CADTH Reimbursement Review

Upadacitinib (Rinvoq)

Sponsor: AbbVie Corporation

Therapeutic area: Psoriatic arthritis

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Clinical Review

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Introduction

Psoriatic arthritis (PsA) is an inflammatory musculoskeletal disease with a heterogenous presentation and disease course. While it is associated with psoriasis, PsA also presents with variable clinical features involving multiple domains, including peripheral arthritis, enthesitis, dactylitis, and axial disease.^1,2 Patients with PsA also present with psoriatic skin lesions and are usually seronegative for rheumatoid factor (95%).^2,3 Pain and stiffness of the affected joints are the most predominant presenting symptoms, with fatigue also occurring in many patients.¹ The prevalence of PsA varies, depending on the case definition and geography, and is estimated to be 1 to 2 per 1,000 in the general population.¹ A population-based Canadian study estimated that the age- and sex-standardized cumulative prevalence of PsA in Ontario ranged from 0.09% in 2008 to 0.15% in 2015. The same study estimated the age- and sex-standardized incidence in 2015 was 14 per 100,000.⁴ These figures may vary; for example, in another reference, the estimated annual incidence of PsA was reported to be 6 per 100,000 per year.¹ Over time, PsA can lead to deformities and joint damage.² This can lead to significant functional impairment, which in turn can affect work productivity and reduce health-related quality of life (HRQoL).^2,3

Several drug classes are employed in the pharmacologic treatment of PsA, including nonsteroidal anti-inflammatory drugs (NSAIDs); conventional synthetic disease-modifying antirheumatic drugs (csDMARDs) such as methotrexate, sulfasalazine, and leflunomide; biologic disease-modifying antirheumatic drugs (bDMARDs) such as tumour necrosis factor (TNF) inhibitors; interleukin (IL)-23, IL-12/23, and IL-17 inhibitors; and targeted synthetic disease-modifying antirheumatic drugs (tsDMARDs) that inhibit phosphodiesterase type 4 (PDE4) such as apremilast or Janus kinases (JAKs) (tofacitinib).⁵ In the treatment of PsA, csDMARDs, typically methotrexate, are recommended as first-line therapy or after a short course of NSAIDs in patients with polyarthritis. Some guidelines also recommend first-line treatment with a TNF inhibitor, particularly in patients with severe PsA or psoriasis.^6,7 In patients with an inadequate response to at least 1 csDMARD, a bDMARD may be started. In the case of biologic drug treatment failure, due to either lack of efficacy or adverse events (AEs), treatment guidelines recommend switching to an alternative biologic drug within a drug class, or to a drug with a different mode of action.^5-7 Specific treatment recommendations are also available for other scenarios, such as for patients with unequivocal enthesitis or predominantly axial disease. Treatment choice is individualized based on numerous factors, including severity and manifestations of disease, contraindications, concomitant conditions (e.g., active inflammatory bowel disease), and patient preference (e.g., route of administration or dosing frequency).⁷

Upadacitinib is an oral JAK inhibitor that is selective for JAK1. By inhibiting JAKs, upadacitinib modulates intracellular signalling pathways of cytokines and growth factors involved in a broad range of cellular processes, such as inflammatory responses, hematopoiesis, and immune surveillance.⁸ It is approved for the treatment of adults with active PsA who have had an inadequate response or intolerance to methotrexate or other disease-modifying antirheumatic drugs (DMARDs). The recommended dosage of upadacitinib for treatment of PsA is 15 mg orally once a day, used as monotherapy or in combination with methotrexate.⁸

The objective of this CADTH Drug Reimbursement Review, as established before the granting of the Notice of Compliance, is to perform a systematic review of the beneficial and harmful effects of upadacitinib 15 mg oral extended-release tablets as monotherapy or in combination with non-bDMARDs for the treatment of active PsA in adult patients who have responded inadequately or who are intolerant to 1 or more DMARDs.

Stakeholder Perspectives

The information in this section is a summary of input provided by the patient groups who responded to CADTH’s call for patient input and from clinical expert(s) consulted by CADTH for the purpose of this review.

Patient Input

Four inputs were submitted for this review from 6 different patient groups: Arthritis Consumer Experts (ACE), the Canadian Spondylitis Association (CSA), the Canadian Association of Psoriasis Patients (CAPP) in partnership with the Canadian Psoriasis Network (CPN), and the Canadian Arthritis Patient Alliance (CAPA) in partnership with the Arthritis Society. Five of these patient organizations (the Arthritis Society, CSA, CAPP, CPN, and CAPA) collaboratively developed a survey that was shared with their respective memberships or patient communities. Overall, the survey drew 94 responses. The CSA also conducted a telephone interview with a patient on upadacitinib. Arthritis Consumer Experts gathered its own data from 5 patients who completed a patient input survey.

Respondents to the surveys emphasized pain, stiffness, lack of mobility, and fatigue, all of which affect activities associated with daily living and family life, and their ability to work and maintain certain hobbies. Other impacts included embarrassment and self-consciousness from symptoms caused by PsA. Respondents reported difficulties contributing and participating at school or work due to their symptoms. The impact of PsA extends to others within a person’s support circle, including caregivers such as spouses, partners, or children who may have to take on additional roles or tasks to support the person living with PsA.

The patient input noted that those living with psoriatic disease often try a succession of treatments throughout their lives. Patients’ responses to medication can vary significantly, and treatments that are initially effective can become less effective over time. As a result, patients need several treatment options to effectively manage their disease. Outcomes that were identified as important to patients with PsA include the route of drug administration (oral versus infusion versus self-injection), a reduction in pain and fatigue, treatments that are effective for psoriasis as well as PsA, increased mobility, ability to work and be productive at work, ability to carry out activities of daily living, ability to effectively carry out parenting tasks and other important social roles, reduced infection rates, affordability of the medication, and improved HRQoL. According to the input from the CSA, new treatment options and different classes of medications fill a void in the unmet needs of patients and prescribers; the oral formulation of upadacitinib provides another option for administration and may help lead to improved adherence and ultimately better outcomes.

Clinician Input

Input from Clinical Experts Consulted by CADTH

The clinical expert consulted by CADTH for this review identified an unmet need in the treatment of psoriatic disease as some patients may not respond to any treatment, and only a minority achieve minimal disease activity (MDA). In the treatment of PsA, numerous domains of disease activity that might not be accomplished by a single drug need to be addressed. In patients who do not respond or become refractory to treatment, a switch to treatment with a different mechanism of action will be necessary.

The clinical expert indicated that any patient with peripheral joint and skin disease that does not respond to csDMARDs would be eligible for upadacitinib, barring contraindications. Tumour necrosis factor inhibitors and IL-17 inhibitors will generally be prescribed before upadacitinib. However, the clinical expert noted that upadacitinib may become a first-line treatment for PsA as clinicians become more experienced with upadacitinib and long-term safety is confirmed. The caveat with this assumption is that longer-term observation in patients on upadacitinib will be needed to confirm the durability of benefit and safety. The clinical expert also identified the oral route of upadacitinib as an advantage, with the improved convenience over subcutaneous injections or IV infusions expected to enhance treatment adherence. It is also expected that a benefit from a JAK inhibitor will become apparent sooner than from TNF inhibitors, and a lack of response and/or side effects will result in the discontinuation of treatment.

According to the clinical expert, the swollen joint count (SJC) is the most likely measure used in clinical practice to assess response, with a reduction in joint count reflecting a meaningful response. Other clinically meaningful responses may be measured using achievement of MDA or patient-reported outcomes.

Clinician Group Input

No clinician group input was received for this reimbursement review.

Drug Program Input

Input was obtained from the jurisdictions participating in CADTH reimbursement reviews. The following were identified as key factors that could impact the implementation:

• The place in therapy of upadacitinib relative to currently available treatments for PsA.

• The significance of potential AEs associated with JAK inhibitors.

• The expected dose of upadacitinib used and any potential for dose escalation.

The clinical expert consulted by CADTH provided responses that can be found in the Drug Program Input section.

Clinical Evidence

Pivotal Studies and Protocol-Selected Studies

Description of Studies

Two multi-centre, phase III, randomized, double-blind, placebo-controlled trials, SELECT-PsA1 and SELECT-PsA2, met the inclusion criteria for this review. Both SELECT studies enrolled adult patients with an established diagnosis of moderately to severely active PsA who had been previously treated with a DMARD. SELECT-PsA1 was conducted in patients who had an insufficient response or were intolerant to a non-bDMARD, whereas PsA2 included patients who had an insufficient response or were intolerant to a bDMARD. Both trials investigated 2 dosages of oral upadacitinib: 15 mg once daily and 30 mg once daily; however, to align with the Health Canada–recommended dosage, only results for upadacitinib 15 mg once daily are presented in this review.

Efficacy and safety of upadacitinib were compared with placebo in both studies; SELECT-PsA1 also included adalimumab as an active comparator. Both studies consisted of 2 periods, and at the end of week 24 in period 1, all patients on placebo were switched to upadacitinib. In SELECT-PsA1, period 1 was 56 weeks in duration and included a 24-week double-blind placebo and active comparator–controlled period followed by 32 weeks of blinded active comparator–controlled treatment. SELECT-PsA2 also consisted of period 1, which was 56 weeks in duration and included 24 weeks of a double-blinded placebo-controlled phase followed by a 32-week non-comparative treatment phase. Period 2 is an ongoing, open-label, long-term treatment extension of up to approximately 5 years for PsA1 and 3 years for PsA2.

In SELECT-PsA1 (N = 1,705), eligible participants were randomized at a 2:2:2:1:1 ratio to 1 of 5 treatment groups: upadacitinib 15 mg once daily, upadacitinib 30 mg once daily, adalimumab 40 mg subcutaneous every other week, and placebo followed by upadacitinib 15 mg once daily or placebo followed by upadacitinib 30 mg once daily. Randomization was stratified by extent of psoriasis (≥ 3% body surface area [BSA] or < 3% BSA), current use of at least 1 DMARD, presence of dactylitis, and presence of enthesitis. Patients enrolled in SELECT-PsA2 (N = 642) were randomized in a 2:2:1:1 ratio to 1 of 4 treatment groups similar to SELECT-PsA1, but without the adalimumab treatment group: upadacitinib 15 mg once daily, upadacitinib 30 mg once daily, and placebo followed by either upadacitinib 15 mg once daily or upadacitinib 30 mg once daily. Randomization was stratified by extent of psoriasis, current use of at least 1 DMARD, and number of prior failed bDMARDs (1 versus > 1). Patients were permitted to continue their stable background non-bDMARD therapy. Both studies had an appropriate randomization strategy, and treatment groups within each study were generally well balanced. Compared to patients in SELECT-PsA1, those in SELECT-PsA-2 had PsA for longer with more significant disease.

The primary end point for both SELECT-PsA1 and SELECT-PsA2 was the proportion of patients who achieved at least a 20% improvement in American College of Rheumatology response criteria (ACR20) at week 12. The primary and major secondary efficacy outcomes were assessed using a hierarchical testing procedure to control the overall type I error rate. The multiplicity-adjusted testing hierarchy included the primary end point plus 14 ranked key secondary end points in SELECT-PsA1, and 7 ranked key secondary end points in SELECT-PsA2. Several additional end points that were not part of the multiplicity-adjusted analyses but were identified in the CADTH systematic review protocol are also discussed in this report.

Efficacy Results

Efficacy results are summarized in Table 2. Results of the primary outcome, key secondary efficacy outcomes, outcomes identified in the review protocol (Table 5), and those considered important by patient groups are reported. Results of efficacy outcomes that were not included in the multiplicity-controlled analyses are described; however, they are considered inconclusive because of the potential for inflated type I error.

In the SELECT-PsA1 trial, upadacitinib 15 mg did not demonstrate superiority over adalimumab in ACR20 response at week 12. As such, statistically significant differences in the key secondary end points lower in the testing hierarchy were not tested because of the multiplicity control strategy.

Clinical Responses in Psoriatic Arthritis Symptoms

Clinical response in PsA symptoms or overall disease activity were measured using ACR20, MDA, and modified Psoriatic Arthritis Response Criteria (PsARC). In SELECT-PsA1, 70.6% and 36.2% of patients treated with upadacitinib 15 mg and placebo, respectively, achieved an ACR20 response, and the difference between the upadacitinib 15 mg group and placebo treatment group was 34.5% (95% confidence interval [CI], 28.2 to 40.7; P < 0.0001), which was clinically relevant and statistically significant in favour of upadacitinib 15 mg. In SELECT-PsA2, 56.9% and 24.1% of patients treated with upadacitinib 15 mg and placebo, respectively, achieved an ACR20 response; the difference between the upadacitinib 15 mg group and the placebo treatment group was 32.8% (95% CI, 24.0 to 41.6; P < 0.0001), which was clinically relevant and statistically significant in favour of upadacitinib 15 mg. Results of the pre-specified subgroup analyses by current use of non-bDMARD, number of prior non-bDMARD (SELECT-PsA1) and number of prior failed bDMARDs (SELECT-PsA2) were consistent with results from the overall population for the primary end point of an ACR20 at week 12; however, these analyses were not included in the hierarchical statistical analysis and should be interpreted with caution because of the potential for inflated type I error. The clinical expert consulted for this review noted that the differences in ACR20 responses compared with placebo were clinically meaningful.

In SELECT-PsA1, the proportion of patients achieving an ACR20 at week 12 with upadacitinib treatment compared to adalimumab was tested for noninferiority and superiority as key secondary end points. An ACR20 response was achieved by 70.6% of the upadacitinib 15 mg group and by 65.0% of patients in the adalimumab group. The difference between the upadacitinib 15 mg group and the adalimumab treatment group was 5.6% (95% CI, −0.6 to 11.8). The adalimumab effect preservation, calculated by (upadacitinib – placebo)/(adalimumab – placebo), was 119.4% (95% CI, 98.0 to 147.9); the lower bound of the 95% CI exceeded the pre-specified noninferiority ratio of at least 50% of the placebo-subtracted adalimumab effect, indicating that upadacitinib 15 mg daily was noninferior to adalimumab 40 mg every other week. In the subsequent testing of superiority, upadacitinib 15 mg was not found to be superior compared to adalimumab 40 mg, as it did not meet the statistical significance for superiority.

For clinical responses measured with the MDA criteria, patients treated with upadacitinib 15 mg had higher response rates compared to placebo at week 24 in both SELECT-PsA1 (36.6% for upadacitinib 15 mg and 12.3% for placebo) and SELECT-PsA2 (25.1% for upadacitinib 15 mg and 2.8% for placebo). The between-group differences were 24.3% (95% CI, 18.8 to 29.8; P = 0.0004) in the SELECT-PsA1 trial and 22.3% (95% CI, 16.0 to 28.6; P < 0.0001) in SELECT-PsA2. In both trials, the between-group differences were statistically significant in favour of upadacitinib 15 mg.

For modified PsARC response at week 24, a higher proportion of patients treated with upadacitinib achieved a response compared to patients randomized to adalimumab or placebo in both studies (SELECT-PsA1: 83.7% for upadacitinib 15 mg, 76.6% for adalimumab 40 mg, and 59.3% for placebo; SELECT-PsA2: 68.2% for upadacitinib 15 mg and 36.3% for placebo). In SELECT-PsA1, the response-rate difference between the upadacitinib and adalimumab groups was 7.0% (95% CI, 1.7 to 12.3), whereas the difference between upadacitinib and placebo was 24.3% (95% CI, 18.5 to 30.2). In SELECT-PsA2 the difference between upadacitinib and placebo was 31.9 (95% CI, 22.9 to 40.9). These analyses were not included in the hierarchical statistical analysis.

Measurement of Function and Disability

The improvement in physical function at week 12 as measured by the Health Assessment Questionnaire–Disability Index (HAQ-DI) was statistically significant. The change in scores from baseline in upadacitinib 15 mg and placebo were −0.42 and −0.14, respectively, in SELECT-PsA1, and −0.30 and −0.10, respectively, in SELECT-PsA2. The differences in change from baseline between upadacitinib 15 mg and placebo were −0.28 (95% CI, −0.35 to −0.22; P < 0.0001) in SELECT-PsA1 and −0.21 (95% CI, −0.30 to −0.12; P < 0.0001) in SELECT-PsA2. While in both studies, the between-group differences in the improvement of the HAQ-DI scores did not exceed the estimated minimal important difference (MID) of 0.35 found in the literature, the proportions of patients who achieved a clinically meaningful improvement in HAQ-DI at week 12 in the SELECT-PsA1 study were 33.4%, 47.2%, and 57.9% in the placebo, adalimumab 40 mg, and upadacitinib 15 mg treatment groups, respectively, while the proportions of patients who achieved a clinically meaningful improvement in HAQ-DI at week 12 in SELECT-PsA2 were 27.2% and 44.6% in the placebo and upadacitinib 15 mg treatment groups, respectively.

Work productivity was measured by the Work Productivity and Activity Impairment (WPAI) questionnaire in a portion of study participants in both studies. Numerically greater reductions in work or activity impairment due to disease were observed for the upadacitinib 15 mg group compared to placebo at week 24. Although it appears the suggested MID was achieved by SELECT-PsA1 patients in the upadacitinib group for improvement in presenteeism (≥ 20%) and activity impairment (≥ 20%), the between-group differences in change from baseline compared to placebo or adalimumab did not exceed this threshold. The least squares (LS) mean difference in the change in scores between upadacitinib and adalimumab was −2.5 (95% CI, −6.2 to 1.2), whereas the LS mean differences between upadacitinib and placebo were −13.4 (95% CI, −17.1 to −9.7) in SELECT-PsA1 and −12.2 (95% CI, −18.8 to −5.6) in SELECT-PsA2. With a smaller number of patients included in the analysis, and the lack of a confirmed MID for the WPAI instrument, it remains unclear whether the differences were clinically meaningful. This was identified as an important outcome by the patient groups, but as it was an exploratory variable in both SELECT-PsA1 and SELECT-PsA2 it was not included in the multiplicity-controlled analyses.

