Joint Ventures Recaps Influential EULAR Studies in SLE, RA, and More

In this episode of Joint Ventures, hosts Jack Arnold, MBBS, PhD, academic clinical lecturer in rheumatology, University of Leeds, and Rihards Buss, MD, consultant rheumatologist, Freeman Hospital, Newcastle, record a post-meeting debrief of the 2026 European Alliance of Associations for Rheumatology (EULAR) Annual Meeting held in London this past June, working through the abstracts each identified as the meeting's most consequential. The conversation spans cellular therapies in autoimmune disease, proteomic biomarkers for treatment selection in rheumatoid arthritis (RA), a landmark head-to-head trial in psoriatic arthritis (PsA), evolving classification criteria for axial spondyloarthritis, the first phase 3 randomized controlled trial of a JAK inhibitor in polymyalgia rheumatica (PMR), and more. Check out the studies highlighted below.

AB-101 Allogeneic NK Cell Therapy in Rheumatologic Diseases (LB0003)

AB-101 is a non-genetically modified, allogeneic, off-the-shelf cryopreserved NK cell therapy derived from cord blood, administered in combination with rituximab following low-dose cyclophosphamide and fludarabine conditioning — entirely in the outpatient setting. The phase 2a basket study enrolled 31 patients with treatment-refractory RA, Sjögren disease, or systemic sclerosis. At 6 months, RA patients showed a mean 73% reduction in swollen joint counts, with 71% achieving ACR50. Sjögren patients achieved clinically meaningful reductions in ClinESSDAI and ESSPRI, and all four evaluable systemic sclerosis patients achieved an rCRISS25 response. No CRS, ICANS, or hypogammaglobulinemia were observed.¹

Arnold's enthusiasm is as much for the trial design as the results. The basket approach — grouping patients across diagnostic labels by shared B cell-driven pathology — is well established in oncology but remains underused in rheumatology. He argues that 2 patients with Sjögren disease may have less in common therapeutically than 1 Sjögren patient and 1 lupus patient, and that the field needs to move beyond siloing patients into legacy diagnoses when investigating shared mechanisms. Buss agrees the design broadens access to trials for patients with rarer conditions who would otherwise struggle to reach trial-eligible numbers. "It just will allow to enroll more patients in these trials, giving rarity of some of the conditions," he notes.

CARLYSLE: Obe-Cel CAR T in Severe Refractory SLE (OP0341)

Obecabtagene autoleucel (obe-cel) is an autologous CD19-directed CAR T-cell therapy with a fast off-rate binding domain designed to reduce immunotoxicity, previously validated in relapsed/refractory B-cell acute lymphoblastic leukemia. The Phase I CARLYSLE study enrolled patients with severe refractory systemic lupus erythematosus (SLE); 9 adults were infused at either 50×10⁶ (50M) or 100×10⁶ (100M) CAR T-cells. In the 50M cohort, 5/6 patients achieved DORIS remission at a median onset of 5.1 months, 3/6 achieved complete renal response at month one, and no ICANS or grade ≥2 CRS events occurred. Early 100M cohort data were consistent.²

Arnold highlights the favorable safety signal and the evidence of immunological reset — B-cell reconstitution in the 50M cohort was predominantly with transitional and naïve phenotypes. He tempers this with practical questions the data cannot yet answer: what to do with patients during the 4 to 6 weeks of autologous manufacturing, and how to bridge them in that window. "More data in SLE, impressive initial results, but we just have to see what comes out of it."

Ex Vivo Proteomic Signatures of JAK and TNF Inhibitors Predict Treatment Response in RA (OP048)

A group from the University of Cordoba exposed immune cells from 48 DMARD-naive RA patients to baricitinib or etanercept ex vivo, then profiled the resulting protein signatures across a 92-protein inflammation panel. These drug-specific signatures were then tested in baseline serum from 223 patients initiating either drug. Patients whose baseline serum resembled the baricitinib-induced signature responded better to baricitinib, and vice versa — with cross-therapy analysis showing high expression of one drug's signature predicted non-response to the alternative. A machine learning neural network incorporating 7 proteins achieved an AUC of 0.75 for predicting treatment-specific remission. The effect was only present in DMARD-naive patients; prior exposure to DMARDs abolished it.³

Buss finds the signal genuinely exciting and describes it as "a further step closer to true precision medicine in rheumatology." He argues the finding that the model only works in naive patients is itself clinically important: once the immune system has been altered by prior treatment, the original disease state is lost. Arnold agrees, but cautions that translating ex vivo assays into scalable clinical tools will require further simplification and validation. Both see this as the direction the field needs to go — moving away from the current try-and-fail approach toward upfront treatment matching.

