, and ultrasound scores of PsA patients in the ORCHESTRA cohort and across different definitions.

Characteristic	Total patients, ORCHESTRA, N: 76	EULAR			GRAPPA
Characteristic	Total patients, ORCHESTRA, N: 76	Non-D2M/TR-PsA, N: 60	D2M, N: 16	TR-PsA, N: 14	Non-C2M/TR-PsA, N: 41	C2M, N: 35	C2M without TR-PsA, N: 21	TR-PsA, N: 14
Sex, male^#	34 (44.7%)	24 (40%)	10 (62.5%)	9 (64.3%)	17 (41.5%)	17 (48.6%)	8 (38.1%)	9 (64.3%)
Age^$	48.25 ± 12.72	46.67 ± 12.42	54.33 ± 12.40	52.46 ± 12.08	46.08 ± 12.71	50.74 ± 12.44	49.67 ± 12.84	52.46 ± 12.08
Body mass index*	30.96 (26.51–38.98)	32.06 (26.56–38.98)	29.29 (24.27–40.45)	29.29 (24.51–38.73)	30.50 (26.54–40.09)	31.40 (24.39–37.92)	32.70 (23.20–37.96)	29.29 (24.50–38.73)
Documented Depression	20 (26.7%)	17 (28.3%)	3 (20%)	3 (21.4%)	13 (31.7%)	7 (20.6%)	4 (20%)	3 (21.4%)
Fibromyalgia	9 (11.8%)	7 (11.7%)	2 (12.5%)	2 (14.3%)	4 (9.8%)	5 (14.3%)	3 (14.3%)	2 (14.3%)
Symmetric Polyarthritis Pattern^#	64 (84.2%)	48 (80%)	16 (100%)	14 (100%)	32 (78%)	32 (91.4%)	18 (85.7%)	14 (100%)
Oligoarthritis/Monoarthritis Pattern^#	7 (9.2%)	7 (11.7%)	0 (0%)	0 (0%)	5 (12.2%)	2 (5.7%)	2 (9.5%)	0 (0%)
DIP Predominant Pattern^#	2 (2.6%)	2 (3.3%)	0 (0%)	0 (0%)	2 (4.9%)	0 (0%)	0 (0%)	0 (0%)
Axial Disease Pattern^#	16 (21.1%)	13 (21.7%)	3 (18.8%)	2 (14.3%)	9 (22%)	7 (20%)	5 (23.8%)	2 (14.3%)
Arthritis Mutilans Pattern^#	0 (0%)	0(0%)	0(0%)	0(0%)	0(0%)	0 (0%)	0 (0%)	0(0%)
C-reactive protein, mg/l*	3.7 (1.38–10.38)	4.3 (1.8–13.2)	1.8 (0.7–5.2)	2.3 (0.9–6)	5.8 (1.8–12.9)	3 (0.9–7.8)	3 (0.9–13.7)	2.3 (0.9–6)
Erythrocyte sedimentation rate, mm/h*	15 (7–29)	15 (7–34.5)	14 (5–22)	13 (4.5–21)	16 (7–33)	13.5 (6.3–24.8)	14 (7–45)	13 (4.5–21)
Disease duration, years*	4 (1–14.75)	3.5 (1–9.5)	16.5 (4.3–24)	16.5 (4.8–24)	2 (0.8–5)	8 (3–18)	7 (1.5–15)	16.5 (4.8–24)
Symptoms duration, years*	8 (3–18)	6.5 (2–14.8)	20.5 (8–27.8)	20.5 (7.8–25.3)	5 (1.8–10)	14 (8–24)	12 (6–22)	20.5 (7.8–25.3)
Enthesitis (Ever)^#	39 (51.3%)	29 (48.3%)	10 (62.5%)	9 (64.3%)	21 (51.2%)	18 (51.4%)	9 (42.9%)	9 (64.3%)
Dactylitis (Ever) ^#	42 (55.3%)	36 (60%)	6 (37.5%)	5 (35.7%)	30 (73.2%)	12 (34.3%)	7 (33.3%)	5 (35.7%)
Nail psoriasis (Ever) ^#	39 (51.3%)	29 (48.3%)	10 (62.5%)	8 (57.1%)	17 (41.5%)	22 (62.9%)	14 (66.7%)	8 (57.1%)
Skin psoriasis (Ever) ^#	63 (82.9%)	47 (78.3%)	16 (100%)	14 (100%)	31 (75.6%)	32 (91.4%)	18 (85.7%)	14 (100%)