Measurement of PsA Symptoms

Symptoms of PsA such as fatigue and pain were reported in both studies. A statistically greater improvement in fatigue from baseline, measured using the Functional Assessment of Chronic Illness Therapy–Fatigue (FACIT-F), was seen at week 12 with upadacitinib 15 mg compared to placebo in both studies. The mean changes from baseline were 6.3 for upadacitinib 15 mg and 2.8 for placebo in SELECT-PsA1 (between-group difference of 3.5; 95% CI, 2.4 to 4.7; P = 0.0004) and 5.0 for upadacitinib 15 mg and 1.3 for placebo in SELECT-PsA2 (between-group difference of 3.7; 95% CI, 2.0 to 5.4; P < 0.0001). The between-group difference in the improvement in FACIT-F score at week 12 exceeded the estimated MID (3.1 points) in both studies. The impact of upadacitinib on pain is uncertain as this end point was not part of the hierarchical analysis, and no MID was identified for the patient’s assessment of pain numeric rating scale (NRS) in patients with PsA.

Health-Related Quality of Life

Health-related quality of life was measured by the Short Form (36) Health Survey (SF-36) and EuroQol 5-Dimensions 5-Levels questionnaire (EQ-5D-5L) in SELECT-PsA1 and PsA2. Only the physical component summary (PCS) of the SF-36 was part of the multiplicity-adjusted testing hierarchy in the SELECT studies, and the differences between the 2 groups were statistically significant. In SELECT-PsA1, the difference in mean change from baseline for upadacitinib 15 mg versus placebo was 4.67 (95% CI, 3.67 to 5.67; P = 0.0004) in favour of upadacitinib 15 mg; in SELECT-PsA2 the difference in mean change from baseline for upadacitinib 15 mg versus placebo was 3.52 (95% CI, 2.07 to 4.98; P < 0.0001) in favour of upadacitinib 15 mg. For the mental component summary (MCS), a numerically greater improvement from baseline was seen for upadacitinib compared to placebo in both trials; the difference in mean change from baseline between the 2 treatment groups was 1.70 (95% CI, 0.58 to 2.82) in SELECT-PsA1 and 2.98 (95% CI, 1.44 to 4.52) in SELECT-PsA2. The results from the EQ-5D-5L suggest that there were greater improvements in the utility index and the Visual Analogue Scale (VAS) scores from baseline to week 24 in the upadacitinib treatment group compared to patients randomized to placebo in both studies, as well as adalimumab in PsA1. For the utility index, the difference in mean change from baseline between upadacitinib and adalimumab was 0.03 (95% CI, 0.00 to 0.05), whereas the difference in mean change from baseline between upadacitinib and placebo was 0.09 (95% CI, 0.06 to 0.11) in SELECT-PsA1 and 0.10 (95% CI, 0.06 to 0.14) in SELECT-PsA2. For the VAS, the difference in mean change from baseline between upadacitinib and adalimumab was 2.8 (95% CI, 0.0 to 5.6), whereas the difference in mean change from baseline between upadacitinib and placebo was 10.9 (95% CI, 8.0 to 13.7) in SELECT-PsA1 and 6.8 (95% CI, 2.5 to 11.1) in SELECT-PsA2. For the comparison of upadacitinib to placebo in both studies, the mean between-group differences in the EQ-5D-5L utility index reached the MID threshold identified in the literature for the general Canadian population (summarized mean of 0.056; standard deviation [SD] = 0.011). These results suggest that treatment with upadacitinib 15 mg was associated with improved HRQoL. Although patient groups identified HRQoL as an important outcome, the EQ-5D-5L and the MCS of the SF-36 were not part of the hierarchical analysis plan and were not adjusted for multiple comparisons; therefore, the results should be interpreted with caution due to the risk of inflated type I error.

Measurement of Skin Disease

Extent and severity of skin disease was measured in both studies using the PASI, sIGA, and SAPS. Only patients with psoriasis involving a 3% or greater BSA baseline had a PASI assessment. In SELECT-PsA1, the proportion of patients achieving a PASI response of 75 in the upadacitinib 15 mg treatment group was 62.6% compared to 21.3% in the placebo group, and the difference between the upadacitinib 15 mg and placebo groups was 41.3% (95% CI, 32.8 to 49.8; P < 0.0001), which was statistically significant in favour of upadacitinib 15 mg. In PsA2, the proportion of patients achieving a 75% reduction in PASI score (PASI 75) in the upadacitinib treatment group was 52.3% compared to 16.0% in the placebo group, and the difference between the upadacitinib 15 mg group and placebo group was 36.3% (95% CI, 25.6 to 46.9; P < 0.001), which was statistically significant in favour of upadacitinib 15 mg. The clinical expert consulted for this review indicated that the between-group differences in PASI 75 were considered clinically relevant, although the true effect should be derived from separate studies that are specifically designed for patients with skin disease.

Only patients with an sIGA score of 2 or greater at baseline, and an improvement of at least 2 points from baseline at week 16 were included in the assessment. In both SELECT-PsA1 and SELECT-PsA2, a statistically significant difference was seen in the proportion of patients achieving a response (an sIGA of psoriasis score of 0 or 1) in favour of upadacitinib. At week 16, the proportions of responders were 41.9% for upadacitinib 15 mg and 10.9% for placebo (between-group difference of 31.1%; 95% CI, 24.7 to 37.5; P < 0.0001) in SELECT-PsA1, and 36.8% for upadacitinib 15 mg and 9.2% for placebo (between-group difference of 27.6%; 95% CI, 19.2 to 36.1; P < 0.0001) in SELECT-PsA2.

A greater reduction in SAPS score from baseline was reported for patients in the upadacitinib group compared to placebo at week 16. In PsA1, the difference in LS mean change from baseline between upadacitinib and placebo was −17.1 (95% CI, −19.6 to −14.6) for upadacitinib 15 mg versus placebo. Testing for superiority of upadacitinib compared to placebo was part of the multiplicity-adjusted analyses in PsA1; however, because it was ranked after the point at which the hierarchical analysis failed and was stopped, no appropriate statistical comparisons can be made. In SELECT-PsA2, the difference between groups in the LS mean change from baseline in SAPS scores was statistically significant, favouring upadacitinib compared to placebo (−22.9; 95% CI, −27.4 to −18.4; P < 0.0001).

Measurement of Other Musculoskeletal Disease

Impact of treatment on musculoskeletal disease was assessed by measuring the resolution of enthesitis with the Leeds Enthesitis Index (LEI), resolution of dactylitis with the Leeds Dactylitis Index (LDI), and change in axial disease using the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI). For SELECT-PsA1 patients with enthesitis at baseline, resolution of enthesitis (LEI = 0) was achieved by a statistically significantly higher proportion of patients in the upadacitinib 15 mg treatment group (53.7%) compared to placebo (32.4%) at week 24 (between-group difference of 21.3%; 95% CI, 13.0 to 29.7; P = 0.0004). In SELECT-PsA2, a numerically higher proportion of patients in the upadacitinib 15 mg treatment group achieved resolution of enthesitis at week 24 compared to patients in the placebo group, with a difference of 27.6% (95% CI, 17.3 to 37.8); however, this end point was not part of the multiplicity-controlled analyses in PsA2. As there is a risk of inflated type I error, no appropriate statistical comparisons can be made. Resolution of dactylitis (LDI = 0) was achieved by a numerically higher proportion of patients in the upadacitinib group compared to the placebo group at week 24 in both trials. The differences between the 2 treatment groups were 36.8% (95% CI, 25.7 to 47.9) in SELECT-PsA1 and 30.1% (95% CI, 13.0 to 47.1) in SELECT-PsA2. In SELECT-PsA1, this end point was included in the hierarchical statistical analysis; however, it was ranked after the point at which the hierarchical analysis failed and was stopped. In SELECT-PsA2, this end point was not part of the multiplicity-controlled analyses. Results for this end point are therefore considered exploratory in both trials.

Change in axial disease was assessed in patients with the presence of psoriatic spondylitis at baseline. The improvement in BASDAI score from baseline to week 24 numerically favoured the upadacitinib 15 mg treatment group compared to the placebo group in both studies and compared to the adalimumab group in SELECT-PsA1. In SELECT-PsA1, the difference in the LS mean change in scores from baseline between upadacitinib and adalimumab was −0.57 (95% CI, −1.09 to −0.05), and between upadacitinib and placebo it was −1.42 (95% CI, −1.94 to −0.90). In PsA2, the difference between upadacitinib and placebo was −1.85 (95% CI, −2.55 to −1.15). However, this outcome assessment was not included in the hierarchical statistical analysis and should be considered inconclusive because of the potential for inflated type I error.

Radiographic Changes

Radiographic change was assessed only in SELECT-PsA1 using the Sharp/van der Heijde Score (SHS). At week 24, the differences in LS mean change from baseline in SHS was statistically significant, favouring the upadacitinib 15 mg treatment group over placebo (−0.29; 95% CI, −0.44 to −0.14; P = 0.0004). According to the clinical expert consulted for this review, these numerically small changes are unlikely to be clinically meaningful to patients over a period of only 24 weeks and it is difficult to extrapolate the significance of these changes over the lifetime of a patient with PsA. In particular, it is uncertain whether the radiographic changes seen in SELECT-PsA1 correlate with a direct and meaningful improvement in a patient’s physical function, quality of life, or permanent disability. However, the observations satisfy the regulatory requirement that upadacitinib can inhibit radiographic progression.

Harms Results

Safety data are summarized in Table 2. By week 24, the proportion of patients in SELECT-PsA1 who experienced a treatment-emergent adverse event (TEAE) was higher in the upadacitinib 15 mg and adalimumab treatment groups compared to the placebo group. In PsA2, the proportion of patients who experienced a TEAE was similar between the upadacitinib and placebo groups. Generally, the majority of AEs were mild or moderate in severity, and the most frequently reported AE in both studies was an upper respiratory tract infection. The frequency of serious adverse events (SAEs) and withdrawals due to adverse events (WDAEs) were low across all treatment groups and generally below 5%, with the exception of the upadacitinib 15 mg treatment group of SELECT-PsA2, which had the highest proportion of patients experiencing an SAE (5.7%) or WDAE (7.1%) across both studies. None of the specific SAEs were reported by more than 2 patients. Two treatment-emergent deaths were reported by week 24, both in the placebo group. One non–treatment-emergent death (i.e., occurring more than 30 days after the last dose) was reported in the upadacitinib 15 mg group.

Critical Appraisal

A few major limitations and sources of bias are described in this section. Further details for each point, as well as a complete list of limitations and sources of bias, are available in the Critical Appraisal subsection of the Clinical Evidence – Results section.

• Key end points comparing upadacitinib to adalimumab were measured at week 12. According to the clinical expert consulted for this review, this may not have provided enough time for adalimumab to show maximal benefit. The benefit of JAK inhibitors is thought to be seen generally sooner than that of TNF inhibitors. As such, end points measured at week 12 may be biased in favour of upadacitinib. While results of both upadacitinib and adalimumab were consistent until 24 weeks, it is uncertain whether upadacitinib 15 mg is noninferior to adalimumab due to the lack of statistical testing at week 24. Also, the noninferiority and superiority comparison between upadacitinib and adalimumab was conducted only for the ACR20 efficacy outcome, making it unclear whether upadacitinib would be noninferior to adalimumab for other outcome measures.

• Not all end points measured in this trial may be clinically meaningful to patients, despite showing statistically significant differences in the trials. For example, subjective measures such as fatigue or the small changes seen in SHS may not reflect clinically meaningful improvement, particularly when measured over such a short length of time relative to the long disease course. The clinical expert consulted for this review noted that it is difficult to extrapolate the significance of these changes over the lifetime of a patient with PsA. Also, several outcomes that were identified in the CADTH review protocol and reported in the studies fell outside the statistical testing hierarchy and should therefore be interpreted with consideration of type I error. Given the large number of comparisons in the study, a statistically significant finding may be attributable to inflated type I error. Furthermore, the results of the pre-specified subgroup analyses performed for the primary end point should be interpreted with caution due to the small sample sizes and lack of control for type I error, and also because the trial was not powered to test specific hypotheses in subgroups. As with the end points, which were not part of the statistical testing hierarchy, the results of these subgroup analyses should be interpreted with caution.

• SELECT-PsA1 required patients to have the presence of either at least 1 erosion on an X-ray or high-sensitivity C-reactive protein (hs-CRP) levels greater than the upper limit of normal (ULN) for inclusion in the study, and this may affect the generalizability of the study’s results. According to the clinical expert consulted for this review, a substantial proportion of patients seen in clinical practice generally do not have evident erosion or inflammatory markers elevated to this degree and yet still require treatment with bDMARDs.

• Although long-term data were reported for up to week 56 in both studies, placebo-controlled data for upadacitinib exist only up to week 24.

• Upadacitinib was compared to active treatment (adalimumab) only in patients with no prior exposure to bDMARD treatment. It is unknown if the same relative benefit can be expected from patients who have failed prior treatment with bDMARDs.

Indirect Comparisons

Description of Studies

Other than the inclusion of upadacitinib in SELECT-PsA1 and SELECT-PsA2, there are no studies in which upadacitinib is compared directly to other bDMARDs or tsDMARDs. Therefore, the sponsor conducted an indirect comparison that comprised a network meta-analysis (NMA) that compared the efficacy of upadacitinib to that of TNF inhibitors (adalimumab, certolizumab pegol, etanercept, golimumab, and infliximab), IL-17 inhibitors (secukinumab and ixekizumab), an IL-12/23 inhibitor (ustekinumab), an IL-23 inhibitor (guselkumab), cytotoxic T lymphocyte–associated antigen-4 immunoglobulin (abatacept), a JAK inhibitor (tofacitinib), and a PDE4 inhibitor (apremilast). Results from the indirect treatment comparison (ITC) are summarized in this section only for relevant comparators identified in the CADTH systematic review. Efficacy was compared at 12 and 24 weeks, and the ITC reported results for bDMARD-naive and bDMARD-experienced patients separately.

Efficacy Results

Overall, in biologic-naive patients, the NMA suggests that upadacitinib 15 mg is more efficacious for an ACR20 at week 12 compared to some comparators, specifically an IL-17 inhibitor (secukinumab 15 mg), IL-12/23 inhibitor (ustekinumab 45 mg), and IL-23 inhibitor (guselkumab), but this advantage was only seen for the IL-12/23 inhibitor at 24 weeks. Upadacitinib was also shown to be more efficacious than etanercept at both weeks 12 and 24 for a PASI response; however, this was not seen with other TNF inhibitors. The IL-17 inhibitors (secukinumab 300 mg and ixekizumab) and IL-23 inhibitor (guselkumab) appear to be more efficacious than upadacitinib for PASI responses at week 12, although only the IL-23 inhibitor was favoured over upadacitinib at week 24. For PsARC, upadacitinib was more efficacious than tofacitinib but only at week 12; this was not seen at week 24. For HAQ-DI measured in PsARC responders at week 12, etanercept was more efficacious than upadacitinib; this benefit was not seen with other TNF inhibitors. At week 24, adalimumab appeared to be more efficacious than upadacitinib. The number of comparators included in some analyses (i.e., HAQ-DI at 24 weeks) was limited. Because no difference was seen between upadacitinib and the relevant comparators in other analyses, no consistent benefit of upadacitinib over bDMARDs or tsDMARDs was demonstrated across all measured end points (weeks 12 and 24).

In biologic-experienced patients, upadacitinib 15 mg was favoured only when compared to tofacitinib (a JAK inhibitor) in PASI response at week 12; this comparison was not performed at week 24. No difference in treatment effect was demonstrated in all other comparisons between upadacitinib and the included IL inhibitors. Not all IL inhibitors were included in every analyses, and the IL-23 inhibitor was absent from many comparisons. Furthermore, TNF inhibitors were not included in any of the NMA analyses as there were insufficient eligible data in the bDMARD-experienced patient population, and therefore no conclusions can be drawn on the comparative efficacy of upadacitinib in these patients. Also, because JAK inhibitors were not included in any of the week 24 analyses, the long-term comparative efficacy of upadacitinib compared to tofacitinib is unknown.

Harms Results

The sponsor’s submitted ITC did not report safety outcomes.

Critical Appraisal

Several limitations increase the uncertainty of the results of the ITC discussed in this review. The included studies were highly heterogeneous in terms of inclusion criteria and patient characteristics. Significant differences were noted in potential effect modifiers, such as duration of disease, use of prior DMARDs, and disease severity. Although these factors are heightened due to the variation in inclusion and exclusion criteria across included studies, no sensitivity analysis or subgroup analysis was conducted to assess the impact of these potential effect modifiers on the comparison of upadacitinib and other biologics. The ITC also did not include any analyses of other clinically meaningful outcomes, such as PsA symptoms (e.g., pain and fatigue), HRQoL, or safety.

Overall, there is uncertainty due to the inherent heterogeneity across trials in the networks. The robustness of the comparative efficacy was further compromised by the lack of precision in the findings, and results from the sponsor-submitted ITC must therefore be interpreted with caution.

Other Relevant Evidence

Description of Studies

Each of the SELECT studies included 2 study periods. At the time of this review, data up to the end of period 1 (week 56) was available. Period 1 for SELECT-PsA1 included 24 weeks of randomized, double-blind, placebo- and active comparator–controlled treatment followed by an additional 32 weeks of blinded active comparator–controlled treatment. Period 1 for SELECT-PsA2 included 24 weeks of randomized, double-blind, placebo-controlled treatment followed by an additional 32 weeks of upadacitinib treatment. In both studies, all patients assigned to placebo were switched to pre-assigned upadacitinib 15 mg or 30 mg daily in a 1:1 ratio at week 24.¹¹ Data reported at week 56 used the as-observed dataset, and no adjustments for multiple testing were employed.

