FormalPara Key Summary Points

Why carry out this study?

In addition to reduced mobility and the presence of objective signs and symptoms of inflammation, patients with non-radiographic axial spondyloarthritis (nr-axSpA) frequently experience reduced health-related quality of life (HRQoL) and work productivity.

This study evaluated the impact of Janus kinase inhibitor, upadacitinib, versus placebo on response rates and clinically meaningful improvements in patient-reported outcomes (PROs) in patients with active nr-axSpA in SELECT-AXIS 2.

What was learned from the study?

Upadacitinib improved mean HRQoL and work productivity measures from baseline to week 14 compared to placebo (ASQoL and ASAS HI, ranked, P < 0.001; SF-36 PCS and three of four WPAI domains, nominal P < 0.05).

Higher proportions of patients achieved clinically meaningful improvements in HRQoL measures, including ASQoL (62.6 vs. 40.9%), ASAS HI (44.8 vs. 28.8%), and SF-36 PCS (69.3 vs. 52.0%) at week 14 with upadacitinib vs. placebo (nominal P ≤ 0.01).

Improvements in HRQoL occurred as early as week 2 for ASAS HI and week 4 for ASQoL and SF-36 PCS.

Introduction

Axial spondyloarthritis (axSpA) is a chronic inflammatory condition with manifestations primarily involving the sacroiliac joints (SIJs) and spine [1,2,3]. Axial SpA comprises two subsets of disease, classified as radiographic axSpA (r-axSpA, historically known as ankylosing spondylitis [AS]) and non-radiographic axSpA (nr-axSpA), with r-axSpA or AS presenting with structural damage of the SIJ visible on X-ray and nr-axSpA presenting with no definitive radiographic evidence of SIJ structural damage [2, 4, 5]. Additionally, patients with nr-axSpA compared to AS are more frequently female, have lower levels of inflammation as measured via C-reactive protein (CRP), and have a shorter disease duration [6,7,8].

Inflammatory back pain is a hallmark of both AS and nr-axSpA, frequently contributing to functional impairments and work disability [2, 9, 10]. Patients have rated spinal pain as the most important aspect influencing the burden of axSpA [11]. Additionally, patients with axSpA experience a significant disease burden on their health-related quality of life (HRQoL) and ability to perform daily activities and an increased prevalence of comorbidities contributing to higher healthcare costs [2, 12]. Disease burden in terms of disease activity and impairments in physical activity as well as HRQoL was shown to be similarly affected in nr-axSpA and AS [13].

The goal of treatment includes maximizing HRQoL through control of inflammation, prevention of progressive structural damage, and normalization of function [2, 3, 14,15,16]. The Assessment of SpondyloArthritis international Society (ASAS) and the European Alliance of Associations for Rheumatology (EULAR) now recommend biological disease-modifying antirheumatic drugs (bDMARDs) such as tumor necrosis factor (TNF) inhibitors (TNFis) and interleukin (IL)-17 inhibitors, or Janus kinase (JAK) inhibitors, as treatment in patients with axSpA and persistently high disease activity despite prior use of non-steroidal anti-inflammatory drugs (NSAIDs) [14, 16]. The American College of Rheumatology, Spondylitis Association of America, and Spondyloarthritis Research and Treatment Network recommend a similar treatment approach [15]. Approved treatment options for nr-axSpA are limited and until recently included only TNFis and IL-17 inhibitors, highlighting a need for an oral treatment option for the entire spectrum of axSpA.

Upadacitinib, an oral JAK inhibitor, was recently approved for the treatment of active nr-axSpA in adult patients with objective signs of inflammation in the European Union and the United States, based on data from the SELECT-AXIS 2 study, which evaluated the safety and efficacy of upadacitinib vs. placebo in patients with active nr-axSpA with or without prior exposure to bDMARDs [17]. This analysis evaluated response rates and clinically meaningful improvements in HRQoL and work productivity in patients with active nr-axSpA treated with upadacitinib 15 mg vs. placebo in the SELECT-AXIS 2 trial.

Methods

Study Design and Participants

The study design and primary results through week 14 from the global, multicenter, phase 3 clinical trial SELECT-AXIS 2 were previously reported [17]. Patients in this 52-week randomized, double-blind, parallel-group, placebo-controlled study were randomized 1:1 to receive upadacitinib 15 mg once daily or placebo, followed by a 52-week open-label extension period. The primary endpoint was the ASAS40 response rate at week 14. Here we report improvement in HRQoL and work productivity through 14 weeks in patients with active nr-axSpA enrolled in the SELECT-AXIS 2 trial.

