Reimbursement Review

Bimekizumab (Bimzelx)

Sponsor: UCB Canada Inc.

Therapeutic area: Ankylosing spondylitis

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.

Table 1: Background Information of Application Submitted for Review

NOC = Notice of Compliance.

Introduction

Ankylosing spondylitis (AS) is a chronic, inflammatory, and heterogeneous disease that places a significant burden on patients that is driven by pain, fatigue, and stiffness.^1,2 Axial spondyloarthritis (axSpA) encompasses radiographic axial spondyloarthritis (r-axSpA, also known as AS) and nonradiographic axial spondyloarthritis (nr-axSpA). Although nr-axSpA shares several features with AS, advanced sacroiliac joint damage and spine ankylosis are absent.³ Patients with uncontrolled inflammation may progress to irreversible axial structural damage,⁴ spinal fractures, and severe spinal cord injury.^5,6 Patients may also experience extramusculoskeletal manifestations, such as uveitis.^6,7 A population-based study of the incidence and prevalence of AS in 2010 using Ontario provincial health administrative databases found age- and sex-standardized prevalence and incidence rates of 0.213% and 0.015%, respectively.^6,7 AS was estimated to affect 300,000 patients in Canada in 2019.⁸

Nonsteroidal anti-inflammatory drugs (NSAIDs) are the first-line treatment for adult patients with active AS.^4,9,10 After NSAIDs, advanced therapy consists of biologic disease-modifying antirheumatologic drugs (bDMARDs) or targeted DMARDs, respectively. There are currently 2 classes of bDMARDs available in Canada for AS, tumour necrosis factor (TNF) inhibitors and interleukin-17A (IL-17A) inhibitors. Janus kinase (JAK) inhibitors are the only class of targeted DMARD available for the treatment of AS in Canada, and are indicated after a patient’s condition has shown an inadequate response to a bDMARD. Many patients with AS receiving advanced therapy will experience treatment failure.^11-13 When failure of advanced therapies occurs, it is recommended to switch to another advanced therapy, either within the same class or to another class.^4,10 There is very little evidence to guide switching between advanced therapies; therefore, when treatment failure occurs, guidelines recommend switching to a different therapy that is either within or between treatment classes.^4,10

Bimekizumab is a humanized immunoglobulin G1 kappa monoclonal antibody that neutralizes IL-17A, IL-17F, and IL-17AF cytokines, thereby potentially blocking proinflammation and pathological bone formation of axSpA.

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of bimekizumab (Bimzelx) 160 mg/mL solution for subcutaneous injection in the treatment of adult patients with active AS.

Stakeholder Perspectives

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

Patient Input

A total of 2 inputs were received for this review. One was submitted by Arthritis Consumer Experts (ACE), and the other was a joint submission by 4 patient groups, the Canadian Spondyloarthritis Association, Canadian Arthritis Patient Alliance, Arthritis Society Canada, and Creaky Joints. ACE conducted an online survey between 2019 and 2022 to gather information from patients (n = 4) with AS. The joint input from the 4 patient groups was prepared based on an online survey conducted from September to October 2023 among patients with AS (n = 109).

According to the joint input from the 4 patient groups, the majority of patients with AS experience back pain (90.48%), joint stiffness (79.05%), fatigue (77.14%), and hip pain (71.43%); have difficulties exercising or being active (80.77%); have challenges with sleep (73.08%); and have an impaired ability to work (57.69%) and make social connections (53.85%). In addition, patients living with AS require help with daily activities and emotional support from caregivers. The input from ACE echoes the patient experiences reported in the joint input, and added flare-ups, deconditioning, anxiety, and mood changes as other impacts of AS on patients’ daily lives. Outcomes of interest to patients mentioned in the joint input were improved symptoms (71%), such as less fatigue, pain, and stiffness; better quality of life (QoL) (67%), including an ability to socialize more and better mental well-being; affordability in managing AS (66%); reduced side effects of medications (48%); and convenience (36%) in terms of drug-dosing schedules, route of administration, or formulations. The ACE input agrees with these outcomes of interest and added that ease of movement, ability to exercise more, control of back spasms and inflammation, and less weight gain are other outcomes of interest.

The joint input emphasized that approximately half of patients become resistant to their treatments within 5 years; therefore, access to new treatment options is essential. Of note, the 4 patient groups pointed out that for patients in Canada, it takes an average of 7 to 10 years from the onset of symptoms to be diagnosed with AS. Delayed diagnosis and treatment almost always lead to irreversible damage and a negative impact on mental health. According to the input, patients with AS experience significantly impacted QoL and frustration during this time.

Clinician Input

Input From the Clinical Expert Consulted by CADTH

The clinical expert consulted by CADTH indicated that the goals of treatment are to control pain and inflammation and prevent radiographic damage and disability related to AS. The clinical expert stated that the treatment of AS is tailored according to the current manifestations of the disease (axial, peripheral, entheseal, extra-articular symptoms and signs), level of current symptoms, clinical findings and prognostic indicators, disease activity, pain, physical function, structural damage of joints (especially hip involvement and spinal deformities), comorbidities, concomitant drugs, and the wishes and expectations of the patient. The clinical expert consulted by CADTH indicated that unmet needs in the management of AS included the following:

• a lack of response to available treatments once initiated (primary failure) in some patients

• many patients developing active disease after initially experiencing a response to treatment (secondary failure)

• limited access to early diagnosis and treatment

• choosing the right drug for the right patient at the right time (precision medicine) due to the availability of relatively few targeted therapies (TNF, IL-17A, and JAK inhibitors)

• safety concerns for most DMARDs as well as NSAIDs.

According to the clinical expert, these safety concerns include infections with most drugs, new onset or worsening of associated diseases (uveitis, inflammatory bowel disease [IBD], and psoriasis), and comorbidities; thus, treatments that are safe, effective for all manifestations, and well tolerated by most patients are needed. Though the efficacy of various drugs on the musculoskeletal manifestations are similar, no drug is equally effective for all manifestations. A drug’s effect on associated diseases may vary, according to the feedback from the clinical expert.

The clinical expert consulted by CADTH indicated that in clinical practice, bimekizumab would be used after failure of NSAIDs, either by itself or in combination with NSAIDs. The clinical expert stated they would not reserve bimekizumab for patients with refractory disease or patients who are intolerant to other therapies, as no other drugs target both IL-17A and IL-17F cytokines. The clinical expert stated that given bimekizumab’s efficacy in both musculoskeletal and skin disease, it may be the drug of choice following treatment with NSAIDs in patients with severe skin psoriasis who do not have IBD.

Patients with a personal or family history of IBD may not be candidates for treatment with bimekizumab. This is because, according to the clinical expert consulted by CADTH, the use of IL-17 inhibitors would increase the risk of IBD flares, based on the expert’s experience in using DMARDs that target IL-17A in patients with IBD. The clinical expert consulted by CADTH stated that patients who experience an inadequate response to currently available DMARDs most need an additional treatment option. The clinical expert indicated that the patients best suited for treatment with bimekizumab are generally identified by clinician examination and judgment.

According to the clinical expert consulted by CADTH, clinical response is determined by change in severity of back pain as assessed by patient-reported questionnaires, including total back pain score and the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI). More objective measures, such as the Ankylosing Spondylitis Disease Activity Score (ASDAS), are used in tertiary care centres. Other measures include improvements in enthesitis counts, in the number of tender and swollen joints, and in skin psoriasis. These measures align with the assessments used in the clinical trials. The clinical expert consulted by CADTH indicated that a BASDAI score at 3 to 6 months would be used to assess response. A reduction of at least 50% in the BASDAI score (BASDAI50), or an absolute reduction of at least 2 points in the BASDAI score, is usually required to suggest clinically significant improvement.

According to the clinical expert consulted by CADTH, a lack of response in back pain (given that other causes of back pain are excluded) and secondary treatment failure (relapse) are the most important factors to consider when deciding to discontinue treatment with bimekizumab. The clinical expert indicated that recurrent infections and the occurrence of IBD would require discontinuation of bimekizumab. The clinical expert indicated that discontinuing treatment with bimekizumab is determined by clinical evaluation by a rheumatologist, sometimes involving MRI.

The clinical expert stated that because rheumatologists are trained to identify inflammatory sacroiliitis and spondylitis, they should make the diagnosis. According to the clinical expert consulted by CADTH, patients with AS are usually treated in an outpatient setting, including community clinics and clinics attached to community and academic hospitals. In rare instances, severe disease, including skin, eye, and bowel disease, may warrant admission to a hospital. A rheumatologist is required to diagnose, treat, and monitor patients with AS. Since uveitis, IBD, and skin psoriasis are present with AS, ophthalmologists, gastroenterologists, and dermatologists are also relevant to disease management.

The clinical expert consulted by CADTH stressed that the treatment options for patients with active AS are limited; thus, bimekizumab provides an additional treatment option for such patients.

Clinician Group Input

No clinician group input was submitted for this review.

Drug Program Input

Input was obtained from the drug programs that participate in the CADTH reimbursement review process. The following were identified as key factors that could potentially affect the implementation of a CADTH recommendation for bimekizumab: relevant comparators, consideration for initiation of therapy, considerations for continuation or renewal of therapy, considerations for discontinuation of therapy, considerations for prescribing of therapy, generalizability, care provision issues, and system and economic issues.

The clinical expert consulted by CADTH provided advice on the potential implementation issues raised by the drug programs. Refer to Table 4 for more details.

Clinical Evidence

Systematic Review

Description of Studies

One pivotal trial (BE MOBILE 2) was included in the sponsor’s systematic review. The BE MOBILE 2 trial was a phase III, multicentre, randomized, double-blind, placebo-controlled study evaluating the efficacy and safety of bimekizumab in patients with active AS compared with placebo. This study enrolled adults with active AS (i.e., r-axSpA) and fulfilled the modified New York (mNY) criteria. Eligible study participants (N = 332) were randomized 2:1 to receive bimekizumab (n = 221) 160 mg/mL or placebo (n = 111) subcutaneously every 4 weeks. The mean age of all study participants was 40.4 years with a range of 19 to 80 years. Treatment groups were generally well balanced with respect to AS-related and other baseline disease characteristics. At baseline, the majority of study participants were using NSAID therapies (79.8%), and prior anti-TNF therapy was used by 16.3% of all study participants. The primary objective of the BE MOBILE 2 trial was to demonstrate the efficacy of bimekizumab administered subcutaneously every 4 weeks compared with placebo in the treatment of patients with active AS. The primary end point of the study was an improvement of 40% or more in the Assessment of SpondyloArthritis International Society (ASAS40). Secondary end points included scores for the BASDAI, Bath Ankylosing Spondylitis Functional Index (BASFI), nocturnal spinal pain (NSP) based on a numeric rating scale (NRS), the Maastricht Ankylosing Spondylitis Enthesitis Score (MASES), and health-related quality of life (HRQoL) scores using the Ankylosing Spondylitis Quality of Life (ASQoL) and Work Productivity and Activity Impairment–Specific Health Problem (WPAI-SHP) scales.

Efficacy Results

Improvement of 40% or More in the Assessment of SpondyloArthritis International Society

At week 16 of the double-blind treatment period, patients in the bimekizumab group reported a higher adjusted ASAS40 response rate (41.5%) compared with the placebo group (19.8%), with a between-group difference of 21.8% (95% confidence interval [CI], 11.4% to 32.1%). This corresponded to an odds ratio (OR) of 2.88 (95% CI, 1.71 to 4.87; P < 0.001) in favour of bimekizumab. No estimate of a between-group minimal important difference (MID) was identified by CADTH, but the clinical expert input suggested the absolute difference between groups was clinically important based on a 15% threshold. The ASAS40 response in the bimekizumab group was also observed at weeks 24, 36, and 52. Prespecified subgroup analyses of ASAS40 response rate at week 16 were generally consistent with the primary analysis. At week 16 of the double-blind treatment period, patients in the bimekizumab group who were either TNF alpha inhibitor–naive or –experienced reported a higher adjusted ASAS40 response rate compared with those in the placebo group (45.7% and 40.5%, respectively, for bimekizumab versus 23.4% and 17.6%, respectively, for placebo). The results of sensitivity and supportive analyses, including the tipping-point analyses, were in line with the primary efficacy results.

Bath Ankylosing Spondylitis Disease Activity Index

At week 16, patients in the bimekizumab group had a greater reduction from baseline in the least squares (LS) mean (reductions reflect improvement) in BASDAI score compared with patients in the placebo group (LS mean of −2.7 for bimekizumab versus −1.7 for placebo). An estimated median MID of 1.4 points (range, 0.9 to 1.8) was identified in the literature.¹⁴ The clinical expert consulted by CADTH indicated they would consider a 1-point difference between groups as clinically meaningful. The difference in LS mean between treatment groups was −1.04 (95% CI, −1.5 to −0.6; P < 0.001) in favour of bimekizumab. Generally, the treatment effects of bimekizumab on the BASDAI were observed at weeks 24, 36, and 52.

Bath Ankylosing Spondylitis Functional Index

At week 16, patients in the bimekizumab group had a greater reduction from baseline in the LS mean (reductions reflect improvement in physical function) in BASFI score compared with patients in the placebo group, which worsened (LS mean of −1.9 for bimekizumab versus −1.0 for placebo). An estimated median MID of 1.1 points (range, 1.0 to 1.4) was identified in the literature.¹⁴ The difference in LS mean between treatment groups was −1.1 (95% CI, −1.5 to −0.6; P < 0.001) in favour of bimekizumab. The clinical expert consulted by CADTH suggested an MID of 1 point for between-group difference. Generally, the treatment effects of bimekizumab on the BASFI were observed at weeks 24, 36, and 52.

Nocturnal Spinal Pain

At week 16, patients in the bimekizumab group had a greater reduction from baseline in LS mean (reductions reflect improvement) in NSP (based on an NRS) score compared with patients in the placebo group, which worsened (LS mean of −3.2 for bimekizumab versus −1.7 for placebo). An estimated median MID of 1.5 points (range, 1.1 to 2.3) was identified in the literature.¹⁴ The difference in LS mean between treatment groups was −1.5 (95% CI, −2.0 to −1.0; P < 0.001) in favour of bimekizumab. The clinical expert consulted by CADTH identified an MID of 1 point for between-group difference. Generally, the treatment effects of bimekizumab on the NSP were observed at weeks 24, 36, and 52.

Enthesitis-Free State Based on the MASES in Patients With Enthesitis at Baseline

At week 16 of the double-blind treatment period, patients with enthesitis at baseline in the bimekizumab group reported a higher adjusted enthesitis-free rate (43.8%) compared with those in the placebo group (23.9%), with a between-group difference of 19.8% (95% CI, 6.3% to 33.4%). This corresponded to an OR of 2.47 (95% CI, 1.30 to 4.68) in favour of bimekizumab. No estimate of a between-group MID was identified by CADTH, but the clinical expert suggested a 15% difference would be clinically important; therefore, the absolute difference between groups was clinically important. Generally, the treatment effects of bimekizumab on the enthesitis-free rate were observed at weeks 24 and 52. The outcome for enthesitis-free state was not controlled for type I error; thus, these data should be interpreted as supportive evidence only.

Ankylosing Spondylitis Quality of Life

At week 16, patients in the bimekizumab group had a greater reduction from baseline in LS mean (reductions reflect improvement) in ASQoL score compared with patients in the placebo group, which worsened (LS mean of −4.6 for bimekizumab versus −3.1 for placebo, respectively). An MID of 1 unit of worsening (i.e., + 1) or 2 units of improvement (i.e., −2) was identified in the literature.¹⁵ The difference in LS mean between treatment groups was −1.5 (95% CI, −2.4 to −0.7; P < 0.001) in favour of bimekizumab. The clinical expert consulted by CADTH identified an MID of 2 points for between-group difference. Generally, the treatment effects of bimekizumab on the ASQoL were observed at weeks 24, 36, and 52.

Work Productivity and Activity Impairment–Specific Health Problem

At week 16, the bimekizumab group compared with the placebo group had a greater mean reduction (improvement) from baseline in the WPAI-SHP score for the following:

• Percent time missed due to disease-related problems: –5.5 for bimekizumab versus −1.2 for placebo, with a between-group difference of −2.9 (95% CI, –6.9 to 1.1)

• Percent impairment while working due to disease-related problems: –20.8 for bimekizumab versus −6.1 for placebo, with a between-group difference of −12.5 (95% CI, –18.2 to −6.9)

• Percent overall work impairment due to disease-related problems: –22.2 for bimekizumab versus –6.7 for placebo, with a between-group difference of −12.8 (95% CI, –18.7 to −6.9)

• Percent activity impairment due to disease-related problems: –23.3 for bimekizumab versus −14.4 for placebo, with a between-group difference of −9.4 (95% CI, –13.9 to −4.9).

No MIDs for WPAI-SHP were identified in the literature. Generally, the treatment effects of bimekizumab on the WPAI-SHP domains were observed at weeks 24, 36, and 52, except for percent activity impairment due to disease-related problems where patients reported similar results between groups. The WPAI-SHP outcome was not controlled for type I error; thus, these data should be interpreted as supportive evidence only.

Harms Results

An adverse event (AE) was reported among 54.3% of patients in the bimekizumab group and 43.2% of patients in the placebo group at week 16. The most commonly reported AEs (i.e., reported by ≥ 5% of patients in either group) were infections and infestations (28.1% for bimekizumab versus 22.5% for placebo), gastrointestinal disorders (13.1% versus 9.9%), nervous system disorders (8.1% versus 4.5%), upper respiratory tract infection (2.7% versus 7.2%), and eye disorders (2.3% versus 6.3%).

Serious adverse events (SAEs) were reported among 2.3% of patients in the bimekizumab group and 0.9% of patients in the placebo group at week 16. The following SAEs were commonly reported in the bimekizumab group but not in any patients in the placebo group: goitre (0.5%), colitis ulcerative (0.5%), Crohn disease (0.5%), cholelithiasis (0.5%), and hepatitis A (0.5%).

Discontinuation due to AEs was reported among 2.7% of patients in the bimekizumab group but among no patients in the placebo group at week 16. The most commonly reported AEs that led to study discontinuation in the bimekizumab group were abnormal psychiatric evaluation (0.9%), lymphoid tissue hyperplasia (0.5%), Crohn disease (0.5%), oral candidiasis (0.5%), and rash (0.5%). No deaths due to AEs were reported during the double-blind treatment period in the BE MOBILE 2 trial.

Serious infections, fungal infections, opportunistic infections, malignancies, major adverse cardiac events, neutropenia, suicidal ideation and behaviours, IBD, hypersensitivity reactions, and liver injuries or disorders were considered notable harms by the sponsor and/or the clinical expert consulted by CADTH. The commonly reported notable harms were hypersensitivity reactions (7.7% for bimekizumab versus 1.8 for placebo), fungal infections (6.3% versus 0), liver injuries or disorders (4.5% versus 3.6%), IBD (1.8% versus 0), neutropenia (0.5% versus 0), and serious infections (0.5% versus 0.9%).

Critical Appraisal

Internal Validity

The CADTH review team noted there were no comparative data available beyond week 16, as patients in the placebo group were reallocated to receive bimekizumab during the 36-week maintenance period after finishing all assessments at the end of the 16-week double-blind treatment period; therefore, the direct comparative efficacy and safety of bimekizumab after week 16 are uncertain. For the analysis of the primary and key secondary end points, a fixed sequence testing procedure was employed to adjust for multiple comparisons across multiple end points, thereby controlling the type I error. The CADTH review team noted that the analyses of the enthesitis-free state based on the MASES index were not included in the fixed sequence testing hierarchy; thus, the results should be considered as supportive evidence. Although the subgroup analyses were prespecified, the BE MOBILE 2 trial was not powered to detect any change in the ASAS40 response rate between bimekizumab and placebo in subgroup analyses, except for the subgroup of patients who are TNF inhibitor–naive; additionally, no formal statistical tests for interaction between subgroups were conducted. There were 2 protocol amendments regarding eligibility criteria made after the enrolment of the first patient (April 25, 2019). The CADTH review team considered that these 2 protocol amendments may increase patient heterogeneity and introduce bias. The direction of the bias is uncertain, as there were no data reported on the numbers of patients with psoriatic arthritis or patients who had experienced treatment failure with NSAIDs other than the 2 NSAIDs included in the trial. HRQoL is considered a relevant outcome by patients with active AS and the clinical expert consulted by CADTH. However, the assessment of the ability to return to normal activities and/or functioning using the WPAI-SHP was not controlled for multiplicity; thus, it should be considered as exploratory and supportive.