Efficacy Results

In both studies, results from the end of period 1 (week 56 data) suggest that the improvements in clinical and patient-reported outcomes observed at week 24 in patients who received upadacitinib 15 mg once daily starting at day 1 were maintained throughout the 56-week blinded treatment period. Patients who switched from placebo to upadacitinib 15 mg once daily at week 24 also showed improvements in clinical and patient-reported outcomes at week 56; the trajectory for achievement of response or improvement in end points after starting upadacitinib was similar to those observed in patients who started upadacitinib on day 1 of both studies. Numerically greater improvement with upadacitinib compared to adalimumab was also demonstrated for several end points in SELECT-PsA1. For example, the difference in the ACR20 response rate between the upadacitinib 15 mg treatment group (including those switched to upadacitinib 15 mg from placebo), and adalimumab was 6.3% (95% CI, 0.3 to 12.2), and the difference in the proportion of patients achieving MDA was 7.6% (95% CI, 0.4 to 14.8).

Harms Results

The safety profile of oral upadacitinib 15 mg once daily over 56 weeks was consistent with that observed during the 24-week double-blind period in both SELECT-PsA1 and PsA2, with no unexpected safety signals reported. Harms for the analysis at week 56 were presented as exposure-adjusted event rates and were also pooled such that data reported for the upadacitinib exposure combined the upadacitinib 15 mg and placebo-switched-to-upadacitinib 15 mg groups. In SELECT-PsA1, 1 or more AEs were reported at an exposure-adjusted incidence of 265.9 events per 100 patient-years (PYs) in the adalimumab group and 281.1 events per 100 PY in the upadacitinib group. In SELECT-PsA2, 1 or more AEs were reported at a rate of 260.6 events per 100 PY (pooled upadacitinib group). With longer exposure to treatment, a greater proportion of patients treated with upadacitinib compared to adalimumab experienced infectious AEs, including the following, which are presented as events per 100 PY: urinary tract infections (3.6 for adalimumab and 6.7 for upadacitinib in SELECT-PsA1, and 9.8 for upadacitinib in SELECT-PsA2), bronchitis (2.9 for adalimumab and 5.7 for upadacitinib in SELECT-PsA1, and 8.8 for upadacitinib in SELECT-PsA2), hypertension (2.7 for adalimumab and 5.6 for upadacitinib in SELECT-PsA1, and 5.7 for upadacitinib in SELECT-PsA2), and influenza (0.8 for adalimumab and 3.2 for upadacitinib in SELECT-PsA1, and 5.2 for upadacitinib in SELECT-PsA2). Herpes zoster was also reported in a higher proportion of patients treated with upadacitinib across both studies (3.9 per 100 PY and 3.8 per 100 PY in SELECT-PsA1 and SELECT-PsA2, respectively), compared to those treated with adalimumab (0.5 per 100 PY). Other notable AEs that showed an imbalance in groups include elevated creatine phosphokinase (CPK) levels and hepatic disorder, which were reported at a higher incidence by both upadacitinib and adalimumab treatment groups in SELECT-PsA1 compared to SELECT-PsA2. Elevated CPK levels were reported at an incidence rate of 7.3 per 100 PY for adalimumab, 11.9 for upadacitinib in SELECT-PsA1, and 5.2 for upadacitinib in SELECT-PsA2. Hepatic disorder was reported in a higher proportion of patients treated with adalimumab (24.9 per 100 PY for adalimumab and 19.1 per 100 PY for upadacitinib in SELECT-PsA1, and 4.8 per 100 PY for upadacitinib in SELECT-PsA2), although this may be confounded overall by the higher usage of concomitant methotrexate treatment in SELECT-PsA1 patients. Withdrawal of treatment due to AEs was reported at an incidence of 7.4 per 100 PY for adalimumab, and 4.6 per 100 PY and 10.0 per 100 PY for upadacitinib in SELECT-PsA1 and SELECT-PsA2, respectively. In total, 5 deaths occurred in the relevant treatment groups by the end of period 1, inclusive of those counted under week 24 data. These include both treatment-emergent deaths (occurring within 30 days of last dose for upadacitinib or 70 days for adalimumab) and non–treatment-emergent deaths. One treatment-emergent death occurred in the adalimumab group, and 2 non-treatment-emergent deaths occurred in the upadacitinib 15 mg treatment group. The remaining 2 deaths occurred in the placebo groups.

Critical Appraisal

Interpretation of the long-term efficacy and safety outcomes at week 56 is limited by the lack of a placebo control in both SELECT-PsA1 and SELECT-PsA2, as well as the lack of a comparator in SELECT-PsA2. Also, because background therapies were allowed to be modified, it is difficult to disentangle the drug effect from the changes in the background therapies on the reported outcomes. Furthermore, given that all patients were aware that they were receiving an active treatment (upadacitinib or adalimumab), results for patient-reported outcomes may be subject to bias. Because no adjustment was made for multiplicity to evaluate long-term data, and given the large number of analyses performed, there is a risk of inflated type I error. As such, the week 56 data should be interpreted with caution.

Conclusions

Psoriatic arthritis is a complex disease due to the numerous domains of disease activity that need to be addressed with treatment. Based on the double-blind portion of both phase III randomized controlled trials (RCTs) (SELECT-PsA1 and SELECT-PsA2), treatment with oral upadacitinib 15 mg once daily is associated with statistically significant and clinically meaningful improvement compared to placebo in the clinical response of PsA symptoms, as measured by the primary efficacy outcome of an ACR20 response at week 12. In bDMARD-naive patients (SELECT-PsA1), upadacitinib 15 mg orally once daily was no worse than (i.e., noninferior to) adalimumab 40 mg administered subcutaneously every other week in achievement of an ACR20 at week 12. The efficacy of upadacitinib compared to adalimumab in bDMARD-experienced patients is unknown.

Additionally, an overall consistent effect of upadacitinib 15 mg compared to placebo was demonstrated for numerous clinically relevant manifestations of PsA, including function and disability, PsA symptoms (pain and fatigue), HRQoL or patient-reported health outcomes, skin disease or psoriasis, musculoskeletal symptoms (enthesitis, dactylitis, and spinal symptoms), and other measures of clinical response or disease control such as MDA. Improvement in these measures of treatment response with upadacitinib 15 mg compared to placebo was demonstrated across both studies in patients with inadequate response or intolerance to non-bDMARDs (SELECT-PsA1) and bDMARDs (SELECT-PsA2). Efficacy of upadacitinib in radiographic changes was only studied in bDMARD-naive patients, and the clinical meaningfulness of the small improvement seen versus placebo over this short duration is uncertain. Furthermore, an estimated MID for many of the measures used to assess continuous end points was not identified for patients with PsA, making the clinical significance of the numerical improvements seen in some end points, notably the patient’s assessment of pain NRS, uncertain.

Findings from the end of period 1 of SELECT-PsA1 and SELECT-PsA2 suggest that the improvements in outcomes observed at week 24 in the upadacitinib 15 mg treatment group were maintained throughout the 56-week extension.

By week 24, the proportion of patients who experienced a TEAE with upadacitinib in SELECT-PsA1 was comparable to the proportion treated with adalimumab but higher relative to the placebo group. In SELECT-PsA2, the proportion of patients who experienced a TEAE was similar between the upadacitinib and placebo groups. Upper respiratory tract infection was most commonly reported in both studies. The safety profile of upadacitinib 15 mg once daily over 56 weeks was consistent with that observed during the 24-week double-blind period, with no unexpected safety signals reported. However, long-term data from the ongoing extension phase of both SELECT-PsA1 (up to 5 years) and SELECT-PsA2 (up to 3 years) will help better characterize the efficacy and safety of upadacitinib in the treatment of this chronic condition.

No direct comparative evidence for upadacitinib 15 mg versus bDMARDs or tsDMARDs other than adalimumab was identified. A sponsor-submitted ITC comparing upadacitinib 15 mg to bDMARDs or tsDMARDs suggested that in both bDMARD-naive and DMARD-experienced patients, upadacitinib does not show either a consistent or distinct difference in efficacy as measured by ACR20, PASI, PsARC, or HAQ-DI when compared to bDMARDs or tsDMARDs. The value of the ITC results is uncertain due to the inherent heterogeneity across trials in the networks. Moreover, no information was obtained regarding safety compared to other bDMARDs or tsDMARDs. In addition, no conclusion could be drawn on the HRQoL outcomes.

Introduction

Disease Background

Psoriatic arthritis is an inflammatory musculoskeletal disease with a heterogenous presentation and disease course. While it is associated with psoriasis, PsA also presents with variable clinical features involving multiple domains, including peripheral arthritis, enthesitis, dactylitis, and axial disease.^1,2 Diagnosis of PsA can be a challenge, as there is no gold-standard diagnostic test; diagnoses are typically based on clinical findings and imaging features that evaluate specific patterns of joint inflammation or involvement of the different domains. Patients with PsA also present with psoriatic skin lesions and most (95%) are usually seronegative for rheumatoid factor.^2,3 Pain and stiffness of the affected joints are the most predominant presenting symptoms, with fatigue also occurring in many patients.¹

The prevalence of PsA varies, depending on the case definition and geography, and is estimated to be 1 to 2 per 1,000 in the general population.¹ A population-based Canadian study estimated that the age- and sex-standardized cumulative prevalence of PsA in Ontario ranged from 0.09% in 2008 to 0.15% in 2015. The same study estimated the age- and sex-standardized incidence in 2015 to be 14 per 100,000.⁴ These figures may vary. For example, another study found the estimated annual incidence of PsA was 6 per 100,000 per year.¹

About 30% of patients with psoriasis develop PsA; skin disease usually precedes manifestations of PsA by several years (10 years on average), although both can occur simultaneously in some individuals or PsA may occur before the onset of psoriasis.² A Canadian prospective cohort study estimated the annual incidence of PsA was 2.7 cases per 100 psoriasis patients.¹² Over time, PsA can lead to deformities and joint damage.² This can lead to significant functional impairment, which in turn can affect work productivity and reduce HRQoL.^2,3

Standards of Therapy

Treatment goals for patients with PsA include achieving the lowest possible level of disease activity in all domains of disease; optimizing functional status, improving quality of life and well‐being; preventing structural damage to the greatest extent possible; and avoiding or minimizing complications, both from untreated active disease and from therapy. Because this disease affects more than just the joints of the patient, treatment is individualized based on various factors, including disease activity, structural damage, comorbid conditions, and previous therapies.⁶ Accordingly, treatment effects need to be evaluated in different domains involving the musculoskeletal system (e.g., in dactylitis, enthesitis, and axial disease) as well as extra-articular manifestations (e.g., nails, skin, eyes, and the gastrointestinal tract). Several drug classes are employed in the pharmacologic treatment of PsA, including NSAIDs, csDMARDs (methotrexate, sulfasalazine, and leflunomide), bDMARDs (TNF inhibitors, IL-23 inhibitors, IL-12/23 inhibitors, and IL-17 inhibitors), and tsDMARDs (apremilast or tofacitinib).⁵

In the treatment of PsA, NSAIDs are generally used for symptomatic treatment. Despite the lack of evidence from RCTs, guidelines often recommend using csDMARDs, typically methotrexate, as the primary treatment in first-line therapy or after a short course of NSAIDs in patients with polyarthritis.^5,6 These recommendations were based on data from primarily observational studies, the low costs and universal access associated with csDMARDs, and the lack of evidence that a short time delay in the introduction of more effective therapies would affect long-term function and quality of life.⁶ Some guidelines also recommend first-line treatment with a TNF inhibitor, particularly in patients with severe PsA or psoriasis.^6,7 In patients with an inadequate response to at least 1 csDMARD, the European League Against Rheumatism recommends starting a bDMARD. The latest recommendations do not distinguish among TNF inhibitors, IL-17 inhibitors, and IL-12/23 inhibitors, and the choice is individualized based on numerous factors, including cost. However, an IL-17 or IL-12/23 inhibitor may be preferred in patients with relevant skin involvement.⁵

In the case of biologic drug treatment failure due to either lack of efficacy or AEs, switching to an alternative biologic drug within a class or with a different mode of action was recommended in treatment guidelines.^5-7 Tofacitinib (a pan-JAK inhibitor) may be considered in patients who have an inadequate response or intolerance to at least 1 csDMARD and at least 1 bDMARD, or when a bDMARD is not appropriate.⁵ It is given in combination with methotrexate or another csDMARD.¹³ Thereafter, European League Against Rheumatism recommendations suggest considering apremilast (a PDE4 inhibitor) in patients with mild disease and an inadequate response to at least 1 csDMARD, and for whom neither a bDMARD nor a JAK inhibitor is appropriate. Abatacept, a cytotoxic T lymphocyte–associated antigen-4 immunoglobulin co-stimulation modulator, is another potential option in the treatment of PsA; however, its use is generally limited due to its relatively low efficacy, and it is also not funded by public drug plans.⁵ Specific treatment recommendations are also available for other scenarios; for example, in patients with unequivocal enthesitis or predominantly axial disease, for whom bDMARDs are generally considered after insufficient response to NSAIDs.⁵ In addition to the severity and manifestations of disease, treatment choice may vary depending on contraindications, concomitant conditions (e.g., active inflammatory bowel disease), and patient preference (e.g., route of administration or dosing frequency).⁷

Although there is no specific Canadian treatment guideline for the management of PsA, the Canadian Rheumatology Association/Spondyloarthritis Research Consortium of Canada Treatment Recommendations for the Management of Spondyloarthritis from 2014 include the following recommendations: (1) methotrexate, sulfasalazine, and leflunomide may be considered in patients with peripheral arthritis, however these treatments have only minimal to moderate evidence of efficacy (on peripheral joints and no efficacy on the spine); (2) combination therapy with DMARDs should be considered in peripheral arthritis, particularly in patients with moderate to high disease activity, poor prognostic features, and in patients with recent-onset disease, and combination therapy should also be considered in patients with inadequate response to monotherapy; (3) TNF inhibitors should be offered to those with persistent inflammation despite a trial of NSAID and 1 csDMARD in patients with predominantly peripheral arthritis; and (4) TNF inhibitors should be offered to patients with refractory enthesitis or dactylitis accompanied by persistent inflammation. The recommendations on the use of csDMARDs and TNF inhibitors in peripheral arthritis were based on PsA data.¹⁴

Currently, a single JAK inhibitor is available in Canada. Tofacitinib, a pan-JAK inhibitor, is indicated for the treatment of adult patients with active PsA when the response to previous DMARD therapy has been inadequate.¹³ However, it has not undergone review by CADTH and is not publicly funded in Canada. Some safety signals have been identified for tofacitinib. Most recently, preliminary results for a post-marketing safety trial comparing tofacitinib to TNF inhibitors identified an increased risk of major adverse cardiovascular events (MACEs). The trial enrolled patients with rheumatoid arthritis who were 50 years of age or older and had at least 1 additional cardiovascular risk factor. Results suggest that risks are associated with dosages of both 5 mg twice daily and 10 mg twice daily.¹⁵

The input received from the clinical expert consulted by CADTH for this review is consistent with the guidelines. According to the clinical expert, treatment of PsA is directed at fundamental disease mechanisms, and is complex because decisions are based on managing the diverse domains of psoriatic disease (i.e., arthritis, spondylitis, enthesitis, dactylitis, and skin and nail dysfunction). Treatment decisions are further influenced by the number of swollen peripheral joints (oligoarthritis versus polyarthritis), by the magnitude of skin disease, and by the presence of associated conditions, such as inflammatory bowel disease and uveitis. For treatment goals in PsA, the clinical expert noted that a treat-to-target strategy is advocated. The recommended target for arthritis is MDA, a high-level end point of almost no disease activity. Targets involving peripheral joint activity may also be acceptable. Control of musculoskeletal disease is expected to prevent joint damage and deformity. In turn, control of articular disease would be expected to improve quality of life, restore functional capacity, and enhance work attendance and productivity. The clinical expert indicated that the current treatment plan in Canada for patients with PsA begins with a DMARD. Trials of at least 2 csDMARDs as monotherapy (methotrexate, leflunomide, or sulfasalazine) lasting 3 to 4 months. In patients with inadequate responses to csDMARDs, treatment is escalated to a TNF inhibitor. Patients with inadequate response to TNF inhibitors would be treated with an IL-17 inhibitor or IL-12/23 inhibitor next, and the oral JAK inhibitor tofacitinib is currently considered as last-line therapy. Oral apremilast, a PDE4 inhibitor, is considered in a minority of patients, typically those with less joint and skin disease; however, it is not reimbursed by public drug plans. The clinical expert noted that patients with spinal disease are treated first with NSAIDs, followed by biologics when NSAIDs fail. Oral and IV steroids are not indicated in patients with spondylitis and used only with caution and at low doses for patients with peripheral arthritis. Intra-articular steroid injections can be used in patients with oligoarthritis and dactylitis.

Drug

Upadacitinib is an oral JAK inhibitor that has greater inhibitory potency on JAK1 proteins relative to JAK2, JAK3, and TYK2. By inhibiting JAKs, upadacitinib modulates the intracellular signalling pathways of cytokines or growth factors involved in a broad range of cellular processes, such as inflammatory responses, hematopoiesis, and immune surveillance. Specifically, upadacitinib prevents the phosphorylation and activation of signal transducers and activators of transcription by JAKs, blocking intracellular activity, including gene expression. Upadacitinib is available as 15 mg extended-release tablets.⁸

The recommended dosage of upadacitinib for treatment of PsA is 15 mg orally once a day. Upadacitinib may be used as monotherapy or in combination with methotrexate.⁸

Upadacitinib is approved by Health Canada for the treatment of adults with active PsA who have had an inadequate response or are intolerant to methotrexate or other DMARDs. The requested reimbursement criteria align overall with the Health Canada indication. AbbVie Inc., the sponsor, is requesting reimbursement for the treatment of active PsA in adult patients who have responded inadequately or who are intolerant to 1 or more DMARDs. The approved indication states that upadacitinib may be used as monotherapy or in combination with methotrexate. The European Commission has also granted marketing authorization for upadacitinib for the same indication.¹⁶

Upadacitinib has been previously reviewed by CADTH for the treatment of adults with moderately to severely active rheumatoid arthritis who have had an inadequate response or are intolerant to methotrexate.¹⁷

In addition to upadacitinib, various treatments for PsA are currently approved in Canada (Table 3).