Inflammatory bowel disease^#	8 (10.7%)	4 (6.7%)	4 (26.7%)	3 (21.4%)	3 (7.3%)	5 (14.7%)	2 (9.5%)	3 (21.4%)
Uveitis^#	6 (7.9%)	4 (6.7%)	2 (12.5%)	2 (14.3%)	4 (9.8%)	2 (5.7%)	0 (0%)	2 (14.3%)
Gout^#	4 (5.3%)	1 (1.7%)	3 (18.8%)	3 (21.4%)	0 (0%)	4 (11.4%)	1 (4.8%)	3 (21.4%)
Previous number of csDMARDs*	2 (1–3)	2 (1–3)	2 (1–3)	2 (1–3)	2 (1–2.5)	2 (1–3)	2 (1.5–3)	2 (1–3)
Previous number of advanced therapies*	1 (0–2)	0 (0–1)	3 (2–4)	3 (2–4)	0 (0–0)	2 (1–3)	1 (1–1.5)	3 (2–4)
Morning stiffness^#	73 (96.1%)	57 (95%)	16 (100%)	14 (100%)	39 (95.1%)	34 (97.1%)	20 (95.2%)	14 (100%)
Duration of morning stiffness*	1 (0.5-2)	1 (0.6-2)	1 (0.5–6)	1 (0.5–4.5)	1 (0.5–2)	1 (0.8–3)	1.5 (1–3)	1 (0.5–4.5)
Swollen joint count (66)*	5 (2–8)	5 (2–7.8)	6 (2–9)	6 (2–9)	4 (1.5–7.5)	5 (2–8)	5 (1.5–8)	6 (2–9)
Tender joint count (68)*	9 (4-18.75)	9.5 (4–18)	8.5 (3–19)	8.5 (3–19)	8 (3–15.5)	12 (6–23)	17 (6.5–24)	8.5 (3–19)
Pain VAS*	6 (4-7)	6 (4–7)	6 (5–7)	6 (5–6.5)	5 (4–7)	6 (5–7)	6 (5–7)	6 (5–6.5)
Patient VAS*	6 (5–7.75)	6 (5–7.8)	6 (5.3–7.8)	6 (5–7.3)	6 (4–7)	6 (6–8)	7 (6–8.5)	6 (5–7.3)
Physician VAS*	6 (4.25–7)	5 (4–7)	6 (5–7)	6 (5–7)	5 (3.5–6)	6 (5–7)	6 (5–7)	6 (5–7)
HAQ^$	1.01 ± 0.73	1.02 ± 0.77	0.96 ± 0.55	0.87 ± 0.53	0.83 ± 0.71	1.22 ± 0.71	1.44 ± 0.72	0.87 ± 0.53
SPARCC*	2 (0–4)	2 (0–4)	2 (0.3–4)	2 (0–4.5)	2 (0–3.8)	2 (0.8–6)	2 (1–6.8)	2 (0–4.5)
BASDAI*	6.05 (4.28–7.21)	6.05 (3.95–7.2)	6.05 (4.62–7.36)	5.65 (4.43–7.06)	5.6 (3.5–7.1)	6.4 (5.3–7.4)	7 (5.5–7.6)	5.7 (4.4–7.1)
BSA Psoriasis*	0.5 (0–2)	1 (0–2)	0 (0–1)	0.5 (0–1.3)	0 (0–2)	1 (0–2)	1 (0–2)	0.5 (0–1.3)
B-mode score*	19 (10–27)	19 (8–27)	18 (11.8–33.5)	18 (11.3–32)	19 (8–26)	18 (11.3–32.8)	18 (11.3–34.3)	18 (11.3–32)
Doppler score*	3 (0–6.5)	3 (0–6.3)	3 (0–8)	4 (0–10.5)	3 (0–5)	2.5 (0–8.5)	2 (1–9)	4 (0–10.5)
GLOESS score*	19 (11–28)	19 (10–27)	19.5 (11.3–32)	19.5 (10.8–30)	18.5 (9.3–26.8)	21 (11–33)	21 (11.5–35.5)	19.5 (10.8–30)
Enthesis-inflammation score*	6 (2–11)	5 (1–11)	8 (5–12.5)	8 (5–14)	5 (2.5–11)	7.5 (1.8–11)	6.5 (0–10.8)	8 (5–14)
Enthesis Doppler score*	0 (0–1)	0 (0–1.8)	0 (0–0)	0 (0–0.5)	0 (0–2)	0 (0–0)	0 (0–0)	0 (0–0.5)
Enthesis chronicity score*	10.5 (3–15)	10 (2.8–13.5)	11 (3.3–17.8)	11 (3.8–17.3)	11 (3–13)	10 (3–17.3)	8 (2–16.5)	11 (3.8–17.3)
Enthesis total score*	15 (6–24.5)	15 (5.5–24)	18.5 (6.8–36.3)	18.5 (8.3–31.3)	15 (6–24)	15 (4.8–25.8)	13 (3.3–23.5)	18.5 (8.3–31.3)