External Validity

The BE MOBILE 2 trial used placebo as the comparator group. According to the clinical expert consulted by CADTH, an anti-TNF biosimilar monoclonal antibody would be an appropriate comparator for bimekizumab. The clinical expert indicated placebo is not an appropriate comparator; head-to-head studies with an active drug would be ideal. The BE MOBILE 2 trial excluded patients who had been treated with more than 1 TNF alpha inhibitor and/or more than 2 additional non–TNF alpha biological-response modifiers, or any IL-17 biological-response modifier at any time. The clinical expert indicated that these patients should be considered eligible for bimekizumab. Although the response rate might be lower, some patients’ conditions do respond to bDMARDs after previously failing to respond to TNF inhibitors and IL-17 inhibitors; therefore, the clinical expert stated they would switch treatments within the same class due to the relatively limited treatment options. According to the clinical expert, the study results would not be generalizable to the previously mentioned patients, as it is expected that response rates will be lower in that patient population, which tends to have lower response rates with subsequent treatments in clinical practice. In addition, there was no study site in Canada in the BE MOBILE 2 trial, which may compromise the generalizability of the study results to clinical practice in Canada.

GRADE Summary of Findings and Certainty of the Evidence

For the pivotal BE MOBILE 2 trial identified in the sponsor’s systematic review, the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) was used to assess the certainty of the evidence for outcomes considered most relevant to inform CADTH’s expert committee deliberations, and a final certainty rating was determined as outlined by the GRADE Working Group.^16,17 Following the GRADE approach, evidence from randomized controlled trials (RCTs) started as high-certainty evidence and could be rated down for concerns related to study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, or publication bias.

The selection of outcomes for GRADE assessment was based on the sponsor’s summary of clinical evidence, consultation with a clinical expert, and the input received from the patient groups and the public drug plans. The following outcomes were finalized in consultation with expert committee members: ASAS40, BASDAI, BASFI, NSP, MASES, HRQoL (ASQoL and WPAI-SHP), and SAEs.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty-of-evidence assessment was based on the point estimate and where it was located relative to the threshold for a clinically important effect (when a threshold was available) or to the null. The target of the certainty-of-evidence assessment was the presence or absence of a clinically important effect based on thresholds informed by the clinical expert consulted by CADTH for this review for ASAS40, BASDAI, BASFI, NSP, MASES, ASQoL, and SAEs. There is no established MID for the WPAI-SHP, and the clinical expert consulted by CADTH could not provide an MID threshold, so the target of the certainty-of-evidence assessment was the presence or absence of any (non-null) effect.

Table 2: Summary of Findings for Bimekizumab Versus Placebo for Patients With Active Ankylosing Spondylitis

ASAS40 = improvement of 40% or more in the Assessment of SpondyloArthritis International Society; ASQoL = Ankylosing Spondylitis Quality of Life; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASFI = Bath Ankylosing Spondylitis Functional Index; CI = confidence interval; MASES = Maastricht Ankylosing Spondylitis Enthesitis Score; MID = minimal important difference; NR = not reported; NRS = numeric rating scale; NSP = nocturnal spinal pain; OR = odds ratio; RCT = randomized controlled trial; WPAI-SHP = Work Productivity and Activity Impairment Questionnaire–Specific Health Problem.

Note: Study limitations (which refer to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias were considered when assessing the certainty of the evidence. All serious concerns in these domains that led to the rating down of the level of certainty are documented in the table footnotes.

^aRated down 1 level for serious imprecision. There is no established MID but the clinical expert consulted by CADTH considered that a 15% difference between groups in the adjusted ASAS40 response rate and the adjusted enthesitis-free rate at week 16 could be considered a threshold of clinical importance. For both outcomes, the point estimate and the upper bound of the 95% CI for the between-group difference suggested a clinically important difference for bimekizumab vs. placebo, while the lower bound of the 95% CI suggested no clinically important difference between the 2 groups.

^bRated down 1 level for serious imprecision. There is no established between-group MID but the estimated median MID for the change from baseline is 1.4 points (range, 0.9 to 1.8). The clinical expert consulted by CADTH considered that a 1-point difference between groups in the change from baseline in BASDAI total score at week 16 could be considered a threshold of clinical importance. The point estimate and the upper bound of the 95% CI for the between-group difference suggested a clinically important difference for bimekizumab vs. placebo based on a 1-point threshold, while the lower bound of the 95% CI suggested no clinically important difference between the 2 groups.

^cRated down 1 level for serious imprecision. There is no established MID for between-group difference but the estimated median MID for the change from baseline is 1.1 points (range, 1.0 to 1.4). The clinical expert consulted by CADTH considered that a 1-point difference between groups in the change from baseline in BASFI at week 16 could be considered a threshold of clinical importance. The point estimate and the upper bound of the 95% CI for the between-group difference suggested a clinically important difference for bimekizumab vs. placebo based on a 1-point threshold, while the lower bound of the 95% CI suggested no clinically important difference between the 2 groups.

^dImprecision was not rated down. There is no established between-group MID, but the estimated median MID for the change from baseline is 1.5 points (range, 1.1 to 2.3). The clinical expert consulted by CADTH considered that a 1-point difference between groups in the change from baseline in NSP score (based on an NRS) at week 16 could be considered a threshold of clinical importance. The point estimate and the 95% CI for the between-group difference suggested a clinically important difference for bimekizumab vs. placebo based on a 1-point threshold.

^eThe statistical testing was not adjusted for multiplicity in the trial and should be considered as supportive evidence.

^fRated down 1 level for serious imprecision. There is no established between-group MID, but the estimated median MID for the change from baseline is −2 points for improvement. The clinical expert consulted by CADTH considered that a 2-point difference between groups in the change from baseline in ASQoL total score at week 16 could be considered a threshold of clinical importance. The point estimate and the upper bound of the 95% CI for the between-group difference suggested no clinically important difference for bimekizumab vs. placebo, while the lower bound of the 95% CI suggested a clinically important difference between the 2 groups based on a 2-point threshold.

^gThere is no established MID and the clinical expert consulted by CADTH could not provide a threshold of important difference, so the target of the certainty appraisal was any effect for the change from baseline in the WPAI-SHP at week 16. For percent time missed due to problems (related to disease), impression was rated down for 2 levels, as the CADTH review team judged that the point estimate suggested a possibility of benefit, but the 95% CI for the between-group difference included the possibility of both benefit and harm (fewer benefits) for bimekizumab vs. placebo. For percent impairment while working due to (disease-related) problems, percent overall work impairment due to problems, and percent activity impairment due to problems, impression was not rated down, as the CADTH review team judged that the point estimate and the 95% CI for the between-group difference suggested no clinically important difference for bimekizumab vs. placebo.

^hRated down 2 levels for very serious imprecision. The CADTH review team considered the 16-week double-blind follow-up period to not be long enough to assess comparative long-term harms. The lower bound of the 95% CI for the between-group difference was below zero while the upper bound was above zero, suggesting no clinically important difference between the 2 groups. Additionally, the rate of SAEs was relatively low in both treatment groups based on a small sample size.

Sources: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸ The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

Long-Term Extension Studies

Description of Studies

One single-arm, phase II, open-label extension study, BE AGILE 2, was submitted by the sponsor as supporting evidence. Patients who had been enrolled in and completed the BE AGILE trial were rolled over to the BE AGILE 2 study (N = 255), which was conducted in European countries and the US. All patients in the BE AGILE 2 trial received open-label bimekizumab 160 mg every 4 weeks for up to 204 weeks, which meant a total possible exposure of 252 weeks for those who received bimekizumab in the parent trial, BE AGILE.

Efficacy Results

ASAS40 response was sustained up to week 208 in the BE AGILE 2 trial, with response rates of 59% (147 out of 249) using nonresponder imputation (NRI) and 73.1% (147 out of 201) using observed case data. The mean BASDAI score (n = 249) decreased from baseline and was sustained at week 208 (decrease of −4.01; standard error [SE] of 0.13 versus an MID range of 0.9 to 1.8 points). The mean BASFI score (n = 249) decreased from baseline (−3.1; SE = 0.15) and was sustained at week 208. Relative to baseline, the NSP score (n = 249) decreased and was maintained at week 208 (−4.55; SE = 0.16) versus an MID range of 1.1 to 2.3 points. Also, the mean ASQoL score (n = 249) decreased from baseline (−5.9; SE = 0.3) and was maintained through week 208 versus an MID ranging from an increase of 1 (worsening) to a decrease of 2 (improvement) units. Among patients with enthesitis at baseline, the mean MASES (n = 164) decreased by −0.37 (SE = 0.23) and maintained improvement up to week 208 in the BE AGILE 2 trial. WPAI-SHP was not assessed in the BE AGILE 2 study.

Harms Results

A total of 237 (92.9%) study participants reported a treatment-emergent adverse event (TEAE) during the BE AGILE 2 trial. The most commonly reported TEAEs were nasopharyngitis (18%), upper respiratory tract infection and COVID-19 infection (12.9% each), and bronchitis (8.6%). There were 46 patients (18.0%) who experienced at least 1 SAE, with COVID-19 infection and pneumonia being the most common (1.2% each). Twenty-one patients (8.2%) discontinued study treatment due to a TEAE, mostly due to alanine aminotransferase (1.2%) and aspartate aminotransferase (0.8%) elevation. Two fatal TEAEs were reported during the study, 1 incident due to a road traffic accident and another incident due to cardiorespiratory arrest. Fungal infection (18.4%) and hypersensitivity (11.4%) were the most common AEs of special interest reported during the BE AGILE 2 study, where the vast majority of fungal infections did not lead to treatment discontinuation. (A single patient discontinued due to perirectal abscess.)

Critical Appraisal

The lack of a control group, open-label design, and selective patient population are the major limitations of the BE AGILE 2 extension study. An open-label design without a comparator arm can overestimate results for efficacy outcomes, especially patient-reported outcomes. Moreover, a risk of selection bias was noted for the BE AGILE 2 study, since patients who have experienced a response to bimekizumab and tolerated any side effects are more likely to continue the extension period.

Indirect Comparisons

Description of Studies

Network meta-analyses (NMAs) were performed to determine the clinical efficacy and safety of bimekizumab at weeks 12 and 16 compared with other relevant interventions for the treatment of patients with AS. The NMAs were conducted on 3 different networks: purely naive (100% bDMARD-naive; 24 studies, 4,145 patients), predominantly naive (approximately 90% bDMARD-naive; 26 studies, 5,271 patients), and purely experienced (100% bDMARD-experienced; 9 studies, 1,048 patients).

Unanchored matching-adjusted indirect comparisons (MAICs) were performed to establish the long-term relative clinical efficacy of bimekizumab compared with other IL-17A inhibitors in patients with AS at week 52.

Efficacy Results

Network Meta-Analyses

In the bDMARD purely naive network, for most comparisons between bimekizumab versus TNF, IL-17, or JAK inhibitors, there were no clear differences observed. The exceptions to this were 2 findings in which bimekizumab showed statistically significant improvement in the Short Form (36) Health Survey physical component summary (SF-36 PCS) results compared with adalimumab and compared with secukinumab, but the differences observed were not clinically significant and the credible intervals (CrIs) were wide, indicating uncertainty.

In the bDMARD predominantly naive network, for most comparisons between bimekizumab versus TNF, IL-17, or JAK inhibitors, there were no clear differences observed. There were some exceptions to this general observation that were statistically significant differences, but these were not clinically significant and the CrIs were wide, indicating uncertainty. Bimekizumab improved the SF-36 PCS and partial remission in the Assessment of SpondyloArthritis International Society (ASAS-PR) results compared with secukinumab. Results favoured etanercept compared with bimekizumab for the BASDAI50 and BASFI. Results favoured golimumab IV compared with bimekizumab for BASFI and ASQoL. Results favoured adalimumab and certolizumab over bimekizumab for Ankylosing Spondylitis Disease Activity Score inactive disease state (ASDAS-ID). Results favoured tofacitinib over bimekizumab for the Bath Ankylosing Spondylitis Metrology Index. Results favoured upadacitinib over bimekizumab for ASDAS-ID.

In the bDMARD purely experienced network, for most comparisons between bimekizumab versus TNF, IL-17, or JAK inhibitors, there were no clear differences observed. The exceptions to this were 2 findings in which results favoured certolizumab over bimekizumab for the ASQoL and SF-36 PCS. In these 2 instances, the difference may be clinically significant, but the CrIs were wide, indicating uncertainty, and these results were not confirmed in the other networks.

Matching-Adjusted Indirect Comparison

The MAIC analyses suggested that bimekizumab 160 mg every 4 weeks had statistically significantly better results at week 52 compared with ixekizumab 80 mg every 4 weeks for the following: improvement of 20% or more in the Assessment of SpondyloArthritis International Society (ASAS20), ASAS40, BASDAI change from baseline, and BASDAI50. Results also favoured bimekizumab over secukinumab 150 mg every 4 weeks for ASAS40 and BASDAI change from baseline. However, there were significant limitations to the MAIC that preclude making claims of superiority of bimekizumab over comparators.

Harms Results

Network Meta-Analyses

The sponsor conducted NMAs of bimekizumab compared with other medications in the context of axSpA for 2 harms outcomes, discontinuation due to any reason and SAEs. The comparators of interest with data available for this NMA were 2 IL-17A inhibitors (ixekizumab and secukinumab); TNF alpha inhibitors adalimumab, certolizumab pegol, etanercept, golimumab (subcutaneous or IV routes), and infliximab (IV); and JAK inhibitors tofacitinib and upadacitinib.

The network for analysis of discontinuation due to any reason contained 18 studies. Study discontinuation rates were low in all trials (range, 0 to 14 patients per treatment arm). Time points between 12 and 16 weeks were used for this analysis. The CrIs were very wide for most estimates. There were no clear differences observed between bimekizumab and any other treatment. There was 1 finding in which bimekizumab had a higher risk of study discontinuation compared with tofacitinib; however, the uncertainty around this estimate was high, as reflected by a wide CrI.

The network for analysis of SAEs contained 18 studies. SAE rates were low in all studies (range, 0 to 10 SAEs per treatment arm). Time points between 12 and 16 weeks were used for this analysis. The CrIs were very wide for most estimates. There were no clear differences between bimekizumab and any other treatments in the network.

Matching-Adjusted Indirect Comparison

There were no harms outcomes assessed in the MAIC.

Critical Appraisal

Network Meta-Analyses

The sponsor conducted an NMA using a Bayesian approach. This was a reasonable method to apply, given the common comparator of placebo. The sponsor’s decision to perform 3 separate NMA analyses based on the potential effect modifier of prior exposure to bDMARDs was appropriate. Some networks had a large number of trials and a large number of patients, which was considered a strength of the NMA analyses. The sponsor did not perform sensitivity analyses in the NMA and did not attempt to identify and adjust for effect modifiers, despite the availability of a large number of trials for some of the networks. The time point of 12 to 16 weeks that was selected for the outcome analyses was reasonable and clinically relevant for efficacy but was not as meaningful for harms since an assessment of long-term harms was lacking.

The CIs and CrIs were wide for many estimates in the NMA. Despite the large number of trials, the number of patients and events in some analyses were small, precluding the possibility of detecting a difference between treatments. For example, the incidence of harms outcomes was small, resulting in very wide CrIs around the estimates. For this reason, the results of the harms analyses were not informative and did not serve to illuminate the risk of harms for bimekizumab relative to other treatments.

Matching-Adjusted Indirect Comparison

The sponsor performed an unanchored MAIC because of the lack of a placebo arm beyond week 16 for bimekizumab and comparators. This was an adequate justification for performing an MAIC. The selection of comparators from the same pharmacologic group (IL-17A inhibitors) was a rational approach, but comparisons with other biologics would also have been of interest. The MAIC allowed a comparison of 52 weeks of clinical data. The MAIC analyses suggested there were some differences favouring bimekizumab compared with secukinumab and ixekizumab for ASAS20, ASAS40, BASDAI change from baseline, and BASDAI50, but several limitations of the MAIC prevent drawing strong conclusions regarding the comparative effectiveness of bimekizumab. For example, there were important differences between the studies included in the MAIC that did not account for several of the prognostic factors that were deemed important by the authors of the MAIC and were not used in the weighting adjustments of the MAIC. There were notable differences in the study populations before and after adjustment. In the MAIC analyses, the effective sample size (ESS) for the bimekizumab group was reduced to 80% for the comparison with secukinumab 150 mg, to 51% for the comparison with secukinumab 300 mg, and to 20% for the comparison with ixekizumab. Not all matching variables that were deemed important were used in the weighting adjustments of the MAIC analyses. Regarding the MAIC analyses, the sponsor noted that “the amount of bias in the indirect comparisons is likely to be substantial,”¹⁹ and the CADTH reviewers agree with this assessment.

Summary

The results of the sponsor’s NMA did not show consistent differences for efficacy or harms outcomes between bimekizumab and comparators in the networks. While differences were reported in a small number of comparisons in some populations, these were associated with wide 95% CrIs for many comparisons, indicating imprecision of the results.

Results of the sponsor’s MAIC favoured bimekizumab for some outcomes, but there were significant limitations. The limitations include differences in study design and providing models with partial adjustments of prognostic and effect modifiers rather than fully adjusted models. These limitations, in addition to the substantial reduction in the ESSs, undermine any claims of superior performance of bimekizumab over comparators in the MAIC.

Neither the NMA nor the MAIC provided clear evidence of a difference in efficacy or harm outcomes for bimekizumab compared to comparators.

Studies Addressing Gaps in the Evidence From the Systematic Review

The BE MOBILE 1 trial was a phase III, multicentre, randomized, double-blind, placebo-controlled study evaluating the efficacy and safety of bimekizumab in patients with active nr-axSpA. The sponsor identified BE MOBILE 1 as a study addressing the gap in the evidence on the efficacy and safety of bimekizumab. However, the CADTH review team considered the BE MOBILE 1 trial irrelevant to this review, as patients with active nr-axSpA are different from patients with active AS, the population for the indication being reviewed. Therefore, the CADTH review team found that no studies addressing gaps in the systematic review evidence had been identified for this review.

Conclusions

Patients and clinicians highlighted the need for new effective treatments for active AS that control disease and symptoms and improve QoL compared with current treatments.

One phase III, multicentre, randomized, double-blind, placebo-controlled trial (BE MOBILE 2) comparing bimekizumab with placebo in treating patients with moderate to severe AS demonstrated that bimekizumab increased the adjusted ASAS40 response rate at week 16. Likewise, the results of the analysis of BASDAI total score indicated that patients treated with bimekizumab had a greater improvement in disease control than patients who received placebo at week 16. A GRADE assessment of the sponsor-submitted systematic review, which included only the BE MOBILE 2 trial, suggested that bimekizumab likely results in a clinically important improvement in the ASAS40 response rate and in BASDAI total score compared with placebo.

Compared with placebo, bimekizumab results in a clinically important reduction in NSP score (based on an NRS) and likely results in a clinically important reduction in the BASFI, the adjusted enthesitis-free rate based on the MASES index, and the ASQoL total score. With regard to the WPAI-SHP, bimekizumab may result in a reduction in percent time missed due to problems (related to disease) and does result in a reduction in percent impairment while working due to problems, percent overall work impairment due to problems, and percent activity impairment due to problems. The CADTH review team noted that the clinical importance of the reduction in the WPAI-SHP is unclear because a clinically meaningful threshold could not be determined.