Stakeholder Perspectives

Summary of Patient Group Input

This section was prepared by CADTH staff based on the input provided by patient groups.

About the Patient Group(s) and Information Gathered

Four inputs were submitted for this review from 6 different patient groups: ACE, the CSA, the CAPP in partnership with the CPN, and the CAPA in partnership with the Arthritis Society.

Canada’s largest and longest-running national arthritis patient organization, ACE is headquartered in Vancouver, British Columbia, and has 50,000 members from coast to coast. It provides free, science-based information and education programs to people with arthritis and those who care for and support them.

The CSA is a national not-for-profit organization federally registered in Canada. Its mission is to be the leading voice for the spondyloarthritis community in Canada, raise awareness, and provide support, education and advocacy for patients, caregivers, and health professionals.

The CAPP is a national not-for-profit organization formed to better serve the needs of psoriasis patients across the country. Its mission is to be a resource and advocate for psoriatic patients and their families to improve patient care and quality of life.

The CPN is a national, not-for-profit organization dedicated to improving the quality of life of people in Canada who live with psoriasis and PsA. The CPN does this by providing current information on research and treatment options and by working with others to build awareness and advocacy about the complexities of these conditions.

The CAPA is a grassroots, patient-driven and managed, independent, national education and advocacy organization with members and supporters across Canada. It creates links between Canadians with arthritis, helps them become more effective advocates, and seeks to improve the quality of life of all people living with the disease.

The Arthritis Society is dedicated to a vision of living in a world in which people are free from the devastating effects of arthritis. The Arthritis Society is Canada’s principal health charity, providing education, programs, and support to the more than 6 million Canadians living with arthritis.

Five patient organizations collaborated by collectively developing survey questions using SurveyMonkey for the submitted inputs. The survey was launched by the Arthritis Society and the raw data were shared with all collaborating organizations. Each organization shared the surveys with its respective memberships or patient communities via newsletters, social media channels, and websites. The CSA conducted a telephone interview with 1 patient on upadacitinib (Rinvoq). Notes from this interview were shared with the CAPP, CPN, CAPA, and Arthritis Society.

The survey was translated into French to reach a broader national audience. The surveys were available from December 15, 2020, to January 18, 2021. Overall, there were 85 responses to the English survey and 9 responses to the French survey. Of those who provided their demographic information, 10 respondents were from British Columbia (10.8%), 6 from Alberta (6.5%), 3 from Saskatchewan (3.2%), 17 from Manitoba (19.3%), 30 from Ontario (32.3%), 13 from Quebec (14.0%), 3 from New Brunswick (3.2%), 5 from Nova Scotia (5.4%), and 4 from Newfoundland and Labrador (4.3%). Only 1 response was from someone outside Canada. In addition to PsA, 90.7% of 54 respondents (n = 49) across the 2 surveys indicated they live with psoriasis, 16.7% (n = 9) live with another inflammatory condition, 35.2% (n = 19) also live with another type of arthritis, and 33.3% (n = 18) live with at least 1 other condition, including fibromyalgia, borderline personality disorder, eczema, Raynaud’s syndrome, scoliosis, hypothyroidism, allergies, asthma, high blood pressure, or bladder conditions. Fifty-four respondents provided information about their age across the English and French surveys. These respondents ranged in age from 26 to 80 years old.

In addition, ACE gathered data from 5 patients who completed a patient input survey through SurveyMonkey from December 18, 2020, to January 26, 2021.

Disease Experience

Psoriatic arthritis is an inflammatory disease that causes swelling and pain in multiple joints and can sometimes result in permanent and debilitating joint damage. It is also a systemic disease affecting other parts of the body, including the eyes and heart, and can vary in severity from mild to very severe. There is currently no cure for PsA. Respondents to the survey emphasized pain, stiffness, lack of mobility, and fatigue, which affect their daily living activities, their family lives, and their ability to work and maintain certain hobbies. People can live in constant pain without remission.

When asked about the most significant impacts of PsA on their daily quality of life, respondents from the collaborative input expressed that PsA interfered with social connections (78%), self-esteem (69%), family life (66%), mental health (66%), work (60%), friendships (50%), intimacy (44%), and parenting (15%). Other impacts included embarrassment and self-consciousness from symptoms caused by PsA. Respondents reported difficulty contributing and participating at school or work due to the fatigue, pain, and other symptoms of PsA. The following quotes from respondents illustrate how living with PsA, and its symptoms, affects their lives:

EXTREMELY painful joints. To the point that walking, climbing stairs (to the 2nd floor where the master bedroom was located), etc. was a challenge. Pain killers and NSAIDs were a minimal daily requirement just to get out of bed and make it through the day... Sometimes not being able to go to work because of the pain. When I was first diagnosed with PsA, my son had just been born. I was not even able to do what many fathers do, and take for granted, and that was hold my son up over my head with my arms, smiling up to him as he would be laughing down to me... That was almost 30 years ago. My son passed in March of 2019 and to this day, I still regret not being able to do that simple thing, that “[rite] of passage”... Holding my infant so up in the air. PsA took that from me...

People living with the disease are also at risk of comorbidities, such as depression and mental health issues, diabetes, and cardiovascular disease. For some people, flares can be incapacitating. Due to their unpredictability, flares must be dealt with reactively. According to the joint input from CAPA and the Arthritis Society, the unpredictable nature of PsA may often make it feel like a person is not in control of their disease and can affect their ability to carry out day-to-day activities and life roles, such as contributing to the workplace. The following respondent quote indicates how chronic fatigue influences daily activities:

Chronic fatigue — daily. Worse in the winter than summer months. Energy levels decline dramatically starting 4pm. I am unable to do much socializing after 6pm due to lack of energy. Dinners out are a rare occurrence. Enthesitis — in my wrists, elbows,—. My hands are now weak (I'm 50) that carrying grocery bags is uncomfortable. I take a rolling cart whenever I have to carry more than 5l lbs. Flares — occur when there is hot weather or I have over-exerted myself exercising. This causes the enthesitis to become painful and I experience extreme body aches and fatigue. Like the flu. This lasts 1 to 3 days. Gluten and dairy flares me — so I avoid them in my diet. I normally have to pack food when travelling as it's difficult to find gf [gluten-free] options.

The impact of PsA extends to others within a person’s support circle, including caregivers such as spouses/partners and children. The CAPA and the Arthritis Society note that these people often take on additional chores or tasks such as cooking, cleaning, and shopping to support the person living with PsA, and family roles change as spouses and partners take on more tasks, such as supporting their partners in getting to and from medical appointments. The following are respondent quotes indicating how PsA affects their caregivers: “I'm sick of the Pain and Fatigue!! I feel like a Burden to my Husband!” and “I had struggled with severe pain for 30 plus years, was very hard to do everyday tasks, my son thank god for him was always there to help me as opening a simple bottle of water was at times very hard to even open.”

Patients who submitted input to ACE’s patient survey spoke to the following disease experiences, and certain aspects of the disease that are more important to control than others:

• “Was diagnosed with PsA in 2015, but started to show symptoms in mid-eighties.” They experience pain in many joints and require 45 minutes warm-up exercises every morning before their day starts. They also experience big toes problems and have “psoriasis in scalp and on face” that requires lotions.

• “Maintaining treatment schedule. Missing treatments, particularly my weekly methotrexate shot catches up with me…aches, stiffness, inflammation.”

Experiences With Currently Available Treatments

Patients living with psoriatic disease often try a succession of treatments throughout their lives. Due to the inflammatory nature of the diseases, treatments that are initially effective can become less effective over time. It is important to note that many people who live with PsA also live with psoriasis. As some drugs are indicated for both diseases, it is not uncommon for people with psoriasis to have some experience taking a treatment indicated for PsA and vice versa. For patients who take treatments for psoriasis, these same treatments can but do not always adequately manage their PsA.

Effective treatments mean that people with PsA do not need to live with the permanent damage, high medical costs (e.g., surgery, mobility aids, and accessible housing), and disability. Early intervention is critical to allow people with PsA the opportunity to fully participate in all aspects of life. Although numerous medication options exist, patients’ responses to medication can vary significantly. As a result, patients need a number of medication options to effectively manage their disease throughout their lives. There are also no specific tests that identify which medication will be effective for a person living with PsA. This means that a person with the disease will need to take 1 or more medications on a trial-and-error basis to find an effective medication. It is also an anxious and stressful experience if medications are not effective and cost thousands of dollars out of pocket. Oftentimes, people with PsA need to make difficult financial choices to pay for their medications.

Of individuals who had taken the treatment and responded to the question from the collaborative input, biologics had the highest proportion of patients, with 50% (18 of 36) indicating the treatment was “very effective” and 22% (n = 8) “mildly effective.” Non-bDMARDs were mildly effective for 36% (16 of 44 respondents) and very effective for 11% (n = 5). For 36% (n = 16) of respondents, DMARDs were either very or mildly ineffective. Nonsteroidal anti-inflammatory drugs were mildly effective for 36% (19 of 53 of respondents) but ineffective for 38% (n = 20), and for 19% (n = 10) there was no difference in symptoms. Forty percent (6 of 15) of those taking leflunomide reported it to be very ineffective, with no respondents reporting it was very effective. Apremilast was very ineffective for 55% of patients (6 of 11 respondents) and mildly effective for 27% (3 of 11 respondents). Tofacitinib was very ineffective for 71% (5 of 7 respondents), with no respondents reporting it to be very or mildly effective.

Many patients have experience with DMARDs. Fifty-four respondents had experience with non-bDMARDs, including combinations containing methotrexate: apremilast (18.5%; n = 10), methotrexate (74.1%, n = 40), azathioprine (0.02%, n = 1), cyclosporine (13.0%, n = 7), hydroxychloroquine (13.0%, n = 7), leflunomide (14.8%, n = 8), sulfasalazine (22.2%, n = 12), and Salazopyrin (0.04%, n = 2). Several respondents had experience with methotrexate and noted a variety of side effects, including nausea, raised liver enzyme levels, headaches, and a sore mouth. Patient experiences with methotrexate include “worsening my brain fog” and “feeling ‘worse’ on it in other ways” are reflected in the following statement:

I've only tried methotrexate and am currently on it. It manages my PsA quite well but I'm still susceptible to joint injuries and cuts and abrasions can still lead to new psoriatic lesions. The side effects I get are poor sleep, nausea, poor balance, mental fog, runny stool, flatulence, and some GI discomfort. The needs not being met by methotrexate are that it doesn't completely put my arthritis into remission and it doesn't completely eliminate my psoriatic lesions. Wishful thinking, I know.

The side effects of sulfasalazine were noted by several respondents. One reported, “I had fever, chills, a horrible hot tingling rash and fatigue, ended up going off all medications including my anti-inflammatory. So I spent 2 weeks [with] no medications and my pain increased more and more each day till I was able to resume my anti-inflammatory…. My PsA remains uncontrolled.” Another noted that “sulphasalazine almost cost me my marriage due to mood swings.”

Cyclosporine was also noted to affect other organs: “When I took cyclosporine my kidney function became a problem. I have not been able to achieve comfort with any treatment.” Some respondents noted that apremilast (Otezla) helped their psoriasis but did not improve their PsA, and came with challenging side effects, including increased heart rate, nausea, diarrhea, depression, and moodiness.

Several respondents also had experience with different biologics, but the benefits do not last forever: “The biologics do a good job until they fail. Then the search is on for the next one that will work” and “on my 5th biologic since 2007.”

According to the input from CSA, new treatment options and different classes of medications fill unmet needs of patients and prescriber. The Rinvoq oral formulation is an exciting option for patients. It provides another option for administration and may help lead to improved adherence and ultimately better outcomes.

The input from CAPA and the Arthritis Society stated that patients may also pursue medical cannabis and/or non-pharmacological approaches to manage PsA symptoms, such as physiotherapy, occupational therapy, massage therapy, counselling, or acupuncture. These approaches can often help address the symptoms of the disease, such as pain and fatigue. However, the patient groups noted that there are significant unmet patient needs in terms of accessing non-pharmacological treatments, often because they are not reimbursed through provincial health care systems, the treatment options are simply not offered, or there are lengthy waits.

Patients identified a number of issues in accessing treatment options. Expense, travel, and time required for treatment were all cited as being prohibitive. Some patients also identified a lack of access to specialists and general practitioners, and/or restrictions associated with the coronavirus disease 2019 pandemic.

One patient who filled out ACE’s patient survey spoke to experiences taking infliximab (Remicade) as an infusion every 7 weeks combined with a weekly dose of methotrexate. They experienced nausea for a couple of days after taking methotrexate and higher levels of fatigue. When asked if there are any needs that are not met by current therapy, this patient stated: “Not for me to say – but many friends and relatives have asked about my treatments and their inability to access them or their acceptance of arthritis/pain as part of getting old.” While this patient had no hardships in accessing current therapies, they added: “But I have private coverage for Remicade, which is roughly $3K every 7 weeks – roughly $21-25,000.”

Improved Outcomes

Through their joint input, the CAPA and the Arthritis Society identified several outcomes that are important to patients with PsA and that should be considered when evaluating new therapies. These include the route of drug administration (oral versus infusion versus self-injections), reduction in pain and fatigue, effectiveness for psoriasis symptoms as well as PsA symptoms, increased mobility, ability to work and be productive at work, ability to carry out activities of daily living, ability to effectively carry out parenting tasks and other important social roles, reduced infection rates, affordability of the medication, and increased quality of life.

People living with PsA shared the following experiences:

• “I would hope [a new treatment] would not have side effects and I wish it would work on my psoriasis as well as my arthritis. I would hope to gain more confidence in my appearance (psoriasis plaques lessened) so that I would desire to be more social. Being more social would get me out of the house and more active which would help with the everyday aches and pains of arthritis. It's all linked...you cannot separate the arthritis and the psoriasis. They work together against the body and mind unfortunately.”

• “The drug needs to be 8-%+ to be effective. No side effects would be the goal, I have had enough to them. Better quality of life so that the morning you wish you were alive rather than dead. I do mean that literally.”

Experience With Drug Under Review

Of those surveyed in the joint patient survey, 6 respondents indicated that they have experience taking upadacitinib for PsA. Four of these respondents were from Manitoba and 2 were from Quebec. Five indicated that they live with psoriasis. Five of the 6 respondents who have experience with upadacitinib indicated that they have used DMARDs. No patients from ACE’s survey had experience with upadacitinib. When asked about their positive and negative experiences with upadacitinib, 1 of these respondents noted: “[The] only negative [side effect] as mention[ed] before is a feeling of being cold all the time. I have had 80-90% relief of pain.” Another reported that “It has helped a lot with my pain and joint swelling, I really haven’t noticed any side effects. It’s definitely better than the injections I take, I hope that taking this now will prevent worsening problems on my joints as I age. It makes day to day easier and less painful.” A third reported, “Getting up every morning with pain is a challenge! In contrast, with upadacitinib, pain lasts less when you wake up. The look of others, when I had psoriasis on hands was a challenge, now I have beautiful nails and beautiful hands. Maintain[ing] clean[lines] is much easier. Besides, shortness of breath is the only side effect I have observed.”

When asked about the impact of upadacitinib (Rinvoq) on caregivers and their family and their day-to-day activities, these respondents answered with: “Yes. I am able to move more freely and exercise more often,” “I have more energy to do activities with my partner. I had support from my partner, he does the shopping. The support of the research centre team was very important to me, [to address] my anxieties and my fears about the disease,” and “Greater self-esteem and easier to get around.”

The CSA conducted 1 interview with a female patient who has lived with PsA for 30 years, has had severe psoriasis her entire life, and who also had experience with upadacitinib through a clinical trial. Her PsA caused pain and reduced her mobility to the point that she had to retire from her job. Her past medication history included methotrexate, “every topical under the sun,” and 2 biologics that she injected but that were not effective for her. Her experience with upadacitinib was positive: within 2 weeks of beginning treatment, she was able to walk around her home without walking aids and her pain was significantly better. Her skin was almost entirely clear, with the exception of a couple of small spots. She noted no side effects, despite previously being very prone to them and unable to tolerate other medications. She has been able to return to work part-time. She commented that for her, this drug is life-altering, has given her freedom, and she can’t believe how much she can move around and not be in pain all the time.

Additional Information

According to the joint input from CAPP and CPN, PsA is a disease that often “falls through the cracks.” Some patients are seen by a dermatologist while others are seen by rheumatologists. Joint pain is not always discussed with a dermatologist and plaques on the skin are not always discussed with rheumatologists. These challenges often lead to delays in diagnosis and consequently severe and irreversible damage to the joints. Roughly 30% of people with psoriasis will develop PsA. Most (80%) of the time, psoriasis comes first, but it remains difficult to predict whether a person living with psoriasis will later develop PsA, despite research advances that have identified a number of biomarkers associated with PsA and led to predictive screening tools.

The patient input from the CSA states that effective treatments give Canadians the opportunity to regain self-confidence and re-integrate into society and personal relationships. The input from this group noted that, for those yet to find relief, it can be a life of darkness, isolation, and desperation, leaving people spiralling into further depression and anxiousness and robbing them of life. The input emphasized it is important for CADTH take into consideration that many patients over the course of their journey have tried several, and in some cases all, the options currently available.