BE BOLD: Bimekizumab vs Risankizumab in PsA (LB0001)

BE BOLD is the first head-to-head trial to demonstrate statistically significant superiority on a joint endpoint between 2 biologics in PsA. In 553 adults with active PsA, bimekizumab — which inhibits both IL-17A and IL-17F — was superior to the IL-23 inhibitor risankizumab on the primary endpoint of ACR50 at week 16 (49.1% vs 38.4%; P =0.0078), with response curves diverging as early as week four. The second ranked endpoint, minimal disease activity, was numerically higher with bimekizumab (43.0% vs 39.9%) but did not reach significance, stopping sequential testing. Candida infections were more frequent with bimekizumab, as expected from the mechanism, but all were mild or moderate with no discontinuations.⁴

Both hosts welcome the data, with Arnold noting that companies taking the risk of head-to-head trials produces more clinically useful information than placebo-controlled data. Buss contextualises this alongside the real-world POS0483 poster,⁵ which found risankizumab non-inferior to TNF inhibitors and IL-17 inhibitors on joint counts in routine practice — a different setting and a different question, and both speakers are careful not to draw direct comparisons between the two. The dual IL-17A/F mechanism of bimekizumab versus single-pathway IL-23 inhibition appears to drive its faster onset and superior joint response, though how this translates to clinical decision-making will depend on the full dataset and patient profile.

2025 ASAS-SPARTAN Revised AxSpA Classification Criteria (OP0236, OP0239, POS0190, POS1341)

Buss devotes his final segment to 4 abstracts collectively addressing the revision of axSpA classification criteria. The 2009 ASAS criteria failed to meet the prespecified performance targets of ≥75% sensitivity and ≥90% specificity in the global CLASSIC inception cohort, prompting the ASAS-SPARTAN initiative. The resulting 2025 revised criteria achieved 79.5% sensitivity and 90.4% specificity in the CLASSIC validation dataset, with imaging — specifically MRI with global assessment of both active and structural sacroiliac joint lesions — carrying the greatest independent weight. Dactylitis, NSAID response, and family history have been removed; a patient who is imaging-negative cannot be classified as axSpA on clinical features alone.^6-9

Three external cohorts validated the pattern: high specificity (94–99%) at the cost of substantially reduced sensitivity (35–61%). In the RABBIT-SpA real-world registry, 25% of rheumatologist-diagnosed axSpA patients met neither set of criteria, and the clinical arm of the 2025 criteria was fulfilled by only 3% of eligible patients versus 58% under the 2009 criteria. Buss's primary concern is the downstream impact on clinical trial populations: studies enrolling to the new criteria will capture a more imaging-positive, potentially more severe cohort, reducing generalisability to the broader real-world population. He also flags the increasing emphasis on MRI quality and reader expertise as a practical implementation challenge. Arnold notes the volume of axSpA imaging work at the meeting as a signal the field is actively grappling with these questions.

JAK-SPARE: Baricitinib in Early Polymyalgia Rheumatica (LB0005)

JAK-SPARE is the first phase 3 randomized controlled trial of a JAK inhibitor in PMR. Forty-six patients with early-onset PMR were randomized 1:1 to baricitinib 4 mg or placebo, both combined with a rapid glucocorticoid taper from 20 mg to zero over 11 weeks. Glucocorticoid-free remission at week 16 — the primary endpoint — was achieved in 65% of baricitinib patients versus 17% of placebo patients (P =0.002), with significantly longer time to first relapse and lower cumulative steroid dose at week 16 (962 mg vs 1183 mg; P =0.007). After crossover, the placebo-to-baricitinib group reached 65% glucocorticoid-free remission by week 28, consistent with a catch-up effect.¹⁰

Arnold is enthusiastic about what the data could mean for this population — PMR is extremely common, steroid toxicity in this older population is a real clinical problem, and proper phase 3 data in this space is overdue. He is, although, quick to caution that the sample is small, the steroid taper is more aggressive than most clinicians would use in practice, and the cardiovascular risk profile of JAK inhibitors in an elderly cohort warrants continued scrutiny. One serious adverse event occurred in the baricitinib arm (prostate adenocarcinoma, week 18). He warns against directly transposing the rapid taper protocol to clinical practice, noting that trial designs drive aggressive tapering to generate flares that demonstrate a between-group difference. "You have to be aware that these trials are trying to see an effect," he notes.