Compared with placebo, there is low-certainty evidence that bimekizumab may result in an increase in the percentage of patients who experienced SAEs at week 16. No new safety signals were identified in the long term BE AGILE and BE AGILE 2 trials in patients with AS.

The sponsor-submitted indirect treatment comparison analyses did not provide clear evidence of a difference in efficacy or harms outcomes for bimekizumab relative to other treatments.

Introduction

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of bimekizumab (Bimzelx) 160 mg/mL solution for subcutaneous injection in the treatment of adult patients with active AS.

Disease Background

Contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following has been summarized and validated by the CADTH review team.

AS is a chronic, inflammatory, and heterogeneous disease that places a significant burden on patients that is driven by pain, fatigue, and stiffness.^1,2 Axial spondyloarthritis encompasses r-axSpA (also known as AS) and nr-axSpA. AS is characterized by inflammatory back pain, sacroiliitis, a high prevalence of human leukocyte antigen B27 (HLA-B27), and excess bone formation that can lead to spinal ankylosis.⁴ Key symptoms of AS include fatigue and stiffness with loss of physical function.^2,3,20

The symptoms and progression of AS can vary between patients.²¹ The initial symptoms of AS often begin in young adults (45 years and younger), with the peak onset between the ages of 20 and 30.^2,22 Patients with AS typically present first with chronic back pain (≥ 3 months in duration) of insidious onset and an inflammatory nature that improves with exercise.^1,2 This chronic back pain is caused by inflammation in the sacroiliac joints and/or spine. Over several years, as the disease progresses and inflammation continues, affected joints can become damaged and new bone formation can develop, leading to ankylosis of the sacroiliac joints and formation of syndesmophytes in the spine.^21,23 (Syndesmophytes are bony growths in ligaments in the intervertebral joints that cause irreversible impairment of spinal mobility.²³) Patients with uncontrolled inflammation may progress to irreversible axial structural damage,^5,6 spinal fractures, and severe spinal cord injury,²⁴ which is associated with uncontrolled inflammation.²⁵ Therefore, inflammation is predictive of structural progression in patients with axSpA.²⁵ Beyond the key axial symptoms, patients with AS can experience enthesitis, dactylitis, peripheral arthritis, and extramusculoskeletal manifestations, such as uveitis, IBD, and psoriasis and hidradenitis suppurativa.^6,7

As a result of both the axial and nonaxial manifestations, patients with AS experience significant disease burden that impacts their daily lives, reduces their QoL, and leads to a reduction in work productivity.²⁶ Because AS usually starts before age 45, the disease has a considerable effect on different aspects of life (career, family, and social life).²⁷

A clinical diagnosis of AS is based on clinical presentation, in combination with laboratory and imaging tests of the spine and/or the sacroiliac joints.^2,3,28 While there are no official diagnostic criteria for AS, the mNY classification criteria for AS or Assessment of SpondyloArthritis International Society (ASAS) classification criteria for axSpA are often referenced to aid in diagnosis.^2,3 To determine whether patients meet the ASAS axSpA criteria, clinicians rely on physical examinations (to identify dactylitis, enthesitis, and psoriasis), blood testing (to measure C-reactive protein levels and detect HLA-B27), and imaging (such as X-rays or MRI scans to view structural damage or inflammation of the sacroiliac joint, a hallmark feature of AS).^2,3,10 Elevated C-reactive protein is indicative of inflammation, which can be a sign of AS.² HLA-B27 is considered to be a biomarker for AS, as 80% to 90% of patients with AS have tested positive for HLA-B27; therefore, testing for HLA-B27 presence is a key tool in AS diagnosis.³ A diagnosis of AS is confirmed once the patient exhibits radiographic abnormalities consistent with sacroiliitis.^2,3,28 According to the recommendations of the American College of Rheumatology, Spondylitis Association of America, and Spondyloarthritis Research and Treatment Network, patients are considered to have active AS when symptoms are unacceptably bothersome to the patient and judged by the examining clinician to be due to inflammation.⁴ According to the clinical expert’s input, the patient is judged to have active AS based on a BASDAI score of 4 or greater and a rheumatologist assessment.

A population-based study of the incidence and prevalence of AS in 2010 using Ontario provincial health administrative databases found age- and sex-standardized prevalence and incidence rates of 0.213% and 0.015%, respectively.²⁹ AS was estimated to affect 300,000 patients in Canada in 2019.⁸

Standards of Therapy

Contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following has been summarized and validated by the CADTH review team.

International guidelines from the American College of Rheumatology, Spondylitis Association of America, and Spondyloarthritis Research and Treatment Network (published in 2019)⁴ and the Assessment of SpondyloArthritis International Society and European Alliance of Associations for Rheumatology (updated in 2022)¹⁰ are available to guide the treatment of AS.³⁰ Updated Canadian guidelines from the Canadian Rheumatology Association and Spondyloarthritis Research Consortium of Canada were published in July 2024 were published in July 2024.³²

According to the previously mentioned guidelines, treatment goals for patients with AS are to control symptoms and inflammation, prevent progressive structural damage, preserve or normalize physical function and social participation, decrease disease complications, and maintain ability to work.^4,10 Similarly, the clinical expert consulted by CADTH stated that treatment goals include controlling pain and inflammation and preventing radiographic damage and disability.

The choice of treatment should be individualized to each patient to meet treatment goals.¹⁰ Clinicians in Canada select treatment based on the current signs and symptoms of disease (axial, peripheral, extra-articular manifestations) and patient preference (e.g., oral versus injection, frequency of injection) to ensure treatment goals can be met.³⁰ The clinical expert consulted by CADTH agrees that treatment of AS is tailored according to current manifestations of the disease (axial, peripheral, entheseal, extra-articular symptoms and signs; level of current symptoms; clinical findings and prognostic indicators; disease activity, pain, and physical function; and structural damage at a joint, especially hip involvement and spinal deformities), comorbidities, concomitant drugs, and the wishes and expectations of the patient.

According to the clinical expert consulted by CADTH, physical therapy and NSAIDs are usually recommended to patients after the diagnosis is made. In Canada, several drug classes are used in the pharmacologic therapy of AS. NSAIDs are the first line of treatment for adult patients with active AS.^4,9,10 If a patient experiences an inadequate response to NSAIDs, or if there are contraindications, advanced therapies are the next line of treatment.^4,9,10 Advanced therapy consists of bDMARDs or targeted DMARDs. Currently, 2 classes of bDMARDs are available in Canada for AS, TNF inhibitors and IL-17A inhibitors. JAK inhibitors are the only class of targeted DMARDs available for the treatment of AS in Canada, and they are indicated after a patient has experienced an inadequate response to a bDMARD (Figure 1). In the Assessment of SpondyloArthritis International Society–European Alliance of Associations for Rheumatology 2022 update, TNF, IL-17, and JAK inhibitors are all recommended for patients when their condition has failed to respond to first-line NSAID therapy, with a preference to start advanced therapy with a TNF or IL-17 inhibitor.¹⁰ The clinical expert consulted by CADTH has a different opinion in that, unless contraindicated, a TNF inhibitor biosimilar should be the first biologic of choice for AS followed by another TNF inhibitor, then an IL-17 inhibitor, and then a JAK inhibitor. For patients with a history of recurrent uveitis or active IBD, a monoclonal antibody TNF inhibitor (adalimumab, certolizumab pegol, golimumab, or infliximab) is preferred, while for patients with significant psoriasis, an IL-17 inhibitor is preferred.¹⁰ The clinical expert CADTH consulted added that these drugs (especially the TNF inhibitors) improve the symptoms and signs of AS and, with prolonged treatment, lead to reduced spinal ankylosis (reduced progression as assessed by the modified Stoke Ankylosing Spondylitis Spinal Score), especially in those with early disease.

Many patients with AS receiving advanced therapy will experience treatment failure.^11-13 When advanced therapies fail, it is recommended that patients switch to another advanced therapy, either within the same class or to another class.^4,10 There is very little evidence to guide switching between advanced therapies; therefore, when treatment failure occurs, guidelines recommend switching to a different therapy that is either within or between treatment classes.^4,10

Figure 1: Treatment Paradigm for Patients With Ankylosing Spondylitis in Canada

Source: The sponsor’s summary of clinical evidence.¹⁹

Drug Under Review

Key characteristics of bimekizumab and other treatments available for adult patients with active AS are summarized in Table 3.

Bimekizumab (Bimzelx) 160 mg/mL solution for subcutaneous injection is available in a prefilled syringe or autoinjector for single use only. The recommended dosage for adult patients with active axSpA (including AS and nr-axSpA) is 160 mg given as 1 subcutaneous injection every 4 weeks.³³ After proper training in subcutaneous injection technique, patients may self-inject, if their physician determines it is appropriate, with medical follow-up as necessary.³³

Bimekizumab was approved by Health Canada for the treatment of adult patients with active AS and the treatment of adult patients with active nr-axSpA on March 11, 2024. The reimbursement request for the review by CADTH is for the treatment of adult patients with active AS. Bimekizumab was previously approved by Health Canada on February 14, 2022, for moderate to severe plaque psoriasis in adult patients who are candidates for systemic therapy or phototherapy. On April 18, 2022, CADTH recommended bimekizumab for reimbursement with conditions for this indication.³⁴

Bimekizumab is a humanized immunoglobulin G1 kappa monoclonal antibody. It has 2 identical antigen binding regions that bind and neutralize IL-17A, IL-17F, and IL-17AF cytokines, blocking their interaction with the IL-17RA and IL-17RC receptor complex.³³ Elevations of IL-17A and IL-17F levels are each independent pivotal drivers of inflammation and pathological bone formation in in vitro models of axSpA. Therefore, inhibiting IL-17F in addition to IL-17A is expected to provide greater resolution of inflammation and more inhibition of pathological bone formation than blocking IL-17A alone.

Table 3: Key Characteristics of Bimekizumab, TNF Inhibitors, Other IL-17 Inhibitors, and JAK Inhibitors

AS = ankylosing spondylitis; bDMARD = biologic disease-modifying antirheumatologic drug; CYP = cytochrome; Fc region = fragment crystallizable region; IBD = inflammatory bowel disease; IgG1 = immunoglobulin G1; IL = interleukin; JAK = Janus kinase; MACE = major adverse cardiovascular event; SC = subcutaneous; TNF = tumour necrosis factor; TyK2 = tyrosine kinase 2.

^aHealth Canada–approved indication.

Sources: Sponsor’s summary of clinical evidence¹⁹ and product monographs for Bimzelx, Humira, Cimzia, Enbrel, Simponi, Remicade, Taltz, Cosentyx, Xeljanz, and Rinvoq.^45-54

Stakeholder Perspectives

Patient Group Input

This section was prepared by the CADTH review team based on the input provided by patient groups. The full original patient inputs received by CADTH have been included in the stakeholder section of this report.

A total of 2 inputs were submitted for this review. One was from ACE, Canada’s largest and longest-running national arthritis patient organization. The other was a joint submission by 4 patient groups:

• the Canadian Spondyloarthritis Association, the only patient-led organization in Canada dedicated solely to people living with spondyloarthritis

• the Canadian Arthritis Patient Alliance, a grassroots, patient-driven, independent, and national education and advocacy organization

• Arthritis Society Canada, the largest charitable funder of cutting-edge arthritis research

• Creaky Joints, a digital community for arthritis patients and caregivers worldwide.

ACE conducted an online survey between 2019 and 2022 to gather information from patients with AS (n = 4). The joint input by the 4 other patient groups was prepared based on an online survey conducted from September to October 2023 among patients with AS (n = 109).

According to the joint input from the 4 patient groups, the majority of patients with AS experience back pain (90.48%), joint stiffness (79.05%), fatigue (77.14%), and hip pain (71.43%); have difficulties exercising or being active (80.77%); challenges with sleep (73.08%); and have an impaired ability to work (57.69%) and make social connections (53.85%). In addition, patients living with AS require help with daily activities and emotional support from caregivers. Even though 30% of respondents stated they do not typically miss school or work due to AS symptoms or side effects of medications, 18% of patients stated they do not work or attend school due to AS and 22% of patients missed an average of 1 to 5 days of work per month due to AS. Even though 73.4% of participants said they did not require support from caregivers, others needed help with climbing, carrying, long-distance walking, shopping and housework, meal preparation, and needed emotional support from family, friends, or a home-care worker. The input from ACE echoes the patient experiences reported in the joint input, and added flare-ups, deconditioning, anxiety, and mood changes as other impacts of AS on patients’ daily lives. Outcomes of interest mentioned in the joint input were improved symptoms (71%), such as less fatigue, pain, and stiffness; better QoL (67%), including an ability to socialize more and better mental well-being; affordability in managing AS (66%); reduced side effects of medications (48%); and convenience (36%) in terms of drug-dosing schedules, route of administration, or formulations. The ACE input agrees with these outcomes of interest and added that ease of movement and ability to exercise more, control of back spasms and inflammation, and less weight gain are other outcomes of interest.

The joint input emphasized that approximately half of patients become resistant to their treatments within 5 years; therefore, access to new treatment option is essential. Of note, the 4 patient groups pointed out that for patients in Canada, it takes an average of 7 to 10 years from the onset of symptoms to be diagnosed with AS. Delayed diagnosis and treatment almost always lead to irreversible damage and a negative impact on mental health. According to the input, patients with AS experience significantly impacted QoL and frustration during this time.

Clinician Input

Input From the Clinical Expert Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise regarding 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 adult patients with active AS.

Unmet Needs

The clinical expert consulted by CADTH indicated that the goals of treatment are to control pain and inflammation and prevent radiographic damage and disability related to AS. The clinical expert stated that the treatment of AS is tailored according to current manifestations of the disease (axial, peripheral, entheseal, extra-articular symptoms and signs); level of current symptoms; clinical findings and prognostic indicators; disease activity, pain, and physical function; and structural damage at a joint, especially hip involvement and spinal deformities), comorbidities, concomitant drugs, and the wishes and expectations of the patient. In addition, the clinical expert commented that the optimal management of AS requires a combination of nonpharmacological and pharmacological treatments; nonpharmacological treatment of AS should include patient education and regular exercise. Individual and group physical therapy should be considered.

The clinical expert consulted by CADTH indicated that unmet needs in the management of AS included the following:

• a lack of response to available treatments once initiated (primary failure) in some patients

• many patients developing active disease after initially experiencing a response to treatment (secondary failure)

• limited access to early diagnosis and treatment

• choosing the right drug for the right patient at the right time (precision medicine) due to the availability of relatively few targeted therapies (TNF, IL-17A, and JAK inhibitors)

• safety concerns for most DMARDs as well as NSAIDs.

According to the clinical expert, these safety concerns include infections with most drugs, new onset or worsening of associated diseases (uveitis, IBD, and psoriasis), and comorbidities; thus, treatments that are safe, effective for all manifestations, and well tolerated by most patients are needed. Current therapies include those administered orally, subcutaneously, and intravenously; patients thus have options when deciding which therapy to choose. Though the efficacy of various drugs on musculoskeletal manifestations are similar, no drug is equally effective for all manifestations and a drug’s effect on associated diseases may vary, according to the feedback from the clinical expert.

Place in Therapy

According to the clinical expert consulted by CADTH, current drugs used for the treatment of AS include NSAIDs and DMARDs including TNF inhibitors (i.e., etanercept, infliximab, adalimumab, golimumab and certolizumab), IL-17A inhibitors (i.e., secukinumab, ixekizumab) and JAK inhibitors (i.e., upadacitinib). These drugs (especially the TNF inhibitors) improve the symptoms and signs of AS and, with prolonged treatment, lead to reduced spinal ankylosis (reduced progression as assessed by the modified Stoke Ankylosing Spondylitis Spinal Score), especially in those with early disease. After the diagnosis is made, physical therapy and NSAIDs are recommended to patients. If the patient continues to have active disease, as judged by a BASDAI score greater than 4 and by a rheumatologist, then treatment with DMARDs is indicated. DMARDs have been shown to improve signs and symptoms, reduce spinal damage progression, and improve HRQoL and function. These drugs may also reduce the need for joint replacement, especially of hip joints.

The clinical expert consulted by CADTH indicated that in clinical practice, bimekizumab would be used after the failure of NSAIDs, either by itself or in combination with NSAIDs. The clinical expert stated they would not reserve bimekizumab for patients with refractory disease or patients who are intolerant to other therapies, as there are no other drugs targeting both IL-17A and IL-17F cytokines. The clinical expert stated that given bimekizumab’s efficacy in both musculoskeletal and skin disease, it may be the drug of choice following treatment with NSAIDs in patients with severe skin psoriasis who do not have IBD.

Patient Population

Patients with a personal or family history of IBD may not be candidates for treatment with bimekizumab. This is because, according to the clinical expert consulted by CADTH, the use of IL-17 inhibitors increase the risk of IBD flares, based on the expert’s experience in using DMARDs that target IL-17A in patients with IBD. However, one should note that the incidence of flares or new-onset IBD was very low, as patients with active IBD were excluded from the clinical trials with bimekizumab. The presence of IBD or severe uveitis and active infections, especially fungal infections, would be contraindications for treatment with bimekizumab.

The clinical expert consulted by CADTH stated that patients experiencing an inadequate response to currently available DMARDs are most in need of an additional treatment option. The clinical expert indicated that the patients best suited for treatment with bimekizumab are generally identified by clinician examination and judgment. There are no biomarkers for treatment response or AEs. The clinical expert indicated that patients with objective evidence of inflammation, such as a positive MRI or abnormal C-reactive protein level, are most likely to respond to DMARD therapy.

Assessing the Response to Treatment

According to the clinical expert consulted by CADTH, clinical response is determined by a change in the severity of back pain as assessed by patient-reported questionnaires, including total back pain score and the BASDAI. More objective measures, such as the ASDAS, are used in tertiary care centres. Other measures include improvements in enthesitis counts, in the number of tender and swollen joints, and in skin psoriasis. These measures are aligned with the assessments used in clinical trials.

The clinical expert consulted by CADTH indicated that a BASDAI score at 3 to 6 months would be used to assess response. A reduction in the BASDAI50 score, or an absolute reduction of at least 2 points in the BASDAI score, is usually required to suggest clinically significant improvement. The clinical expert stated that an improvement in physical function may be desired but may be delayed. The clinical expert mentioned that patients also report improvement in pain and fatigue, but such improvement may be difficult to achieve in longstanding disease. According to the clinical expert, the response magnitude is unlikely to vary between rheumatologists. The clinical expert commented that, with longer follow-up assessments, the lack of spinal radiographic progression may be demonstrated, but that is not usual clinical practice.

Discontinuing Treatment

According to the clinical expert consulted by CADTH, the lack of response in back pain (given that other causes of back pain are excluded) and secondary failure (relapse) are the most important factors to consider when deciding to discontinue treatment with bimekizumab. The clinical expert indicated that recurrent infections and the occurrence of IBD would require discontinuation of bimekizumab. The clinical expert indicated that discontinuing treatment with bimekizumab is determined through a clinical evaluation by a rheumatologist and sometimes involves MRI scans.

Prescribing Considerations

The clinical expert consulted by CADTH indicated that, currently, there are no markers that will help predict which patients will do well or not when using any drug, including bimekizumab. Diagnosing AS by primary care practitioners is difficult due to the lack of biomarkers, difficulty in interpreting MRI scans, and the overlap of symptoms with common musculoskeletal symptoms (such as back pain). The clinical expert stated that rheumatologists are trained to identify inflammatory sacroiliitis and spondylitis; therefore, they should make the diagnosis. Misdiagnosis is more likely in primary care than in rheumatology clinics.

