Predicting Persistent Arthritis After ICI Therapy: Kate Harnden, MBChB, PhD

Immune checkpoint inhibitors (ICIs) now treat a growing share of solid tumors, but inflammatory arthritis and arthralgia remain among their most common rheumatologic complications. Up to 6 times more patients develop arthralgia alone than frank inflammatory arthritis, yet most outcome data have focused only on the latter group.¹ Clinicians weigh this risk against uncontrolled joint disease, since immunosuppression needed to treat arthritis can in theory blunt the antitumor effect of the checkpoint inhibitor itself.

A Leeds cohort of more than 200 patients with ICI-induced arthritis or arthralgia, followed for up to 5 years, found 57% had persistent symptoms at a median follow-up of 16 months.¹ Persistence rates were similar whether patients presented with arthralgia or inflammatory arthritis, and 2 multivariable models linked elevated CRP, tenosynovitis on imaging, and a rheumatoid arthritis-like peripheral MRI pattern to persistence and DMARD requirement. Continuing versus stopping the checkpoint inhibitor showed no significant association with outcome.

The findings were presented as a poster at the EULAR 2026 Congress in London.¹ The same cohort data are also being presented at ACR Convergence 2026, held November 6-11 in Orlando, Florida, in a plenary session on immunologic complications of medical therapy.² Flagging predictors early, before DMARDs become necessary, could help rheumatology and oncology teams decide when immunosuppression is warranted without reflexively holding effective cancer therapy.

Kate Harnden, PhD, MBChB, MRCP, of the Leeds Institute of Rheumatic and Musculoskeletal Medicine at the University of Leeds, led the research. In the following interview, recorded as part of a Joint Ventures episode at EULAR 2026 in London with host Rihards Buss, MD, consultant rheumatologist at Freeman Hospital in Newcastle, Harnden discusses the predictive models and what comes next for multicenter validation.

Rihards Buss, MD: Can you describe your research and the findings from your poster?

Kate Harnden, PhD, MBChB: My research is on ICI-induced arthritis and arthralgia. In Leeds, over the last 5 years, we’ve developed a cohort of over 200 patients with arthralgia or arthritis following ICIs, which are increasingly used to treat cancer but cause a lot of immune-related adverse events (AEs). Inflammatory arthritis is very common, but arthralgia without obvious swelling may be up to 6 times more common, and our cohort is fairly unique in that we include the arthralgia patients as well, not just the ones with obvious inflammatory arthritis.

It’s a really complicated area, because any immunosuppression we give these patients will theoretically work against what the ICI is doing for their cancer, so it’s a difficult balance to treat them. My poster has been looking at long-term arthritis outcomes in these patients, specifically whether we can predict their requirement for a DMARD and whether they’re going to have persistent symptoms.

Early research suggested around 50% of patients with inflammatory arthritis have persistent symptoms despite stopping the immunotherapy, which is different from most of the other organ toxicities, which tend to resolve once the checkpoint inhibitor stops. For arthritis, for some reason we’re not sure of, it seems to persist more. We took all the patients who’d been in our study for at least 6 months and invited them back for a final follow-up visit, where we assessed whether they had persistent symptoms and whether they’d needed a DMARD, or weren’t able to taper their steroids below 10 mg of prednisone or equivalent, which is the level at which we’d consider starting a DMARD.

From there we developed 2 multivariable predictive models for disease persistence. Our median follow-up was about 16 months, and we found that, in keeping with the literature, around 57% of patients had persistent symptoms. Interestingly, patients with arthralgia had the same levels of persistence as those with inflammatory arthritis, which is important because there may be up to 6 times more of these patients, and it’s a growing burden in rheumatology services.

We found certain factors that predicted both disease persistence and the need for a DMARD: having an elevated CRP, having tenosynovitis on imaging, and having a specific peripheral inflammatory arthritis MRI pattern similar to what we see in rheumatoid arthritis. Those are the factors associated with both DMARD requirement and symptom persistence.

HCPLive: Does stopping or continuing the ICI affect whether arthritis symptoms persist?

Kate Harnden, MBChB, PhD: There’s only 1 other study that’s looked at persistence rates using multivariable analysis, and it only included patients with inflammatory arthritis who had to stop the immunotherapy, finding that 50% persisted despite stopping. We included all patients, regardless of whether they’d stopped the immunotherapy, but we did factor into our model whether they were on immunotherapy at baseline or had stopped it within 3 months of baseline, and we didn’t find that was significantly associated with persistence rates. We had patients whose arthritis symptoms resolved while still on immunotherapy, and patients who persisted even though they’d stopped immunotherapy a long time ago, so in our model it didn’t seem to make a difference. Other smaller studies have also reported that in some patients arthritis goes away despite stopping immunotherapy, which is an interesting finding.

HCPLive: What are your plans for future work in this area?

Kate Harnden, MBChB, PhD: We want to develop these multivariable models further, looking specifically at whether there are certain biomarkers we can add to improve their accuracy. We’re also keen to collaborate with other centers, to see if these models hold up in different populations, and to build bigger collaborations more broadly.

Editor’s Note: This transcript has been edited for grammar and clarity using artificial intelligence tools.

REFERENCES:

1. Harnden K, Sharrack S, Sugden H-J, et al. Predicting the outcomes of immune checkpoint inhibitor-induced inflammatory arthritis and arthralgia. Ann Rheum Dis. 2026;85(suppl 1):POS0147. doi:10.1136/annrheumdis-2026-eular.B.1618

2. Harnden K. Predicting the outcome of immune checkpoint inhibitor-induced inflammatory arthritis and arthralgia: development of multivariable models to predict disease persistence [plenary presentation]. Presented at: ACR Convergence 2026; November 10, 2026; Orlando, FL.