Median (IQR).

Mean ± SD.

n (%).

BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; BSA, body surface area; csDMARDs, conventional synthetic disease-modifying antirheumatic drugs; GLOESS, global OMERACT–EULAR synovitis score; HAQ, health assessment questionnaire; SPARCC, Spondyloarthritis Research Consortium of Canada enthesitis index; VAS, visual analog scale.

EULAR and GRAPPA definitions’ prevalence, and agreements

According to the EULAR definition, 16 patients (21.1%) met D2M-PsA criteria, including one patient (1.3%) with D2M alone and 14 (18.4%) who met the criteria for TR-PsA. Overall, 60 (78.9%) were non-D2M-PsA. Based on the GRAPPA definition, 35 patients (46.1%) were categorized as having C2M-PsA, including 21 (27.6%) who had C2M without objective inflammation (non-TR-PsA), and 14 (18.4%) who met the criteria for TR-PsA, whereas the remaining 41 (53.9%) were classified as non-C2M PsA.

There was moderate agreement between the EULAR D2M-PsA and GRAPPA C2M-PsA classifications (κ = 0.476). All patients classified as D2M-PsA by EULAR were also categorized as C2M-PsA by GRAPPA; however, 19 of the 35 patients (54.2%) identified as C2M-PsA by GRAPPA were not identified as D2M-PsA under EULAR criteria because they did not meet the required entry criteria of failure of at least two advanced therapies with different MOA. The absolute agreement between the two definitions was 75% (Table 3).

Table 3.

Agreement between GRAPPA C2M-PsA and EULAR D2M-PsA definitions, and between TR-PsA GRAPPA and EULAR definitions.

The agreement of EULAR D2M-PsA and GRAPPA C2M-PsA definitions
		EULAR		Kappa
		Non-D2M-PsA	D2M-PsA
GRAPPA	Non-C2M-PsA	41	0	0.476
	C2M-PsA	19	16
The agreement of EULAR and GRAPPA TR-PsA definitions
		EULAR		Kappa
		Non-TR-PsA	TR-PsA
GRAPPA	Non-TR-PsA	62	0	1.000
	TR-PsA	0	14

C2M-PsA, complex-to-manage psoriatic arthritis; D2M-PsA, difficult-to-manage psoriatic arthritis; EULAR, European alliance of associations for rheumatism; GRAPPA, group for research and assessment of psoriasis and psoriatic arthritis; TR-PsA, treatment-refractory psoriatic arthritis.

In contrast, nearly perfect agreement was observed for the classification of TR-PsA between EULAR and GRAPPA frameworks (κ = 1.00). The same 14 patients were identified as TR-PsA by both EULAR and GRAPPA, yielding 100% absolute agreement (Table 3). The degree of overlap between EULAR and GRAPPA classification frameworks is depicted in Figure 1.

Figure 1.

Nested Overlap of EULAR and GRAPPA Definitions of D2M/C2M PsA and TR-PsA Within the ORCHESTRA Cohort. Numbers indicate the number of patients at each step of the nested overlap (none, GRAPPA-only, D2M-only, and TR-PsA overlap).

Comparative characteristics by D2M-PsA, C2M-PsA, and TR-PsA status

Baseline demographic, clinical, and ultrasound characteristics according to EULAR D2M-PsA and GRAPPA C2M-PsA classifications are summarized in Table 2, with unadjusted comparisons available in Table 4.

Table 4.

Comparison of demographic, clinical, and ultrasound characteristics between D2M/C2M PsA patients according to EULAR and GRAPPA definitions.