According to the clinical expert consulted by CADTH, patients with AS are usually treated in an outpatient setting, including community clinics and clinics attached to community and academic hospitals. In rare instances, severe disease, including skin, eye, and bowel disease, may warrant admission to a hospital. A rheumatologist is required to diagnose, treat, and monitor patients with AS. Since uveitis, IBD, and skin psoriasis are present with AS, ophthalmologists, gastroenterologists, and dermatologists are also relevant to disease management.

Additional Considerations

The clinical expert consulted by CADTH stressed that the treatment options for patients with active AS are limited; thus, bimekizumab provides an additional treatment option for such patients.

Clinician Group Input

No clinician group input was submitted for this review.

Drug Program Input

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

Table 4: Summary of Drug Plan Input and Clinical Expert Response

AS = ankylosing spondylitis; ASAS40 = improvement of 40% or more in the Assessment of SpondyloArthritis International Society; ASAS-EULAR = Assessment of SpondyloArthritis International Society–European Alliance of Associations for Rheumatology; ASDAS = Ankylosing Spondylitis Disease Activity Score; axSpA = axial spondyloarthritis; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; CDEC = Canadian Drug Expert Committee; CRP = C-reactive protein; DMARD = disease-modifying antirheumatic drug; IBD = inflammatory bowel disease; IL-17 = interleukin-17; IL-17i = interleukin-17 inhibitor; JAKi = Janus kinase inhibitor; LTBI = latent tuberculosis infection; NSAID = nonsteroidal anti-inflammatory drug; SpA = spondyloarthritis; tDMARD = targeted disease-modifying antirheumatic drug; TNFi = tumour necrosis factor inhibitor.

Clinical Evidence

The objective of this Clinical Review Report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects bimekizumab 160 mg/mL solution for subcutaneous injection in the treatment of active AS in adult patients. The focus of this report is on comparing bimekizumab with relevant comparators and identifying gaps in the current evidence.

A summary of the clinical evidence included by the sponsor in the review of bimekizumab is presented in 3 sections, with CADTH’s critical appraisal of the evidence included at the end of each section. The first section, the systematic review, includes pivotal studies and RCTs that were selected according to the sponsor’s systematic review protocol. CADTH’s assessment of the certainty of the evidence in this first section uses the GRADE approach and follows the critical appraisal of the evidence. The second section includes sponsor-submitted long-term extension studies. The third section includes indirect evidence from the sponsor.

Included Studies

Clinical evidence from the following is included in the CADTH review and appraised in this document:

• One pivotal study: One randomized, double-blind, placebo-controlled, multicentre phase III trial (BE MOBILE 2) identified in the systematic review

• One long-term extension study: 1 single-arm phase II trial (BE AGILE) and its open-label extension study (BE AGILE 2)

• Four indirect treatment comparisons: 3 NMAs and 1 MAIC.

Systematic Review

Contents within this section have been informed by materials submitted by the sponsor. The following has been summarized and validated by the CADTH review team.

Description of Studies

Characteristics of the included studies are summarized in Table 5.

One pivotal trial (BE MOBILE 2) was included in the sponsor’s systematic review. The BE MOBILE 2 trial was a phase III, multicentre, randomized, double-blind, placebo-controlled study evaluating the efficacy and safety of bimekizumab in patients with active AS compared with placebo. The primary objective of the BE MOBILE 2 trial was to demonstrate the efficacy of bimekizumab administered subcutaneously every 4 weeks compared with placebo in the treatment of patients with active AS. This study enrolled adults who had active AS (i.e., r-axSpA) and fulfilled the mNY criteria.

Table 5: Details of Studies Included in the Systematic Review

AS = ankylosing spondylitis; ASAS5/6 = an improvement of at least 20% in 5 of 6 Assessment of SpondyloArthritis International Society domains; ASAS20 = Improvement of 20% or more in the Assessment of SpondyloArthritis International Society; ASAS40 = improvement of 40% or more in the Assessment of SpondyloArthritis International Society; ASAS-PR = partial remission in the Assessment of SpondyloArthritis International Society; ASDAS-MI = major improvement in Ankylosing Spondylitis Disease Activity Score; ASQoL = Ankylosing Spondylitis Quality of Life; AxSpA = axial spondyloarthritis; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index, BASFI = Bath Ankylosing Spondylitis Functional Index, BASMI = Bath Ankylosing Spondylitis Metrology Index; IL-17 = interleukin-17; MASES = Maastricht Ankylosing Spondylitis Enthesitis Score; mNY = modified New York criteria; NRS = numeric rating scale; NSAID = nonsteroidal anti-inflammatory drug; NSP = nocturnal spinal pain; q.4.w. = every 4 weeks; SF-36 PCS = Short Form (36) Health Survey physical component summary; TNF = tumour necrosis factor.

Sources: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸ The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

Patients were enrolled between April 25, 2019, and August 8, 2022. The final data cut-off date was September 9, 2022. Eligible study participants (N = 332) were randomized 2:1 to receive bimekizumab 160 mg/mL (n = 221) or placebo (n = 111) subcutaneously every 4 weeks. Randomization was performed via an interactive voice or web response system and stratified according to region and prior exposure to TNF alpha inhibitors (yes versus no).

The BE MOBILE trial included a screening period, a 16-week double-blind treatment period, a 36-week maintenance period, and a 20-week safety follow-up period (Figure 2). Assessments of eligibility were initiated during the screening period with a minimum duration of approximately 14 days and a maximum duration of up to 35 days. The screening period also enabled the washout of medications not permitted for use during the study, allowed initiation of treatment for latent tuberculosis where necessary, and allowed completion of the imaging assessments required to determine eligibility, including time for the reading of test results by a central laboratory.

At the end of the 16-week double-blind treatment period, after all assessments had been completed, patients in the placebo group were reallocated to receive bimekizumab 160 mg subcutaneous every 4 weeks during the 36-week maintenance period. Starting at week 20, nonbiologic rescue therapy for axSpA could have been adjusted or added. The maximum study duration per study participant was up to 73 weeks. Patients who completed the BE MOBILE 2 trial (52 weeks) were potentially eligible to be enrolled in a separate ongoing open-label extension study (AS0014; NCT04436640).

Figure 2: Study Schematic

ASAS40 = improvement of 40% or more in the Assessment of SpondyloArthritis International Society; BKZ = bimekizumab; BL = baseline; IMP = investigational medicinal product; q.4.w. = every 4 weeks; sc = subcutaneous; SFU = safety follow-up; W = week.

Note: The planned enrolment was approximately 300 participants.

^a Study participants were eligible for nonbiologic rescue therapy starting at week 20 with treatment at the discretion of the investigator while continuing to receive BKZ. Treatment with non-BKZ biologics or prohibited treatment led to BKZ discontinuation, if applicable, per withdrawal criteria.

Source: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸

Populations

Inclusion and Exclusion Criteria

Study participants were required to be adults with a diagnosis of active AS based on radiologic evidence fulfilling the mNY criteria for AS, including experiencing 3 or more months of symptoms and younger than age 45 at symptom onset, with moderate to severe active disease at baseline (defined as BASDAI ≥ 4 plus spinal pain ≥ 4 on a 0 to 10 NRS from BASDAI item 2) and intolerant to NSAIDs or with disease that has failed to respond to NSAIDs. Patients were allowed to have previously received treatment with a TNF alpha inhibitor or to be biologic treatment–naive; the enrolment of patients experienced with TNF alpha inhibitors was limited to 30% of the total study population. Patients were excluded if they had received more than 1 TNF alpha inhibitor and/or more than 2 additional non–TNF alpha inhibitor biological-response modifiers, or any IL-17 biological-response modifiers at any time. Patients were also excluded if they were diagnosed with an inflammatory condition other than axSpA (e.g., rheumatoid arthritis). Patients with a diagnosis of Crohn disease, or ulcerative colitis or another IBD were allowed to participate if they did not have active symptomatic disease at baseline.

Interventions

Eligible patients were randomized 2:1 to receive bimekizumab (160 mg subcutaneously every 4 weeks) or placebo (subcutaneously every 4 weeks) during the double-blind treatment period. Patients in the placebo group were then reassigned to receive bimekizumab treatment (160 mg subcutaneously every 4 weeks) during the 36-week maintenance period. Bimekizumab was supplied in a 1 mL prefilled syringe at a nominal formulation of 160 mg/mL for subcutaneous injection. Placebo (preservative-free physiological saline) was also supplied in a 1 mL prefilled syringe.

Special precaution was taken during the double-blind treatment period up to week 16; given the differences in presentation between the delivery systems, the bimekizumab and placebo treatments were administered at the investigational sites by unblinded study personnel. The unblinded personnel were not involved in the study in any other way; to preserve patient and physician blinding, these staff members did not assess the patients or communicate the treatments being administered.

Patients were allowed to remain on their background medication if the dose and regimen was stable before baseline and was to be maintained until week 16. No medication increases or additions for treating AS were permitted until after the week 20 visit assessments. However, a decrease in the dose or dosing frequency of any medication was permitted at any time for reasons of intolerance, AEs, and/or side effects. All concomitant medications, including over-the-counter products, herbal and traditional remedies, vitamin and mineral supplements, other dietary supplements, “nutraceuticals,” and hormones were recorded in the study participant’s source documentation (e.g., clinical chart) and on the electronic case report form. Nonbiologic rescue therapies (e.g., acetaminophen, paracetamol, NSAIDs, cyclooxygenase-2 inhibitors, corticosteroids, methotrexate, leflunomide, hydroxychloroquine, sulfasalazine, and apremilast) could be used as add-on therapy to bimekizumab at any time from week 20 or later at the investigator’s discretion.

Protocol Amendment

A total of 4 protocol amendments were reported in the BE MOBILE 2 trial. Of these, 2 amendments were particularly noteworthy and impactful. Protocol amendment 1, made September 11, 2019, removed eligibility for patients with psoriatic arthritis from study enrolment. Protocol amendment 2, made October 17, 2019, restricted eligibility to patients whose disease had failed to respond to 2 different NSAIDs that had been taken at the maximum tolerated dose for a total of 4 weeks.

Outcomes

A list of efficacy end points assessed in this Clinical Review Report is provided in Table 6, followed by descriptions of the outcome measures. The summarized end points are based on outcomes included in the sponsor’s summary of clinical evidence as well as any outcomes identified as important to this review according to the clinical expert consulted by CADTH and the input from the patient and clinician groups and public drug plans. Using the same considerations, the CADTH review team selected end points that were considered the most relevant to inform CADTH’s expert committee deliberations, and finalized this list of end points in consultation with members of the expert committee. All summarized efficacy end points were assessed using GRADE. Select notable harms outcomes considered important for informing CADTH’s expert committee deliberations were also assessed using GRADE.

Table 6: Outcomes Summarized From the Studies Included in the Sponsor’s Systematic Review

ASAS40 = improvement of 40% or more in the Assessment of SpondyloArthritis International Society; ASQoL = Ankylosing Spondylitis Quality of Life; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASFI = Bath Ankylosing Spondylitis Functional Index; MASES = Maastricht Ankylosing Spondylitis Enthesitis Score; NRS = numeric rating scale, NSP = nocturnal spinal pain; WPAI-SHP = Work Productivity and Activity Impairment Questionnaire–Specific Health Problem.

^aStatistical testing for these end points was adjusted for multiple comparisons (i.e., sequential testing procedure).

Sources: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸ The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

A summary of instruments used in the BE MOBILE 2 trial are provided in Table 7.

Improvement of 40% or More in the ASAS

The primary end point in the BE MOBILE 2 trial was the ASAS40 response rate at week 16. An ASAS40 response is defined as a relative improvement of at least 40% and an absolute improvement from baseline of at least 2 units (range of 0 to 10 units on an NRS) in at least 3 of the 4 main domains (Patient’s Global Assessment of Disease Activity [PGADA], pain assessment, BASFI, and BASDAI) without any worsening in the remaining domains.^18,19 ASAS40 response is considered important to patients, according to patient group input.

Bath Ankylosing Spondylitis Disease Activity Index

BASDAI was assessed as a secondary outcome in the BE MOBILE 2 trial. The BASDAI consists of 6 horizontal NRSs and was used to measure the disease activity of AS from the study participant’s perspective. Each NRS contains 10 units to measure the severity and duration of the 5 major symptoms: fatigue, spinal pain, peripheral joint pain and swelling, enthesitis, and morning stiffness over the last week. Scores range from 0 to 10, with lower scores indicating lower disease activity; a reduction in the BASDAI score is considered improvement. The ASAS inflammation component is calculated as the average of the 2 scores relating to morning stiffness measurements (i.e., question 5, “How would you describe the overall level of morning stiffness you have had from the time you wake up?” and question 6, “How long does your morning stiffness last from the time you wake up?”). The ASAS inflammation score ranges from 0 to 10. If 1 of the 2 morning stiffness measurements was missing, the other was used as the ASAS inflammation score. If both measurements were missing, the ASAS inflammation score was set to missing. An estimated median MID of 1.4 points (range, 0.9 to 1.8) has been reported for patients with active AS using anchor-based and distribution-based methods.¹⁴ The clinical expert consulted by CADTH indicated they would consider a 2-point improvement from baseline as clinically meaningful for within-group difference. The BASDAI was commonly used in clinical practice to assess patients with AS and considered an important outcome by patients. In addition, BASDAI data were the source of key inputs to the pharmacoeconomic model that was submitted to CADTH.

Bath Ankylosing Spondylitis Functional Index

BASFI was assessed as a secondary outcome in the BE MOBILE 2 trial.

The BASFI contains 10 questions and was measured by the study participant to assess their physical functions during the past week. The first 8 questions evaluate activities related to functional anatomic limitations due to the course of AS. The final 2 questions evaluate the study participants’ ability to cope with everyday life. An NRS ranging from 0 (easy) to 10 (impossible) was used to answer the questions on the test. The arithmetic mean of the 10 scores gives the BASFI score, which is a value between 0 and 10, with lower scores indicating better physical function. In case of missing answers to 1 or 2 of the single items within the questionnaire, the BASFI score was calculated by imputing the missing items using the mean of the completed items. Then, the BASFI score was calculated as described previously. If more than 2 of the items were missing, the BASFI score was shown as missing. An estimated median MID of 1.1 points (range, 1.0 to 1.4) has been reported for patients with AS using anchor-based and distribution-based methods.¹⁴ According to patient input, the BASFI has been commonly used in clinical practice to assess patients with AS and is considered an important outcome by patients.

Nocturnal Spinal Pain

NSP (based on an NRS) was assessed as a secondary outcome in the BE MOBILE 2 trial. The NSP score was analyzed as a secondary end point. The spinal pain experienced by participants was measured using 2 separate questions: total pain in the spine due to AS, and pain in the spine at night (i.e., the NSP) due to AS. Participants considered the average amount of pain in the preceding week and reported it on a scale of 0 (no pain) to 10 (most severe pain).^18,19 Using anchor-based and distribution-based methods, an estimated median MID of 1.5 points (range, 1.1 to 2.3) has been reported for patients with active AS.¹⁴ According to patient input, the NSP has been commonly used in clinical practice to assess patients with AS and is considered an important outcome by patients.

Maastricht Ankylosing Spondylitis Enthesitis Score

MASES was assessed as a secondary outcome in the BE MOBILE 2 trial. The MASES is an index that measures the severity (i.e., intensity and extent) of enthesitis through the assessment of 13 entheses (bilateral costochondral [1], costochondral [7], anterior superior iliac spine, posterior iliac spine, iliac crest and proximal insertion of the Achilles tendon sites, and the fifth lumbar vertebral body spinous process). Each result is scored as 0 or 1 and then summed for a possible score of 0 to 13, with higher scores indicating worse enthesitis. The enthesitis-free state is based on the MASES index and is defined as study participants having achieved a MASES index of 0 at week 16.^18,19 According to patient input, the MASES has been commonly used in clinical practice to assess patients with AS and is considered an important outcome by patients.

Ankylosing Spondylitis Quality of Life

ASQoL was assessed as a secondary outcome in the BE MOBILE 2 trial. The ASQoL is an 18-item questionnaire that was developed to measure HRQoL in patients with active AS. Each item on the questionnaire is given a score of 1 (yes, QoL impaired) or 0 (no, QoL not impaired). Total scores range from 0 to 18, with higher scores indicating worse HRQoL.^18,19 An estimated MID of 1 unit of worsening (i.e., change of + 1) or 2 units of improvement (i.e., change of −2) in patients with AS was identified in the literature.¹⁵ The ASQoL is considered an important outcome by patients with AS, according to the patient input..

Work Productivity and Activity Impairment Questionnaire–Specific Health Problem

The WPAI-SHP (version 2.0) is a patient-reported questionnaire that assesses patient’s employment status, work absenteeism, work impairment while working (presenteeism), overall work, and daily activity impairment attributable to a specific health problem.^56,57 Five out of the 6 items of the WPAI-SHP are regrouped into 4 dimensions, with scores expressed as percentages, where higher numbers indicate greater impairment and less productivity, i.e., worse outcomes, as described in the WPAI-SHP scoring rules. The WPAI-SHP is considered an important outcome by patients with AS, according to the patient input.

Safety Outcomes

In the BE MOBILE 2 trial, patients who experienced serious and nonserious AEs during the double-blind treatment period had their experiences recorded and reported using the Medical Dictionary for Regulatory Activities (version 19.0). Every AE was followed until it resolved or stabilized, the investigator determined it was no longer clinically significant, or the patient was lost to follow-up. If a patient’s AE was ongoing at the end of the study, follow-up was to be provided until resolution or a stable level of sequelae was achieved, the investigator no longer deemed the AE clinically significant, or the patient was lost to follow-up. If no follow-up was provided, the investigator had to justify. The follow-up was usually continued for 20 weeks after the patient discontinued treatment with bimekizumab. Notable harms included serious infection, fungal infection, opportunistic infection, malignancy, major adverse cardiac event, neutropenia, suicidal ideation and behaviour, IBD, hypersensitivity reaction, and liver injury or disorder.

Table 7: Summary of Outcome Measures and Their Measurement Properties

AS = ankylosing spondylitis; ASAS = Assessment of SpondyloArthritis International Society; ASAS20 = improvement of 20% or more in the Assessment of SpondyloArthritis International Society; ASAS40 = improvement of 40% or more in the Assessment of SpondyloArthritis International Society; ASQoL = Ankylosing Spondylitis Quality of Life; axSpA = axial spondylarthritis; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASFI = Bath Ankylosing Spondylitis Functional Index; BASMI = Bath Ankylosing Spondylitis Metrology Index; CI = confidence interval; ES = effect size; HRQoL = health-related quality of life; ICC = intraclass coefficient; MASES = Maastricht Ankylosing Spondylitis Enthesitis Score; MEI = Mander Enthesitis Index; MID = minimal important difference; NRS = numeric rating scale; NSP = nocturnal spinal pain; PGADA = Patient’s Global Assessment of Disease Activity; QoL = quality of life; VAS = visual analogue scale; WPAI-SHP = Work Productivity and Activity Impairment Questionnaire–Specific Health Problem.

^aAnalyses of the MASES index were restricted to the subset of study participants with enthesitis at baseline (i.e., a MASES index score > 0). If ≥ 7 items were available, MASES was derived by dividing the sum score by the number of available assessments and multiplying the result by 13. If < 7 items were available, MASES was treated as missing. An enthesitis-free state based on the MASES index was a categorical status defined as achieving a MASES index score of 0 (i.e., complete resolution of enthesitis) at week 16 in the BE MOBILE 2 trial.

Source: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸

Statistical Analysis

Details of the statistical analysis methods used in the BE MOBILE 2 trial are summarized in Table 8.

Sample Size and Power Calculation

The sample size and power calculations were done at a significance level of 0.05 in a 2-sided 2-sample Chi square test with continuity correction, in addition to using software (nQuery Advisor 7.0). Approximately 300 study participants were planned to be randomly assigned (2:1 ratio) to either the treatment group (200 patients) or the placebo group (100 patients). The following assumptions were made regarding sample size and power calculations:

• Using the ASAS40 response data from the phase IIb bimekizumab study (AS0008), it was assumed there would be a 46.7% rate of response to bimekizumab 160 mg every 4 weeks at week 12.⁷¹

• It was assumed the treatment responses at week 12 and week 16 would be the same.

