This Week in Rheumatology — Oct 1, 2026

This AudioScholar briefing is generated by artificial intelligence for healthcare professionals and trainees. It is not medical advice.

Welcome to This Week in Rheumatology. This week we're covering 10 notable papers spanning lupus treatment and prognosis, risk prediction across vasculitis, giant cell arteritis and gout, and treatment strategy questions in rheumatoid arthritis, juvenile arthritis and hand osteoarthritis. Let's dive in.

We start with lupus, where two papers address the two questions that dominate clinic: what to induce with, and how to measure what you're doing. In RMD Open, Ueno and colleagues report from the LOOPS registry a retrospective comparison of single-agent induction in active lupus nephritis, pitting rituximab against mycophenolate mofetil against intravenous cyclophosphamide, each on a background of glucocorticoids [1]. At 52 weeks there were no significant differences between the three strategies in complete renal response or in improvement in proteinuria, and no differences in estimated glomerular filtration rate, disease activity, or concomitant steroid dose. The immunophenotyping is the interesting part: rituximab predictably narrowed B-cell subsets while mycophenolate and cyclophosphamide produced broader reductions that included activated T cells, but the formal test for differences between strategies over time was not significant, so the authors are careful to call these within-strategy observations rather than proof of divergent mechanisms. This is registry data, not randomised, with the confounding by indication that implies, so it supports equipoise between agents rather than settling it. Alongside that, in Rheumatology, Andraos and colleagues followed 335 patients across nearly 3,600 outpatient visits at a Swedish centre over seven years, asking whether the newer SLE Disease Activity Score predicts irreversible organ damage better than the older SLEDAI-2K [6]. Time-averaged SLE-DAS independently predicted damage accrual at three, five and seven years, whereas SLEDAI-2K only predicted it at five years, and the model-fit statistics favoured SLE-DAS at every time point. That is observational single-centre evidence, and SLE-DAS was scored retrospectively here, but it strengthens the case that a more granular activity index tracks the thing patients actually care about, which is damage.

The second theme is risk prediction, and this week it runs from the bedside score to the methylome. In RMD Open, Thielmann and colleagues took 70 patients from the prospective PREDICT-GCA study, all of whom had a temporal artery biopsy in addition to ultrasound and clinical workup, and applied four published clinical probability rules for giant cell arteritis [5]. The Ing model performed best, with the highest area under the curve at about 0.84 and a negative predictive value of around 87 percent; the Southend score, the Bhavsar-Khalidi score and the PREDICT score all performed less well. The uncomfortable finding is that in close to four patients in ten, the different scores disagreed clinically, classifying the same person as low risk by one rule and non-low risk by another. Those discordant patients less often had cranial symptoms, but their rate of visual impairment was no lower, which is a caution against treating any single pretest probability score as a safe rule-out in an atypical presentation. The cohort is small and the study is ongoing, so this is a signal about score heterogeneity rather than an endorsement of one instrument. Staying in large-vessel and small-vessel disease, Rheumatology also publishes an external validation by Delestre and colleagues of relapse and infection prediction models in ANCA-associated vasculitis, using 217 patients from the MAINRITSAN trials who received 18 months of rituximab maintenance [3]. Relapse occurred in 76 patients, with relapse-free survival at six years of about 55 percent, while serious infections occurred in 33 patients, a cumulative incidence of around one in five. The relapse model reproduced reasonably well and was most informative in the PR3-ANCA subgroup; the infection model performed more strongly, roughly tripling the identified risk of serious infection in those flagged as high risk. The authors frame this as support for individualised conversations about extending rituximab beyond 18 months, weighing relapse risk against infection risk, rather than as a protocol. And in Annals of the Rheumatic Diseases, Chen and colleagues tackle the oldest question in gout: why only a minority of hyperuricaemic people ever get an attack [10]. In a two-stage case-control epigenetic study of 124 patients with gout, 121 with asymptomatic hyperuricaemia and 124 normouricaemic controls, peripheral blood DNA methylation signatures mapped onto 84 genes overlapping known gout and urate risk loci, including HLA-DRB1 and TAP2. A methylation-only classifier separated gout from normouricaemic controls very well, and for the hardest distinction, gout versus asymptomatic hyperuricaemia, adding methylation to clinical variables lifted discrimination from mediocre to good. The authors themselves say prospective validation is needed before any clinical deployment, so treat this as mechanistic biology with a plausible stratification future, not a test to order.