As previously described, eligible patients were adults with a clinical diagnosis of active nr-axSpA, fulfilling the 2009 ASAS classification criteria [5] but not the radiological criterion of the modified New York criteria [18], and an inadequate response or intolerance to ≥ 2 NSAIDs [17]. Included patients had objective signs of inflammation-based magnetic resonance imaging (MRI) of the SIJ or elevated CRP [17]. Patients who were MRI-negative but had elevated CRP were eligible for enrolment. Prior treatment with no more than one bDMARD (TNFi or IL-17 inhibitor) was permitted for at least 20% but no more than 35% of patients who discontinued prior bDMARD due to lack of efficacy (after at least 12 weeks of treatment at an adequate dose) or intolerance (irrespective of treatment duration). Patients who had a lack of efficacy for both a TNFi and IL-17 inhibitor were not eligible for enrolment. In addition, previous exposure to a JAK inhibitor was prohibited.

Compliance with Ethics Guidelines

The SELECT-AXIS 2 trial was conducted in compliance with the International Conference on Harmonization guidelines, local regulations and guidelines governing clinical study conduct, and the ethical principles of the Declaration of Helsinki. As per Good Clinical Practice, the protocol was approved by an ethics committee or institutional review board at all study sites (Supplemental Table S1). All patients provided informed consent before study participation.

Outcomes

This analysis assessed mean change from baseline as well as the proportion of patients reporting clinically meaningful improvements across multiple measures of HRQoL and work productivity.

The mean change from baseline through 14 weeks was assessed for the following measures of HRQoL: Ankylosing Spondylitis Quality of Life (ASQoL) [19], ASAS Health Index (ASAS HI) [20], and Short Form-36 (SF-36) Physical Component Summary (PCS) score [21]; and for work productivity, by the Work Productivity and Activity Impairment (WPAI) Questionnaire [22]. WPAI consists of the following four domains: absenteeism (work missed), presenteeism (reduced work effectiveness), overall work impairment, and activity impairment.

Clinically meaningful improvements were assessed for key patient-reported measures that assess HRQoL and physical function with validated minimum clinically important difference (MCID) thresholds, including ASQoL (MCID ≥ 3-point decrease [2]), ASAS HI (MCID ≥ 3-point decrease [23]), and SF-36 PCS (MCID ≥ 2.5-point increase [24]).

Post-baseline assessments of ASQoL, ASAS-HI, and SF-36 PCS began at week 1, with ASQoL and SF-36 PCS assessed at weeks 1, 4, 8, and 14 and ASAS HI at weeks 1, 2, 4, 8, and 14. Post-baseline assessments for WPAI began at week 14. Mean change from baseline was assessed at each week when assessment data was collected through 14 weeks. Clinically meaningful improvements ≥ MCID were assessed at week 14. WPAI absenteeism, presenteeism, and overall work impairment domains were assessed in patients employed at baseline, while activity impairment was assessed in all patients.

Statistical Analyses

Analyses were conducted in the full analysis set, which included all randomized patients who received at least one dose of study drug. Baseline characteristics are summarized with descriptive statistics using means and standard deviation for continuous endpoints and frequency (n) and percentages for categorical endpoints. For ranked endpoints (mean change from baseline to week 14 in ASQoL and ASAS HI), results were obtained via sequential multiple testing procedure controlling the overall type I error rate of all primary and multiplicity-adjusted secondary endpoints at the significance level of 0.05. Other outcomes were not multiplicity adjusted and all other reported P values were nominal.

Mean change from baseline to week 14 along with 95% confidence intervals (CIs) and nominal P values were based on mixed-effects repeated measures model (MMRM) analysis, including stratification factors of treatment, visit, and treatment-by-visit interaction as fixed factors and baseline value as a covariate for all patient-reported outcomes (PROs) except WPAI. MRI and screening high-sensitivity CRP (hsCRP) status were also included in the model as main stratification factors. All observed patient data were included in the MMRM. For WPAI, within-group mean change from baseline and 95% CI and between-group least squares mean, 95% CI, and nominal P value for each visit were based on an analysis of covariance model, including treatment and baseline value as a covariate; MRI and screening hsCRP status were also included in the model.