Clinician Input

Input from Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise in the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol; assisting in the critical appraisal of clinical evidence; interpreting the clinical relevance of the results; and providing guidance on the potential place in therapy). The following input was provided by 1 clinical specialist with expertise in the diagnosis and management of PsA.

Unmet Needs

There are unmet needs in the treatment of psoriatic disease because some patients may not respond to any treatment, and only a minority achieve MDA. The treatment of PsA involves numerous domains of disease activity, not all of which are likely to be addressed by a single agent. Further durability of any therapy is limited and switching to treatment with a different mechanism of action will be necessary when a patient becomes refractory to current treatment.

Place in Therapy

Most rheumatologists will generally use TNF inhibitors and IL-17 inhibitors before upadacitinib. However, as experience is accrued with upadacitinib, and when safety has been confirmed (i.e., when data show a lack of association with AEs of special interest, such as bowel perforations, venous thromboembolic events, MACEs, and malignancy), it is anticipated that upadacitinib could become a first-line treatment for PsA. The caveat with this assumption is that longer-term observation of patients on upadacitinib will be needed to confirm the durability of efficacy and safety.

In general, oral agents are considered more convenient than therapies administered intravenously or subcutaneously and are expected to demonstrate enhanced adherence to treatment. Upadacitinib would have the advantage of oral administration, which adds further support to its appeal.

Patient Population

Any patient with peripheral joint and skin disease who has not responded to conventional synthetic DMARDs would be eligible for upadacitinib. Studies in spondylitis will address its use in this domain. Currently, some rheumatologists may not prescribe upadacitinib to patients with a recent history of thromboembolic disease, and patients with active infection would also not be treated with this drug. Additional data may reveal other contraindications to therapy.

Assessing Response to Treatment

In clinical practice, the SJC is the most likely measure of response, with a reduction in joint count reflecting a meaningful response. Some rheumatologists may use the achievement of MDA as a meaningful response. Other clinically meaningful responses to treatment may involve patient-reported outcomes such as a health assessment questionnaire, patient’s global assessment, and patient’s assessment of pain; however, these outcomes are subjective and prone to bias. Outcome measures of RCTs are seldom used in clinical practice.

It is expected that benefit from a JAK inhibitor will be readily apparent by week 12, and most rheumatologists will decide whether to continue with therapy at that point.

Discontinuing Treatment

Lack of response and/or side effects will result in a decision to discontinue treatment. Blood tests will be monitored every 1 to 3 months.

Prescribing Conditions

Upadacitinib should be prescribed by specialists who treat PsA. Diagnosis, prescribing of treatment, and continued patient monitoring should be performed by specialists such as rheumatologists who treat PsA.

Clinician Group Input

No registered clinician group input was received for this reimbursement review.

Drug Program Input

The drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may affect their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 4.

Clinical Evidence Selection

The clinical evidence included in the review of upadacitinib is presented in 3 sections. The first section, the systematic review, includes pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those studies selected according to an a priori protocol. The second section includes indirect evidence from the sponsor and indirect evidence selected from the literature that met the selection criteria specified in the review. The third section includes sponsor-submitted long-term extension studies and additional relevant studies that were considered to address important gaps in the evidence included in the systematic review.

Systematic Review (Pivotal and Protocol-Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of upadacitinib 15 mg oral extended-release tablets as monotherapy or in combination with non-bDMARDs for the treatment of active PsA in adult patients who have responded inadequately or who are intolerant to 1 or more DMARDs.

Methods

Studies selected for inclusion in the systematic review include pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 5. Outcomes included in the CADTH review protocol reflect outcomes considered important by patients, clinicians, and drug plans.

The systematic review protocol was established before the granting of a Notice of Compliance from Health Canada.

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy according to the Peer Review of Electronic Search Strategies checklist (https://www.cadth.ca/resources/finding-evidence/press).²⁸

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946‒) via Ovid and Embase (1974‒) via Ovid. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concept was upadacitinib (Rinvoq). Clinical trials registries searched included the US National Institutes of Health’s clinicaltrials.gov, WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

No filters were applied to limit the retrieval by study type. Retrieval was not limited by publication date or by language. Conference abstracts were excluded from the search results. Appendix 1 provides detailed search strategies.

The initial search was completed on February 1, 2021. Regular alerts updated the search until the meeting of the CADTH Canadian Drug Expert Committee on June 16, 2021.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature checklist.²⁹ Included in this search were the websites of regulatory agencies (FDA and European Medicines Agency). Google was used to search for additional internet-based materials. See Appendix 1 for more information on the grey literature search strategy. These searches were supplemented by reviewing bibliographies of key papers and through contacts with appropriate experts. In addition, the manufacturer of the drug was contacted for information regarding unpublished studies.

Two CADTH clinical reviewers independently selected studies for inclusion in the review based on titles and abstracts, according to the predetermined protocol. Full-text articles of all citations considered potentially relevant by at least 1 reviewer were acquired. Reviewers independently made the final selection of studies to be included in the review, and differences were resolved through discussion.

A focused literature search for NMAs dealing with PsA was run in MEDLINE All (1946–) on February 1, 2021.

Findings From the Literature

A total of 2 studies were identified from the literature for inclusion in the systematic review (Figure 1). The included studies are summarized in Table 6. A list of excluded studies is presented in Appendix 2.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Description of Studies

Two phase III, multi-centre, double-blind RCTs met the inclusion criteria for this systematic review.^9,10 The study designs of SELECT-PsA1 and SELECT-PsA2 are shown in Figure 2 and Figure 3. Both SELECT-PsA1 and PsA2 were funded by AbbVie, Inc.

Both SELECT studies enrolled adults with moderate to severely active PsA who had been previously treated with a DMARD; however, patients enrolled in PsA1 were bDMARD-naive. A screening period running 35 days before randomization was used to assess patients’ eligibility, during which their medical history, treatment history, and current medical condition were assessed, and relevant laboratory tests including tuberculosis skin tests and radiographic examinations were performed. Both studies consisted of 2 periods, and at the end of period 1, all patients on placebo were switched to upadacitinib. In SELECT-PsA1, period 1 lasted for 56 weeks and included a 24-week double-blind placebo and active comparator–controlled period followed by 32 weeks of blinded active comparator–controlled treatment. In SELECT-PsA2, period 1 was also 56 weeks in duration and included 24 weeks of a double-blinded placebo-controlled phase, followed by 32 weeks of a non-comparative treatment phase. Period 2 is an ongoing open-label long-term extension for up to approximately 5 years in PsA1 and 3 years in PsA2.

The primary objective of period 1 in SELECT-PsA1 was to compare the efficacy of upadacitinib versus placebo and versus adalimumab for the treatment of signs and symptoms of moderately to severely active PsA in patients who have an inadequate response or are intolerant to at least 1 non-bDMARD. Prevention of structural progression and safety were secondary objectives. The objective for period 1 in SELECT-PsA2 was to compare the efficacy and safety of upadacitinib versus placebo for the treatment of signs and symptoms of moderately to severely active PsA in patients who had an inadequate response or are intolerant to at least 1 bDMARD. The objective of period 2 in both SELECT studies was to evaluate the long-term safety, tolerability, and efficacy of upadacitinib in patients who completed period 1. In both studies, period 1 was designed to test the superiority of upadacitinib versus placebo in achieving the primary end point (ACR20 at week 12) and other efficacy parameters at week 12 to 24. In addition, period 1 in PsA1 was designed to test the noninferiority of upadacitinib versus adalimumab (ACR20 at week 12). Results from both studies were presented using interim analyses, with a data cut-off of December 13, 2019, for PsA1 and October 9, 2019, for PsA2. At data cut-off, all patients had completed week 24 of period 1 or had discontinued from the study. Both studies are ongoing.

In SELECT-PsA1 (N = 1,705), eligible participants were randomized after the screening phase in a 2:2:2:1:1 ratio to 1 of 5 treatment groups: upadacitinib 15 mg orally once daily, upadacitinib 30 mg orally once daily, adalimumab 40 mg subcutaneous every other week, and placebo followed by either upadacitinib 15 mg daily or upadacitinib 30 mg daily (Figure 2). Randomization was performed through a centralized interactive voice response system and stratified by extent of psoriasis (≥ 3% BSA or < 3% BSA), current use of at least 1 DMARD, presence of dactylitis, and presence of enthesitis. The exception was for patients from China or Japan, where randomization was stratified only by the extent of psoriasis.

In SELECT-PsA2 (N = 642), eligible participants were randomized after the screening phase in a 2:2:1:1 ratio to 1 of 4 treatment groups: upadacitinib 15 mg orally once daily, upadacitinib 30 mg orally once daily, and placebo followed by either upadacitinib 15 mg daily or upadacitinib 30 mg daily (Figure 3). Randomization was performed through a centralized interactive voice response system and stratified by extent of psoriasis (≥ 3% or < 3% BSA), current use of at least 1 DMARD, and number of prior failed (i.e., inadequate response to) bDMARDs (1 versus > 1). The exception was for patients from China or Japan, where randomization was stratified only by the extent of psoriasis.

In both SELECT-PsA1 and SELECT-PsA2, at week 24, patients randomized to placebo at the start of the study were switched to pre-assigned upadacitinib 15 mg or 30 mg daily in a 1:1 ratio, regardless of response. Patients, study-site personnel, the investigator, and the sponsor study team were blinded to study drug assignment, until all patients completed the week 24 visit. Thereafter, an unblinded analysis was conducted by the sponsor but the individual investigators, study sites, and patients remained blinded. When the last patient completed the last visit of period 1 (week 56), the study drug assignments were unblinded and treatment was continued in an open-label manner until the end of period 2.

Starting at week 36 in both studies, patients who did not show an improvement of at least 20% from baseline in either or both TJC and SJC at 2 consecutive visits were discontinued from study treatment but had the option of continuing to participate in the study for continued follow-up.

The primary end point for both SELECT-PsA1 and SELECT-PsA2 was the proportion of patients who achieve an ACR20 at week 12. The Health Canada–approved indication is for upadacitinib 15 mg once daily. Data from the SELECT studies for the 30 mg dose are not presented in this review.

Figure 2: Study Design for SELECT-PsA1

ABT-494 = upadacitinib; ACR20 = American College of Rheumatology 20% improvement in rheumatoid arthritis; BL = baseline; e.o.w. = every other week; PBO = placebo; QD = once daily; vs = versus; Wk = week.

Note: Sample sizes represent the actual number of patients randomized to each treatment sequence.

^a All patients received X-rays of hands and feet at screening, week 24, week 56, week 104, week 152, and at premature discontinuation.

^b At week 16, rescue therapy was offered to patients classified as nonresponders (defined as not achieving an improvement of at least 20% in either or both tender joint count and swollen joint count at both week 12 and week 16).

^c At week 24, all placebo patients were switched to upadacitinib 15 mg once daily or 30 mg once daily (1:1 ratio) regardless of response.

Source: SELECT-PsA1 Clinical Study Report.⁹

Figure 3: Study Design for SELECT-PsA2

ACR20 = American College of Rheumatology 20% improvement in rheumatoid arthritis; BL = baseline; PBO = placebo; QD = once daily; vs = versus; Wk = week.

^a At week 16, rescue therapy was offered to patients classified as nonresponders (defined as not achieving an improvement of at least 20% in either or both tender joint count and swollen joint count at both week 12 and week 16).

^b At week 24, all placebo patients switched to upadacitinib 15 mg once daily or 30 mg once daily (1:1 ratio) regardless of response.

Source: SELECT-PsA2 Clinical Study Report.¹⁰

Populations

Inclusion and Exclusion Criteria

To be eligible for both SELECT-PsA1 and SELECT-PsA2, patients were required to be at least 18 years of age and have an established diagnosis of PsA with at least 3 tender joints and 3 swollen joints. The patients were required to have active or a documented history of plaque psoriasis. Patients enrolled in SELECT-PsA1 were also required to exhibit either 1 or more examples of erosion on X-ray, or hs-CRP levels greater than the ULN. Patients in PsA1 were bDMARD-naive but had been previously treated with at least 1 non-bDMARD. Patients in PsA2 were all bDMARD-experienced.

Patients were excluded if they had prior exposure to a JAK inhibitor, were on concomitant treatment for psoriasis, or were receiving more than 2 concurrent non-bDMARDs at baseline. Patients on concomitant non-biologic treatment must have been on a stable dose for at least 4 weeks. The use of methotrexate in combination with leflunomide was also not permitted. Patients with extra-articular manifestation of PsA who were not clinically stable for 30 days or longer were excluded. In SELECT-PsA1, patients were also excluded if they had prior exposure to biologic immunomodulators or a history of demyelinating disease.

Baseline Characteristics

Across both studies, in the treatment groups of interest to this review, the mean age ranged from 50 to 54 years, and the majority of patients were White. The mean duration of PsA diagnosis was shorter in the SELECT-PsA1 trial and ranged from 5.9 to 6.2 years in SELECT-PsA1 and from 9.6 to 11.0 years in SELECT-PsA2. The mean TJCs were 20 in SELECT-PsA1 and 25 in SELECT-PsA2. The mean SJC was between 11 and 12 in both SELECT-PsA1 and SELECT-PsA2. The proportion of patients with psoriasis affecting 3% or more of their BSA was lower in SELECT-PsA1 (50%) than in SELECT-PsA2 (61%). Enthesitis at baseline was reported in 77% of patients in SELECT-PsA1 and 82% of patients in SELECT-PsA2, while dactylitis at baseline was reported in 30% to 32% of patients in SELECT-PsA1 and ranged from 26% to 30% of patients in SELECT-PsA2. Patients in SELECT-PsA1 had a lower mean LDI baseline score, ranging from 88 to 99, compared to patents in SELECT-PsA2, who had a baseline score of between 96 and 124. In SELECT-PsA1, more than 99% of patients had received a prior non-bDMARD compared to 74% to 81% of patients in SELECT-PsA2, and most patients had received methotrexate. In SELECT-PsA1, most patients (91%) were enrolled due to inadequate response to prior non-bDMARDs (including patients who had lack of efficacy as well as those who had both a lack of efficacy and intolerance), whereas 7% were enrolled due to an intolerance or contraindication to non-bDMARDs.³⁵ As per protocol, 100% of patients in SELECT-PsA2 had received a prior bDMARD; most (62%) had failed 1 prior biologic treatment, whereas 8% were enrolled due to an intolerance to a bDMARD. At baseline, more patients in SELECT-PsA1 were concurrently using a non-bDMARD (81% to 82%) compared to patients enrolled in SELECT-PsA2 (46 to 47%).

The baseline demographic and disease characteristics were generally similar across the treatment groups within SELECT-PsA1 and SELECT-PsA2. There were some minor imbalances between the treatment groups in PsA2; for example, the placebo group had a slightly higher mean LDI (124.19 versus 96.27 for the upadacitinib 15 mg group) and sIGA score of 0 (8.0% versus 4.3% for upadacitinib 15 mg), and a larger share of patients who had received no prior non-bDMARD (25.9% versus 19.4% for upadacitinib 15 mg). A slightly higher proportion of patients in the upadacitinib 15 mg group had a baseline sIGA score of 2 compared to patients in the placebo group (27.8% versus 38.9% upadacitinib 15 mg group).

Interventions

While both included studies randomized patients to 15 mg oral upadacitinib or placebo once daily, the SELECT-PsA1 study also included an active comparator arm (adalimumab 40 mg subcutaneous every other week). Both studies also randomized patients to a 30 mg upadacitinib arm; however, further description of the 30 mg upadacitinib group is not included in this systematic review as it is higher than the dose approved by Health Canada. All patients in the SELECT-PsA1 study received matching placebo injections or tablets. In this double-dummy design, patients received both the oral study drug (upadacitinib or matching placebo) once daily and the subcutaneous study drug (adalimumab or matching placebo) every other week. Adalimumab injections were administered by medical staff or, after training, self-administered or administered by a designee, including a friend or family member. In both SELECT-PsA1 and SELECT-PsA2, all patients randomized to placebo were switched to upadacitinib 15 mg or 30 mg once daily (1:1 ratio) at week 24. No background non-bDMARD therapy was required for participation in either trial, but patients were permitted to continue their stable background non-bDMARD. Patients taking methotrexate were to also take oral folic acid supplements.

In both trials, starting at week 16, rescue therapy was offered to nonresponders, defined as patients who did not achieve an improvement of at least 20% in both TJC and SJC at both week 12 and week 16. Rescue therapy included addition or dose modification of non-bDMARDs, NSAIDs, acetaminophen, low-potency opioids, or oral corticosteroids. Alternatively, 1 intra-articular, trigger point or tender point, intra-bursa, or intra-tendon-sheath corticosteroid injection was permitted to the corresponding affected location. After week 36, patients were permitted to start or change background PsA medications (i.e., NSAIDs, acetaminophen, low-potency opiates), and non-bDMARDs (concomitant use of up to 2 non-bDMARDs except the combination of methotrexate and leflunomide). Any treatment for psoriasis, except non-bDMARDs, was also permitted to start after week 16; initiation or change in dose of non-bDMARD was permitted at week 36.

Prohibited therapies in both studies included treatment with any bDMARD therapy for PsA or non-bDMARD other than methotrexate, leflunomide, sulfasalazine, hydroxychloroquine, or apremilast. The combination of methotrexate with leflunomide was not permitted, nor was the concomitant use of more than 2 non-bDMARDs.

Outcomes

Details of outcome measures are provided in Appendix 4. Only the efficacy outcomes identified in the review protocol included in the multiplicity-controlled analyses (primary and major secondary efficacy outcomes) and those indicated as important outcomes by patient groups are presented in Table 5. In both SELECT-PsA1 and PsA2, the following assessments were performed by an independent and blinded assessor if possible: PASI, BSA, sIGA, TJC and SJC assessments, dactylitis, and enthesitis. In total, 20 end points in 7 categories are discussed in this review and summarized in Table 8.