Characteristic	EULAR D2M-PsA		p	GRAPPA C2M-PsA		p	EULAR/GRAPPA TR-PsA		p
Characteristic	Yes, N: 16	No, N: 60	p	Yes, N: 35	No, N: 41	p	Yes, N: 14	No, N: 62	p
Sex, male^#	10 (62.5%)	24 (40%)	0.108	17 (48.6%)	17 (41.5%)	0.534	9 (64.3%)	25 (40.3%)	0.103
Age^$	54.33 ± 12.40	46.67 ± 12.42	0.037	50.74 ± 12.44	46.08 ± 12.71	0.119	52.46 ± 12.08	47.33 ± 12.77	0.189
Body mass index*	29.29 (24.27–40.45)	32.06 (26.56–38.98)	0.703	31.40 (24.39–37.92)	30.50 (26.54–40.09)	0.841	29.29 (24.51–38.73)	32.06 (26.53–39.02)	0.679
Documented Depression	3 (20%)	17 (28.3%)	0.746	7 (20.6%)	13 (31.7%)	0.278	3 (21.4%)	17 (27.9%)	0747
Fibromyalgia	2 (12.5%)	7 (11.7%)	1.000	5 (14.3%)	4 (9.8%)	0.725	2 (14.3%)	7 (11.3%)	0.668
Symmetric Polyarthritis Pattern^#	16 (100%)	48 (80%)	0.060	32 (91.4%)	32 (78%)	0.111	14 (100%)	50 (80.6%)	0.108
Oligoarthritis/Monoarthritis Pattern^#	0 (0%)	7 (11.7%)	0.334	2 (5.7%)	5 (12.2%)	0.442	0 (0%)	7 (11.3%)	0.337
DIP Predominant Pattern^#	0 (0%)	2 (3.3%)	1.000	0 (0%)	2 (4.9%)	0.496	0 (0%)	2 (3.2%)	1.000
Axial Disease Pattern^#	3 (18.8%)	13 (21.7%)	1.000	7 (20%)	9 (22%)	0.835	2 (14.3%)	14 (22.6%)	0.721
Arthritis Mutilans Pattern^#	0(0%)	0(0%)	NA	0(0%)	0(0%)	NA	0(0%)	0(0%)	NA
C-reactive protein, mg/l*	1.8 (0.7–5.2)	4.3 (1.8–13.2)	0.025	3 (0.9–7.8)	5.8 (1.8–12.9)	0.076	2.3 (0.9–6)	4 (1.6–12.6)	0.112
Erythrocyte sedimentation rate, mm/h*	14 (5–22)	15 (7–34.5)	0.413	13.5 (6.3–24.8)	16 (7–33)	0.555	13 (4.5–21)	15 (7–33.5)	0.291
Disease duration, years*	16.5 (4.3–24)	3.5 (1–9.5)	0.001	8 (3–18)	2 (0.8–5)	<0.001	16.5 (4.8–24)	3.5 (1–10.3)	0.001
Symptoms duration, years*	20.5 (8–27.8)	6.5 (2–14.8)	0.002	14 (8–24)	5 (1.8–10)	0.001	20.5 (7.8–25.3)	7 (2–15)	0.014
Enthesitis (Ever) ^#	10 (62.5%)	29 (48.3%)	0.314	18 (51.4%)	21 (51.2%)	0.985	9 (64.3%)	30 (48.4%)	0.282
Dactylitis (Ever) ^#	6 (37.5%)	36 (60%)	0.108	12 (34.3%)	30 (73.2%)	0.001	5 (35.7%)	37 (59.7%)	0.103
Nail psoriasis (Ever) ^#	10 (62.5%)	29 (48.3%)	0.314	22 (62.9%)	17 (41.5%)	0.063	8 (57.1%)	31 (50%)	0.629
Skin psoriasis (Ever) ^#	16 (100%)	47 (78.3%)	0.058	32 (91.4%)	31 (75.6%)	0.068	14 (100%)	49 (79%)	0.110
Inflammatory bowel disease^#	4 (26.7%)	4 (6.7%)	0.046	5 (14.7%)	3 (7.3%)	0.456	3 (21.4%)	5 (8.2%)	0.164
Uveitis^#	2 (12.5%)	4 (6.7%)	0.600	2 (5.7%)	4 (9.8%)	0.681	2 (14.3%)	4 (6.5%)	0.304
Gout^#	3 (18.8%)	1 (1.7%)	0.028	4 (11.4%)	0 (0%)	0.041	3 (21.4%)	1 (1.6%)	0.018
Previous number of csDMARDs*	2 (1–3)	2 (1–3)	0.853	2 (1–3)	2 (1–2.5)	0.106	2 (1–3)	2 (1–3)	0.545
Previous number of advanced therapies*	3 (2–4)	0 (0–1)	<0.001	2 (1–3)	0 (0–0)	<0.001	3 (2–4)	0 (0–1)	<0.001
Morning stiffness^#	16 (100%)	57 (95%)	1.000	34 (97.1%)	39 (95.1%)	1.000	14 (100%)	59 (95.2%)	1.000
Duration of morning stiffness*	1 (0.5–6)	1 (0.6–2)	0.847	1 (0.8–3)	1 (0.5–2)	0.499	1 (0.5–4.5)	1.3 (0.9–2)	0.684
Swollen joint count (66)*	6 (2–9)	5 (2–7.8)	0.364	5 (2–8)	4 (1.5–7.5)	0.272	6 (2–9)	5 (1.8–8)	0.353
Tender joint count (68)*	8.5 (3–19)	9.5 (4–18)	0.878	12 (6–23)	8 (3–15.5)	0.064	8.5 (3–19)	9.5 (4–18.3)	0.799
Pain VAS*	6 (5–7)	6 (4–7)	0.558	6 (5–7)	5 (4–7)	0.157	6 (5–6.5)	6 (4–7)	0.724
Patient VAS*	6 (5.3–7.8)	6 (5–7.8)	0.718	6 (6–8)	6 (4–7)	0.046	6 (5–7.3)	6 (5–8)	0.903
Physician VAS*	6 (5–7)	5 (4–7)	0.448	6 (5–7)	5 (3.5–6)	0.035	6 (5–7)	5 (4–7)	0.376
HAQ^$	0.96±0.55	1.02±0.77	0.747	1.22±0.71	0.83±0.71	0.025	0.87±0.53	1.04±0.77	0.453
SPARCC*	2(0.3–4)	2 (0–4)	0.888	2 (0.8–6)	2 (0–3.8)	0.466	2(0–4.5)	2 (0–4)	0.938
BASDAI*	6.05 (4.62–7.36)	6.05 (3.95–7.2)	0.454	6.4 (5.3–7.4)	5.6 (3.5–7.1)	0.054	5.65 (4.43–7.06)	6.2 (4.08–7.28)	0.928
BSA Psoriasis*	0 (0–1)	1 (0–2)	0.505	1 (0–2)	0 (0–2)	0.702	0.5 (0–1.3)	0.5 (0–2)	0.879
B-mode score*	18 (11.8–33.5)	19 (8–27)	0.515	18 (11.3–32.8)	19 (8–26)	0.442	18 (11.3–32)	19 (8–27)	0.667
Doppler score*	3 (0–8)	3 (0–6.3)	0.989	2.5 (0–8.5)	3 (0–5)	0.638	4 (0–10.5)	3 (0–5.8)	0.609
GLOESS score*	19.5 (11.3–32)	19 (10–27)	0.703	21 (11–33)	18.5 (9.3–26.8)	0.422	19.5 (10.8–30)	19 (10.5–27.5)	0.849
Enthesis-inflammation score*	8 (5–12.5)	5 (1–11)	0.132	7.5 (1.8–11)	5 (2.5–11)	0.398	8 (5–14)	5 (1–11)	0.131
Enthesis Doppler score*	0 (0–0)	0 (0–1.8)	0.286	0 (0–0)	0 (0–2)	0.122	0 (0–0.5)	0 (0–1.3)	0.458
Enthesis chronicity score*	11 (3.3–17.8)	10 (2.8–13.5)	0.465	10 (3–17.3)	11 (3–13)	0.684	11 (3.8–17.3)	10 (3–14.5)	0.599
Enthesis total score*	18.5 (6.8–36.3)	15 (5.5–24)	0.304	15 (4.8–25.8)	15 (6–24)	0.761	18.5 (8.3–31.3)	15 (5–24)	0.330

