Effects of upadacitinib on patient-reported outcomes: results from SELECT-BEYOND, a phase 3 randomized trial in patients with rheumatoid arthritis and inadequate responses to biologic disease-modifying antirheumatic drugs

Background Patient-reported outcomes (PROs) are important when evaluating treatment benefits in rheumatoid arthritis (RA). We compared upadacitinib, an oral, selective JAK-1 inhibitor, with placebo to assess clinically meaningful improvements in PROs in patients with RA who have had inadequate responses to biologic disease-modifying antirheumatic drugs (bDMARD-IR). Methods PRO responses between upadacitinib 15 mg or 30 mg and placebo were evaluated at week 12 from the SELECT-BEYOND trial. Improvement was determined by measuring Patient Global Assessment of Disease Activity (PtGA), pain, Health Assessment Questionnaire Disability Index (HAQ-DI), Short Form-36 Health Survey (SF-36), duration and severity of morning (AM) stiffness, and Insomnia Severity Index (ISI). Least squares mean changes and percentage of patients reporting improvements ≥ minimum clinically important differences (MCID) and scores greater than or equal to normative values were determined. The number needed to treat (NNT) to achieve clinically meaningful improvements was calculated. Results In 498 patients, both upadacitinib doses resulted in statistically significant changes from baseline versus placebo in PtGA, pain, HAQ-DI, SF-36 Physical Component Summary (PCS), 7 of 8 SF-36 domains (15 mg), 6 of 8 SF-36 domains (30 mg), and AM stiffness duration and severity. Compared with placebo, more upadacitinib-treated patients reported improvements ≥ MCID in PtGA, pain, HAQ-DI, SF-36 PCS, 7 of 8 SF-36 domains (15 mg), 5 of 8 SF-36 domains (30 mg), AM stiffness duration and severity, and ISI (30 mg) and scores ≥ normative values in HAQ-DI and SF-36 domains. Across most PROs, NNTs to achieve MCID with upadacitinib ranged from 4 to 7 patients. Conclusions In bDMARD-IR RA patients, upadacitinib (15 mg or 30 mg) improved multiple aspects of quality of life, and more patients reached clinically meaningful improvements approaching normative values compared with placebo. Trial registration The trial is registered with ClinicalTrials.gov (NCT02706847), registered 6 March 2016.


Background
Rheumatoid arthritis (RA) is a chronic, inflammatory, and destructive disease of the synovial joints that is associated with substantial clinical burden, reduced healthrelated quality of life (HRQOL), and shortened life expectancy [1,2]. Most patients with RA experience impaired physical functioning, chronic pain, fatigue, and morning (AM) stiffness, which affect their HRQOL [2][3][4][5][6][7][8]. In addition, patients with RA have reported that sleep disturbances are a key determinant of their well-being [3,[9][10][11]. These reports are supported by cross-sectional studies, which have demonstrated that sleep disturbance correlates with greater pain, disease activity, and fatigue in RA [12][13][14][15]. Thus, understanding the patient's perspective of how a therapy impacts multiple aspects of HRQOL is crucial when evaluating the efficacy of treatments for RA [16][17][18].
SELECT-BEYOND is a phase 3, randomized controlled trial (RCT) of upadacitinib 15 mg or 30 mg once daily in patients with RA who are bDMARD-IR and receiving background csDMARDs. The RCT demonstrated that significantly more upadacitinib-treated patients had American College of Rheumatology 20% improvement (ACR20) responses and lower disease activity than placebo [25]. The objectives of the following analyses are to compare response rates and numbers needed to treat (NNTs) with upadacitinib versus placebo for PROs and assess the achievement of clinically meaningful improvements and normative values in patients with treatmentrefractory RA.

Study design and participants
The full study design of SELECT-BEYOND, a phase 3 RCT (NCT02706847), has been published previously [24,25]. Patients aged 18 years and older with moderate to severe RA for ≥ 3 months were randomized 1:1:1 to receive oral upadacitinib 15 or 30 mg once daily, or placebo, for 12 weeks. Patients were excluded if they had prior exposure to a JAK inhibitor. The study was approved by independent ethics committees or institutional review boards at each study site and conducted in accordance with ethical principles outlined in the current Declaration of Helsinki and consistent with International Conference on Harmonisation Good Clinical Practice and Good Epidemiology Practices, along with all applicable local regulatory requirements. All patient data were deidentified and complied with patient confidentiality requirements.