The third theme is treatment strategy and drug safety across inflammatory and degenerative arthritis. In RMD Open, Molina-Collada and colleagues report the nationwide Spanish ARCA registry of 728 patients with rheumatoid arthritis and a prior cancer, followed over more than 11,000 person-years, with 80 new cancer events giving an incidence of about 7 per 1,000 person-years [2]. Compared with conventional synthetic DMARDs, there was no detectable difference in the risk of new primary cancer, locoregional progression or metastasis for TNF inhibitors, interleukin-6 inhibitors, rituximab, targeted synthetic DMARDs or abatacept, and that held after adjusting for exposure duration, line of therapy and prior cancer status. The confidence around the targeted synthetic and abatacept estimates is wide and the design is retrospective, so this is reassurance rather than exoneration for a group routinely excluded from trials. In Rheumatology, Zalcman and colleagues looked at what to do when a child on a biologic flares, across four Israeli tertiary centres, covering 65 children with non-systemic polyarticular juvenile idiopathic arthritis and 116 flares [7]. Overall, roughly seven in ten flares met the success definition of regaining inactive disease within three months and holding it. Intra-articular corticosteroid injection succeeded in about 82 percent of flares and biologic dose optimisation in about three quarters, while immediately switching biologics succeeded in only about a third, and switching remained an independent predictor of both lower remission and slower time to remission after adjustment. This is retrospective with obvious channelling, since the sickest flares are the ones that get switched, but it does support a stepwise approach before cycling biologics. Osteoarthritis and Cartilage carries a meta-analysis by Wen and colleagues of 34 randomised trials and 3,763 participants covering 15 guideline-endorsed drugs for hand osteoarthritis [4]. Topical NSAIDs produced the largest overall treatment effects on pain, function and stiffness, but when the placebo arm is accounted for they beat placebo only for function, and around 70 percent of the apparent benefit was contextual rather than pharmacological. There were no head-to-head comparisons, so the authors urge caution, but it is a sober reminder of how much of what we see in hand osteoarthritis consultations is context. Finally, the Lancet Rheumatology publishes a post-hoc pharmacodynamic analysis by Benschop and colleagues of the phase 2a trial of peresolimab, the PD-1 agonist antibody, in 98 adults with rheumatoid arthritis [8]. Receptor occupancy exceeded 90 percent on circulating CD4 effector memory T cells at both doses, PD-1 high-expressing CD4 T cells fell by more than half, neutrophils fell modestly, and proteomics showed down-modulation of interleukin signalling pathways including IL-6 from week 8. The target engagement story is clean, but the paper also records that development of peresolimab was discontinued after the subsequent phase 2b dose-ranging study, so this stands as mechanistic proof of concept for PD-1 agonism rather than a drug in waiting.

One more paper, on comorbidity. In Rheumatology, Bae and colleagues used a worldwide physician-reported registry of 1,868 patients with anti-synthetase syndrome and 2,043 controls across 103 centres to quantify pulmonary hypertension [9]. Pulmonary hypertension was present in about 9 percent of patients with anti-synthetase syndrome, higher than in non-anti-synthetase myositis, rheumatoid arthritis or interstitial pneumonia with autoimmune features, and not significantly different from systemic sclerosis or lupus. Five-year survival was 94 percent overall but fell to 88 percent in those with pulmonary hypertension, which translated to roughly two and a half times the mortality risk independent of age, sex and interstitial lung disease.

If you only have time for one paper this week, make it the anti-synthetase pulmonary hypertension study in Rheumatology [9]. It puts a number on a complication many of us suspected but could not quantify, and it places anti-synthetase syndrome much closer to scleroderma than to other myositis in pulmonary vascular risk, which reopens the question of who warrants systematic screening.

Here is what this week's evidence adds up to in Rheumatology. First, in active lupus nephritis, registry data give no signal that rituximab, mycophenolate or cyclophosphamide induction differ at one year, which supports choosing on comorbidity, fertility and tolerability grounds while we wait for randomised head-to-head comparisons. Second, measurement matters: SLE-DAS outperformed SLEDAI-2K for predicting damage accrual in a real-world cohort, though this is single-centre and partly retrospective. Third, risk prediction tools are maturing unevenly — the infection model in ANCA-associated vasculitis validated more convincingly than the relapse model, while the four giant cell arteritis probability scores disagreed with each other in a substantial minority of patients. Fourth, for patients with rheumatoid arthritis and a prior cancer, nationwide registry data found no difference in cancer events across biologic and targeted synthetic strategies, which is reassuring but not definitive. And fifth, in hand osteoarthritis, most of the apparent benefit of topical NSAIDs in trials appears to come from the therapeutic context rather than the drug, a finding that should temper expectations rather than dictate prescribing.

That's your roundup for This Week in Rheumatology. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And if it saves you time, a quick rating in your podcast app helps other physicians find the show.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.