Median (IQR).

Mean ± SD.

n (%).

Statistically significant values (p < 0.05) are shown in bold.

BASDAI, bath ankylosing spondylitis disease activity index; B-mode, gray-scale ultrasound score; BSA, body surface area; C2M-PsA, complex-to-manage psoriatic arthritis; CRP, C-reactive protein; csDMARDs, conventional synthetic disease-modifying antirheumatic drugs; D2M-PsA, difficult-to-manage psoriatic arthritis; DIP, distal interphalangeal; ESR, erythrocyte sedimentation rate; EULAR, European alliance of associations for rheumatology; GLOESS, global OMERACT–EULAR synovitis score; GRAPPA, group for research and assessment of psoriasis and psoriatic arthritis; HAQ, health assessment questionnaire; MDA, minimal disease activity; NiPsA, non-inflammatory psoriatic arthritis; PiPsA, persistent inflammatory psoriatic arthritis; PsA, psoriatic arthritis; SPARCC, spondyloarthritis research consortium of Canada score; TR-PsA, treatment-refractory psoriatic arthritis; VAS, visual analog scale.

Across both frameworks, D2M-PsA and C2M-PsA patients had significantly longer disease and symptom durations and greater prior exposure to advanced therapies than those not meeting the definitions (all p ⩽ 0.002; Table 4). No significant differences were observed in age, sex, body mass index, fibromyalgia, depression, inflammatory markers, clinical joint counts, patient- and physician-reported disease activity, functional status, or ultrasound measures of synovitis and enthesitis across either definition.

Cutaneous and nail psoriasis were numerically more frequent among D2M or C2M patients, whereas dactylitis was more commonly observed in non-D2M patients, particularly in the GRAPPA comparison.

Because EULAR and GRAPPA identified an identical group of patients with TR-PsA, baseline characteristics for TR-PsA were examined relative to non-TR-PsA patients irrespective of framework (Table 4). As expected, TR-PsA patients also demonstrated longer disease and symptom duration and greater prior exposure to advanced therapies. In contrast, measures of inflammation, including clinical assessment, US, and laboratory testing, were largely comparable between TR-PsA and non-TR-PsA patients.

Then we further analyzed patients captured by GRAPPA C2M-PsA but not by EULAR (the discordant group). This discordant group exhibited worse patient- and physician-reported global assessments and greater functional impairment compared with patients not meeting either definition (non-D2T disease). On the other hand, no differences were observed in clinical and ultrasound measures of activity between the discordant and EULAR D2M-PsA groups. Together, these patterns suggest that the discordant GRAPPA-C2M group is more similar to the D2M phenotype rather than to non-D2T disease. Details are presented in the Table S1.

Discussion

Despite the increasing number of therapy options becoming available over the last two decades, D2T-PsA remains a major challenge in clinical practice with a substantial burden on healthcare systems.^21,22 Few prior studies have examined D2T-PsA^6,7,9–11; however, most have relied on adaptations of the D2T-RA framework.^8,23 Such adaptations do not fully address the multifaceted nature of PsA, including its multidomain involvement and high burden of comorbidities, nor do they adequately distinguish inflammatory from non-inflammatory disease drivers. These limitations may hinder the implementation of targeted, holistic management approaches and contribute to suboptimal achievement of treatment targets. To address this unmet need, EULAR and GRAPPA have recently developed PsA-specific D2T frameworks to stratify complex cases by disease drivers, sharing a common conceptual goal but differing in their operationalization.^12,13 To the best of our knowledge, this study is the first to apply both the EULAR and GRAPPA consensus definitions of D2T-PsA in a real-world cohort, providing empirical insight into their clinical applicability, overlap, and performance.