Patient-reported outcome assessment
Several clinically relevant PROs and assessments were used to evaluate the potential impact of upadacitinib on patient HRQOL and disease burden. These include Patient Global Assessment of Disease Activity (PtGA); pain on a visual analog scale (VAS); Health Assessment Questionnaire Disability Index (HAQ-DI) [27]; HRQOL using the Short Form 36 Health Survey (SF-36), which includes 8 domains scored 0-100 (physical functioning, role-physical, bodily pain, general health, vitality, social functioning, role-emotional, and mental health) and 2 aggregate Physical (PCS) and Mental Component Summary (MCS) scores (range of 0-100) [28,29] with higher scores indicating better HRQOL; AM stiffness severity measured by a numeric rating scale ranging from 0 to 10 with higher scores indicating greater severity, duration of AM stiffness measured in minutes [8,30,31]; the Insomnia Severity Index (ISI) to identify and grade insomnia severity (scores range from 0 to 28 where higher scores indicate increased insomnia) [32][33][34]; and the Euro Qol 5-Dimension 5-Level Questionnaire (EQ-5D-5L), which includes a general health status index measured by a 0-100 VAS and consists of 5 health states [35].
Clinically meaningful responses for each PRO were defined as changes from baseline greater than or equal to minimal clinically important difference (MCID) or greater than or equal to normative values. MCIDs were defined as ≥ 10-point decreases for PtGA [30], ≥ 10point decreases for pain VAS [30], ≥ 0.22-point decreases for HAQ-DI [30], ≥ 2.5-point increases in SF-36 PCS and MCS [36], ≥ 5-point increases in SF-36 domains [36], ≥ 8.4-point decreases for ISI (moderate improvement) [32], and ≥ 0.05-cm increases for EQ-5D-5L [35]. Owing to the lack of a predefined MCID for AM stiffness in the literature, the minimum important difference (MID) was defined as a reduction of ≥ 1 point for severity and one half standard deviation of the mean baseline values for duration of AM stiffness [8,30,31]. Normative values for PROs were obtained from the literature and defined as the following: ≤ 20 for PtGA [37], 0.25 for HAQ-DI [38], 50 for SF-36 PCS and MCS [39], 0-7 for ISI [33], and 0.915 for EQ-5D-5L [40]. Normative values are not available for pain VAS or AM stiffness [8,30,31,41]. The eight SF-36 domains were compared with age-and gender-matched (A/G) normative US population values as a benchmark [42].

Statistical analyses
All patients included in the intention-to-treat population of the BEYOND RCT were eligible for this post hoc analysis. Mean changes were calculated as changes in least squares mean (LSM) from baseline to week 12 for upadacitinib versus placebo based on a mixed effect for repeated measures model. Responses for each PRO were estimated at weeks 1, 4, and 12. Non-responder imputation (NRI) was used to impute missing responses. Between-group differences in responses were assessed using chi-square tests. Time to response was assessed by the Kaplan-Meier analysis and was compared using logrank test. "Spydergrams" were used to compare SF-36 domain responses [42]. In this study, NNTs were measured as the number of patients needed to achieve one additional responder, defined as the reciprocal of the response rate difference between treatment and placebo groups, and were calculated for each PRO at week 12.

Results
Patient disposition and demographic information have been published [24]. Disease characteristics and baseline PRO values across treatment groups were well-balanced (Tables 1 and 2). Decrements in PROs from normative values at baseline indicate that patients in this RCT reported substantial impairments in HRQOL. SF-36 mean baseline domain scores were approximately 25-50 points lower than in the A/G normative US population ( Fig. 1).
Compared with placebo, statistically significant improvements at week 12 were evident with both upadacitinib 15 mg and 30 mg for PtGA, pain VAS, HAQ-DI, PCS, and AM stiffness (all P ≤ 0.001, Table 2). Duration of AM stiffness was reduced from baseline by 43% and 58% in the upadacitinib 15 mg and 30 mg groups, respectively, versus 11% in the placebo group, and 72% and 80% of patients receiving upadacitinib 15 mg and 30 mg, respectively, versus 52% of patients receiving placebo reported a reduction in severity greater than or equal to MID. MCS baseline values were close to normative values, and although changes from baseline were numerically greater with upadacitinib 15 mg and 30 mg (4.54 and 3.37, respectively) compared with placebo (3.01), they were not statistically significant (P = 0.52). Changes from baseline in SF-36 domain scores with upadacitinib 15 mg exceeded placebo across all eight domains and were statistically significant except in the SF-36 mental health domain. For upadacitinib 30 mg, changes from baseline were statistically significant across all domains except mental health and role-emotional. All mean improvements for upadacitinib were clinically meaningful (Fig. 2). NNTs for upadacitinib 15 mg versus placebo ranged from 3 to 4 for PtGA, pain VAS, and HAQ-DI; 4 to 5 for PCS and AM stiffness severity; and 5 to 7 for seven of eight SF-36 domains. Similar results were reported with upadacitinib 30 mg.
In both upadacitinib groups, median time to response was 2 weeks for pain VAS and HAQ-DI compared with 4 weeks for these scores in the placebo group. Time to response for severity of AM stiffness was 1 week for both upadacitinib cohorts versus 2 weeks for the placebo cohort. At week 12, a significantly greater percentage of patients in the upadacitinib 15 mg group versus the placebo group reported scores greater than or equal to normative values in PtGA (28% vs 15%); HAQ-DI (16% vs 7%); the role-physical (15% vs 7%), bodily pain (24% vs  Fig. 3).

