Upadacitinib withdrawal following axSpA remission: Clinical outcomes from SELECT-AXIS 2

Key findings from EULAR 2026 outline whether axSpA remission is maintained a year following upadacitinib withdrawal.

SELECT-AXIS 2 (NCT04169373) was a placebo-controlled, double-blind phase 3 study investigating the efficacy and safety of upadacitinib in patients with axial spondyloarthritis (axSpA). The findings from SELECT-AXIS 2 supported the approval of upadacitinib as the first Janus Kinase (JAK) inhibitor indicated for the treatment of axSpA.

In this interview we spoke with Prof. Atul Deodhar (Oregon Health & Science University, Portland, OR, USA) about his analysis of the SELECT-AXIS 2 study, looking at clinical outcomes following upadacitinib withdrawal in patients who achieved remission at week 104 in the phase 3 study.

Abstract: Clinical Outcomes Following Withdrawal of Upadacitinib in Patients With Axial Spondyloarthritis Who Achieved Remission: Findings From the Phase 3, Placebo-Controlled, Double-Blind SELECT-AXIS 2 Trial. EULAR 2026, June 3– 6, London, UK.

touchIMMUNOLOGY coverage of EULAR 2026

My name is Atul Deodhar, I’m a rheumatologist and Professor of Medicine at Oregon Health & Science University in Portland, Oregon.

Where does upadacitinib fit in the treatment paradigm for active axSpA?

We have a very limited number of agents available for axial spondyloarthritis. There are four types of drug therapies, the oldest being non-steroidal anti-inflammatory drugs (NSAIDs). Of the novel therapies, there are three classes: TNF inhibitors, IL-17 inhibitors and more recently JAK inhibitors.

In Europe, the EMA recommendations are that JAK inhibitors can be used as a start therapy after NSAIDs. In the US, we cannot use them unless the patient has failed TNF inhibition. Typically, patients would be given NSAIDs first with physical therapy followed by TNF inhibitors, which have the longest record for safety and efficacy.

In the 2019 American College of Rheumatology SAA-SPARTAN treatment recommendations, we put JAK inhibitors third after IL-17 inhibitors. The 2026 update to the treatment guidelines will be presented for the first time at the ACR meeting later this year, and there you will see that we have moved JAK inhibitors up a little bit. After TNF failure, you will be able to select JAK inhibition or IL-17 inhibition, depending on the patient’s characteristics and their phenotype.

Could you give a brief overview of the SELECT-AXIS 2 study?

SELECT-AXIS 2 was a phase 3, placebo-controlled, double-blind trial, which lasted 2 years. The first group enrolled included patients with axial spondyloarthritis fulfilling modified New York criteria for radiographic axial spondyloarthritis (r-axSpA), and they had to have an inadequate response to or be intolerant to biologic DMARDs. The second group enrolled patients with non-radiographic axial spondyloarthritis (nr-axSpA), fulfilling the 2009 classification criteria.

The endpoint was to assess what percentage of patients achieved clinical remission defined as Ankylosing Spondylitis Disease Activity Score (ASDAS) < 2.1 at Week 88 and ASDAS <1.3 at Week 104; both of those had to be met. Low disease activity at Week 88 and inactive disease at Week 104. Then at Week 104, we would say that they are in remission.

What were the objectives of your analysis in patients who achieved remission?

A question that is commonly asked by patients is now that I am doing well, do I have to take this medication for the rest of my life? It is expensive and there are side effects, can I stop the medication? The EMA in Europe have made it mandatory for pharmaceutical companies to do a withdrawal study, whereas the FDA hasn’t mandated that in the US.

The withdrawal study is important but can only take place in patients who are doing remarkably well, and that’s why we chose to do this study in patients with low disease activity at Week 88 and then inactive disease at Week 104. The first goal was to see how long patients would remain in an inactive disease state once they stopped the drug. Secondly, for patients who flared, we wanted to see how quickly they flared. Then thirdly, for patients who flared, what happens if you re-challenge them with the same drug, will they respond and will they get to the same state as before? The objectives were to answer those real-world questions our patients have.

What was the methodology of the withdrawal analysis?

In the original SELECT-AXIS 2 study, 734 patients participated. Of those, 194 patients (26.4%) were in remission at 2 years, meaning that they had low disease activity or inactive disease and they were willing to participate in this withdrawal study. We enrolled them into this study at 104 weeks where they completely stopped the drug in an open-label fashion. The withdrawal study then continued for another year.

What were the outcomes following upadacitinib withdrawal?

We found that 22.2% (43/194) of the patients maintained the ASDAS inactive disease state a year after complete withdrawal. This to me is remarkable because this is the closest we have come to a cure, and I’m not saying that we have cured these people, but drug-free remission is as close to cure as you get. If we followed these patients for another year, there is a possibility they might flare. But the next time a patient asks me if they can stop the drug, I can tell them that one-fifth of patients can stay in this disease state for another year.

What were the efficacy and safety findings following retreatment with upadacitinib?

As part of the clinical trial, patients who flared following withdrawal received retreatment with upadacitinib. To our relief, by Week 24, 89% of them achieved their low disease activity status again, ASDAS LDA <2.1.

The safety profile remained consistent with the previous reports and there were no new safety signals in the year within the 78% of patients who went back on the drug.

Looking ahead, how will these findings be used to inform treatment decisions, and what further studies are needed?

These findings are important, as sometimes patients have to stop taking treatments for reasons such as intercurrent illnesses, COVID-19, major surgery, or pregnancy. Now, a percentage could theoretically stop the drug completely and be in this disease free, drug-free remission state. Looking at this glass half full, although 78% of the patients flared after withdrawal, I want to look at the small percentage who remained in drug-free status for at least a year, which is remarkable.

We also need to identify the clinical characteristics of the patients who withdraw and remain in low disease activity or remission – is this affected by sex, age, BMI, markers such as HLA B27or high CRP? This will also give us an idea as to which patients are more likely to flare. Then in the clinic, we can make an educated judgment of whether the patient can come off the drug or not.

Another aspect, which looking back we should have assessed in the study, would be to have a third group of patients receiving half the dose. We could have tried reducing the frequency to every other day or a few times a week. Many times, patients do this without telling us they forget to take the pill. There are other similar studies in biologics, for instance in a study of certolizumab there was a group which completely stopped the therapy, one group which continued, and the third group received half the dose in a blinded fashion.