Our findings revealed key differences in patient capture, particularly within the broader, non–inflammation-restricted classifications. The GRAPPA C2M-PsA framework, by requiring failure of only one advanced therapy, identified a larger proportion of patients. In contrast, the EULAR D2M-PsA criteria applied stricter entry thresholds, resulting in a smaller cohort and failing to capture more than half of the patients identified by GRAPPA, yielding only moderate agreement between the two definitions. Importantly, all patients classified by EULAR were also captured by GRAPPA, indicating complete inclusion of the EULAR-defined group within the GRAPPA framework.

When comparing patients identified by the two frameworks (EULAR D2M-PsA and GRAPPA C2M-PsA), notable differences emerged. Within the more stringent EULAR D2M-PsA cohort, the majority of patients were also classified as TR-PsA (87.5%), whereas only 42% of GRAPPA C2M-PsA patients met TR-PsA criteria. This finding indicates that the EULAR definition preferentially enriches for patients with true inflammatory refractoriness, while the broader GRAPPA framework captures a wider spectrum of disease complexity beyond biologic non-response. Moreover, the trend toward the EULAR D2M-PsA group showing a greater concentration of multidomain PsA manifestations, including enthesitis, dactylitis, uveitis, psoriasis, inflammatory bowel disease, and gout, than the GRAPPA C2M-PsA group, may suggest that patients meeting the more stringent EULAR definition could represent a more biologically complex and systemically involved PsA phenotype. However, given the smaller size of the EULAR-defined group, this apparent enrichment may partly reflect a denominator effect and should be interpreted cautiously as hypothesis-generating.

Taken together, these findings raise the question of whether the GRAPPA framework may be overly inclusive. To explore this, we examined the discordant group captured by GRAPPA but not by EULAR and assessed whether this subset more closely resembled patients with D2M disease or those without D2T features. We found that the discordant group exhibited characteristics similar to those of the D2M-PsA patients with respect to clinical disease activity and ultrasound inflammation. In contrast, compared with non-D2T patients, the discordant group had worse patient- and physician-reported outcomes and functional impairment, despite similar objective inflammatory measures. This pattern supports the interpretation that GRAPPA may identify complex patients earlier along a disease trajectory, rather than inappropriately inflating the D2T population. Indeed, this broader inclusiveness is consistent with the GRAPPA task force’s stated intent, which acknowledges C2M-PsA as a dynamic, heterogeneous, and multifactorial state, and by requiring failure of only one approved therapy, the GRAPPA definition intentionally captures a spectrum ranging from patients with longstanding complexity to those who may demonstrate unexpectedly favorable responses to intervention early on.

Therefore, in clinical practice, EULAR and GRAPPA definitions appear to serve complementary rather than interchangeable roles. The GRAPPA C2M-PsA framework is more inclusive and potentially clinically meaningful for real-world practice, on the basis of discordant patients being more similar to D2M patients for the patient-reported outcomes and functional status. This framework may be capturing patients with early or emerging disease complexity driven by multidomain involvement, comorbidities, and functional impairment, even after limited therapeutic exposure. In routine care, this broader framework may help flag patients who require multidisciplinary management, reassessment of treatment goals, and optimization of non-inflammatory contributors before further immunologic escalation. In contrast, the EULAR D2M-PsA framework preferentially enriches for patients with multidomain disease in whom inflammatory refractoriness remains clinically relevant. This greater stringency may be advantageous by concentrating patients with D2T characteristics, thereby yielding a more homogeneous cohort and facilitating patient selection in research, consistent with the EULAR main objective in developing these criteria. On the other hand, in clinical practice, it may simply indicate that patients warrant more urgent reassessment of treatment response and consideration of alternative therapeutic strategies when similar patient profiles are encountered. Notably, this increased precision comes at the cost of reduced inclusiveness, underscoring the importance of selecting definitions based on the intended clinical or research context.

With respect to definitions intended to identify true inflammatory biologic non-response, both EULAR and GRAPPA demonstrated perfect agreement in classifying patients as TR-PsA. Despite this concordance, the two frameworks are not necessarily interchangeable, as they differ in how objective inflammation is conceptualized and operationalized. EULAR permits broader clinical indicators of inflammation, including tender or swollen joints, tender entheses, and inflammatory back pain, whereas GRAPPA emphasizes more specific clinical findings such as swollen joints, dactylitis, or defined psoriasis body surface area thresholds, and relies more heavily on imaging for domains such as axial disease and enthesitis, where clinical examination alone may lack specificity. In the present cohort, the absence of patients with isolated axial or entheseal-predominant disease, together with the modest sample size, may have limited our ability to detect meaningful differences arising from these operational distinctions. Larger and more heterogeneous cohorts enriched for axial and entheseal phenotypes are needed to determine whether these definitional differences translate into clinically meaningful variation in patient classification.