Discussion
Although there are a number of first-line therapies that improve RA outcomes, these same therapies may not be efficacious as second-or third-line therapy [19]. Additionally, not only is it essential to demonstrate that new treatments reduce the signs and symptoms of active RA in patients with refractory disease, but it is also necessary to show that these new treatments improve HRQOL outcomes from the patient's perspective. We analyzed data collected on several measures during SELECT-BEYOND to gain insight into the benefits of upadacitinib on HRQOL in patients with treatmentrefractory RA. Upadacitinib treatment resulted in rapid and clinically meaningful improvements in outcomes of   importance to patients with refractory disease: disease activity, pain, physical function, and AM stiffness, even in a difficult to treat population. Upadacitinib 15 mg and 30 mg had an equally effective impact on most of the PROs in this bDMARD-IR RA population. Upadacitinibtreated patients reported a substantial reduction in duration of AM stiffness and a significant decrease in severity as early as week 1. Reduced sleep quality is common among patients with RA and is associated with disease activity, pain, and functional disability [43,44]. Consistent with this finding, upadacitinib-treated patients with improvements in PtGA, pain VAS, and HAQ-DI scores also reported significant improvement in ISI scores. These clinically relevant improvements in PROs are consistent with the positive efficacy findings reported in this csDMARD-IR population [25,45] and suggest that upadacitinib may be an important treatment option in patients with active and refractory RA. Although a number of studies have shown that patients with RA who fail to respond to one anti-TNF agent may benefit from treatment with a second or third anti-TNF agent [46][47][48][49][50][51], the response is likely to decline as the number of anti-TNF agents increases [52]. Less information is available for the more recently approved JAK inhibitors and interleukin-6 receptor antagonist. In ORAL-STEP [53], a phase 3 RCT of tofacitinib in bDMARD-IR patients, the percentage of tofacitinibtreated (5 mg) patients reporting improvements ≥ MCID was less than in upadacitinib-treated (15 mg) patients in PtGA (65% vs 73%) and pain (69% vs 74%), similar in HAQ-DI (61% vs 63%) and SF-36 MCS (54% each), and greater in SF-36 PCS (68% vs 60%). Different results were observed in RA-BEACON [54], a phase 3 RCT of baricitinib in bDMARD-IR patients, where fewer baricitinib-treated (2 mg) patients reported improvements ≥ MCID than those treated with upadacitinib (15 mg) in SF-36 PCS (49% vs 60%) and MCS (33% vs 54%). In TARGET [55], a phase 3 RCT of sarilumab in bDMARD-IR patients, fewer sarilumab-treated (200 mg) patients reported improvements ≥ MCID than those treated with upadacitinib in PtGA (25% vs 73%), pain (26% vs 74%), HAQ-DI (21% vs 63%), SF-36 PCS (21%    [53][54][55].
There are several strengths in these analyses. Data were collected during a phase 3 RCT, which ensures patients are closely followed for an extended period and PROs are consistently measured. The randomized and blinded study design mitigates biases that may arise due to differences between treatment groups without knowledge of treatment allocation. The blinded study design also allows for unbiased reporting from a patient's perspective. The validated PROs used in these analyses evaluate different aspects of the patient's experience which may shed light on how patients perceive the effects of upadacitinib on a wide range of typical RA-related impediments. Use of MCID and normative values to measure responses makes these data clinically meaningful and interpretable for patients.
There are also limitations in these results. PROs were collected at fixed visits, and therefore, response could not be assessed at time intervals between visits. Because the duration of treatment was relatively short (12 weeks), additional studies are needed to determine if the improvements observed with upadacitinib treatment are maintained longterm. Furthermore, results may not be generalizable to all patients with RA, as clinical trial participants are selected based on specific inclusion/exclusion criteria and may differ from the broader RA patient population in clinical practice. The patients were bDMARD treatment failures, and results could be different in other scenarios. Lastly, the method used to impute missing data (NRI) assumes that missing PRO scores are associated with non-response, which may underestimate the true rate of response.

Conclusion
In conclusion, over 12 weeks, results from the SELECT-BEYOND trial demonstrated that in difficult-to-treat bDMARD-IR patients with active RA, treatment with upadacitinib compared with placebo resulted in significantly more patients with clinically meaningful improvements in PROs and responses that approached normative values. Furthermore, the NNTs to achieve clinically meaningful responses were ≤ 10, which are generally considered favorable [56]. Upadacitinib may be a treatment option for bDMARD-IR patients with RA providing clinically significant relief from symptoms that impair HRQOL.