• The ASAS40 response rate for placebo at week 16 was assumed to be 15%.

In addition, the observed ASAS40 response rates were adjusted to account for a higher number of study participants with prior exposure to TNF alpha inhibitors and a higher number of study participants with early withdrawal.

Of note, the BE MOBILE 2 trial was not powered to detect any change in the ASAS40 response rate between bimekizumab and placebo in subgroup analyses, except for the subgroup of patients who are TNF inhibitor–naive.

Statistical Test or Model

The primary end point of ASAS40 response at week 16 was analyzed for all study participants in the randomized set. The statistical null hypothesis for the ASAS40 response at week 16 was that there is no difference between the bimekizumab treatment and placebo groups in the proportion of study participants with an ASAS40 response (i.e., the conditional OR for ASAS40 response in the bimekizumab treatment compared with placebo was equal to 1). The alternative hypothesis was that there is a difference between bimekizumab treatment and placebo. A logistic regression model was used to assess the treatment effect on ASAS40 response at week 16. The model included fixed effects for treatment, and prior exposure to TNF alpha inhibitors (yes versus no) and region as stratification factors. The suitability of including these variables in the model was assessed using the Hosmer-Lemeshow goodness-of-fit test.

Multiple Testing Procedure

A fixed sequence testing procedure was applied for the primary end point of ASAS40 and the key secondary end points. The testing procedure accounted for multiplicity and control of the familywise type I error rate at a 2-sided alpha of 0.05. According to this strategy, the statistical testing of an end point can be investigated only if the null hypothesis for the previous end point has been rejected (i.e., if P < 0.05). Figure 3 shows the testing order for the primary and key secondary end points in the BE MOBILE 2 trial.

Figure 3: Sequential Testing Procedure of Primary and Key Secondary Efficacy End Points From the BE MOBILE 2 Trial

Source: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸

Data Imputation Methods

The primary efficacy analysis evaluated the composite estimand using NRI that combines the clinically meaningful improvement from baseline in ASAS40 response at week 16 and the intercurrent event of not discontinuing early from study treatment for any reason before week 16.

The following 4 attributes describe the composite estimand that was used to define the treatment effect of interest for the primary efficacy analysis:

• Population: Study participants were enrolled according to the protocol-specified inclusion and exclusion criteria and randomized to bimekizumab.

• Study participant-level outcome: ASAS40 at week 16.

• Intercurrent event handling: An intercurrent event was defined as discontinuation of study treatment before week 16. A composite strategy was implemented in which a positive clinical outcome was defined as achieving ASAS40 at week 16 and not discontinuing study treatment through week 16.

• Population-level summary measure: Conditional OR comparing bimekizumab with placebo.

Intercurrent events were acknowledged as an unfavourable outcome for the composite because study participants with intercurrent events were categorized as not having experienced a response to the study treatment (nonresponder). Consequently, if the date of an intercurrent event occurred before or at week 16, study participants were considered nonresponders at week 16. Additionally, missing data at week 16 that were not preceded by an intercurrent event were imputed as nonresponders.

Other methods were also used to assess the effects of missing data as defined in the sensitivity analyses.

Subgroup Analyses

Subgroup analyses were performed on the primary efficacy end point of ASAS40 response at week 16 on the randomized set. Subgroups analyses (except for the subgroup of patients who were TNF inhibitor–naive) did not take multiplicity into account. The following subgroups are considered relevant based on input from the clinical expert consulted by CADTH:

• gender (male, female)

• disease duration (< 2 years, ≥ 2 years)

• body mass index (BMI) (< 18.5, ≥ 18.5 to < 25, ≥ 25 to < 30, ≥ 30)

• high-sensitivity C-reactive protein level (≤ upper limit of normal of 5 mg/L, > upper limit of normal of 5 mg/L)

• prior exposure to TNF alpha inhibitors (yes, no)

• HLA-B27 positivity (yes, no)

For each subgroup analysis, a logistic regression was fitted that involved the same terms from the primary analysis model and additional terms for the subgroup and the subgroup by treatment interaction. However, for the analysis by stratification variables (i.e., prior exposure to TNF alpha inhibitors and region), the analysis model contained the same terms from the primary analysis model plus the subgroup by treatment interaction only. The same imputation method as the 1 used for the primary analysis (i.e., NRI) was used to handle missing data. The covariates were provided in the same order as for the primary analysis model, with the terms for subgroup and for subgroup by treatment interaction added at the end of the model statement. For each subgroup category and each treatment group, the number of responders, the adjusted responder rate with the associated 95% CI, and the adjusted OR (for the comparison of bimekizumab and placebo) and associated 95% CI were provided. The ORs and associated 95% CIs for each subgroup category are displayed on a single forest plot in Appendix 1 (Figure 7).

Sensitivity Analyses

Several planned sensitivity analyses were conducted for the primary outcome and other outcomes using additional analysis sets (per-protocol set, full analysis set, and COVID-19–free set) and alternative methods for handling missing data (multiple imputation, treatment policy strategy, observed case, or tipping-point analysis), as presented in Table 8.

Secondary Outcomes

The change from baseline in BASFI at week 16, change from baseline in NSP score at week 16, and change from baseline in ASQoL total score at week 16 were analyzed using the analysis of covariance with treatment, region, and prior exposure to TNF alpha inhibitors (yes versus no) as fixed effects and the respective baseline value as covariate. Missing data were imputed using reference-based multiple imputation.

The adjusted enthesitis-free rate (based on the MASES index) at week 16 among the study participants with enthesitis at baseline was analyzed using logistic regression for treatment effect and adjusted for prior TNF alpha inhibitor exposure (yes versus no) and region; the comparison was based on the 2-sided Wald test (alpha = 0.05). Missing data were imputed using multiple imputation.

The change from baseline in WPAI-SHP at week 16 was analyzed using the analysis of covariance with treatment, region, and prior TNF alpha inhibitor exposure (yes versus no) as fixed effects and the baseline value as covariate. Work productivity and activity impairment scores were based on 1 item related to activity impairment while working (presenteeism), 2 items related to absenteeism, and multiple items related to overall work productivity. A score could not be calculated if a response to a corresponding item was missing. Because questions 2 to 6 were only relevant when the answer to question 1 (Are you currently employed?) was yes, and because the imputation of results was questionable without knowing the answer to question 1, only observed case data were analyzed for this end point.

Table 8: Statistical Analysis of Efficacy End Points From the BE MOBILE 2 Trial

ANCOVA = analysis of covariance; ASAS40 = improvement of 40% or more in the Assessment of SpondyloArthritis International Society; ASQoL = Ankylosing Spondylitis Quality of Life; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASFI = Bath Ankylosing Spondylitis Functional Index; FAS = full analysis set; MASES = Maastricht Ankylosing Spondylitis Enthesitis Score; MI = multiple imputation; NR = not reported; NSP = nocturnal spinal pain; OC = observed case; PPS = per-protocol set; TNF = tumour necrosis factor; WPAI-SHP = Work Productivity and Activity Impairment Questionnaire–Specific Health Problem.

Sources: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸ The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

Analysis Populations

A summary of analysis populations used in the BE MOBILE 2 trial is provided in Table 9.

Table 9: Analysis Populations of the BE MOBILE 2 Trial

Sources: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸ The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

Results

Patient Disposition

A total of 612 patients were screened; of these, 280 patients (45.8%) did not pass screening. Ineligibility due to not meeting inclusion criterion 4 (diagnosis of AS per mNY criteria, including documented radiologic evidence based on central laboratory reading) was the most common reason (42.2%) for screen failure. A total of 332 patients were randomized and started the double-blind treatment period, 221 patients in the bimekizumab group and 111 patients in the placebo group. The percentages of study participants who completed the double-blind treatment period were similar between the treatment groups: 213 patients (96.4%) in the bimekizumab group and 109 patients (98.2%) in the placebo group. During the double-blind treatment period, a higher percentage of patients in the bimekizumab group (3.6%) discontinued from the study compared with the placebo group (1.8%). The most common primary reasons for discontinuation during the double-blind treatment period were due to an AE (1.4% for bimekizumab versus 0 for placebo) and withdrawal by study participant (1.4% versus 0.9%).

Important Protocol Deviations

The incidence of important protocol deviations was similar between the treatment groups (5.0% for bimekizumab versus 4.5% for placebo) (Table 11). Overall, the most common important protocol deviation was prohibited concomitant medication use (1.4% versus 2.7%), which was reported more frequently in the placebo group compared with the bimekizumab group.

Table 10: Summary of Patient Disposition From the BE MOBILE 2 Trial (Enrolled Set)

AS = ankylosing spondylitis; FAS = full analysis set; mNY criterial = modified New York criteria; PP = per protocol.

^aMost study participants were ineligible because they did not meet inclusion criterion 4 (diagnosis of AS per mNY criteria, including documented radiologic evidence [X-ray] read by a central laboratory).

^bThe cases with “other” as the reason for screening failure relate mainly to the COVID-19 situation, such as hospital closures and UCB Inc.’s decision to stop enrolment during the first pandemic wave.

Sources: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸ The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

Table 11: Summary of Important Protocol Deviations From the BE MOBILE 2 Trial (Randomized Set)

Source: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸

Baseline Characteristics

Overall, demographic characteristics were well balanced between treatment groups. The mean age of all study participants was 40.4 years with a range of 19 to 80 years. The majority of study participants were male (72.3%) and white (80.4%), 27.7% of participants were female, 0.3% of participants were Black, 0.9% participants were from another or a mixed racial group, and the racial group information was missing for 1.2% of participants. Study participants were most commonly enrolled in the following regions: Eastern Europe (49.1%), Western Europe (29.8%), Asia (18.4%), and North America (2.7%).

The treatment groups were generally well balanced at baseline with respect to AS-related and other disease characteristics. At baseline, the majority of all study participants were using NSAID therapies (79.8%) and anti-TNF therapy had previously been used by 16.3% of study participants. Generally, the baseline disease burden factors were well balanced across treatment groups.

The baseline characteristics outlined in Table 12 are limited to those that are most relevant to this review or were considered to affect the outcomes or interpretation of the study results.

Table 12: Summary of Baseline Characteristics From the BE MOBILE 2 Trial (Safety Set)

AS = ankylosing spondylitis; ASAS = Assessment of SpondyloArthritis International Society; ASDAS = Ankylosing Spondylitis Disease Activity Score; ASDAS-CRP = Ankylosing Spondylitis Disease Activity Score using C-reactive protein; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASFI = Bath Ankylosing Spondylitis Functional Index; BMI = body mass index; DMARD = disease-modifying antirheumatic drug; geoCV = geometric coefficient of variation; HLA-B27 = human leukocyte antigen B27; hs-CRP = high-sensitivity C-reactive protein; NRS = numeric rating scale; NSAID = nonsteroidal anti-inflammatory drug; NY = New York; PGADA = Patient's Global Assessment of Disease Activity; SD = standard deviation; TNF = tumour necrosis factor; ULN = upper limit of normal.

^aCriterion A referred to low back pain and stiffness for more than 3 months that improved with exercise but was not relieved by rest.

^bCriterion B referred to limitation of motion of the lumbar spine in both the sagittal and frontal planes.

^cCriterion C referred to limitation of chest expansion relative to normal values corrected for age and sex.

^dRadiologic criterion referred to sacroiliitis grade ≥ 2 bilaterally or grade 3 to 4 unilaterally.

^eCurrent medications were medications concomitant at baseline.

^fThe other category includes participants receiving a concomitant combination of the above conventional synthetic DMARDs.

^gThis was a component of the primary outcome measure of ASAS.

^hThe ULN value for hs-CRP was 5 mg/L.

ⁱASDAS-CRP status definition: Inactive disease is when ASDAS-CRP < 1.3; low disease activity is when ASDAS-CRP ≥ 1.3 to < 2.1; high disease activity is when ASDAS-CRP ≥ 2.1 to ≤ 3.5; and very high disease activity is when ASDAS-CRP > 3.5.

Sources: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸ The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

Prior and Concomitant Disease

In the final analysis safety set (data cut-off date: September 9, 2022), a slightly higher proportion of patients in the bimekizumab group reported a previous and ongoing medical history condition compared with the placebo group (93.2% for bimekizumab versus 90.1% for placebo) (Table 13). Generally, the most frequently reported conditions or diseases (≥ 15% of patients in either treatment group) at baseline were similar between treatment groups, except for tendinitis (27.1% versus 19.8%), which was reported more frequently in the bimekizumab group, and uveitis (10.4% versus 18.9%), which was reported more frequently in the placebo group.

History of Peripheral and Extra-Articular Manifestations

In the final analysis safety set (data cut-off date: September 9, 2022), generally, the most frequently reported disease history of peripheral and extra-articular manifestations at screening or baseline were similar between treatment groups, except for enthesitis, which was reported more frequently in the bimekizumab group (29.0% for bimekizumab versus 21.6% for placebo), and uveitis, which was reported more frequently in the placebo group (14.9% versus 21.6%) (Table 14).

Table 13: Summary of Prior and Concomitant Disease From the BE MOBILE 2 Trial (Safety Set)

Source: BE MOBILE 2 trial final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸

Exposure to Study Treatments

The median time on bimekizumab during the double-blind treatment period was 112 days for both the bimekizumab and placebo groups (Table 15). Nonbiologic rescue therapy was at the investigator’s discretion as an add-on therapy to bimekizumab at any time from week 20; therefore, during the 16-week double-blind treatment period, no patients received rescue therapy. Treatment adherence was similar between treatment groups (97.2% for bimekizumab versus 97.3% for placebo) during the double-blind treatment period.

Table 14: Summary of History of Peripheral and Extra-Articular Manifestations From the BE MOBILE 2 Trial (Safety Set)

Source: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸

Table 15: Summary of Treatment Exposure During Double-Blind Treatment Period From the BE MOBILE 2 Trial

Sources: BE MOBILE 2 trial final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸ The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

Concomitant Medications

The use of concomitant medications was similar between the bimekizumab and placebo groups in the double-blind treatment period (95.0% for bimekizumab versus 93.7% for placebo) (Table 16). The most common concomitant medications (≥ 50% of patients) in the bimekizumab group versus the placebo group were musculoskeletal system medications (86.0% versus 79.3%), nonsteroidal anti-inflammatory and antirheumatic products (85.1% versus 79.3%), and drugs used to treat the alimentary tract and metabolism, which were reported more frequently in the bimekizumab group compared with the placebo group (56.6% versus 50.5%).

The most common concomitant axSpA-related medications used during the double-blind treatment period were NSAIDs (81.9% for bimekizumab versus 78.4% for placebo) and conventional synthetic DMARDs, which were reported more frequently in the bimekizumab group compared with the placebo group (21.3% versus 18.0%). Use of analgesics and opioids was reported by a slightly higher proportion of patients in the placebo group compared with the bimekizumab group (14.9% versus 15.3%) (Table 17). The most common concomitant NSAIDs used during the double-blind treatment period were etoricoxib (17.2% for bimekizumab versus 15.3% for placebo) and meloxicam (15.4% versus 11.7%), which were both reported more frequently in the bimekizumab group compared with the placebo group, and diclofenac sodium, which was reported more frequently in the placebo group compared with the bimekizumab group (7.2% versus 11.7%). No study participants used concomitant bDMARDs (TNF inhibitors or non–TNF inhibitors).

Table 16: Summary of Concomitant Medications During Double-Blind Treatment Period the BE MOBILE 2 Trial (Safety Set)

Source: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸

Table 17: Summary of Concomitant axSpA-Related Medications During the Double-Blind Treatment Period of the BE MOBILE 2 Trial (Safety Set)

axSpA = axial spondyloarthritis; csDMARD = conventional synthetic disease-modifying antirheumatic drug; NSAID = nonsteroidal anti-inflammatory drug.

^aThe unclassified category included the following medications: tizanidine hydrochloride, tolperisone hydrochloride, baclofen, tizanidine, gabapentin, pregabalin, and metamizole.

Source: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸

Efficacy

Details of key efficacy data in the randomized set during the double-blind treatment period are summarized in Table 18.

Improvement of 40% or More in the ASAS

At week 16 of the double-blind treatment period, patients in the bimekizumab group reported a higher adjusted ASAS40 response rate (41.5%) compared with the placebo group (19.8%), with a between-group difference of 21.8% (95% CI, 11.4% to 32.1%). This corresponded to an OR of 2.88 (95% CI, 1.71 to 4.87; P < 0.001) in favour of bimekizumab.

In the exploratory analyses, the ASAS40 response rate for patients in the bimekizumab group increased from week 16 (44.8%) to week 52 (58.4%). In patients who were switched from placebo to bimekizumab at week 16, the adjusted ASAS40 response rate increased from week 16 (22.5%) to week 24 (56.8%), and further increased by week 52 (68.5%). Refer to Appendix 1 for the detailed exploratory analyses data.

Table 18: Summary of Key Efficacy Results From the BE MOBILE 2 Trial (Randomized Set)

ASAS40 = improvement of 40% or more in the Assessment of SpondyloArthritis International Society; ASQoL = Ankylosing Spondylitis Quality of Life; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASFI = Bath Ankylosing Spondylitis Functional Index; CI = confidence interval; LS = least squares; MASES = Maastricht Ankylosing Spondylitis Enthesitis Score; NRS = numeric rating scale; NSP = nocturnal spinal pain; SD = standard deviation; SE = standard error; WPAI-SHP = Work Productivity and Activity Impairment Questionnaire–Specific Health Problem.

^aAt week 16, missing data for ASAS40 were imputed as nonresponse.

^bAdjusted response rate, odds ratio, and P values for the comparison of bimekizumab vs. placebo were calculated using logistic regression with factors for treatment, prior TNF alpha inhibitor exposure, and region.

^cMissing data of change from baseline at week 16 for the BASDAI total score, BASFI, NSP score, and ASQoL total score were imputed using reference-based multiple imputation.

^dMissing data for MASES index change from baseline at week 16 and enthesitis-free state based on the MASES index at week 16 in the subgroup of study participants with enthesitis at baseline were imputed using multiple imputation.

^eThe analyses of change from baseline in WPAI-SHP at week 16 used observed cases.

Sources: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸ The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

Prespecified subgroup analyses of the ASAS40 response rate at week 16 were generally consistent with the primary analysis across all prespecified subgroups, except for subgroups of patients with a BMI of at least 30 kg/m² and patients with a BMI of less than 18.5 kg/m². It is noted that the sample sizes in these subgroups were smaller. At week 16 of the double-blind treatment period, patients in the bimekizumab group who were TNF alpha inhibitor–naive or –experienced reported a higher adjusted ASAS40 response rate compared with those in the placebo group (45.7% and 40.5% for bimekizumab versus 23.4% and 17.6% placebo). This corresponded to ORs of 2.79 (95% CI, 1.59 to 4.91) and 3.48 (95% CI, 0.84 to 14.40) in favour of bimekizumab in the TNF alpha inhibitor–naive and –experienced subgroups, respectively. In the exploratory analyses, the ASAS40 response rate for patients who were TNF alpha inhibitor–naive in the bimekizumab group increased from week 16 (45.7%) to week 52 (58.7%). In patients who were TNF alpha inhibitor–naive and were switched from placebo to bimekizumab at week 16, the response rate in TNF alpha inhibitor–naive participants increased from week 16 (23.4%) to week 24 (59.6%) and further increased at week 52 (71.3%).

