EULAR 2025 round-up

The EULAR 2025 Congress took place in Barcelona, Spain, from 11–14 June

Psoriatic arthritis: Hit hard and early

Data presented at the European Alliance of Associations for Rheumatology (EULAR) 2025 Congress in Barcelona looked at the potential benefit of early biologic disease-modifying anti-rheumatic drugs (bDMARD) use in more severe psoriatic arthritis (PsA). EULAR currently recommends a treat-to-target approach, and suggests more intensive therapy for people with poor prognostic factors, depending on the disease presentation.

Two recent studies suggest there is no significant benefit of early biologics over standard step-up care with methotrexate, but these did not select for poor prognosis. The aim therefore of the SPEED (Severe Psoriatic arthritis –Early intervEntion to control Disease) trial – which was funded by the National Institute for Health Research – was to compare disease activity in 192 PsA patients with poor prognostic factors when treated with one of three regimens: Standard step-up with conventional synthetic DMARD (csDMARD), combination csDMARD, or early induction with a tumour necrosis factor inhibitor (TNFi). The primary endpoint was the mean PsA disease activity score (PASDAS) at 24 weeks.

At week 24, a difference was found in PASDAS mean scores between treatment groups, with both the combination csDMARD and early TNFi groups showing evidence of a difference when compared to standard step-up care. Of note, there was no evidence of a difference between the early TNFi and combination csDMARD groups.

However, by week 48 the benefit compared to standard step-up care was seen only for early TNFi therapy.

Presenting the work, Laura Coates said: “These data show that initial intensive therapy with early biologics or combination csDMARDs are superior for rapid control of early moderate-to-severe PsA. Even with only six months of early biologic therapy, better outcomes are maintained at  one year in those initially receiving a TNF inhibitor.”

A case series also presented at the Congress aimed to describe real-world safety and effectiveness of combined biologic and targeted synthetic DMARD therapy in PsA. Andre Lucas Ribeiro and colleagues analysed prospectively collected data from the University of Toronto PsA cohort, which included 22 people treated with combinations of a bDMARD and either JAKi or TYK2i, with some patients trying multiple combinations. The primary indications for combination therapy were active peripheral arthritis and skin disease, including palmoplantar psoriasis. Results showed numerical improvement across multiple disease-activity measures. In the bDMARD plus JAKi group, the most frequent combination was an IL-17i plus JAKi, and over 10.5 patient–years of exposure only one case of mild infectious stomatitis was reported, which did not result in treatment discontinuation. Additionally, IL-23i plus JAKi were used for 3.7 patient–years without any reported safety events.

For the bDMARD plus TYK2i group, IL-17i plus TYK2i were used for 8.5 patient-years. Combinations of bDMARD plus apremilast were also reported. Overall, the safety profile of bDMARD combinations with JAKi, TYK2i, and apremilast appear favourable. All reported infections were mild, managed without hospitalisation, and rarely led to treatment discontinuation. Furthermore, patients achieved short-term responses, with improvements in both musculoskeletal and skin domains. However, as this is an observational study with short-term follow-up, there is a need for randomised clinical trials to further explore and validate these findings.

Risk factors identified for progression in spondyloarthritis

New research presented at the European Alliance of Associations for Rheumatology 2025 Congress aims to increase the current limited knowledge regarding the predictors of long-term spinal radiographic progression in radiographic axial spondyloarthritis (r-axSpA).

These patients are at risk of spinal ankylosis, and high disease activity has previously been linked to accelerated radiographic spinal progression in patients with early axSpA, but more information is needed.