Notably, compared with non-TR-PsA patients, those classified as TR-PsA did not exhibit higher clinical or ultrasound inflammatory activity. This underscores the need to interpret inflammatory activity within the broader context of disease trajectory, cumulative therapeutic exposure, and patterns of treatment failure—including primary versus secondary non-response—and to more precisely delineate what constitutes true inflammatory activity in PsA across its domains. In this context, recent work by Zabotti et al. introduced an imaging-based framework distinguishing Persistent Inflammatory PsA (PiPsA) and Non-Inflammatory PsA (NiPsA) as mutually exclusive strata based on objective evidence of inflammation on imaging.²⁴ While conceptually aligned with GRAPPA and EULAR, this approach may serve as a subsequent refinement, enabling clearer stratification into exclusively inflammatory versus non-inflammatory disease and facilitating more targeted, mechanism-driven treatment strategies. Accordingly, future work will focus on applying the PiPsA/NiPsA framework within larger, real-world cohorts to evaluate its added value beyond existing D2T classifications.

Limitations

While this study provides valuable real-world insights, several limitations should be acknowledged. First, the modest sample size limits the statistical power and the strength of inferences drawn from subgroup comparisons. In addition, prevalence estimates reflect patients referred for initiation or switching of advanced therapy in a tertiary care setting and may not capture individuals who were unable to start or escalate therapy due to comorbidities, contraindications, or access-related barriers. Finally, although this analysis was conducted within a prospective cohort, application of the D2T definitions was retrospective. As such, pragmatic proxies—such as BASDAI to infer axial involvement—were used in the absence of real-time clinical adjudication, which may be less accurate or may not reliably reflect active inflammatory disease when confounding factors are present. In patients with concurrent activity across multiple disease domains or with psychosocial comorbidities, this approach may further limit the ability to definitively identify the dominant driver of treatment failure or therapy

Conclusion

This study provides the first real-world comparison of the EULAR and GRAPPA frameworks for D2T PsA, demonstrating shared conceptual goals but distinct clinical performance. Differences were most apparent within the non–inflammation-restricted classifications, where GRAPPA identified a broader and clinically meaningful population, reflecting greater inclusiveness and earlier capture of disease complexity, whereas EULAR identified a more concentrated subset enriched for multidomain involvement and inflammatory refractoriness after multiple therapeutic exposures. Although both frameworks identified the same patients when applying inflammation-restricted definitions for TR-PsA in this cohort, differences in their operationalization preclude assuming interchangeability. Larger prospective studies are needed to formally assess performance and to refine stratified, mechanism-driven management strategies in PsA.

Supplemental Material

sj-docx-1-tab-10.1177_1759720X261443557 – Supplemental material for Real-world performance of EULAR and GRAPPA definitions of difficult-to-treat psoriatic arthritis within the ORCHESTRA cohort: an observational cohort study

Supplemental material, sj-docx-1-tab-10.1177_1759720X261443557 for Real-world performance of EULAR and GRAPPA definitions of difficult-to-treat psoriatic arthritis within the ORCHESTRA cohort: an observational cohort study by Rahaf Zyad Attar, Seyyid Bilal Acikgoz, Elliot Hepworth, Ozun Bayindir Tsechelidis, Ricardo Sabido-Sauri, Tara Swami and Sibel Zehra Aydin in Therapeutic Advances in Musculoskeletal Disease

Supplemental Material

sj-docx-2-tab-10.1177_1759720X261443557 – Supplemental material for Real-world performance of EULAR and GRAPPA definitions of difficult-to-treat psoriatic arthritis within the ORCHESTRA cohort: an observational cohort study

Supplemental material, sj-docx-2-tab-10.1177_1759720X261443557 for Real-world performance of EULAR and GRAPPA definitions of difficult-to-treat psoriatic arthritis within the ORCHESTRA cohort: an observational cohort study by Rahaf Zyad Attar, Seyyid Bilal Acikgoz, Elliot Hepworth, Ozun Bayindir Tsechelidis, Ricardo Sabido-Sauri, Tara Swami and Sibel Zehra Aydin in Therapeutic Advances in Musculoskeletal Disease

Footnotes

Acknowledgements

The authors would like to express their sincere gratitude to all patients who contributed their time and data to the ORCHESTRA registry, making this research possible. We also thank the dedicated clinical and research teams involved in data collection, management, and coordination of the registry for their ongoing commitment to advancing inflammatory arthritis research.

Declarations

ORCID iDs

Rahaf Zyad Attar

Seyyid Bilal Acikgoz

Supplemental material

Supplemental material for this article is available online.

References

FitzGerald

Ogdie

Chandran

, et al. Psoriatic arthritis. Nat Rev Dis Primer 2021; 7(1): 59.

Gupta

Syrimi

Hughes

, et al. Comorbidities in psoriatic arthritis: a systematic review and meta-analysis. Rheumatol Int 2021; 41(2): 275.