The number and percentage of patients with improvements in each PRO ≥ MCID are summarized at week 14. Comparisons of the proportion of patients achieving MCID were made between upadacitinib and the placebo group using the Cochran–Mantel–Haenszel test (CMH). The point estimate and 95% CIs estimated using normal approximation and nominal P values for the treatment comparisons are presented. The number needed to treat (NNT), defined as the number of patients who need to receive treatment to achieve one additional ‘responder’ on a PRO of interest vs. placebo, was calculated. Specifically, NNTs for each PRO with an MCID were calculated as follows: 1/ (MCID response rate in the treatment group—MCID response rate in the placebo group), and the NNTs per MCID response were compared between upadacitinib and placebo using the CMH test at week 14. For all MCID analyses, missing data were reported using non-responder imputation (NRI) incorporating multiple imputation (MI) to handle missing data due to COVID-19 or NRI if there were no missing data due to COVID-19. The number of responders was based on the total number of patients and MI-aggregated response rates.

The proportion of patients who achieved Ankylosing Spondylitis Disease Activity Score (ASDAS) inactive disease (ASDAS CRP < 1.3) at week 14 among those who achieved MCID for ASAS HI, ASQoL, and SF-36 PCS, separately, at week 14 was calculated for the full analysis set. MCID and NNT analyses were repeated in a post hoc subgroup analysis of patients stratified by prior TNFi status. Patients were categorized as TNFi-naïve if they had no prior exposure to TNFis, and patients who had prior exposure, including either intolerance or lack of efficacy to TNFis were categorized as TNFi-IR at baseline.

Results

Several key baseline characteristics were previously reported for the SELECT-AXIS 2 study [17]. A total of 313 patients with active nr-axSpA were included in the analysis, of whom 156 and 157 were randomized to upadacitinib or placebo, respectively. Most patients were female (58.5%), with a mean age of approximately 42 years. Baseline impairments were observed in patient-reported measures of HRQoL, as indicated by the high scores in ASAS HI and ASQoL and low scores in SF-36 PCS (Table 1). Baseline scores in WPAI overall work impairment (mean scores: 60.4 placebo and 54.6 upadacitinib 15 mg) indicate a loss in work productivity in these patients. Overall, baseline characteristics and PRO scores were balanced across treatment groups (Table 1).

Table 1 Baseline demographics and clinical characteristics
Full size table

Significant improvements in mean change from baseline to week 14 were reported with upadacitinib 15 mg compared with placebo across key ranked secondary endpoints ASQoL and ASAS HI (P < 0.001); improvements from baseline in SF-36 PCS were also greater with upadacitinib compared to placebo (nominal P < 0.001) (Fig. 1). Improvements from baseline with upadacitinib 15 mg differentiated from placebo as early as week 2 for ASAS HI and week 4 for ASQoL and SF-36 PCS and continued to week 14 (Fig. 1). Furthermore, improvement in mean change from baseline to week 14 across WPAI domains of presenteeism, overall work impairment, and activity impairment was greater with upadacitinib 15 mg compared with placebo (nominal P < 0.05); however, improvement in absenteeism was similar between treatment groups (nominal P = 0.3387) (Fig. 2).

Fig. 1
figure 1

Change from baseline through week 14 in A ASQoL, B ASAS HI, and C SF-36 PCS (MMRM). *Nominal P value ≤ 0.05; **nominal P value ≤ 0.01; ***nominal P value ≤ 0.001. aSignificant after multiplicity adjustment. MMRM analysis included treatment, visit, and treatment-by-visit interaction as fixed factors and baseline value as covariate. Stratification factor hsCRP level (≤ ULN vs. > ULN) was also included in the model. ASAS HI Assessment of SpondyloArthritis international Society Health Index, ASQoL Ankylosing Spondylitis Quality of Life, CI confidence interval, hsCRP high-sensitivity C-reactive protein, MMRM mixed-effect model for repeated measures, PBO placebo, PCS Physical Component Summary, SF-36 36-Item Short Form Health Survey, ULN upper limit of normal, UPA upadacitinib

Full size image
Fig. 2
figure 2

Mean change (95% CI) from baseline in WPAI at week 14. *Nominal P value ≤ 0.05; **nominal P value ≤ 0.01. aIncludes patients currently employed. Within-group mean and 95% CI and between-group mean, 95% CI, and nominal P value for each visit are based on analysis of covariance model including treatment and main stratification factors MRI and screening hsCRP status (MRI + /hsCRP > ULN, MRI + /hsCRP ≤ ULN, and MRI–/hsCRP > ULN) as fixed factors and baseline value as covariate for all WPAI domains. CI confidence interval, hsCRP high-sensitivity C-reactive protein, MRI magnetic resonance imaging, PBO placebo, ULN upper limit of normal, UPA upadacitinib, WPAI Work Productivity and Activity Impairment

Full size image

At week 14, a higher proportion of patients reported clinically meaningful improvements ≥ MCID with upadacitinib 15 mg compared with placebo in key PRO measures of ASQoL (62.6 vs. 40.9%; nominal P ≤ 0.001), ASAS HI (44.8 vs. 28.8%; nominal P ≤ 0.01), and SF-36 PCS (69.3 vs. 52.0%; nominal P ≤ 0.01) (Fig. 3). The NNTs were < 10, ranging from 4.6 to 6.2 for all PROs assessed, suggesting a clinically relevant difference [25] in measures of HRQoL (Fig. 3).