The data presented described a potential set of predictors for spinal progression, based on data collected over 13 years in 176 patients with r-axSpA. Patients fulfilling the modified New York criteria for r-axSpA (previously termed ankylosing spondylitis) were included in this longitudinal study, and everyone was examined at baseline, and then again at five and 13 years. Spinal radiographs were graded according to the modified Stoke AS Spinal Score (mSASSS), and predictors for continuous change were assessed over various time periods. Of the 176 patients included at baseline, 166 and 136 were assessed at the five- and 13-year follow-up, respectively, and 126 participated on all three occasions. Over time, there was a significant increase in spinal ankylosis, with a mean increase of 1.6 points from baseline to five years, and 2.7 from five- to 13-year follow-up – reflecting a yearly average increase in mSASSS of 0.29 and 0.34, respectively. Sex-stratified analyses showed significant increases in both males and females. Several factors were flagged as being significant associations when the researchers looked for predictors of progression. For example, higher C-reactive protein and being overweight or obese predicted progression in both sexes. Higher mSASSS at the start of follow-up and exposure to tumour necrosis factor inhibitors (TNFi) or bisphosphonates were associated with progression in females, whereas smoking and carrying the HLA-B27 gene were significant predictors in males. Anna Deminger, lead author for the study, said: “On a group level, the progression in spinal pathological new bone formation was slow, but continued over 13 years in patients with long-standing r-axSpA.”

This study highlights the importance of inflammation as a negative prognostic marker for spinal radiographic progression. Other modifiable adverse prognostic markers were smoking and having a body mass index over 25. Of note, the higher risk of spinal radiographic progression in women exposed to TNFi or bisphosphonates needs to be further investigated.

Pregnancy outcomes in autoinflammatory disease

To address the lack of prospective large cohort data on pregnancy outcomes in women with autoinflammatory disease – particularly in familial Mediterranean fever (FMF), a French multi-centre prospective pregnancy observational cohort was set up to analyse disease activity, treatment, pregnancy outcomes, delivery, and neonatal health.

The work – presented at the European Alliance of Associations for Rheumatology 2025 Congress – showcased findings from 97 women with an autoinflammatory disease, which highlighted the need for close monitoring of these patients and collaboration between their clinicians.

The most common diagnoses were FMF (81 per cent), followed by undifferentiated systemic autoinflammatory diseases (USAID), tumour necrosis factor receptor-associated periodic syndrome, cryopyrin-associated periodic syndromes (CAPS), Still’s disease, recurrent pericarditis, mevalonate kinase deficiency, A20 haploinsufficiency, and other rare diseases.

Over a period from 2016 to 2024, the 97 women studied carried 115 pregnancies, including five sets of twins. Sixteen pregnancies were terminated before the 37th week of gestation, including two foetal deaths in utero, two therapeutic abortions due to chromosomal abnormalities, and one spontaneous abortion. In the year prior to pregnancy, 57.7 per cent of women had signs of disease activity, and 59.1 per cent experienced flares during their pregnancy.

In women with FMF, the median age at onset was six years, and they were 31 when they got pregnant – with 15 per cent needing assisted reproduction. They reported a mean of four disease flares per year in the year prior to pregnancy. Inflammatory symptoms during pregnancy happened in 65.7 per cent of cases, including three prolonged febrile myalgia syndromes.

Pregnancy complications included one risk of preterm delivery in a twin pregnancy; one patient experienced one anamnios and one oligohydramnios during both of her pregnancies.

Overall, 17 per cent of women delivered their baby before 37 weeks – higher than the 7 per cent expected in the general population for France – and 22.4 per cent of singleton babies had a birth weight below the 10th percentile, again compared to only 7.1 per cent in the general population. In women with USAID, the median age at disease onset was 14, and they were 30 when they got pregnant. Half were receiving colchicine, and biologic therapy was discontinued in two women upon discovery of pregnancy. Only one had a birth weight less than the 10th percentile.

When looking at markers of inflammation, the group found that women with USAID had mean C-reactive protein at enrolment of 8.9mg/dL, compared to 22.5mg/dL for women with FMF.

AI and rheumatology: Multiple applications

The European Alliance of Associations for Rheumatology 2025 Congress showcased a number of studies showing how artificial intelligence (AI) is influencing different areas in rheumatology – from diagnosis through to monitoring, risk prediction, and patient communication. 