Castañeda

Martín-Martínez

González-Juanatey

, et al. Cardiovascular morbidity and associated risk factors in Spanish patients with chronic inflammatory rheumatic diseases attending rheumatology clinics: baseline data of the CARMA Project. Semin Arthritis Rheum 2015; 44(6): 618–626.

Urruticoechea-Arana

Castañeda

Otón

, et al. Prevalence of metabolic syndrome in psoriatic arthritis: systematic literature review and results from the CARMA cohort. JCR: J Clin Rheumatol 2022; 28: e388–e396.

Torre-Alonso

Carmona

Moreno

, et al. Identification and management of comorbidity in psoriatic arthritis: evidence- and expert-based recommendations from a multidisciplinary panel from Spain. Rheumatol Int 2017; 37(8): 1239–1248.

Kumthekar

Ashrafi

Deodhar

Difficult to treat psoriatic arthritis—how should we manage?

Clin Rheumatol 2023; 42(9): 2251–2265.

Singla

Ribeiro

Torgutalp

, et al. Difficult-to-treat psoriatic arthritis (D2T PsA): a scoping literature review informing a GRAPPA research project. RMD Open 2024; 10(1): e003809.

Nagy

Roodenrijs

Welsing

, et al. EULAR definition of difficult-to-treat rheumatoid arthritis. Ann Rheum Dis 2021; 80(1): 31–35.

Favalli

Biganzoli

Cincinelli

, et al. Difficult-to-treat psoriatic arthritis: refining the definition using a statistical model in a real-life cohort. Front Med 2024; 11: 1509082.

10.

Perrotta

Gentileschi

Scriffignano

, et al. Difficult-to-treat concept in psoriatic arthritis: analysis of 2 potential definitions in a large group of patients. A cross-sectional study. J Rheumatol 2024; 51(10): 985–990.

11.

Philippoteaux

Marty-Ane

Cailliau

, et al. Characteristics of difficult-to-treat psoriatic arthritis: a comparative analysis. Semin Arthritis Rheum 2023; 63: 152275.

12.

Marzo-Ortega

Harrison

Fragoulis

, et al. EULAR points to consider and consensus definitions for difficult-to-manage and treatment-refractory psoriatic arthritis. Ann Rheum Dis. Epub ahead of print 29 October 2025, https://ard.eular.org/article/S0003-4967(25)04433-4/abstract (2025, accessed 22 December 2025)

13.

Proft

Ribeiro

Singla

, et al. Consensus definitions of complex-to-manage and treatment-refractory psoriatic arthritis: a GRAPPA initiative. Nat Rev Rheumatol 2026; 22(2): 132–144.

14.

Vandenbroucke

Elm

von Altman

, et al. Strengthening the reporting of observational studies in epidemiology (STROBE): explanation and elaboration. PLoS Med 2007; 4(10): e297.

15.

Coates

Corp

van der Windt

, et al. GRAPPA treatment recommendations: 2021 update. J Rheumatol 2022; 49(6 Suppl 1): 52–54.

16.

Gossec

Kerschbaumer

Ferreira

RJO

Aletaha

Baraliakos

Bertheussen

, et al. EULAR recommendations for the management of psoriatic arthritis with pharmacological therapies: 2023 update. Ann Rheum Dis 2024; 83(6): 706–719.

17.

Bruyn

Iagnocco

Naredo

, et al. OMERACT definitions for ultrasonographic pathologies and elementary lesions of rheumatic disorders 15 years on. J Rheumatol 2019; 46(10): 1388–1393.

18.

Terslev

Naredo

Aegerter

, et al. Scoring ultrasound synovitis in rheumatoid arthritis: a EULAR-OMERACT ultrasound taskforce-part 2: reliability and application to multiple joints of a standardised consensus-based scoring system. RMD Open 2017; 3(1): e000427.

19.

Bakirci

Solmaz

Stephenson

, et al. Entheseal changes in response to age, body mass index, and physical activity: an ultrasound study in healthy people. J Rheumatol 2020; 47(7): 968–972.

20.

Landis

Koch

GG.

The measurement of observer agreement for categorical data. Biometrics 1977; 33(1): 159–174.

21.

Kharouf

Gladman

DD.

Advances in the management of psoriatic arthritis in adults. BMJ 2024; 387: e081860.

22.

McHugh

Maguire

Handel

, et al. Evaluation of the economic burden of psoriatic arthritis and the relationship between functional status and healthcare costs. J Rheumatol 2020; 47(5): 701–707.

23.

Nagy

Roodenrijs

NMT

Welsing

PMJ

, et al. EULAR points to consider for the management of difficult-to-treat rheumatoid arthritis. Ann Rheum Dis 2022; 81(1): 20–33.

24.

Zabotti

Aydin

David

, et al. Delineating inflammatory from non-inflammatory mechanisms for therapy optimization in psoriatic arthritis. Nat Rev Rheumatol 2025; 21(4): 237–248.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.02 MB

0.03 MB

0.00 MB