American College of Rheumatology Measures

The American College of Rheumatology (ACR) criteria for assessing joint status provide a composite measure of 20%, 50%, or 70% improvement from baseline in both SJCs and TJCs and at least 3 of 5 additional core set disease criteria including: patient global assessment of disease activity, physician global assessment of disease activity, patient’s assessment of pain, patient assessment of physical function (e.g., HAQ-DI score), and acute-phase reactant (e.g., hs-CRP). The ACR20 is generally accepted as an indicator reflecting important difference in response to treatment, while the American College of Rheumatology 50% improvement in rheumatoid arthritis (ACR50) and American College of Rheumatology 70% improvement in rheumatoid arthritis (ACR70) more likely reflect truly important change in the long term management of arthropathy. As an example, if a study reported a 50% ACR20, this means that half of the study patients achieved a minimal 20% improvement in TJCs and SJCs and a 20% improvement in at least 3 of the other 5 criteria. This a general measure of clinical response of peripheral joint disease and does not include assessment of enthesitis, dactylitis, the spine, or the skin. The proportion of patients achieving an ACR20 at week 12 was the primary outcome in both SELECT-PsA1 and SELECT-PsA2 and tested in the multiplicity-controlled analyses. Responses of 50% and 70% at week 12 were secondary end points in SELECT-PsA1 and SELECT-PsA2, but they were not included in the multiplicity-controlled analyses.

In both studies, the sponsor noted that a “joint evaluator” assessed whether a particular joint was “tender or painful.” Total TJC was based on 68 joints (TJC68), and the total SJC was based on 66 joints (SJC66). Aspects related to the patient evaluation (e.g., pain and global assessment) was completed by the patient, while a physician’s global assessment was completed by the physician. Global assessments took into consideration both arthritis and psoriasis activity. The patient and physician’s global assessment of disease activity on the NRS as well as patient’s assessment of pain NRS were measured on a scale of 0 to 10, with higher scores indicating greater disease activity. No specific measures beyond the established allocation concealment described earlier were taken to maintain blinding in the ACR-related outcomes assessments.

Minimal Disease Activity

Minimal disease activity is a composite outcome measure that was developed as a target of treatment for patients with PsA that encompasses different aspects of disease domains. In SELECT-PsA1 and SELECT-PsA2, a patient was classified as a responder, or as achieving MDA when 5 out of the following 7 outcome measures were fulfilled: a TJC68 of no more than 1, SJC66 of no more than 1, PASI less than or equal to 1, BSA psoriasis of 3% or less, patient’s assessment of pain of no more than 1.5 (0 to 10 NRS), patient’s global assessment of disease activity of 2 or less (0 to 10 NRS), HAQ-DI score of up to 0.5, and LEI of no more than 1. Patients who did not meet at least 3 of the 7 criteria were considered nonresponders. The proportion of patients achieving MDA at week 24 was a major secondary end point in both studies and was therefore included in the multiplicity-controlled analyses.

Psoriatic Arthritis Response Criteria

The PsARC is a composite responder index that measures signs and symptoms of PsA assessed by TJC and/or SJC, physician’s global assessment (0 to 5 Likert scale), and patient’s global assessment (0 to 5 Likert scale). A patient was classified as a responder if 2 of 4 measures was achieved, 1 of which is TJC68 or SJC66, and no worsening of any of the measures. In the SELECT studies, the specific components were at least a 30% reduction in TJC68, at least a 30% reduction in SJC66, improvement in patient’s global assessment of disease activity, or improvement in physician’s global assessment of disease activity. In both studies, the PsARC response was modified by using the physician’s global assessment and the patient’s global assessment on a NRS instead of a 5-point Likert scale in the original criteria. The PsARC has been shown to be a responsive and discriminate outcome instrument in RCTs of PsA. The PsARC does not account for psoriasis severity and is only a general assessment of clinical status. The MID for PsARC is unknown. This was an additional outcome in SELECT-PsA1 and SELECT-PsA2 and was not included in the multiplicity-controlled analyses.

Health Assessment Questionnaire–Disability Index

The HAQ-DI was developed to assess physical disability and pain in rheumatoid arthritis and has been used extensively in arthritis RCTs, including PsA trials. Through a self-assessed questionnaire of 8 functional areas (dressing and grooming, arising, eating, walking, hygiene, reach, grip, and errands/chores), patients’ difficulty in performing these activities over the past week is scored from 0 (without any difficulty) to 3 (unable to do). The maximum score for all questions in each category is considered as the score for the respective category, and the overall HAQ-DI is calculated as the mean score from all the categories, with higher scores reflecting greater disability. Mease et al.³⁶ estimated that the MID for the HAQ-DI in PsA patients using anchor-based methods is 0.35, while the MID was estimated to be 0.131 in PsA patients using an anchor-based approach (equal bidirectional magnitudes for improvement and worsening) by Kwok and Pope.³⁷ Discrepancies in the MID estimates may partly be explained by differences in the HAQ-DI score of the patients studied at baseline.³⁸ In Mease et al.,³⁶ patients had a mean HAQ-DI score at baseline of 1.16, corresponding to moderate functional impairment. In contrast, patients in the study by Kwok and Pope had less functional impairment at baseline, with a mean HAQ-DI score of 0.732.³⁷ In the SELECT studies, the HAQ-DI score ranged from 1.10 to 1.23. The MID of 0.35 estimated by Mease et al.³⁶ was used for this review. The SELECT studies also considered the MID to be 0.35. As change from baseline in the HAQ-DI score at week 12 was a key secondary end point in SELECT-PsA1 and SELECT-PsA2, it was included in the multiplicity-controlled analyses. In SELECT-PsA1, this end point was also hierarchically tested for superiority of upadacitinib versus adalimumab.

Work Productivity and Activity Impairment

Work productivity was measured by the WPAI questionnaire for PsA, version 2.0. This is a patient-administered instrument used to measure overall health and specific impacts on productivity at work and outside of work. The WPAI consists of 6 questions to determine employment status, missed work due to PsA, impaired work due to PsA, and daily activity impairment due to PsA. Four main impairment scores are derived: percentage of absenteeism, percentage of presenteeism (reduced productivity while at work), overall percentage of work impairment that combines absenteeism and presenteeism, and percentage of impairment in activities performed outside of work. Higher scores indicate greater impairment, and a lower WPAI score indicates an improvement. It has been suggested that improvements of 20% in presenteeism, 15% in work productivity loss, and 20% in activity impairment represent clinically meaningful improvements in patients with active PsA.³⁹ Absenteeism, presenteeism, and work impairment were measured only in patients who were employed. Change from baseline in the WPAI questionnaire up to week 24 was an additional outcome in both SELECT-PsA1 and SELECT-PsA2 and therefore was not included in the multiplicity-controlled analyses.

Patient’s Assessment of Pain Numerical Rating Scale

The patient’s assessment of pain NRS is 1 of the 5 ACR core set criteria. Pain severity was scored on a 0 to 10 scale, with a score of 0 indicating “no pain” and 10 indicating “worst possible pain.” No reported MID was identified for patients with PsA. In SELECT-PsA1, superiority of upadacitinib versus adalimumab in the change from baseline in the patient’s assessment of pain NRS at week 12 was a key secondary multiplicity-adjusted end point. In PsA2, change from baseline in the patient’s assessment of pain NRS was measured as an individual component of ACR response at week 12, and was not included in the multiplicity-controlled analysis.

FACIT-F

The FACIT-F questionnaire is a 13-item tool that measures an individual's level of fatigue during usual daily activities over the past week. The instrument measures physical fatigue (e.g., “I feel tired”), functional fatigue (e.g., trouble finishing things), emotional fatigue (e.g., frustration), and social consequences of fatigue (e.g., limits social activity). Level of fatigue is measured on a 4-point scale, with 4 indicating “not at all fatigued” and 0 indicating “very much fatigued.” The overall score ranges from 0 to 52, with higher scores indicating less fatigue. An MID for the FACIT-F total score has been estimated as 3.1 points. As change from baseline in the FACIT-F at week 12 was a key secondary end point in SELECT-PsA1 and SELECT-PsA2, it was included in the multiplicity-controlled analyses.

Short Form (36) Health Survey

Version 2 of the SF-36 is a 36-item, generic health status instrument that has been used extensively in clinical trials in many disease areas.^9,10 It consists of 8 health sub-domains: physical functioning, role physical, bodily pain, general health, vitality, social functioning, role emotional, and mental health. The sub-domains are aggregated to create 2 component summaries: the PCS and the MCS. The PCS constitutes the domains of physical functioning, role physical, bodily pain, and general health and vitality, and has a normative mean value of 50. The scores for each domain range from 0 to 100, with higher scores indicating better health status or HRQoL. The MID for either the PCS or MCS of the SF-36 for the change from baseline is typically between 2.5 and 5 points.^40-42 Leung and colleagues reported MIDs of change scores of 3.74 and 1.77 in PsA patients treated with TNF inhibitors for PCS and MCS, respectively.⁴³ Although the SF-36 was conducted, only the change from baseline in PCS at week 12 was a key secondary end point in SELECT-PsA1 and SELECT-PsA2; it was therefore included in the multiplicity-controlled analyses.

EuroQol 5-Dimensions 5-Levels Questionnaire

The EQ-5D-5L is a generic instrument for HRQoL evaluation. It has been validated in a diverse patient population; however, no studies specifically validating EQ-5D-5L in patients with PsA were identified. It may be applied to a wide range of health conditions and treatments. The 5-level version consists of a 5-dimensional descriptive system and the EuroQol VAS. The descriptive system comprises the following 5 dimensions of health status: mobility, self-care, usual activities, pain/discomfort and anxiety/depression, each with 5 levels. A level 1 response represents “no problems” and level 5 “extreme problems” or “unable to perform.” Results from the EQ-5D-5L descriptive system can be converted into a single index score. A score of 0 represents the health state of “dead” and 1.0 reflects “perfect health.” Negative scores are also possible for those health states that society (not the individual patient) considers to be “worse than dead.” The EuroQol VAS records the respondent’s self-rated health on a vertical line on which the end points are labelled 0 (“the worst health you can imagine”) and 100 (“the best health you can imagine”). The EQ-5D-5L index and VAS scores can be summarized and analyzed as continuous data. In the SELECT-PsA1 and SELECT-PsA2 studies, the UK scoring algorithm was used, with the patient’s preference for the current health state estimated through the index score from published algorithms. The range is between –0.285 and 1, with higher scores indicating better health.^9,10 The MID estimates for the index score in a Canadian population have a summarized mean of 0.056 (SD = 0.011).⁴⁴ The change from baseline in the EQ-5D-5L index value and VAS up to week 24 was an additional outcome in both PsA1 and PsA2 and therefore was not included in the multiplicity-controlled analyses.

Psoriasis Area Severity Index

A widely used instrument in psoriasis trials, the PASI assesses and grades the severity of psoriatic lesions and the patient’s response to treatment. Severity of psoriasis is measured at 4 anatomic sites (head, upper extremities, trunk, and lower extremities) and each is assessed for erythema, induration, and desquamation using a 5-point scale. Scores for each site range from 0 (no symptoms) to 4 (very marketed symptoms). The extent of lesions in a given anatomic site is also ranked from 0 (no involvement) to 6 (90% to 100% involvement). The PASI produces a numeric score ranging from 0 to 72, with the highest score representing complete erythroderma of the most severe degree possible. In general, a PASI score of 3 or less represents mild disease, scores above 3 and up to 15 represent moderate disease, and scores above 15 are considered indicative of severe disease.

A PASI 75 response is the current benchmark for most clinical trials in psoriasis and the criterion for efficacy of new psoriasis treatments approved by the FDA.⁴⁵ The proportion of patients achieving a PASI 75 response at week 16 (restricted to patients with baseline psoriatic lesions involving ≥ 3% BSA) was a key secondary end point in both SELECT-PsA1 and SELECT-PsA2; it was therefore included in the multiplicity-controlled analyses. At week 24, a 75% reduction in Psoriasis Area Severity Index score (PASI 75), 90% reduction in Psoriasis Area Severity Index score (PASI 90), or 100% reduction in Psoriasis Area Severity Index score (PASI 100) from baseline was an additional outcome and was not included in the multiplicity-controlled analyses.

Static Investigator Global Assessment

The sIGA evaluates the severity of psoriasis. It is a 5-point score ranging from 0 (clear) to 4 (severe), based on the investigator’s assessment of the average elevation, erythema, and scaling of all psoriatic lesions. As a static assessment, the patient’s disease state at the time of assessment is not compared to any of the previous measures. A lower score indicates less-severe psoriasis (i.e., less body coverage).

In both SELECT studies, the proportion of patients achieving an sIGA score of 0 or 1 and an improvement of at least 2 points from baseline at week 16 was a key multiplicity-adjusted secondary end point, and was calculated in patients with baseline sIGA scores of 2 or higher.

Self-Assessment of Psoriasis Symptoms

The SAPS questionnaire contains 11 symptom-focused items. Patients self-assess symptoms of psoriasis including pain, itching, redness, scaling, flaking, bleeding, burning, stinging, tenderness, pain due to skin cracking, and joint pain. Each item is scored from 0 (least severe) to 10 (most severe), with the sum of scores forming a total score that ranges from 0 to 110. No reported MID was identified for patients with PsA. As change from baseline in the SAPS questionnaire at week 16 was a key secondary end point in SELECT-PsA1 and SELECT-PsA2, it was included in the multiplicity-controlled analyses.

Leeds Enthesitis Index

The LEI was developed specifically for PsA diagnoses. It measures enthesitis at 6 sites: lateral epicondyle, left and right; medial femoral condyle, left and right; and Achilles tendon insertion, left and right. Tenderness on examination at each site was assigned a score of 0 (absent) or 1 (present); the results from each site were then added to produce an overall score ranging from 0 to 6. Resolution of enthesitis sites was defined as an LEI of 0. An MID was not identified from the literature. The proportion of patients achieving resolution of enthesitis at week 24 (measured in patients with baseline LEI > 0) was a key secondary outcome in SELECT-PsA1 and was therefore included in the multiplicity-controlled analyses. However, in SELECT-PsA2, this end point was considered additional and was not included in the multiplicity-controlled analyses.

Leeds Dactylitis Index

Presence of dactylitis in all 20 digits was assessed using the LDI. Tenderness and circumference of digits were measured with a dactylometer, with presence of a dactylitic digit defined as at least 1 affected and tender digit with a circumference increase of at least 10% over the reference digit. The reference circumference is either the unaffected contralateral digit or derived from a reference table. A digit score is calculated for each of the 20 digits (hands and feet), with a score of 0 representing an unaffected digit. The sum of scores over all 20 digits forms the LDI. Resolution of dactylitis was defined as an LDI of 0. An MID was not identified from the literature. The proportion of patients achieving resolution of dactylitis at week 24 (measured in patients with a baseline LDI > 0) was a key secondary outcome in SELECT-PsA1 and was therefore included in the multiplicity-controlled analyses. However, in SELECT-PsA2, this end point was considered additional and was not included in the multiplicity-controlled analyses.

Bath Ankylosing Spondylitis Disease Activity Index

The BASDAI contains 6 questions pertaining to the 5 major symptoms (domains) of axial disease activity: fatigue, spinal pain, joint pain and/or swelling, areas of localized tenderness, and morning stiffness. A continuous NRS of 0 to 10 is used to measure these disease activities, with 0 indicating no problem and 10 indicating the worst problem. A BASDAI score from 0 to 10 is then calculated. Scores of 4 or greater suggest suboptimal control of disease. The MID for the BASDAI has been determined as a change of −1.96 on the 10-point BASDAI scale.⁴⁶ The change in BASDAI from baseline up to week 24 (in patients who had psoriatic spondylitis at baseline) was assessed as an additional efficacy outcome in both SELECT trials. This outcome was not included in the multiplicity-controlled analyses.

Sharp/van der Heijde Score

The Sharp method as modified by van der Heijde for PsA was used to assess and score radiographic outcomes. The SHS allows for the assessment of 2 different aspects of joint damage in PsA: articular erosions and joint space narrowing in the hands and feet. A total SHS score is obtained by adding together scores for erosions and joint space narrowing in both hands and feet.

Erosions were assessed in 20 locations in each hand (graded from 0 to 5), and 6 joints in each foot (graded from 0 to 10). Joint space narrowing was assessed in 20 joints in each hand and in 6 joints in each foot (graded 0 to 4). A score of 0 denoted no erosion or narrowing. The range of scores is 0 to 320 for erosions, and 0 to 208 for joint space narrowing. The total score ranges from 0 to 528. A higher score indicates greater radiographic damage. Radiographic images were assessed and scored centrally by 2 independent reviewers who were blinded to patient details and treatment allocation, with the SHS calculated using the score of the 2 closest reads. An MID for SHS is unknown in patients with PsA. Only SELECT-PsA1 assessed change from baseline in SHS at week 24. As this was a key secondary efficacy outcome in PsA1, it was included in the multiplicity-controlled analyses.

Safety

Adverse events, SAEs, severe adverse events (≥ grade 3), AEs of special interest, and WDAEs were recorded in in SELECT-PsA1 and SELECT-PsA2. Two sets of safety analyses were planned, the first by week 24, and a long-term safety analysis.

Unless otherwise noted, all AEs presented were treatment-emergent. The definition of TEAEs and treatment-emergent deaths were those with an onset on or after the first dose of the study drug but no more than 30 days after the last dose of upadacitinib or placebo (oral and subcutaneous) and no more than 70 days after the last dose of adalimumab. Reported AEs did not necessarily have a causal relationship with treatment. Several AEs of special interest were specified in the protocol: serious infections, opportunistic infections, malignancy, hepatic disorders, gastrointestinal perforations, anemia, neutropenia, lymphopenia, herpes zoster, creatine phosphokinase elevation, renal dysfunction, tuberculosis, MACEs, and thrombotic and embolic events.