The results of sensitivity and supportive analyses, including the tipping-point analyses, were in line with the primary efficacy results. Patients in the bimekizumab group had higher ASAS40 response rates compared with patients in the placebo group. In the mean change from baseline in the individual components of the ASAS40, at week 16, patients in the bimekizumab group reported a greater reduction from baseline in LS mean (reductions reflect improvement) compared with patients in the placebo group in each of the ASAS40 components: PGADA (−2.7 for bimekizumab versus −1.4 for placebo), total spinal pain (−3.1 versus −1.7), BASFI physical function score (−2.0 versus −1.0), and BASDAI mean score for questions 5 and 6 (inflammation and morning stiffness) (−3.0 versus −1.9).

Bath Ankylosing Spondylitis Disease Activity Index

In the BE MOBILE 2 final analysis, at baseline, the mean BASDAI score was 6.45 in the bimekizumab group and 6.51 in the placebo group. At week 16, patients in the bimekizumab group had a greater reduction from baseline in the LS mean (reductions reflect improvement) in BASDAI score compared with patients in the placebo group (LS mean of −2.7 for bimekizumab versus −1.7 for placebo). An estimated median MID of 1.4 points (range, 0.9 to 1.8) was identified in the literature.¹⁴ The clinical expert consulted by CADTH indicated they would consider a 1-point difference between groups to be clinically meaningful. The difference in LS mean between treatment groups was −1.04 (95% CI, −1.5 to −0.6; P < 0.001) in favour of bimekizumab.

In the exploratory analyses, the mean change from baseline in BASDAI total score for patients in the bimekizumab group decreased up to week 16 (−2.9) and decreased further to week 52 (−3.6) for patients in the bimekizumab group. In patients who were switched from placebo to bimekizumab at week 16, the mean change in BASDAI total score decreased from baseline to week 16 (−1.9) to week 24 (−3.26), and further decreased to week 52 (−4.0). Refer to Appendix 1 for the detailed exploratory analyses data.

Bath Ankylosing Spondylitis Functional Index

In the BE MOBILE 2 final analysis, at baseline, the mean BASFI score was 5.3 in the bimekizumab group and 5.2 in the placebo group. At week 16, patients in the bimekizumab group had a greater reduction from baseline in LS mean (reductions reflect improvement) in BASFI score compared with patients in the placebo group, which worsened (LS mean of −1.9 for bimekizumab versus −1.0 for placebo). An estimated median MID of 1.1 points (range, 1.0 to 1.4) was identified in the literature.¹⁴ The difference in LS mean between treatment groups was −1.1 (95% CI, −1.5 to −0.6; P < 0.001) in favour of bimekizumab. The clinical expert consulted by CADTH identified an MID of 1 point for between-group difference.

In the exploratory analyses, the mean reduction in BASFI score further decreased from baseline to week 16 (−2.2) to week 52 (−2.8) for patients in the bimekizumab group. In patients who were switched from placebo to bimekizumab at week 16, the mean change in BASFI score decreased from baseline to week 16 (−1.1) to week 24 (−2.2), and further decreased at week 52 (−2.8). Refer to Appendix 1 for the detailed exploratory analyses data.

Nocturnal Spinal Pain (Based on an NRS)

In the BE MOBILE 2 final analysis, at baseline, the mean NSP score (based on an NRS) was 6.6 in the bimekizumab group and 6.8 in the placebo group. At week 16, patients in the bimekizumab group had a greater reduction from baseline in LS mean (reductions reflect improvement) in NSP score (based on an NRS) compared with patients in the placebo group, which worsened (LS mean of −3.2 for bimekizumab versus −1.7 for placebo). An estimated median MID of 1.5 points (range, 1.1 to 2.3) was identified in the literature.¹⁴ The difference in LS mean between treatment groups was −1.5 (95% CI, −2.0 to −1.0; P < 0.001) in favour of bimekizumab. The clinical expert consulted by CADTH identified an MID of 1 point between groups.

In the exploratory analyses, the mean reduction in NSP (based on an NRS) further decreased −3.3 at week 16 and −4.1 at week 52 for patients in the bimekizumab group. In patients who were switched from placebo to bimekizumab at week 16, the mean change in NSP (based on an NRS) decreased from −1.9 at week 16 to −3.7 at week 24, and further decreased −4.6 at week 52. Refer to Appendix 1 for the detailed exploratory analyses data.

Enthesitis-Free State Based on MASES in Patients with Enthesitis at Baseline

At week 16, the enthesitis-free state based on a MASES index assessment was performed only in patients with enthesitis at baseline, including 132 patients in the bimekizumab group and 67 patients in the placebo group. At week 16 of the double-blind treatment period, patients with enthesitis at baseline in the bimekizumab group reported a higher adjusted enthesitis-free rate compared with those in the placebo group (43.8% for bimekizumab versus 23.9% for placebo), with a between-group difference of 19.8% (95% CI, 6.3% to 33.4%). This corresponded to an OR of 2.47 (95% CI, 1.30 to 4.68; nominal P = 0.006) in favour of bimekizumab. No estimate of a between-group MID was identified by CADTH, but clinical expert input suggested a 15% difference would be clinically important; therefore, the absolute difference between groups was clinically important. The enthesitis-free state outcome was not controlled for type I error rate; thus, these data should be interpreted as supportive evidence only.

In the exploratory analyses, the enthesitis-free rate for patients with enthesitis at baseline in the bimekizumab group at week 16 and week 52 was 51.5% and 50.8%, respectively. In patients with enthesitis at baseline who were switched from placebo to bimekizumab at week 16, the enthesitis-free rate at week 16 was 32.8% and 49.3% at week 24, and decreased to 46.3% at week 52. Refer to Appendix 1 for the detailed exploratory analyses data.

Ankylosing Spondylitis Quality of Life

In the BE MOBILE 2 final analysis, at baseline, the mean ASQoL score was 9.0 in the bimekizumab group and 8.5 in the placebo group. At week 16, patients in the bimekizumab group had a greater reduction from baseline in LS mean (reductions reflect improvement) in ASQoL score compared with patients in the placebo group, which worsened (LS mean of −4.6 for bimekizumab versus −3.1 for placebo). An MID of 1 unit of worsening (i.e., change of + 1) or 2 units of improvement (i.e., change of −2) was identified in the literature.¹⁵ The difference in LS mean between treatment groups was −1.5 (95% CI, −2.4 to −0.7; P < 0.001) in favour of bimekizumab. The clinical expert consulted by CADTH identified an MID of 2 points between groups.

In the exploratory analyses, the mean reduction in ASQoL further decreased −5.0 from baseline to week 16 and −5.7 from baseline to week 52 for patients in the bimekizumab group. In patients who were switched from placebo to bimekizumab at week 16, the mean change in ASQoL decreased −3.2 from baseline to week 16 and −4.8 from baseline to week 24, and further decreased −5.6 from baseline to week 52. Refer to Appendix 1 for the detailed exploratory analyses data.

Work Productivity and Activity Impairment Questionnaire–Specific Health Problem

At baseline, the mean WPAI-SHP scores were similar between the bimekizumab group and the placebo group for percent time missed due to disease-related problems (11.5 for bimekizumab versus 10.9 for placebo), percent impairment while working due to disease-related problems (46.1 versus 42.3), percent overall work impairment due to disease-related problems (49.2 versus 43.9), and percent activity impairment due to disease-related problems (53.3 versus 54.1). At week 16, the bimekizumab group compared with the placebo group had a greater mean reduction from baseline in the following:

• WPAI-SHP score for percent time missed due to disease-related problems (−5.5 versus −1.2), with a between-group difference of −2.9 (95% CI, –6.9 to 1.1)

• percent impairment while working due to disease-related problems (−20.8 versus −6.1), with a between-group difference of −12.5 (95% CI, –18.2 to −6.9)

• percent overall work impairment due to disease-related problems (−22.2 versus −6.7), with a between-group difference of −12.8 (95% CI, –18.7 to −6.9)

• percent activity impairment due to disease-related problems (−23.3 versus −14.4), with a between-group difference of −9.4 (95% CI, –13.9 to −4.9).

No MIDs for WPAI-SHP were identified in the literature. The WPAI-SHP outcome was not controlled for type I error; thus, these data should be interpreted as supportive evidence only.

In the exploratory analyses, generally, compared with the placebo group, the bimekizumab group had a greater mean reduction from baseline in WPAI-SHP score at weeks 24, 26, and 52 for percent time missed due to problems (related to disease), percent impairment while working due to problems, and percent overall work impairment due to problems. Patients in the bimekizumab and placebo groups reported similar reductions in percent activity impairment due to problems at weeks 24, 26, and 52. Refer to Appendix 1 for the detailed exploratory analyses data.

Harms

Details of AEs, SAEs, withdrawals due to AEs, mortality, and notable harms data in the safety set during the double-blind treatment period are summarized in Table 19.

Adverse Events

During the double-blind treatment period, an AE was reported for 54.3% of patients in the bimekizumab group and 43.2% in the placebo group. The most common AEs (i.e., reported by ≥ 5% of patients in either group) were infections and infestations (28.1% for bimekizumab versus 22.5% for placebo), including nasopharyngitis (7.7% versus 3.6%), gastrointestinal disorders (13.1% versus 9.9%), nervous system disorders (8.1% versus 4.5%), upper respiratory tract infections (2.7% versus 7.2%), and eye disorders (2.3% versus 6.3%).

Serious Adverse Events

During the double-blind treatment period, SAEs were reported for 2.3% of patients in the bimekizumab group and 0.9% of patients in the placebo group. The following SAEs were commonly reported in the bimekizumab group but not reported by any patients in the placebo group: goitre (0.5%), colitis ulcerative (0.5%), Crohn disease (0.5%), cholelithiasis (0.5%), and hepatitis A (0.5%).

Withdrawals Due to Adverse Events

During the double-blind treatment period, 2.7% of patients in the bimekizumab group and no patients in the placebo group discontinued study due to AEs. The commonly reported AEs that led to study discontinuation were abnormal psychiatric evaluation (0.9%), lymphoid tissue hyperplasia (0.5%), Crohn disease (0.5%), oral candidiasis (0.5%), and rash (0.5%).

Mortality

No deaths due to AEs were reported during the double-blind treatment period in the BE MOBILE 2 trial.

Notable Harms

Serious infections, fungal infections, opportunistic infections, malignancies, major adverse cardiac events, neutropenia, suicidal ideation and behaviours, IBD, hypersensitivity reactions, and liver injuries or disorders were considered notable harms by the sponsor and/or the clinical expert consulted by CADTH. The most commonly reported notable harms were hypersensitivity reactions (7.7% for bimekizumab versus 1.8% for placebo), fungal infections (6.3% versus 0%), liver injuries or disorders (4.5% versus 3.6%), IBD (0.9% versus 0%), neutropenia (0.5% versus 0%), and serious infections (0.5% versus 0.9%). No patients in either group reported any opportunistic infection, malignancy, major adverse cardiac event, or suicidal ideation and behaviour.

Table 19: Summary of Harms Results During From the BE MOBILE 2 Trial (Safety Set)

^aIncluding anaphylaxis.

^bIncluding drug related hepatic disorders, excluding liver neoplasms, benign (including cysts and polyps) and liver neoplasms, malignant and unspecified.

^cAdjudicated definite and probable inflammatory bowel disease.

Sources: BE MOBILE 2 final Clinical Study Report (data cut-off date: September 9, 2022).¹⁸ The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

Critical Appraisal

Internal Validity

The BE MOBILE 2 trial was a phase III, multicentre, randomized, double-blind, placebo-controlled study to evaluate the efficacy and safety of bimekizumab 160 mg/mL administered subcutaneously every 4 weeks compared with placebo in treating patients with active AS. Eligible patients were assigned on a 2:1 ratio to a treatment group using an interactive voice or web response system with stratification by region and prior exposure to TNF alpha inhibitors (yes versus no). Generally, baseline characteristics were balanced between the treatment groups, which indicated that the randomization was successful. According to the clinical expert consulted by CADTH, the baseline characteristics of patients in the BE MOBILE 2 trial were reflective of the general population of patients with active AS. More patients discontinued due to AEs in the bimekizumab group compared with the placebo group (1.4% versus 0%). The clinical expert confirmed this is reflective of clinical practice, as they would expect more side effects with an immunomodulatory therapy. The CADTH review team agreed with the clinical expert that the number of patients who discontinued due to AEs was small in the bimekizumab group, and the difference between treatment groups was not significant, but noted there may be a potential risk of unblinding due to the elevated risk of AEs in the bimekizumab group.

The CADTH review team noted there were no comparative data available beyond week 16, as patients in the placebo group were reallocated to receive bimekizumab during the 36-week maintenance period after finishing all assessments at the end of the 16-week double-blind treatment period. Therefore, it is uncertain what the direct comparative efficacy and safety of bimekizumab is after week 16.

The sponsor assumed that treatment response at week 12 and week 16 would be the same in the sample size and power calculations. The clinical expert consulted by CADTH confirmed that the assumption is reflective of clinical practice, as they do not observe much of a difference in treatment effect between week 12 and week 16.

Multiple data imputation methods, such as NRI, reference-based multiple imputation, and observed case, were used in the efficacy analysis based on variable type (i.e., binary and continuous). In the analysis of the primary end point, ASAS40, a composite estimand using NRI was applied, considering the ASAS40 is based on 4 different component scores. Therefore, the CADTH review team considered that missing data for the primary analysis in the BE MOBILE 2 trial were adequately imputed. For the analysis of the primary and key secondary end points, a fixed sequence testing procedure was employed to adjust for multiple comparisons across multiple end points, thereby controlling the type I error. The CADTH review team noted that the analyses of enthesitis-free state based on MASES index was not included in the fixed sequence testing hierarchy; thus, the results should be considered exploratory and supportive. Although the subgroup analyses were prespecified, the BE MOBILE 2 trial was not powered to detect any change in the ASAS40 response rate between bimekizumab and placebo in subgroup analyses, except for the subgroup of patients who are TNF inhibitor–naive; additionally, no formal statistical tests for interaction between subgroups were conducted. More patients in the bimekizumab group used concomitant NSAIDs, conventional synthetic DMARDs, and drugs to treat the alimentary tract and metabolism compared with the placebo group. The CADTH review team agreed with the clinical expert that the differences in the use of concomitant medications between treatment groups were small and less likely to impact treatment effect. In addition, the clinical expert confirmed that the proportions of patients using the previously mentioned concomitant medications were aligned with clinical practice.

In the BE MOBILE 2 final protocol, dated November 26, 2018, psoriatic arthritis was not listed in the inflammatory conditions that excluded patients from the trial, and patients who had intolerance to administration of at least 1 NSAID were eligible for the trial. Protocol amendment 1, made September 11, 2019, removed eligibility for patients with psoriatic arthritis from study enrolment. Protocol amendment 2, made October 17, 2019, restricted eligibility to patients who had failed to respond to 2 different NSAIDs taken at the maximum tolerated dose for a total of 4 weeks. Of note, protocol amendments 1 and 2 were made after the enrolment of the first patient (April 25, 2019). Therefore, the CADTH review team considered that these 2 protocol amendments could increase patient heterogeneity and introduce bias. The direction of the bias is uncertain, as there were no data reported on the numbers of patients in the trial with psoriatic arthritis or whose disease had failed to respond to more than 2 NSAIDs.

HRQoL is considered a relevant outcome by both patients with active AS and the clinical expert consulted by CADTH. However, the assessment of work productivity using the WPAI-SHP was not controlled for multiplicity and thus should be considered only as exploratory and supportive.

External Validity

The BE MOBILE 2 trial used placebo as the comparator group. According to the clinical expert consulted by CADTH, an anti-TNF biosimilar monoclonal antibody would be an appropriate comparator for bimekizumab. The clinical expert indicated placebo is not an appropriate comparator; head-to-head studies with an active drug would be ideal. The BE MOBILE 2 trial excluded patients who had been treated with more than 1 TNF alpha inhibitor and/or more than 2 additional non–TNF alpha biological-response modifiers, or any IL-17 biological-response modifier at any time. The clinical expert indicated that these patients should be considered eligible for bimekizumab. According to the clinical expert, although the response rate might be lower, some patients do response to bDMARDs after the failure of treatment with TNF inhibitors and IL-17 inhibitors; therefore, the clinical expert would switch treatments within the same class due to limited treatment options. According to the clinical expert, the study results would not be generalizable to these previously mentioned patients, as it is expected that the response rates will be lower in this patient population, which tends to have lower response rates with subsequent treatments in clinical practice.

The clinical expert consulted by CADTH stated that IBD is a common inflammatory condition that happens in patients with AS. According to the clinical expert, about 10% to 15% of patients with AS would have active IBD and 60% of patients would have subclinical IBD based on biopsies. Patients with active IBD were not eligible for the BE MOBILE 2 trial; thus, the number of patients with IBD flares in the trial may be lower compared with the general population of patient with AS. The clinical expert indicated that patients with a personal or family history of IBD may not be candidates for treatment with bimekizumab based on their experience with bDMARDs targeting IL-17A in patients with IBD. The presence of IBD or severe uveitis and active infections, especially fungal infections, would be contraindications for treatment with bimekizumab, as per the feedback from the clinical expert. The CADTH review team noted that IBD is a warning in the proposed product monograph. The clinical expert stated that the exclusion of patients with active IBD would not significantly compromise the generalizability of the study results, as most patients with AS would be eligible for treatment with bimekizumab in clinical practice.

In addition, there was no study site in Canada in the BE MOBILE 2 trial, which may compromise the generalizability of the study results to the clinical practice in Canada.

GRADE Summary of Findings and Certainty of the Evidence

Methods for Assessing the Certainty of the Evidence

For the pivotal studies and RCTs identified in the sponsor’s systematic review, GRADE was used to assess the certainty of the evidence for outcomes considered most relevant to inform CADTH’s expert committee deliberations, and a final certainty rating was determined as outlined by the GRADE Working Group.^16,17

• High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.

• Moderate certainty: We are moderately confident in the effect estimate. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. We use the word “likely” for evidence of moderate certainty (e.g., “X intervention likely results in Y outcome”).

• Low certainty: Our confidence in the effect estimate is limited. The true effect may be substantially different from the estimate of the effect. We use the word “may” for evidence of low certainty (e.g., “X intervention may result in Y outcome”).

• Very low certainty: We have very little confidence in the effect estimate. The true effect is likely to be substantially different from the estimate of effect. We describe evidence of very low certainty as “very uncertain.”

Following the GRADE approach, the evidence from the RCTs started as high-certainty evidence and could be rated down for concerns related to study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty-of-evidence assessment was based on the point estimate and where it was located relative to the threshold for a clinically important effect (when a threshold was available) or to the null. The target of the certainty-of-evidence assessment was the presence or absence of a clinically important effect based on thresholds informed by the clinical expert consulted by CADTH for this review for ASAS40, BASDAI, BASFI, NSP, MASES, ASQoL, and SAEs. For WPAI-SHP, there is no established MID and the clinical expert consulted by CADTH could not provide a threshold of important difference, so the target of the certainty-of-evidence assessment was the presence or absence of any (non-null) effect.

Results of GRADE Assessments

Table 2 presents the GRADE summary of findings for bimekizumab versus placebo.

Long-Term Extension Studies

Contents within this section have been informed by materials submitted by the sponsor. The following has summarized and validated by the CADTH review team.

Description of Studies

BE AGILE 2 (AS0009, NCT03355573) is a single-arm phase II long-term extension study, which followed the preceding study, BE AGILE (AS0008, NCT02963506), which was a dose-ranging, randomized, placebo-controlled phase IIb trial.

Of the 303 adult patients living with active AS who had been enrolled in the BE AGILE trial, 255 (84.5%) entered the BE AGILE 2 study, which took place at 50 sites in 10 European countries and the US. The data presented in this report were collected for up to approximately 5 years for the final analyses (data cut-off: October 19, 2022), which combined 48 weeks of treatment in the BE AGILE study and an additional 204 weeks of treatment in the BE AGILE 2 trial.⁷²

Most of the key outcomes measured in the BE MOBILE 2 trial were also evaluated in the BE AGILE 2 trial, such as ASAS40, BASDAI, BASFI, NSP, ASQoL, Bath Ankylosing Spondylitis Metrology Index, MASES, and safety outcomes.