Fig. 3
figure 3

Proportion of patients reporting improvements ≥ MCID and NNTs in PROs at week 14 (NRI-MI). **Nominal P ≤ 0.01 and ***nominal P ≤ 0.001 versus placebo. NRI-MI is non-responder imputation (NRI) incorporating multiple imputation (MI) to handle missing data due to COVID-19. MCID definitions: ≥ 3 points decrease (ASAS HI and ASQoL) and ≥ 2.5-point increase (SF-36 PCS). ASAS HI Assessment of SpondyloArthritis international Society Health Index, ASQoL Ankylosing Spondylitis Quality of Life, MCID minimal clinically important difference, NNT number needed to treat, PBO placebo, PCS Physical Component Summary, PRO patient-reported outcome, SF-36 36-Item Short Form Health Survey, UPA upadacitinib

Full size image

Clinically Meaningful Improvements in PROs by TNF Exposure Status

The subgroup analysis included 226 patients who were TNFi-naïve (n = 114 for placebo and n = 112 for upadacitinib) and 87 patients who were TNFi-IR (n = 43 for placebo and n = 44 for upadacitinib). Patients who had prior exposure to both mechanisms of action (i.e., protocol deviations) are not excluded here; this includes 3 patients randomized to placebo in the TNFi-IR subgroup who also had prior exposure to IL-17 inhibitors. The TNFi-naïve subgroup included 16 patients with prior exposure to IL-17 inhibitors only. In patients who were TNFi-naïve, a greater proportion of patients treated with upadacitinib 15 mg vs. placebo reported achieving clinically meaningful improvements in ASQoL (64.0 vs. 38.7%; nominal P ≤ 0.001), ASAS HI (46.4 vs. 27.3%; nominal P ≤ 0.01), and SF-36 PCS (70.4 vs. 54.1%; nominal P ≤ 0.01) (Supplementary Fig. S1A). Consistent with the full analysis set and TNFi-naïve population, patients with prior exposure to TNFis at baseline reported a numerically higher proportion of patients achieving clinically meaningful improvements in ASQoL (59.1 vs. 46.5%; nominal P = 0.2216), ASAS HI (40.9 vs. 32.6%; nominal P = 0.4166), and SF-36 PCS (66.7 vs. 46.5%; nominal P < 0.05) with upadacitinib 15 mg treatment at week 14 compared with placebo (Supplemental Fig. S1B). Most NNTs were < 10, ranging from 4.0 to 6.1 for all PROs assessed in the TNFi-naïve patients and from 4.8 to 12.0 in the TNFi-IR patients, suggesting a clinically relevant difference in most measures of HRQoL regardless of prior treatment status [25].

Achievement of Inactive Disease Among Those who Achieved Clinically Meaningful Improvement in PROs

A higher proportion of patients treated with upadacitinib 15 mg compared with placebo achieved ASDAS inactive disease among those who achieved clinically meaningful improvements in ASQoL (21.5 vs. 9.4%; nominal P < 0.05), ASAS HI (28.6 vs. 11.1%; nominal P < 0.05), and SF-36 PCS (20.3 vs. 8.6%; nominal P < 0.05) (Supplementary Fig. S2). The percentage of patients who achieved ASDAS inactive disease overall, without consideration for MCID achievement, was 14.1 vs. 5.2% for upadacitinib- vs. placebo-treated patients, respectively [17].

Discussion

The burden of disease in patients with active nr-axSpA is high, with a detrimental impact on pain, physical function, HRQoL, loss of work productivity, and limits to daily activities [2, 26, 27]. The SELECT-AXIS 2 trial is the first to investigate a JAK inhibitor for the treatment of active nr-axSpA. In this study, improvements from baseline were significantly greater in active nr-axSpA patients treated with upadacitinib vs. placebo at week 14 in key measures of HRQoL, ASQoL, and ASAS HI. Additionally, as previously reported, improvements from baseline were also demonstrated to be greater with upadacitinib vs. placebo at week 14 in total and nocturnal back pain and function as assessed by Bath Ankylosing Spondylitis Functional Index (BASFI) [17]. Clinically meaningful improvements were observed in greater proportions of patients treated with upadacitinib vs. placebo across ASQoL, ASAS HI, and SF-36 PCS, with NNTs meeting the clinically meaningful threshold of < 10 [25]. Results were consistent among patients who were TNFi-naïve and those who previously failed TNFi treatment, with NNTs < 10 for most PROs, suggesting upadacitinib may be a beneficial treatment in patients refractory to bDMARDs.