Statistical Analysis

An unblinded analysis was conducted by an independent data monitoring committee after all patients completed week 24. A final analysis is planned after all patients have completed the last visit of the study (period 2). In SELECT-PsA1, an interim utility analysis of unblinded efficacy data was also performed by the committee after 600 patients completed week 12.

Power analysis in SELECT-PsA1 indicated that 1,640 patients randomized 2:2:2:1:1 to upadacitinib 15 mg, upadacitinib 30 mg, adalimumab 40 mg, and placebo would provide at least 90% power to detect a 20% difference in ACR20 response rates at week 12 versus placebo, assuming a 30% placebo response rate, a 10% dropout, and a 2-sided significance level of 0.025. This sample size also provided at least 85% power to test the noninferiority of upadacitinib versus adalimumab, assuming 50% ACR20 response rates for adalimumab and upadacitinib, and 30% ACR20 response rates for placebo. At least 90% power was also to be provided for the majority of key secondary end points. In SELECT-PsA2, a sample size of 630 was thought to provide at least 90% power to detect a 20% difference in ACR20 response rates, with an assumption of a 20% response rate in the placebo group, at a 2-sided alpha of 0.025, and accounting for a 10% dropout rate. This sample size also provided at least 90% power for the majority of key secondary end points. It was not clear what the basis was for the assumed ACR20 responses in either intervention or placebo groups. No rationale was provided regarding the assumption for the sample size.

Treatment comparisons of categorical efficacy outcomes (e.g., ACR20, ACR50, and ACR70) were reported as point estimates and 95% CIs using normal approximations for response rates of each randomized treatment group. In both trials, treatment comparisons were made between each upadacitinib dose and the combined placebo group using the Cochran-Mantel-Haenszel test, adjusting for the main stratification factor of current DMARD use (yes/no); P values were also constructed using the Cochran-Mantel-Haenszel test. The primary analyses for all continuous efficacy outcomes (e.g., HAQ-DI scores) at the specified time points included in the multiplicity-adjustment plan were based on the mixed model for repeated measures (MMRM). The model included treatment, visit, treatment-by-visit interaction, and the stratification factor of current DMARD use (yes/no) as fixed factors, and the continuous baseline measurements as covariates. The LS mean and 95% CI for each randomized treatment group, the LS mean treatment difference (with associated 95% CI), and the P value between each upadacitinib dose group and the combined placebo group were provided. In SELECT-PsA1, superiority of upadacitinib versus adalimumab in change from baseline in patient’s assessment of pain NRS and change from baseline in HAQ-DI at week 12, were also tested using the same MMRM model. In SELECT-PsA1, change from baseline in SHS at week 24 was also assessed using the analysis of covariance model with treatment and the stratification factor of current DMARD use (yes/no) as the fixed factors and the corresponding baseline values as the covariates.

For SELECT-PsA1 and SELECT-PsA2, all efficacy analyses were conducted in the full analysis set (FAS) population. For the primary end point measuring the proportion of patients with ACR20 at week 12 and other binary end points, nonresponder imputation (NRI) was used for missing values. For the NRI approach, patients who were missing binary response data at a time point of interest were considered as nonresponders for that visit, and patients who had discontinued the study drug before the time point of interest were defined as nonresponders for all subsequent visits. For MDA at week 24 in both trials, resolution of enthesitis at week 24, and resolution of dactylitis at week 24 in SELECT-PsA1, in addition to the NRI data handling, patients who met the rescue criteria (i.e., those not achieving an improvement of at least 20% in either or both TJC and SJC at both week 12 and week 16) were also treated as nonresponders. In both studies, continuous data were analyzed through an MMRM approach that included observed data at all visits. Data collected after premature discontinuation of the study drug were excluded. The parameter estimations for the mixed model assumed that data were missing at random and used the method of restrictive maximum likelihood. Missing data in the radiographic outcome of SHS in SELECT-PsA1 were based on linear extrapolation, with the missing X-ray at the time point of interest after rescue or premature discontinuation imputed by assuming a linear relationship across visits. As-observed handling of missing data was used for supplementary analysis of binary and continuous end points. In such analyses, all observed data regardless of premature discontinuation of the study drug or use of rescue medication were included, and values for missing evaluations were excluded from the analysis for the particular visit. An additional tipping-point analysis was performed for ACR20 and HAQ-DI using multiple imputations to explore various missing data assumptions, including missing not at random.

In SELECT-PsA1, noninferiority of upadacitinib versus adalimumab was assessed for ACR20 at week 12. This was performed using the Koch 3-arm test statistic to demonstrate that upadacitinib preserves at least 50% of the placebo-subtracted adalimumab effect. That is, noninferiority was claimed if the multiplicity-adjusted lower confidence limit for the ratio of placebo-subtracted upadacitinib ACR20 rate compared to the placebo-subtracted adalimumab ACR20 rate was at least 50%. No clear justification for the noninferiority margin was provided in the Clinical Study Report. The noninferiority margin was changed in Protocol Amendment 2; the previous noninferiority margin of 15% was derived from prior data (e.g., a meta-analysis of placebo-subtracted effect of anti-TNF biologics in ACR20 among patients who had inadequate response to csDMARDs). In SELECT-PsA1, superiority of upadacitinib versus adalimumab was also planned and hierarchically tested for ACR20 at week 12, as well as change from baseline in patient’s assessment of pain NRS and HAQ-DI at week 12.

In SELECT-PsA1, the overall type I error rate of the primary and 14 ranked key secondary end points was strongly controlled using a 2-part, sequential, graphical multiple-testing procedure. The testing used the end-point sequence of the primary end point followed by the ranked key secondary end points in the order outlined in the Outcomes section and began with testing the primary end point using alpha/2 (where alpha = 0.0499) for each dose. For the interim futility analysis, an alpha of 0.0001 was spent, and the final analysis at week 24 database lock was performed at the 0.0499 level. Continued testing followed a pre-specified alpha transfer path that included downstream transfer along the end-point sequence within each dose as well as cross-dose transfer between the low (15 mg) and high (30 mg) upadacitinib doses. A superiority-testing sequence at level alpha was subsequently performed if all previous end points were significant. The ranked key secondary end points compared upadacitinib versus placebo, unless specified (e.g., superiority testing versus adalimumab). Adjusted P values were provided for the primary and ranked key secondary end points based on the testing procedure.

In SELECT-PsA2, the overall type I error rate of the primary and 7 ranked key secondary end points was also strongly controlled using a graphical multiple-testing procedure. The testing used the end-point sequence of primary end point followed by the ranked key secondary end points in the order outlined in the Outcomes section, and began with testing the primary end point using an alpha of 0.025 for each dose. Continued testing followed a pre-specified alpha transfer path that included downstream transfer along the end-point sequence within each dose as well as cross-dose transfer between the low (15 mg) and high (30 mg) upadacitinib doses. Sequential noninferiority or superiority testing was not performed in SELECT-PsA2. Adjusted P values were provided for the primary and ranked key secondary end points based on the testing procedure.

In both studies, subgroup analysis was planned for the primary efficacy outcome for age, sex, body mass index, race, geographic region, duration of PsA diagnosis, baseline hs-CRP, and current use of non-bDMARDs. SELECT-PsA1 also included a subgroup factor for number of prior non-bDMARDs (≤ 1, > 1) whereas SELECT-PsA2 included a subgroup factor for the number of prior failed bDMARDs ( = 1, > 1). Subgroup analyses were not adjusted for multiple testing. The CADTH systematic review protocol included subgroups by PsA disease severity at baseline, previous exposure to bDMARDs (treatment-naive versus -experienced), and concomitant treatment with a non-bDMARD. Inclusion criteria for SELECT-PsA1 and SELECT-PsA2 allowed examination of treatment effects separately for biologic-naive (SELECT-PsA1) and biologic-experienced (SELECT-PsA2). Neither study provided subgroup data based on disease severity at baseline.

Analysis Populations

The analysis populations were defined in the same way in both studies.

The FAS consisted of all randomized patients who received at least 1 dose of study medication. The FAS was used for all efficacy and baseline analyses.

The per-protocol set (PPS) consisted of patients in the FAS who did not have any major protocol deviations that could have had an impact on the primary efficacy end point up to week 12 in period 1 of the study. The PPS was used in the additional analysis of the primary efficacy end point to evaluate the impact of major protocol deviations.

The safety analysis set consisted of all patients who received at least 1 dose of study drug. Patients were analyzed “as treated,” regardless of the treatment group to which they were randomized. Specifically, patients were analyzed according to the treatment received during the majority of their drug exposure time during the analysis period.

Results

Patient Disposition

Patient disposition is summarized in Table 9. In SELECT-PsA1, a total of 1,282 patients were randomized to upadacitinib 15 mg, adalimumab 40 mg, or placebo at baseline. In SELECT-PsA2, a total of 423 patients were randomized to upadacitinib 15 mg or placebo at baseline.

Overall, the number of premature discontinuations at week 24 was higher in the placebo groups (10.4% and 20.3% in SELECT-PsA1 and SELECT-PsA2, respectively) than in the upadacitinib 15 mg group (7.2% and 12.3% in SELECT-PsA1 and SELECT-PsA2, respectively), and adalimumab 40 mg every other week (8.4%). Adverse events and patient decisions to withdraw were reported most frequently as the primary reason to discontinue treatment, although lack of efficacy was the most frequent reason in the SELECT-PsA2 placebo group.

Exposure to Study Treatments

In SELECT-PsA1, the mean duration of exposure during the double-blind treatment period (up to week 24) was similar in all groups: 161.7 days (SD = 28.57) for upadacitinib, 161.4 days (SD = 28.24) for adalimumab, and 158.5 days (SD = 32.22) for placebo. Treatment compliance was high (mean > 98%) across all treatment groups. In SELECT-PsA2, the mean duration of exposure during the double-blind treatment period (week 24) was longer in the upadacitinib 15 mg daily group (159.1; SD = 29.10) compared to the placebo group (146.7; SD = 48.24). Treatment compliance was also high (mean > 98%) in both treatment groups. Overall, a higher proportion of patients in the SELECT-PsA1 trial received concomitant treatment with a non-bDMARD compared to patients enrolled in SELECT-PsA2.

Efficacy

Numerous end points were measured at various time points across both SELECT studies. In this review, a total of 20 end points were presented from PsA1, and 15 end points were presented from PsA2. Only the efficacy outcomes identified in the review protocol (Table 5) and those indicated as important by patient groups are reported below. Data tables are only presented for end points that were included in the multiplicity-controlled analyses (primary and major secondary efficacy outcomes); data tables for end points that were not part of the statistical testing hierarchy are presented in Appendix 3, as indicated in the respective sections. In the data tables, P values are presented only for end points that were part of the multiplicity-adjusted testing hierarchy; P values are not reported for end points that fell below failure of the hierarchical analysis. A detailed discussion and critical appraisal of the outcome measures is provided in Appendix 4.

In the multiplicity-controlled analyses of SELECT-PsA1, statistically significant differences for the comparisons between upadacitinib 15 mg once daily and placebo were observed for the primary end point of ACR20 at week 12 and most key secondary end points. Upadacitinib 15 mg once daily did not demonstrate statistically significant superiority over adalimumab in ACR20 at week 12. As such, statistically significant differences in the key secondary end points lower in the testing hierarchy were not tested because of the multiplicity control strategy. The affected end points were the proportion of patients achieving resolution of dactylitis at week 24 (versus placebo), change from baseline in patient’s assessment of pain NRS at week 12 (versus adalimumab), change from baseline in HAQ-DI at week 12 (versus adalimumab), and change from baseline in SAPS at week 16 (versus placebo).

In the multiplicity-controlled analyses of SELECT-PsA2, statistically significant differences for the comparisons between upadacitinib 15 mg and placebo were observed for the primary end point of ACR20 at week 12 and all major secondary end points.

Clinical Responses in Psoriatic Arthritis Symptoms

ACR Responses of 20%, 50%, and 70%

The proportion of patients achieving ACR20 at week 12 was the primary end point in both SELECT-PsA1 and SELECT-PsA2. In SELECT-PsA1, a statistically significantly greater proportion of bDMARD-naive patients in the upadacitinib treatment group achieved ACR20 at week 12 compared with placebo (70.6% for upadacitinib 15 mg versus 36.2% for placebo; P < 0.0001). In SELECT-PsA2, a statistically significantly greater proportion of bDMARD-experienced patients in the upadacitinib treatment group achieved an ACR20 at week 12 compared with placebo (56.9% for upadacitinib 15 mg daily versus 24.1% for placebo; P < 0.0001) (Table 11).

Supplemental analyses using as-observed handling of missing data on the FAS and NRI on the PPS were consistent with the primary analysis. Subgroup analysis by current use of non-bDMARDs, number of prior non-bDMARDs for SELECT-PsA1, and number of failed bDMARDs for SELECT-PsA2 were also consistent with the primary analysis. However, these analyses were not included in the hierarchical statistical analysis approach and should be considered inconclusive because of the potential for inflated type I error.

In SELECT-PsA1, the proportion of patients achieving an ACR20 at week 12 with upadacitinib treatment compared to adalimumab was tested for noninferiority and superiority as key secondary end points. A greater proportion of patients in the upadacitinib treatment group achieved an ACR20 response at week 12 compared with adalimumab (70.6% for upadacitinib 15 mg once daily versus 65.0% for adalimumab 40 mg every other week; between-group difference of 5.6%; 95% CI, −0.6 to 11.8). In the noninferiority analysis, upadacitinib 15 mg daily was noninferior to adalimumab 40 mg every other week. The percentages of adalimumab effect preservation, as calculated by (upadacitinib − placebo)/(adalimumab − placebo), was 119.4% (95% CI, 98.0 to 147.9); the lower bound of the 95% CI exceeded the pre-specified noninferiority ratio of at least 50% of the placebo-subtracted adalimumab effect. The analysis of the PPS showed consistent results for both response-rate difference and adalimumab effect preservation. In the subsequent testing of superiority, upadacitinib 15 mg once daily was not found to be superior to adalimumab 40 mg every other week as it did not meet statistical significance for superiority.

In both SELECT-PsA1 and PsA2, additional exploratory end points showed that, numerically, a greater proportion of patients treated with upadacitinib achieved an ACR50 or ACR70 at week 12 compared to patients who were treated with placebo. In PsA1, the response rates were numerically similar between upadacitinib and adalimumab treatment groups. Furthermore, ACR responses of 20%, 50%, and 70% at week 24 were maintained, with a numerically higher proportion of patients treated with upadacitinib 15 mg achieving a response compared to patients who received placebo (PsA1 and PsA2) or adalimumab (PsA1). As these end points were not part of the multiplicity-controlled analyses, no statistical comparisons can be made.

According to the clinical expert consulted for this review, the between-group differences in ACR20 in the FAS population in SELECT-PsA1 and PsA2 are considered clinically important.

Minimal Disease Activity

In both SELECT-PsA1 and SELECT-PsA2, a responder or MDA achievement was met when 5 out of 7 outcome measures (involving TJC68, SJC66, PASI, or BSA psoriasis, patient’s assessment of pain, patient’s global assessment of disease activity, HAQ-DI, and LEI) were fulfilled. In SELECT-PsA1, a statistically significantly higher proportion of patients in the upadacitinib treatment group achieved MDA at week 24 compared with placebo: 36.6% for upadacitinib 15 mg and 12.3% for placebo (between-group difference of 24.3%; 95% CI, 18.8 to 29.8; P = 0.0004). Similarly, in SELECT-PsA2 at week 24, a statistically significantly higher proportion of patients in the upadacitinib treatment groups achieved an MDA compared with placebo: 25.1% for upadacitinib 15 mg and 2.8% for placebo (between-group difference of 22.3%; 95% CI, 16.0 to 28.6; P < 0.0001) (Table 12).

Modified Psoriatic Arthritis Response Criteria

In SELECT-PsA1, a higher proportion of patients treated with upadacitinib achieved a modified PsARC response at week 24 compared to patients randomized to adalimumab or placebo: 83.7% for upadacitinib 15 mg, 76.6% for adalimumab 40 mg, and 59.3% for placebo (Appendix 3,Table 48). Similarly, in SELECT-PsA2, a higher proportion of patients treated with upadacitinib 15 mg achieved a modified PsARC response at week 24 compared with placebo: 68.2% for upadacitinib 15 mg and 36.3% for placebo. This end point was not part of the multiplicity-controlled statistical analyses.

Measure of Function and Disability

Health Assessment Questionnaire–Disability Index

In SELECT-PsA1, a statistically significantly greater reduction from baseline in the HAQ-DI score was achieved in biologic-naive patients in the upadacitinib treatment groups compared with the placebo group at week 12 (Table 13). The differences in change from baseline between upadacitinib 15 mg and placebo was −0.28 (95% CI, −0.35 to −0.22; P < 0.0001). Similarly, in SELECT-PsA2, a statistically significantly greater reduction from baseline in the HAQ-DI score was achieved in bDMARD-experienced patients in the upadacitinib treatment groups compared with placebo at week 12. The differences in change from baseline between upadacitinib 15 mg and placebo was −0.21 (95% CI, −0.30 to −0.12; P < 0.0001). While in both studies, the between-group differences comparing upadacitinib and placebo in the improvement of the HAQ-DI score did not exceed the MID for the HAQ-DI (estimated in the literature at 0.35), the proportions of patients in the SELECT-PsA1 study who achieved a clinically meaningful improvement in HAQ-DI at week 12 were 33.4%, 47.2%, and 57.9% in the placebo, adalimumab 40 mg, and upadacitinib 15 mg treatment groups, respectively. In SELECT-PsA2, the proportions of patients who achieved a clinically meaningful improvement in HAQ-DI at week 12 were 27.2% and 44.6% in the placebo and upadacitinib 15 mg treatment groups, respectively. The clinical expert consulted for this review indicated that the between-group differences in the change from baseline in HAQ-DI were clinically relevant.