Populations

Patients were eligible for the BE AGILE 2 trial if they had completed the BE AGILE study and, in the opinion of the investigator, were expected to benefit from continued treatment with bimekizumab in the BE AGILE 2 trial. Consistent with the inclusion criteria for the BE MOBILE 2 trial, patients were eligible for the BE AGILE 2 study if they were adults with a diagnosis of active AS based on radiographic evidence fulfilling the mNY criteria for AS, including 3 or more months of symptoms, younger than age 45 at symptom onset, with moderate to severe active disease at baseline (defined as BASDAI ≥ 4 plus spinal pain ≥ 4 on a 0 to 10 NRS from BASDAI item 2), and intolerant to or with AS that has failed to respond to NSAIDs. Patients were allowed to have previously received treatment with a TNF alpha inhibitor.

Patients were excluded from the BE AGILE 2 study if they met the withdrawal criteria in the BE AGILE trial or if they had signs or symptoms that could indicate a medically significant active infection or had an infection requiring systemic antibiotics within 2 weeks of study entry.

Interventions

All patients in the BE AGILE 2 trial received open-label treatment with bimekizumab 160 mg every 4 weeks, regardless of their prior dosing regimen in BE AGILE, the parent trial. During the long-term extension period, NSAIDs, DMARDs (methotrexate, sulfasalazine, leflunomide), and/or joint injections (e.g., intra-articular hyaluronic acid and intrabursal corticosteroid) were permitted. However, a concomitant biologic drug was not allowed under any circumstances during the long-term extension study. Medication changes, additions, and a decrease in dose or dosing frequency of any drug (except for bimekizumab) were permitted at any time after enrolment into the open-label extension study.

Outcomes

Most of the key outcomes measured in the BE MOBILE 2 trial were also evaluated in the BE AGILE 2 trial. These outcomes were defined and evaluated in the same manner as in the BE MOBILE 2 trial. For a full description of these outcomes, refer to Table 6.

Statistical Analysis

All outcomes in the BE AGILE 2 trial were descriptively evaluated compared with baseline from the BE AGILE trial (e.g., BE AGILE 2 trial week 208 [total exposure = 252 weeks] versus BE AGILE study week 0 [total exposure = 0 weeks]). As in the BE MOBILE 2 trial, missing data for binary outcomes in the BE AGILE 2 trial were imputed using NRI, while missing data for continuous outcomes were imputed using multiple imputation based on the assumption that data were missing at random. No formal hypothesis testing was conducted in the BE AGILE 2 trial and no adjustment was made for multiple comparisons.

Results

Patient Disposition

Among the 303 patients who were randomized in the BE AGILE trial, 297 patients (98.0%) completed the double-blind period, and 265 patients (87.5%) completed the dose-blind period at week 48. Nine patients did not enter the BE AGILE 2 trial and 1 patient entered the BE AGILE 2 trial but did not receive bimekizumab. A total of 256 study participants enrolled in the BE AGILE 2 trial and 255 (99.6%) participants received at least 1 dose of bimekizumab in the BE AGILE 2 trial. A majority of patients (78.9%) completed the study. The discontinuation rate for BE AGILE 2 was 20.7% and was most frequently related to TEAEs (7.4%) or withdrawal of consent (9.0%) (Table 20).

Table 20: Patient Disposition

AE = adverse event; TEAE = treatment-emergent adverse event.

^aAttended week 208 visit.

Sources: BE AGILE 2 Clinical Study Report (Table 7 to 1).⁷² The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

Exposure to Study Treatments

In the BE AGILE 2 trial safety set, 96.1% of all study participants reported having greater than 75% treatment compliance. The extent of exposure to bimekizumab during the open-label extension period is summarized in Table 21.

Table 21: Summary of Extent of Exposure

SD = standard deviation.

Sources: BE AGILE 2 Clinical Study Report.⁷² The details included in the table are from the sponsor’s summary of clinical evidence.¹⁹

A total of 25 (9.8%) study participants received rescue medication during the BE AGILE 2 trial. The rescue medications used were NSAIDs (6.7%, with the majority having a change in NSAID type or dose), DMARDs (2.4%), analgesics (2.0%), oral corticosteroids (0.8%), and intra-articular corticosteroids and ”unclassified” (0.4%).

Most patients (83.5%) were receiving concomitant during the BE AGILE 2 trial, as described in Table 22.

Table 22: Concomitant Medications

DMARD = disease-modifying antirheumatic drug; NSAID = nonsteroidal anti-inflammatory drug.

Sources: The details included in the table are from the sponsor’s summary of clinical evidence¹⁹ and the BE AGILE 2 Clinical Study Report.⁷²

Baseline Characteristics

Refer to Table 23 for baseline characteristics.

In the BE AGILE 2 trial, most patients were male (85.1%) and were positive for HLA-B27 (91.0%). Mean symptom duration was approximately 14 years (SD = 9.41) and mean disease duration (time since AS diagnosis) was approximately 7.8 years (SD = 8.5). Mean BASDAI score was approximately 7.4 points (SD = 1.4), reflecting moderately severe patient-reported disease activity. Most patients (79.2%) had received NSAIDs for their active AS in the past. Approximately 11% of patients had previously been exposed to treatment with TNF alpha inhibitors.

Table 23: Baseline Characteristics

AS = ankylosing spondylitis; ASDAS-CRP = Ankylosing Spondylitis Disease Activity Score using C-reactive protein; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASFI = Bath Ankylosing Spondylitis Function Index; BMI = body mass index; HLA-B27 = human leukocyte antigen B27; hs-CRP = high-sensitivity C-reactive protein; NSAID = nonsteroidal anti-inflammatory drug; PGADA = Patient’s Global Assessment of Disease Activity; SD = standard deviation; TNFi = tumour necrosis factor inhibitor.

Sources: The details included in the table are from the sponsor’s summary of clinical evidence,¹⁹ the BE AGILE 2 Clinical Study Report,⁷² and Baraliakos et al. (2022).⁷¹

Efficacy

ASAS40 response was sustained up to week 208 in the BE AGILE 2 trial, with response rates of 59% (147 out of 249) using NRI, and 73.1% (147 out of 201) using observed case data. During the BE AGILE 2 trial, the mean BASDAI score (n = 249) decreased from baseline and was sustained at week 208 (decrease of −4.01; SE = 0.13) versus an MID of −2.0 points. The mean BASFI score (n = 249) decreased from baseline by −3.1 (SE = 0.15) and was sustained at week 208. Relative to baseline, the NSP score (n = 249) decreased −4.55 (SE = 0.16) and was maintained at week 208 versus an MID of 1.5 points. Also, the mean ASQoL score (n = 249) decreased from baseline by −5.9 (SE = 0.3) and was maintained at all times by week 208 versus an MID of −5.1 points. Among patients with enthesitis at baseline, the mean MASES (n = 164) decreased −0.36 (SE = 0.23) and maintained improvement up to week 208 in the BE AGILE 2 trial. WPAI-SHP was not assessed in the BE AGILE 2 study.

Harms

Refer to Table 24 for harms data.

A total of 237 (92.9%) study participants reported a TEAE during the BE AGILE 2 study. The most commonly reported TEAEs were nasopharyngitis (18%), upper respiratory tract infection and COVID-19 infection (12.9% each), and bronchitis (8.6%).

There were 46 patients (18.0%) who experienced at least 1 SAE, with COVID-19 infection and pneumonia being the most common (1.2% each). Twenty-one patients (8.2%) discontinued study treatment due to a TEAE, mostly due to alanine aminotransferase (1.2%) and aspartate aminotransferase (0.8%) elevation. Two fatal TEAEs were reported during the study. One study participant experienced a fatal TEAE of road traffic accident, which was considered by study investigators to not be related to bimekizumab. One study participant who had pre-existing cardiovascular risk factors, including hypertension, type 2 diabetes, and left bundle branch block, experienced a fatal TEAE of cardiorespiratory arrest, which was not considered related to bimekizumab, per the investigator’s assessment.

Of note, fungal infection (18.4%) and hypersensitivity (11.4%) were the most common AEs of special interest reported in the safety set during the BE AGILE 2 trial. The fungal infection cases were all judged by the investigators to be mild to moderate in intensity and the vast majority did not lead to treatment discontinuation (a single patient discontinued due to perirectal abscess).

Table 24: Summary of Harms Results From Long-Term Extension Studies

ALT = alanine aminotransferase; AST = aspartate aminotransferase; IBD = inflammatory bowel disease; MACE = major adverse cardiovascular event; SAE = serious adverse event; TEAE = treatment-emergent adverse event.

^aFrequency > 5%.

^bCategory reported in ≥ 2 people.

Sources: The details included in the table are from the sponsor’s summary of clinical evidence¹⁹ and the BE AGILE 2 Clinical Study Report.⁷²

Critical Appraisal

Internal Validity

In general, the results of the efficacy end points assessed in the BE AGILE 2 trial appear to support the long-term effectiveness of bimekizumab. However, the findings from the BE AGILE 2 extension study (exposure of up to 252 weeks with the BE AGILE and BE AGILE 2 trials combined) were limited due to the lack of a control group and the nature of an open-label study. In clinical trials, the efficacy magnitude (particularly in patients’ self-reported outcomes) is often overestimated due to the nature of an open-label trial and the absence of a control group. The long-term outcome efficacy should therefore be interpreted while considering this limitation. Moreover, there is a risk of selection bias, as patients who have responded to bimekizumab and who tolerated side effects, if any, during the 2 parent studies, are more likely to continue during the extension period. This bias would also increase the chance of overestimation of the efficacy end point measures. Even though detailed information on the use of concomitant medications (NSAIDs and conventional synthetic DMARDs) and rescue medications have been provided in the open-label period, the impact of concomitant and rescue medications on the long-term efficacy assessment still remains unclear. Of note, no new safety signals or increased risks were identified with up to 5 years of bimekizumab treatment between the BE AGILE and BE AGILE 2 studies.

External Validity

No Canadian study site was included in the BE AGILE and BE AGILE 2 studies. However, the clinical expert consulted by CADTH did not raise any issues regarding the generalizability of study results to real-world clinical practice in Canada.

Indirect Evidence

Contents within this section have been informed by materials submitted by the sponsor. The following has been summarized and validated by the CADTH review team.

Objectives for the Summary of Indirect Evidence

The objective of this section is to summarize and critically appraise the methods and findings of the indirect evidence submitted by the sponsor. The sponsor submitted indirect evidence in the form of an NMA and an MAIC to address gaps in the evidence.

Description of Indirect Comparisons

The sponsor conducted a systematic literature review (SLR) to identify evidence for the NMAs and the MAIC in October 2013 and included update searches to January 2023. The searches identified 3,104 records that were potentially of interest and, of these, 286 publications were included that described 65 RCTs. Of the 65 unique trials identified, 28 and 8 trials were deemed feasible to be included in the NMA and MAIC, respectively, based on the study selection criteria described in Table 25.

Objectives

Network Meta-Analyses

The NMAs were performed to determine the clinical efficacy and safety of bimekizumab compared with other relevant interventions at weeks 12 to 16 for the treatment of patients with AS. The NMAs were conducted on 3 different networks: purely naive (100% bDMARD-naive), predominantly naive (approximately 90% bDMARD-naive), and purely experienced (100% bDMARD-experienced).

Matching-Adjusted Indirect Comparisons

The MAICs were performed to establish the long-term relative clinical efficacy of bimekizumab, where data availability allowed, compared with other IL-17A inhibitors in patients with AS at week 52.

Indirect Comparison Methods

Study Selection Methods

Network Meta-Analyses

A feasibility assessment was performed to determine which of the 65 unique RCTs identified by the SLR were suitable for inclusion in the NMA. The eligibility criteria for including studies identified by the SLR in the NMA were in addition to the eligibility criteria of the clinical SLR. The NMA eligibility criteria were applied to ensure that the trial data could be synthesized within a meta-analysis framework. Treatments were restricted to doses and schedules with marketing authorization. Of the 65 unique RCTs included in the SLR, 37 met the additional NMA eligibility criteria.

Matching-Adjusted Indirect Comparison

The NMA and MAIC analyses were informed by the same SLR. Studies reporting efficacy outcomes for RCTs of IL-17A inhibitors in AS were identified through the SLR. Of the 65 unique RCTs included in the SLR, 8 were included in the MAIC analyses.

Table 25: Study Selection Criteria and Methods for the NMAs and MAIC Submitted by the Sponsor

AS = ankylosing spondylitis; ASAS = Assessment of SpondyloArthritis International Society; ASAS20 = improvement of 20% or more in the Assessment of SpondyloArthritis International Society; ASAS40 = improvement of 40% or more in the Assessment of SpondyloArthritis International Society; ASAS5/6 = an improvement of at least 20% in 5 of 6 Assessment of SpondyloArthritis International Society domains; ASDAS = Ankylosing Spondylitis Disease Activity Score; ASDAS-CII = clinically important improvement in Ankylosing Spondylitis Disease Activity Score; ASDAS-CRP = Ankylosing Spondylitis Disease Activity Score using C-reactive protein; ASDAS-ID = Ankylosing Spondylitis Disease Activity Score; ASDAS-MI = major improvement in Ankylosing Spondylitis Disease Activity Score; ASQoL = Ankylosing Spondylitis Quality of Life; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASDAI50 = a reduction of at least 50% in the Bath Ankylosing Spondylitis Disease Activity Index score; BASFI = Bath Ankylosing Spondylitis Functional Index; BASMI = Bath Ankylosing Spondylitis Metrology Index; CRP = C-reactive protein; DMARD = disease-modifying antirheumatic drug; G-BA = Gemeinsamer Bundesausschuss; HAS = Haute Autorité de Santé; HRQoL = health-related quality of life; IL-17 = interleukin-17; IQWiG = Institute for Quality and Efficiency in Health Care; JAK = Janus kinase; MASES = Maastricht Ankylosing Spondylitis Enthesitis Score; MCS = mental component summary; NICE = National Institute for Health and Care Excellence; NMA = network meta-analysis; nr-axSpA = nonradiographic axial spondyloarthritis; NRS = numerical rating scale; NSAID = nonsteroidal anti-inflammatory drug; PBAC = Pharmaceutical Benefits Advisory Committee; PCS = physical component summary; PICOS = population, intervention, comparator(s), outcomes, and study design; PGADA = Patient’s Global Assessment of Disease Activity; r-axSpA = radiographic axial spondyloarthritis; RCT = randomized controlled trial; SF-36 = Short Form (36) Health Survey; SLR = systematic literature review; SMC = Scottish Medicines Consortium; SPARCC = Spondyloarthritis Research Consortium of Canada; TNF alpha = tumour necrosis factor alpha.

^aUnless the publication explicitly states that patients are NSAID-naive or still receiving their first NSAID, it was assumed that patients participating in RCTs for IL-17, TNF alpha, or JAK inhibitors experienced treatment failure with at least 1 NSAID.

^bThese outcomes were listed as relevant outcomes in the eligibility criteria table of the previous SLR, but there is no evidence they were ever extracted; where present, they were extracted from publications included in the April 2022 and January 2023 clinical SLR updates.

^cThese outcomes are predefined primary or secondary outcomes from the BE MOBILE 1 and BE MOBILE 2 trials, but there is no evidence they were extracted during the previous SLR; where present, they were extracted from publications included in the 2022 clinical SLR update.

^dPublications pooling data across multiple trials from the same clinical trial program were eligible for inclusion.

^eRelevant SLRs and NMAs were included at the title- or abstract-screening stage so their bibliographic reference lists could be hand-searched for relevant studies; they were then excluded at the full-text screening stage unless they presented novel data.

Source: Sponsor’s summary of clinical evidence.¹⁹

Analysis Methods

Network Meta-Analyses

A Bayesian NMA was conducted and the summary for the methodology can be found in Table 26. The following types of NMA model were employed:

• binomial model with logit link

• binomial model with logit link with log odds of response on placebo arm as interaction term

• normal model with identity link

• normal model with identity link with change from baseline in placebo are as interaction term.

The trial arm data were fitted to a generalized linear model via Bayesian Markov Chain Monte Carlo methods. The WinBUGs model was run for a minimum burn-in of 10,000 iterations to maximize convergence. Subsequently, 3 chains of 1,000 samples were drawn from the posterior distributions. The mean residual deviance (total residual deviance divided by number of data points) and the deviance information criterion output from WinBUGs provided an overall estimate of how well the predicted values fit the observed dataset.

The primary output from WinBUGs is the relative treatment effect compared with the reference treatment, which is placebo for this analysis. From these outputs, other estimates can be calculated, such as the relative treatment effects comparing all treatments in the network. The relative treatment estimates from the logit model are the (log) OR. The relative treatment effect estimates from the continuous models are mean differences in the change from baseline.

Both random-effects models and fixed-effects models were tested. A random-effects NMA allows the true treatment effect (e.g., OR between 2 treatments) to vary between studies due to heterogeneity. In these random-effects models, a uniform (uninformative) prior was used for the between-studies standard deviation, assuming that the heterogeneity is the same across all comparisons.

For the bDMARD-naive networks and for most outcomes, the fixed-effect placebo-adjusted model was favoured, unless the random-effects placebo-adjusted model was clearly better. For the bDMARD-experienced networks and for most outcomes, the fixed-effect model was favoured, unless the fixed effect with placebo adjustment was clearly better. The decision to recommend the random-effects placebo considered a balance of factors, including whether the 95% CrI was within a realistic range of values.

Rankings are provided, as well as the surface under the cumulative ranking curves, to express the percentage of efficacy for each treatment compared with an ideal treatment ranked first without uncertainty. In addition, the ranking was used to calculate the direct probability of bimekizumab ranking better than a comparator, that is, the proportion of simulations where the rank for bimekizumab was better than the comparator. To account for placebo response and the potential impact on the relative effects, a placebo-adjusted analysis was conducted, where feasible. The main analysis was conducted with a separate node for each treatment.

Table 26: Indirect Comparison Analysis Methods

ASAS-PR = partial remission in the Assessment of SpondyloArthritis International Society; ASAS20 = improvement of 20% or more in the Assessment of SpondyloArthritis International Society; ASAS40 = improvement of 40% or more in the Assessment of SpondyloArthritis International Society; ASAS5/6 = an improvement of at least 20% in 5 of 6 Assessment of SpondyloArthritis International Society domains; ASDAS = Ankylosing Spondylitis Disease Activity Score; ASDAS-MI = Ankylosing Spondylitis Disease Activity Score–Major Improvement; ASQoL = Ankylosing Spondylitis Quality of Life; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASFI = Bath Ankylosing Spondylitis Functional Index; BASMI = Bath Ankylosing Spondylitis Metrology Index; bDMARD, biologic disease-modifying antirheumatic drug; CrI = credible interval; DIC = deviation information criterion; MASES = Maastricht Ankylosing Spondylitis Enthesitis Score; SF-36 = Short Form (36) Health Survey physical component summary.

Source: Clinical evidence summary submitted by the sponsor.¹⁹

Matching-Adjusted Indirect Comparison

The MAIC analyses were conducted in accordance with the National Institute for Health and Care Excellence Decision Support Unit Technical Support Document 18.⁷³

To adjust for cross-trial differences, patients from the bimekizumab trials were reweighted to match the baseline characteristics of patients in the comparator trials; weights were determined using a logistic regression based on matching variables selected through reviewing patient characteristics, reviewing existing relevant literature, validation from 2 key opinion leaders, and sensitivity analyses. The following key matching variables were tested for use in this MAIC analysis: age, BASFI at baseline, ASDAS at baseline, proportion male, proportion with prior exposure to TNF inhibitors, BMI or weight, time from diagnosis, time from symptom onset, BASDAI at baseline, PGADA at baseline, proportion white, and proportion who had used sulfasalazine.