Notably, improvements in HRQoL occurred rapidly in patients treated with upadacitinib vs. placebo. Specifically, improvements in disease-specific and general HRQoL, as measured by ASQoL, ASAS HI, and SF-36 PCS, occurred as early as weeks 2–4; as previously reported, improvements in patient-reported measures of pain and physical function (BASFI) also occurred as early as weeks 2–4 [17]. These rapid improvements from the patient perspective are aligned with treatment goals of reaching early status of inactive disease, normalization of function, and improvement in HRQoL [16].

Furthermore, positive improvements in pain, as well as physical function and general HRQoL as measured by SF-36 PCS and ASAS HI, were consistent with previous studies with IL-17 inhibitors [28, 29]. Results were also consistent with studies with TNFi treatments in nr-axSpA [30,31,32]. Importantly, in the current study, comparable or better outcomes were observed in measures of HRQoL irrespective of prior TNFi exposure status, providing additional evidence for upadacitinib as a potential treatment option in patients with active nr-axSpA, including patients with inadequate response to prior advanced therapies.

Several studies have demonstrated that improvements in PROs are associated with improved clinical measures of disease activity and symptoms, which is critical in axSpA given the high disease burden [28, 30, 33,34,35]. A significantly greater proportion of patients with active nr-axSpA treated with upadacitinib vs. placebo achieved the primary endpoint of ASAS40 at week 14 as well as the threshold of ASDAS inactive disease [17]. Although the current study did not assess a statistical correlation between patients who achieved ASDAS inactive disease and PRO outcomes, improvements in measures of HRQoL were observed through week 14 in nr-axSpA patients treated with upadacitinib. Additionally, among patients who achieved improvements in HRQoL ≥ MCID, the percentage who achieved ASDAS inactive disease with upadacitinib treatment ranged from 20 to 28%, which was numerically higher than the proportion of patients who achieved ASDAS inactive disease for the overall population (reported as 15% with upadacitinib) [17]. Given the correlation between PROs and provider-assessed disease activity and symptoms, future studies may be warranted to characterize this relationship in patients with active nr-axSpA treated with upadacitinib. The high placebo responses observed in AS studies [36] also suggest that measuring patients’ perception of symptom improvement is important to better differentiate treatment effects.

In addition to PROs assessing HRQoL, we also evaluated the impact of upadacitinib treatment on patients’ ability to work. Guidelines recommend that work productivity loss be considered in treatment decisions as it is associated with the indirect economic burden of axSpA [14]. Furthermore, compared to patients with AS, patients with nr-axSpA reported greater impacts of disease on presenteeism and overall activity impairment [2]. Greater improvements from baseline to week 14 in patients treated with upadacitinib vs. placebo were observed in the WPAI domains of presenteeism, overall work impairment, and activity impairment, suggesting that upadacitinib may improve patients’ work productivity and ability to perform daily activities. These improvements are encouraging, given that ability to work is influenced by multiple dimensions.

There are strengths and limitations to this current analysis. Strengths include the randomized study design, which mitigates bias that may arise due to unobservable differences between cohorts, as well as the use of validated PROs to assess improvements in patients’ perception of their symptoms. Limitations including small sample size may result in imprecise estimates and limit the ability to make causal inferences, particularly in subgroup analyses. The number of patients exposed to IL-17 inhibitors was small; therefore, no analysis was conducted assessing outcomes specifically in this population. Future real-word studies are required to confirm the generalizability of these results as patients enrolled in clinical trials may differ from patients in the general population. Furthermore, data here are provided for up to 14 weeks; therefore, long-term follow-up is needed to confirm the durability of PRO improvements. Future studies may also be warranted to characterize the relationship between PROs and the achievement of improved clinical signs and symptoms.

Conclusions

Upadacitinib provides rapid and clinically meaningful improvements in HRQoL as assessed by ASQoL, ASAS HI, and SF-36 PCS, as well as work productivity, in patients with active nr-axSpA. Furthermore, clinically important improvements in HRQoL were consistently observed regardless of prior treatment with TNFis, thus providing further evidence substantiating the overarching benefit of upadacitinib in patients with active nr-axSpA. These findings, together with previously reported efficacy and safety in patients with active AS [37], provide further evidence for the benefit of upadacitinib across all active axSpA patients.