High-resolution CT is the standard to diagnose and assess progression in interstitial lung disease (ILD), a key feature in systemic sclerosis (SSc). But AI-assisted interpretation has the potential to improve the quantification and characterisation of SSc-ILD, making it a powerful tool for monitoring. Francesca Motta offered new data from an observational study pitting AI-assisted analysis against two radiologists with expertise in thoracic imaging. Results showed that the AI outperformed visual scoring in assessing the progression of fibrosis in patients with SSc-ILD, and showed more significant correlation with values from pulmonary function tests – enabling detection of subtle changes over time.

Seulkee Lee also presented findings around AI in diagnosis, investigating a deep learning model that integrates inflammatory and structural changes in sacrum MRI to address the gap between detection of bone marrow oedema and clinical diagnosis of axial spondyloarthritis. An end-to-end deep learning framework was developed, using short tau inversion recovery and T1-weighted MRI sequences to reflect inflammatory and structural changes, respectively. Using data from 291 patients, the classification model demonstrated high sensitivity, specificity, and accuracy. Of note, it was able to identify six out of nine patients who met clinical, but not ASAS-defined, positive MRI criteria – indicating its ability to detect features beyond conventional criteria. These findings highlight the potential of AI not only to detect specific imaging features, but also to predict clinical diagnoses.

Also in the field of imaging, Claus Juergen Bauer and colleagues explored the role of a supervised deep learning model in ultrasound – specifically to assist in image classification for the presence or absence of lesions typical of giant cell arteritis. The developmental dataset included 3,800 images from 244 patients. The model outperformed two comparators, with superior diagnostic performance for both the axillary and superficial common temporal arteries – with the exception of smaller branches of the superficial temporal artery that demonstrated lower performance, reflecting inherent diagnostic challenges. Future work will focus on expanding datasets and incorporating multi-centre validation to optimise detection in smaller arteries and enhance model generalisability.

Turning to risk factor identification, Antonio Tonutti showcased results from two machine learning models that were developed and tuned to predict interceptable cancers (those diagnosed synchronously or after the first non-Raynaud symptom) in people with SSc using clinical, serological, and treatment data. Breast cancer was the most common malignancy (32 per cent), followed by lung (16 per cent), gynaecological (8 per cent), colorectal (7.5 per cent), and haematological cancers (7 per cent). The models appeared to have good performance and accuracy of 73-79 per cent, although there were differences in sensitivity, precision, and specificity between the two, with no one model winning out over the other in all parameters. Key predictors identified included baseline ILD, digital ulcers, oesophageal involvement, telangiectasia, and high C-reactive protein, while taking mycophenolate mofetil was protective in both models. Finetuning and validation of these AI models could offer hope for personalised screening strategies to improve early cancer detection in SSc patients.

Another group shared work on using large language models for risk assessments. Pallavi Vij and colleagues evaluated the effectiveness of these models and prompt engineering techniques in delivering osteoporosis care guidance through case-based scenarios to assess capability in risk stratification, treatment recommendations, and referral decisions in accordance with national guidelines. Findings suggest promising utility for risk stratification and referral triage – potentially reducing administrative burden. However, there was lower concordance in treatment recommendations, which highlights the necessity of clinical expertise for therapeutic decision-making. Further validation studies are needed.

Marco Capodiferro presented a piece on digital biomarkers – based on a multicentric European cohort (Lausanne, Bari, Bern) – looking at how advances in deep learning and computer vision might provide opportunities to simplify hand-motion tracking via smartphones – offering significant potential for remote assessment of disease activity in rheumatoid arthritis. Participants in the study performed five rapid finger flexions with their dominant hand while being recorded on a smartphone camera. The algorithm quantified joint angle changes and time to maximal flexion to analyse kinetic variables. Findings suggest significant associations between these kinetic features and clinical measures, and the model was able to robustly predict low disease activity and remission.