Arnold’s disclosures include Alumis, Roche, and Novartis. Buss has no relevant disclosures to report.

References

1. Gaylis N, Brionez T, Valenzuela G, et al. AB-101, an outpatient-administered allogeneic NK cell therapy combined with rituximab, generates robust clinical efficacy responses comparable with autologous CAR T in 31 patients with rheumatologic diseases [abstract LB0003]. Presented at: EULAR Annual Congress 2026; June 3–6, 2026; London, United Kingdom.

2. Leandro M, Pepper R, Parker B, et al. Obecabtagene autoleucel (obe-cel), a CD19-targeting chimeric antigen receptor (CAR) T-cell therapy, in patients with severe, refractory systemic lupus erythematosus (srSLE): initial safety, preliminary efficacy, pharmacokinetics, and biomarker results from the Phase I CARLYSLE study [abstract OP0341]. Presented at: EULAR Annual Congress 2026; June 3–6, 2026; London, United Kingdom.

3. Corrales-Díaz Flores S, Perez-Sanchez C, Cerdó T, et al. Ex vivo immune-derived molecular signatures of JAK and TNF inhibitors predict treatment response in biologic- and targeted synthetic DMARD-naïve rheumatoid arthritis patients [abstract OP048]. Presented at: EULAR Annual Congress 2026; June 3–6, 2026; London, United Kingdom.

4. McInnes IB, Gossec L, Gottlieb AB, et al. Bimekizumab efficacy and safety versus risankizumab in patients with active psoriatic arthritis: 16-week results from a head-to-head, multicentre, randomised, phase 3B study (BE BOLD) [abstract LB0001]. Presented at: EULAR Annual Congress 2026; June 3–6, 2026; London, United Kingdom.

5. Mease PJ, Walsh JA, Ye X, et al. Comparative effectiveness of risankizumab versus tumour necrosis factor inhibitors or interleukin-17 inhibitors on joint outcomes in biologic-naïve patients with psoriatic arthritis: a global real-world study [abstract POS0483]. Presented at: EULAR Annual Congress 2026; June 3–6, 2026; London, United Kingdom.

6. Maksymowych WP, van der Heijde D, Caplan L, et al. The 2025 Assessment of SpondyloArthritis International Society (ASAS) and Spondyloarthritis Research and Treatment Network (SPARTAN) revised classification criteria for axial spondyloarthritis: development and validation in the Classification of AxSpA Inception Cohort Study [abstract OP0236]. Presented at: EULAR Annual Congress 2026; June 3–6, 2026; London, United Kingdom.

7. de Bruin L, Ramiro S, van Lunteren M, et al. High specificity but modest sensitivity of the 2025 ASAS-SPARTAN revised axial spondyloarthritis classification criteria in early axial spondyloarthritis: data from the SPACE cohort [abstract OP0239]. Presented at: EULAR Annual Congress 2026; June 3–6, 2026; London, United Kingdom.

8. Weiß A, Poddubnyy D, Kiltz U, et al. Performance of the 2025 revised ASAS-SPARTAN classification criteria and the 2009 ASAS classification criteria in axSpA: an analysis from the RABBIT-SpA registry [abstract POS0190]. Presented at: EULAR Annual Congress 2026; June 3–6, 2026; London, United Kingdom.

9. Wichuk S, Horbal N, Carmona R, et al. Performance of the 2025 ASAS SPARTAN revised classification criteria for axial spondyloarthritis in a Canadian multicenter inception cohort with psoriasis, uveitis, or colitis, presenting with undiagnosed back pain [abstract POS1341]. Presented at: EULAR Annual Congress 2026; June 3–6, 2026; London, United Kingdom.

10. Lechner-Radner H, Anderle K, Bond M, et al. Baricitinib for remission induction and glucocorticoid sparing in new-onset polymyalgia rheumatica (JAK-SPARE): a phase III randomized controlled trial [abstract LB0005]. Presented at: EULAR Annual Congress 2026; June 3–6, 2026; London, United Kingdom.