Work Productivity

In SELECT-PsA1 and PsA2, numerically greater reductions in work or activity impairment due to disease as measured by the WPAI questionnaire were observed for the upadacitinib group compared with placebo or adalimumab (in SELECT-PsA1) at week 24 (Appendix 3, Table 49). The between-group differences in change from baseline with upadacitinib compared to placebo or adalimumab did not exceed the estimated MID threshold of an improvement of at least 20% for presenteeism and at least 20% for activity impairment. Not all randomized patients participated in this evaluation; absenteeism, presenteeism, and work impairment were measured only in patients who were employed. For these 3 scores, the number of patients evaluated in each treatment group ranged from 42% to 48% in SELECT-PsA1 and 31% to 44% in SELECT-PsA2. More patients were included in the assessment of activity impairment (87% to 90% of patients in each SELECT-PsA1 treatment group; 79% to 87% of patients in each SELECT-PsA2 treatment group). Also, without a confirmed MID for the WPAI instrument, it remains unclear whether differences were clinically meaningful. This was an exploratory variable in both studies and was not included in the multiplicity-controlled statistical analyses.

Measurement of Psoriatic Arthritis Symptoms

Pain

In SELECT-PsA1 and SELECT-PsA2, the mean change in patient’s assessment of pain NRS scores decreased (improved) from baseline to week 12 for all treatment arms (Table 14), and the LS mean reduction in score was the same for upadacitinib and adalimumab in SELECT-PsA1. Specifically, the between-group comparison favoured upadacitinib numerically compared to placebo in both studies, with an LS mean difference of −1.3 (95% CI, −1.6 to −1.0) in SELECT-PsA1 and −1.4 (95% CI, −1.9 to −1.0) in SELECT-PsA2. However, neither treatment was favoured when upadacitinib was compared to adalimumab in SELECT-PsA1 (LS mean difference of 0.0; 95% CI, −0.3 to 0.3). In both trials, the comparison between upadacitinib and placebo was not included in the hierarchical statistical analysis. In SELECT-PsA1, the end point exploring comparisons of upadacitinib and adalimumab was ranked after the point at which the hierarchical analysis failed and was stopped, and no statistical comparisons can be made. An MID was not identified for patients with PsA.

Fatigue

In SELECT-PsA1, a statistically significantly greater improvement from baseline in the FACIT-F score was seen in biologic-naive patients in the upadacitinib treatment groups compared with placebo at week 12 (Table 15). The differences in change from baseline between upadacitinib 15 mg and placebo was 3.5 (95% CI, 2.4 to 4.7; P = 0.0004). Similarly, in SELECT-PsA2, a statistically significantly greater improvement from baseline in the FACIT-F score was seen in bDMARD-experienced patients in the upadacitinib treatment groups compared with placebo at week 12. The differences in change from baseline between upadacitinib 15 mg and placebo was 3.7 (95% CI, 2.0 to 5.4; P < 0.0001). The between-group differences in the improvement in FACIT-F score at week 12 exceeded the estimated MID (3.1 points) in both studies.

Health-Related Quality of Life

Short Form (36) Health Survey

In SELECT-PsA1 at week 12, patients in the upadacitinib group reported higher PCS scores compared with placebo on the SF-36 (Table 16). The LS mean difference for the change in scores from baseline was 4.67 for upadacitinib 15 mg versus placebo (95% CI, 3.67 to 5.67; P = 0.0004). Similarly, in SELECT-PsA2 at week 12, patients in the upadacitinib treatment reported higher PCS scores compared with placebo on the SF-36. The LS mean difference was 3.52 for upadacitinib 15 mg versus placebo (95% CI, 2.07 to 4.98; P < 0.0001). In both studies, the difference versus placebo was more modest for the MCS than for the PCS, with an LS mean difference for upadacitinib 15 mg versus placebo of 1.70 (95% CI, 0.58 to 2.82) in SELECT-PsA1 and 2.98 (95% CI, 1.44 to 4.5) in SELECT-PsA2. The between-group differences achieved the estimated MID (typically between 2.5 and 5 points) for the PCS component in both studies, but the MCS component was only met in PsA2. As reported by Leung et al.,⁴³ when using the MID of 3.74 (for PCS) and 1.77 (for MCS) in patients treated with TNF inhibitors, the between-group differences in patients enrolled in SELECT-PsA2 met the threshold only for the MCS component. Only the PCS component of SF-36 was part of the multiplicity-controlled statistical analyses.

EuroQol 5-Dimensions 5-Levels Questionnaire

In SELECT-PsA1, there were greater improvements in the EQ-5D-5L utility index and VAS scores from baseline to week 24 in the upadacitinib treatment group compared to patients randomized to adalimumab treatment or placebo. The LS mean differences in the utility index were 0.09 (95% CI, 0.06 to 0.11) for upadacitinib 15 mg versus placebo and 0.03 (95% CI, 0.00 to 0.05) for upadacitinib 15 mg versus adalimumab 40 mg. The LS mean differences in the VAS score were 10.9 (95% CI, 8.0 to 13.7) for upadacitinib 15 mg versus placebo and 2.8 (95% CI, 0.0 to 5.6) for upadacitinib 15 mg versus adalimumab 40 mg (Appendix 3, Table 50).

In SELECT-PsA2, there were greater improvements in the EQ-5D-5L utility index and VAS scores from baseline to week 24 in the upadacitinib treatment group compared to the placebo group. The LS mean differences in the utility index was 0.10 (95% CI, 0.06 to 0.14) for upadacitinib 15 mg versus placebo. The LS mean differences in the VAS score was 6.8 (95% CI, 2.5 to 11.1) for upadacitinib 15 mg versus placebo (Appendix 3, Table 50).

For the comparison of upadacitinib to placebo in both SELECT-PSA1 and SELECT-PsA2, the LS mean between-group differences in the EQ-5D-5L utility index reached the MID threshold identified in the literature for the general Canadian population (summarized mean of 0.056; SD = 0.011). This end point was not included in the multiplicity-controlled testing hierarchy.

Measurement of Skin Disease

Psoriasis Area and Severity Index

Patients considered to be PASI 75 responders were those with a 75% improvement from baseline scores. Only patients with a BSA involvement of at least 3% at baseline had a PASI assessment.

In SELECT-PsA1, the proportion of patients achieving a PASI 75 response in the upadacitinib treatment group compared to the placebo group at week 16 was statistically significantly higher: 62.6% for upadacitinib 15 mg and 21.3% for placebo (41.3% difference; 95% CI, 32.8 to 49.8; P < 0.0001). In SELECT-PsA2, the proportion of patients achieving a PASI 75 response in the upadacitinib treatment group compared to placebo at week 16 was also statistically significantly higher: 52.3% for upadacitinib 15 mg and 16.0% for placebo (36.3% difference; 95% CI, 25.6 to 46.9; P < 0.001) (Table 17).

In both SELECT-PsA1 and SELECT-PsA2, additional exploratory end points showed that, numerically, a greater proportion of patients treated with upadacitinib achieved a PASI 75, PASI 90, or PASI 100 score at week 24 compared to patients who were treated with placebo. In SELECT-PsA1, the response rates were numerically similar between upadacitinib and adalimumab treatment groups, although a slightly higher proportion of patients in the adalimumab group achieved PASI 90 or PASI 100 at week 24 (Appendix 3, Table 51).

The clinical expert consulted for this review indicated that the between-group differences in PASI 75 were clinically relevant.

Static Investigator Global Assessment of Psoriasis

Only patients with an sIGA score of at least 2 at baseline, and an improvement of at least 2 points from baseline at week 16 were assessed (Table 18).

In SELECT-PsA1, the proportion of patients achieving a response (an sIGA of psoriasis score of 0 or 1 and an improvement of at least 2 points from baseline) in the upadacitinib treatment group compared to placebo was statistically significantly higher at week 16: 41.9% for upadacitinib 15 mg and 10.9% for placebo (between-group difference 31.1%; 95% CI 24.7 to 37.5; P < 0.0001). In SELECT-PsA2, the proportion of patients achieving response in the upadacitinib treatment group compared to placebo was also statistically significantly higher at week 16: 36.8% for upadacitinib 15 mg and 9.2% for placebo (between-group difference 27.6%; 95% CI, 19.2 to 36.1; P < 0.0001).

Self-Assessment of Psoriasis Symptoms

In SELECT-PsA1 at week 16, patients in the upadacitinib group reported greater reductions in SAPS score compared with placebo (Table 19). The LS mean difference was −17.1 (95% CI, −19.6 to −14.6) for upadacitinib 15 mg versus placebo. Testing for superiority of upadacitinib compared to placebo was part of the multiplicity-adjusted analyses in SELECT-PsA1; however, it was ranked after the point at which the hierarchical analysis failed and was stopped, and no statistical comparisons can be made. In SELECT-PsA2 at week 16, patients in the upadacitinib treatment reported greater reductions in SAPS scores compared to placebo. The LS mean difference was −22.9 for upadacitinib 15 mg versus placebo (95% CI, −27.4 to −18.4; P < 0.0001).

Measurement of Other Musculoskeletal Disease

Enthesitis

The resolution of enthesitis, defined as an LEI score of 0, was included in the hierarchical statistical analysis only in SELECT-PsA1, and was measured in patients who had a baseline LEI greater than 0 (Table 20). Using multiplicity-controlled analyses, the proportion of patients in SELECT-PsA1 achieving resolution of enthesitis was statistically significantly higher at week 24 in the upadacitinib treatment group compared to placebo: 53.7% for upadacitinib 15 mg and 32.4% for placebo (between-group difference 21.3%; 95% CI, 13.0 to 29.7; P = 0.0004). In SELECT-PsA2, a numerically higher proportion of patients in the upadacitinib treatment group achieved resolution of enthesitis at week 24 compared to patients in the placebo group: 42.9% for the upadacitinib 15 mg group, 15.3% for placebo (between-group difference 27.6%; 95% CI, 17.3 to 37.8). This end point was not part of the multiplicity-controlled statistical analyses in PsA2.

Dactylitis

The resolution of dactylitis, defined as an LDI of 0, was included in the hierarchical statistical analysis in SELECT-PsA1 only, and was measured in patients who had a baseline LDI of greater than 0 (Table 21). In SELECT-PsA1, the proportion of patients achieving resolution of dactylitis, defined as an LDI of 0, was numerically higher at week 24 in the upadacitinib treatment group compared to placebo: 76.5% for upadacitinib 15 mg and 39.7% for placebo (between-group difference 36.8%; 95% CI, 25.7 to 47.9). This end point was included in the hierarchical statistical analysis in SELECT-PsA1; however, because it was ranked after the point at which the hierarchical analysis failed and was stopped, no statistical comparisons can be made. In SELECT-PsA2, a numerically higher proportion of patients in the upadacitinib treatment group achieved resolution of dactylitis at week 24 compared to patients in the placebo group: 58.2% for upadacitinib 15 mg and 28.1% placebo (between-group difference 30.1%; 95% CI; 13.0 to 47.1). This end point was not part of the multiplicity-controlled statistical analyses in PsA2.

Axial Arthritis

In SELECT-PsA1 and SELECT-PsA2, change in axial disease was assessed in patients with psoriatic spondylitis at baseline. The number of patients included in this assessment was approximately 30% of those randomized into each treatment group in SELECT-PsA1, and approximately 35% in patients randomized to each treatment in SELECT-PsA2 (Appendix 3, Table 52). In SELECT-PsA1, greater improvement in BASDAI score from baseline to week 24 was observed in the upadacitinib treatment group compared to adalimumab or placebo. The LS mean difference between upadacitinib and placebo in the change in scores was −1.42 (95% CI, −1.94 to −0.90), whereas the LS mean difference between upadacitinib and adalimumab was −0.57 (95% CI, −1.09 to −0.05). In PsA2, greater improvement in the BASDAI score from baseline to week 24 was observed in the upadacitinib 15 mg group compared to the placebo group. The LS mean difference between upadacitinib and placebo in PsA2 was −1.85 (95% CI, −2.55 to −1.15). The difference between treatment groups in the change in scores from baseline did not achieve the MID identified in the literature (a change of −1.96). However, the change in scores from baseline within each active treatment group (upadacitinib and adalimumab) achieved the estimated MID. This end point was not part of the multiplicity-controlled statistical analyses.

Radiographic Changes

Sharp/van der Heijde Score

Radiographic change, as measured by SHS, was only assessed in SELECT-PsA1 (Table 22). At week 24, the difference in mean change from baseline in SHS was statistically significant for the upadacitinib treatment group compared to placebo (−0.29; 95% CI, −0.44 to −0.14; P = 0.0004). An MID for SHS in patients with PsA is unknown; however, the clinical expert consulted for this review noted that the small changes seen are unlikely to be significantly clinically meaningful to patients, and that it is difficult to observe meaningful radiographic changes within 24 weeks in the study population.

Harms

Only those harms identified in the review protocol are reported below. Table 23 provides detailed harms data.

Adverse Events

In SELECT-PsA1, the percentage of patients with TEAEs overall was numerically higher in the upadacitinib and adalimumab treatment groups compared to placebo. One or more AE was reported in 66.9% of patients treated with upadacitinib 15 mg once daily, 64.8% treated with adalimumab 40 mg every other week, and 59.6% of patients in the placebo group during the double-blind treatment period. In SELECT-PsA2, a similar proportion of patients experienced 1 or more AEs (64.0% of patients in the upadacitinib 15 mg once daily group and 65.6% in the placebo group) during the double-blind treatment period. Generally, the majority of AEs were mild or moderate in severity. The most frequently reported AE in both studies was upper respiratory tract infection. Across the 2 studies, a higher proportion of patients in PsA1 experienced upper respiratory tract infection, as well as elevated blood CPK, alanine aminotransferase, and aspartate aminotransferase levels; a higher proportion of patients in SELECT-PsA2 experienced psoriatic arthropathy, which referred to the worsening of the underlying PsA disease.

Serious Adverse Events

In SELECT-PsA1, the proportion of patients who experienced 1 or more SAEs was similar across all treatment groups (3.3% of upadacitinib 15 mg, 3.7% of adalimumab 40 mg, 3.1% in placebo). No more than 1 patient in any treatment group reported SAEs, with the exception of pneumonia (2 patients in the placebo group) and sepsis (2 patients in the upadacitinib 15 mg group).

In SELECT-PsA2, a greater proportion of patients in the upadacitinib 15 mg (5.7%) experienced an SAE compared to patients in the placebo group (1.9%). No more than 1 patient in any treatment group reported SAEs, with the exception of psoriatic arthropathy, and nephrolithiasis (each of which were reported in 2 patients in the upadacitinib 15 mg group). Psoriatic arthropathy referred to worsening of the underlying PsA disease. Details of the reported SAEs are presented in Table 23.

Withdrawals Due to Adverse Events

In SELECT-PsA1, the percentages of patients with AEs leading to discontinuation of study drug were numerically highest in the adalimumab group and similar between the upadacitinib and placebo groups. Up to week 24, 3.0% of patients treated with upadacitinib 15 mg, 5.1% of patients treated with adalimumab 40 mg, and 3.1% of patients in the placebo group discontinued study treatment due to an AE (Table 23). Most AEs leading to study drug discontinuation were mild to moderate in severity. Four patients in the placebo group, 6 patients in the adalimumab group, and 4 patients in the upadacitinib group experienced SAEs that led to study drug discontinuation. Of these, the SAEs that lead to treatment discontinuation and were thought to be reasonably related to the study drug were sepsis (2 patients) and pyrexia (1 patient) in the upadacitinib 15 mg group, and hepatic encephalopathy (1 patient) and portal vein thrombosis (1 patient) in the adalimumab 40 mg group.

In SELECT-PsA2, the percentage of patients who discontinued study treatment due to an AE was numerically higher in the upadacitinib group. Up to week 24, 7.1% of patients treated with upadacitinib 15 mg and 5.2% of patients in the placebo group discontinued study treatment due to an AE (Table 23). Most AEs leading to study drug discontinuation were mild to moderate in severity. Two patients in the placebo group and 6 patients in the upadacitinib 15 mg group experienced an SAE that led to study drug discontinuation. Of these, the SAEs that lead to treatment discontinuation and were reasonably assumed to be related to the study drug were rapidly progressive osteoarthritis, psoriatic arthropathy, blood loss anemia, acute respiratory failure, pulmonary embolism, cardiorespiratory arrest (1 patient each) in the upadacitinib 15 mg group.

Mortality

During the double-blind period (up to week 24), SELECT-PsA1 and SELECT-PsA2 each recorded a single treatment-emergent death. In both studies, the death occurred in a patient in the placebo group.

Notable Harms

Percentages of patients who experienced a notable harm identified for this review can be found in Table 23. Overall, the proportions of patients experiencing notable harms during the double-blind period were low, and no explicit imbalances between the 2 studies were seen, with the exception of a numerically higher proportion of patients who reported CPK elevation and hepatic disorder in SELECT-PsA1. In SELECT-PsA1, 8.9% of patients treated with upadacitinib 15 mg, 5.6% of patients treated with adalimumab 40 mg, and 1.4% of patients in the placebo group experienced elevated CPK levels. Comparatively, in SELECT-PsA2, 1.9% of patients in each treatment group (upadacitinib 15 mg and placebo) experienced elevated CPK levels. Hepatic disorder was reported in 9.1%, 15.6%, and 3.8% of SELECT-PsA1 patients treated with upadacitinib 15 mg, adalimumab 40 mg, and placebo, respectively. In SELECT-PsA2, 1.9% of patients treated with upadacitinib 15 mg and 1.4% of patients in the placebo group were reported as having hepatic disorder as an AE.