Bimekizumab 52-week outcomes were calculated by applying weights from the matching logistic regression with the bimekizumab trial data. These recalculated outcomes were then compared with comparator outcomes via unanchored (non–placebo adjusted) comparisons and were reported as ORs with 95% CIs (based on ESS) for binary outcomes and as differences in means with 95% CIs for continuous outcomes based on ESS. No time points other than week 52 were considered.

All available IL-17 inhibitors used to treat AS were included as comparators due to their similar mechanism of action. Outcomes from the bimekizumab trials at week 52 were compared with outcomes from comparator trials at week 52. The following outcomes were included in the MAIC: ASAS20, ASAS40, ASAS-PR, BASDAI change from baseline, BASDAI50, and an ASDAS of less than 2.1. For binary end points, each MAIC was conducted on NRI data, derived where necessary from the number of observed responses and the number of randomized patients. For continuous end points, published comparator data were reported using a mixed model for repeated measures or observed case data; missing data from the bimekizumab trials were imputed using last observation carried forward. The recalculated bimekizumab week 52 outcomes were compared with those for secukinumab and ixekizumab using unanchored (non–placebo adjusted) comparisons and were reported as ORs or mean differences with 95% CIs.

Results

Summary of Included Studies

Network Meta-Analysis

The SLR identified 28 trials that reported data for the AS population that were deemed feasible to be included in the AS NMA. The NMA included a mixed set of studies in relation to prior bDMARD experience:

• Fifteen studies included 100% bDMARD-naive patients: ATLAS,⁷⁴ Bao (2014),⁷⁵ Calin (2004),⁷⁶ Canadian AS trial (Lambert et al. [2007]),⁷⁷ COAST-V,⁷⁸ Davis (2003),⁷⁹ ETN Study 314,⁸⁰ GO-RAISE,⁸¹ Gorman (2002),⁸² Hu (2012),⁸³ Huang (2014),⁸⁴ Leeds ETN Study,⁸⁵ SELECT-AXIS 1,⁸⁶ SPINE,⁸⁷ van der Heijde (2017).⁸⁸

• Two studies included 100% bDMARD-experienced patients: COAST-W,⁸⁹ SELECT-AXIS 2 (Study 1).⁹⁰

• Eleven studies included patients who had some experience with bDMARDs (between 12% and 39% of the patients in the trial): ASSERT (actual percentage is unknown; assumed to be less than 50% due to the age of the study and to align with previous meta-analyses),⁹¹ ASTRUM,⁹² BE AGILE,⁹³ BE MOBILE 2⁹⁴, Deodhar (2021),⁹⁵ GO-ALIVE,⁹⁶ MEASURE 2⁹⁷, MEASURE 4,⁹⁸ MEASURE 5,⁹⁹ RAPID-axSpA,¹⁰⁰ Xue (2022).¹⁰¹

Evidence networks were constructed in 3 scenarios:

• A purely bDMARD-naive network with 24 studies and 4,145 patients (Figure 4). This network includes either studies where 100% of patients are bDMARD-naive or studies that reported separate data for the bDMARD-naive subgroup:

  • Fifteen studies where 100% of the included patients were bDMARD-naive: ATLAS, Bao (2014), Calin (2004), Canadian AS trial, COAST-V, Davis (2003), ETN Study 314, GO-RAISE, Gorman (2002), Hu (2012), Huang (2014), Leeds ETN Study, SELECT-AXIS 1, SPINE, van der Heijde (2017).

  • Nine studies reporting data for the bDMARD-naive subgroup: ASTRUM, BE AGILE, BE MOBILE 2, Deodhar (2021), MEASURE 2, MEASURE 4, MEASURE 5, RAPID-axSpA, Xue (2022).

• A predominantly bDMARD-naive network with 26 studies and 5,271 patients (Figure 5). This network includes studies where it was either confirmed or can be assumed that more than 50% of the included patients were bDMARD-naive:

  • Fifteen studies where 100% of patients were bDMARD-naive: ATLAS, Bao (2014), Calin (2004), Canadian AS trial, COAST-V, Davis (2003), ETN Study 314, GO-RAISE, Gorman (2002), Hu (2012), Huang (2014), Leeds ETN Study, SELECT-AXIS 1, SPINE, van der Heijde (2017).

  • Ten studies where more than 50% of patients were bDMARD-naive: ASTRUM (71% bDMARD-naive), BE AGILE (89%), BE MOBILE 2 (82%), Deodhar (2021) (76%), GO-ALIVE (85%), MEASURE 2 (61%), MEASURE 4 (72%), MEASURE 5 (79%), RAPID-axSpA (85%), Xue (2022) (88%).

  • One study with an unknown percentage of bDMARD-naive patients, but for which it is reasonable to assume that more than 50% of patients were bDMARD-naive: ASSERT.

• A purely bDMARD-experienced network with 9 studies and 1,048 patients (Figure 6). This network includes either studies where 100% of the included patients were bDMARD-experienced or studies that reported separate data for the bDMARD-experienced subgroup:

  • Two studies where 100% of the included patients were bDMARD-experienced: COAST-W and SELECT-AXIS 2 (Study 1).

  • Seven studies reporting data for the bDMARD-experienced: ASTRUM, BE MOBILE 2, Deodhar (2021), MEASURE 2, MEASURE 4, MEASURE 5, and RAPID-axSpA.

All studies were RCTs with a duration of at least 12 weeks. Study dates were not provided but it was evident that some of the older included studies were performed in the late 1990s or early 2000s. The size of the individual treatment arms ranged from 20 to 305 patients. The majority of patients in the studies were male, representing approximately 63% to 100% of the patients enrolled across treatment arms. The interventions were administered subcutaneously, intravenously, or by mouth. The frequency of the administration of the treatments varied and included daily, weekly, every 2 weeks, and every 4 weeks regimens.

The outcome time frame used for this NMA was 12 to 16 weeks. Preference was given to 16-week data if studies reported measurements at more than 1 time point within this time frame, as this aligned with the follow-up time frame for the primary and secondary end points in BE MOBILE 1 and 2, the phase III trials in which patients with AS (nr-axSpA) received bimekizumab, the key treatment of interest in the networks. Randomization was conducted in appropriately 16 of 28 trials but the randomization methods were unclear in 12 trials. Allocation concealment techniques were adequate for most trials but were unclear for 8 trials. Most studies employed blinding techniques (26 of 28 trials). Five of 28 trials had imbalances in prognostic factors at baseline. All studies had a placebo group. No studies were deemed unsuitable by the sponsor for inclusion in the NMA based on concerns regarding risk of bias. In general, the evaluated trials had a low risk of bias, though the level of risk was unclear for some items.

The majority of the studies included patients with axSpA who have active disease that aligns with the Assessment in Spondyloarthritis International Society–European Alliance of Associations for Rheumatology (ASAS-EULAR) or British Society of Rheumatology criteria for starting TNF inhibitor treatment after NSAID failure. Of the 28 AS studies included in the NMA, enrolment of patients in the majority of trials was based on the 1984 mNY criteria. However, according to the sponsor, the entry criteria for 3 studies aligned with the ASAS classification criteria, which used the mNY criteria for r-axSpA plus additional criteria. Therefore, some patients included in the other trials might not have been eligible for enrolment in the COAST-V, COAST-W, or Xue (2022) studies. It was assumed by the sponsor that these populations were sufficiently similar to enable these studies to be directly compared in the NMA.

The range of the proportion of patients in the trials with HLA-B27–positive status was approximately 75% to 95%. C-reactive protein scores were not consistently reported, with some data missing and other studies reporting measures of mean, geometric mean, or median values, making it difficult to compare between studies. Where C-reactive protein means were reported, the range was approximately 12 mg/L to 32 mg/L. The range in mean baseline values was 5.5 to 7.5 for the BASDAI and 3.4 to 4.2 for the ASDAS using C-reactive protein.

Matching-Adjusted Indirect Comparison

Bimekizumab was compared with secukinumab 150 mg every 4 weeks, secukinumab 300 mg every 4 weeks, and ixekizumab 80 mg every 4 weeks for this analysis. In the analysis for bimekizumab versus secukinumab 150 mg, individual patient data from the BE MOBILE 2 trial were matched to the pooled secukinumab 150 mg loading dose and no-loading summary data from the MEASURE 1, MEASURE 2, MEASURE 3 and MEASURE 4 trials for pairwise comparisons. In the analysis for bimekizumab versus secukinumab 300 mg, individual patient data from the BE MOBILE 2 trial were matched to the secukinumab 300 mg summary data from the MEASURE 3 study for pairwise comparisons. It should be noted that in the MEASURE 3 trial, patients received 3 IV-loading doses of secukinumab 300 mg every 2 weeks, which is off label, before being switched at week 4 to subcutaneous secukinumab 300 mg every 4 weeks; the impact of this on the comparisons could not be determined. For pairwise comparisons in the bimekizumab versus ixekizumab analysis, individual patient data from patients randomized to bimekizumab in the BE MOBILE 2 trial were matched to the summary data from the patients in the COAST-V and COAST-W trials who were randomized to bimekizumab 80 mg every 4 weeks.

Summary of Results

Efficacy

Network Meta-Analysis

In the bDMARD purely naive network, the results favoured bimekizumab over placebo for most outcomes except for MASES and the SF-36 mental component summary (SF-36 MCS). For most comparisons between bimekizumab versus TNF, IL-17, or JAK inhibitors, there were no clear differences observed. The exceptions to this were 2 findings in which bimekizumab showed improvement in SF-36 PCS results compared with adalimumab and secukinumab.

In the bDMARD predominantly naive network, results favoured bimekizumab over placebo for most outcomes except for MASES and SF-36 MCS. For most comparisons between bimekizumab versus TNF, IL-17, or JAK inhibitors, there were no clear differences observed. There were some exceptions to this general observation. Bimekizumab showed improvement in SF-36 PCS and ASAS-PR results compared with secukinumab. Results favoured etanercept compared with bimekizumab for BASDAI50 and BASFI. Results favoured golimumab IV compared with bimekizumab for BASFI and ASQoL. Results favoured adalimumab and certolizumab over bimekizumab for ASDAS-ID. Results favoured tofacitinib over bimekizumab for Bath Ankylosing Spondylitis Metrology Index. Results favoured upadacitinib over bimekizumab for ASDAS-ID.

In the bDMARD purely experienced network, results favoured bimekizumab over placebo for ASAS-40, ASDAS using C-reactive protein, and BASFI. There were no clear differences between bimekizumab and placebo for ASAS20, major improvement in the ASDAS, BASDAI50, BASDAI, and NSP. For most comparisons between bimekizumab versus TNF, IL-17, or JAK inhibitors, there were no clear differences observed. The exceptions to this were 2 findings in which the results favoured certolizumab over bimekizumab for ASQoL and SF-36 PCS.

In the 3 networks, there were no clear differences observed between bimekizumab and other active comparators for the outcomes identified for the GRADE analysis in this report (ASAS40, BASDAI, BASFI, NSP, enthesitis state, ASQoL, WPAI-SHP, and SAEs).

Harms

Network Meta-Analysis

The sponsor conducted NMAs of bimekizumab versus relevant comparators in the context of axSpA for 2 harms outcomes, discontinuation due to any reason and SAEs. The comparators of interest with data available for this NMA were the IL-17A inhibitors ixekizumab and secukinumab; the TNF alpha inhibitors adalimumab, certolizumab pegol, etanercept, golimumab (subcutaneous or IV routes) and infliximab (IV); and the JAK inhibitors tofacitinib and upadacitinib.

The network for the analysis of discontinuation due to any reason contained 18 studies. Study discontinuation rates were low in all trials (range of 0 to 14 patients per treatment arm). Time points between 12 and 16 weeks were used for this analysis. The CrIs were very wide for most estimates. There were no clear differences observed between bimekizumab and any other treatment. There was 1 finding in which bimekizumab had a higher risk of study discontinuation compared with tofacitinib; however, the uncertainty around this estimate was high, as reflected by a wide CrI.

The network for the analysis of SAEs contained 18 studies. SAE rates were low in all studies (range of 0 to 10 SAEs per treatment arm). Time points between 12 and 16 weeks were used for this analysis. The CrIs were very wide for most estimates. There were no clear differences between bimekizumab and any other treatments in the network.

Evidence Networks

Figure 4: Network 1 — Purely bDMARD-Naive^a

ADA = adalimumab; bDMARD = biologic disease-modifying antirheumatic drug; BID = twice a day; BIW = twice a week; BKZ = bimekizumab; CZP = certolizumab pegol; ETN = etanercept; GOL = golimumab; IFX = infliximab; IR = inadequate response; IXE = ixekizumab; NSAID = nonsteroidal anti-inflammatory drug; PBO = placebo; Q2W = twice a week; Q4W = 4 times a week; Q6W = 6 times a week; Q8W = 8 times a week; QD = once daily; QW = once weekly; SC = subcutaneous; SEC = secukinumab; TOF = tofacitinib; UPA = upadacitinib.

^a Study population is IR to ≥ 1 NSAIDs and either 100% bDMARD-naive patients or data available for the subset of patients that are naive to bDMARDs (> 9 patients); studies in grey are studies excluded from this network but included in network 2 (predominantly naive). Light blue indicates interleukin-17A inhibitors, dark blue indicates interleukin-17A and interleukin-17F inhibitors (i.e., bimekizumab), red indicates Janus kinase inhibitors, and green indicates tumour necrosis factor inhibitors.

Source: Clinical evidence summary submitted by the sponsor.¹⁹

Figure 5: Network 2 — Predominantly bDMARD-Naive

ADA = adalimumab; axSpA = axial spondyloarthritis; bDMARD = biologic disease-modifying antirheumatic drug; BID = twice a day; BIW = twice a week; BKZ = bimekizumab; CZP = certolizumab pegol; ETN = etanercept; GOL = golimumab; IFX = infliximab; IR = inadequate response; IXE = ixekizumab; NSAID = nonsteroidal anti-inflammatory drug; PBO = placebo; Q2W = twice a week; Q4W = 4 times a week; Q6W = 6 times a week; Q8W = 8 times a week; QD = once daily; QW = once weekly; SC = subcutaneous; SEC = secukinumab; TOF = tofacitinib; TNF = tumour necrosis factor; UPA = upadacitinib.

Note: Study population is IR to ≥ 1 NSAID studies and 61% to 100% bDMARD-naive patients. Approximately 90% of patients included in this network were bDMARD-naive. Light blue indicates interleukin-17A inhibitors, dark blue indicates interleukin-17A or −17F inhibitors (i.e., bimekizumab), red indicates Janus kinase inhibitors, and green indicates tumour necrosis factor inhibitors.

Source: Clinical evidence summary submitted by the sponsor.¹⁹

Figure 6: Network 3 — Purely bDMARD-Experienced^a

ADA = adalimumab; ASAS40 = improvement of 40% or more in the Assessment of SpondyloArthritis International Society; ASDAS-CRP = Ankylosing Spondylitis Disease Activity Score using C-reactive protein; BASDAI50 = a reduction of at least 50% in the Bath Ankylosing Spondylitis Disease Activity Index score; BASFI = Bath Ankylosing Spondylitis Functional Index; bDMARD = biologic disease-modifying antirheumatic drug; BID = twice a day; BIW = twice a week; BKZ = bimekizumab; CZP = certolizumab pegol; ETN = etanercept; GOL = golimumab; IFX = infliximab; IR = inadequate response; IXE = ixekizumab; NSAID = nonsteroidal anti-inflammatory drug; PBO = placebo; Q2W = twice a week; Q4W = 4 times a week; Q6W = 6 times a week; Q8W = 8 times a week; QD = once daily; QW = once weekly; SC = subcutaneous; SEC = secukinumab; TOF = tofacitinib; UPA = upadacitinib.

Note: Studies in grey are those excluded from this network but included in network 2 (predominantly naive). Light blue indicates interleukin-17A inhibitors, dark blue indicates interleukin-17A and −17F inhibitors (i.e., bimekizumab), red indicates Janus kinase inhibitors, and green indicates tumour necrosis factor inhibitors.

^a Study population of patients who experienced an IR to ≥ 1 NSAIDs who were either 100% bDMARD-experienced or who had had 1 or more prior bDMARDs (> 9 patients).

° Data not included in network, as there were < 10 patients in study arm.

^ Data available for the subset of patients who had received 1 or more prior bDMARD (> 9 patients).

Source: Clinical evidence summary submitted by the sponsor.¹⁹

Table 27: Summary of Patient Characteristics for Studies Included in the NMA (N = 28)

ASDAS-CRP = Ankylosing Spondylitis Disease Activity Score using C-reactive protein; b.i.d. = twice a day; b.i.w. = twice a week; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BKZ = bimekizumab 160 mg q.4.w. SC; Crl = credible interval; CRP = C-reactive protein; CZP = certolizumab pegol 200 mg q.2.w. SC or 400 mg q.4.w. SC; ETN = etanercept 50 mg q.w. SC or 25 mg b.i.w. SC; GOL = golimumab 50 mg q.4.w SC or golimumab 2 mg/kg q.8.w. IV; HLA-B27 = human leukocyte antigen B27; IFX = infliximab 5 mg IV q.6.w.; IL = interleukin; IXE = ixekizumab 80 mg q.4.w. SC; JAK = Janus kinase; MD = mean difference; NA = not applicable (not included in network); NR = not reported; OR = odds ratio; PBO = placebo; q.4.w. = every 4 weeks; q.6.w. = every 6 weeks; q.8.w. = every 8 weeks; q.d. = once daily; q.w. = every week; SC = subcutaneous; SEC (no load) = secukinumab 150 mg q.4.w. SC (no loading); SEC (SC load) = secukinumab 150 mg q.4.w. SC (with SC loading); TNF = tumour necrosis factor; TOF = tofacitinib 5 mg b.i.d. oral; UPA = upadacitinib 15 mg q.d. oral.

Source: Clinical evidence summary submitted by the sponsor.¹⁹

Table 28: Summary of Outcomes at 12 to 16 Weeks —bDMARD Purely Naive Network (1 of 2)^a

ADA = adalimumab 40 mg q.2.w. SC; AS = ankylosing spondylitis; ASAS20 = improvement of ≥ 20% in the Assessment of SpondyloArthritis International Society; ASAS40 = improvement of ≥ 40% in the Assessment of SpondyloArthritis International Society; ASDAS-MI = major improvement in Ankylosing Spondylitis Disease Activity Score; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index; BASFI = Bath Ankylosing Spondylitis Functional Index; bDMARD = biological disease-modifying antirheumatic drug; b.i.d. = twice a day; b.i.w. = twice a week; BKZ = bimekizumab 160 mg q.4.w. SC; Crl = credible interval; CRP = C-reactive protein; CZP = certolizumab pegol 200 mg q.2.w. SC or 400 mg q.4.w. SC; ETN = etanercept 50 mg q.w. SC or 25 mg b.i.w. SC; GOL = golimumab 50 mg q.4.w. SC or golimumab 2 mg/kg q.8.w. IV;IFX = infliximab 5 mg IV q.6.w.; IL-17i = interleukin-17 inhibitor; IXE = ixekizumab 80 mg q.4.w. SC; JAK = Janus kinase; MD = mean difference; NA = not applicable (not included in network); NE = not estimated; NSP = nocturnal spine pain; OR = odds ratio; PBO = placebo; q.4.w. = every 4 weeks; q.6.w. = every 6 weeks; q.8.w. = every 8 weeks; q.d. = once daily; SC = subcutaneous; SEC (no load) = secukinumab 150 mg q.4.w. SC (no loading); SEC (SC load) = secukinumab 150 mg q.4.w. SC (with SC loading); TNFi = tumour necrosis factor inhibitor; TOF = tofacitinib; UPA = upadacitinib 15 mg q.d. oral.

Note: Bold font indicates significant difference, meaning that the CrI does not cross the null.

^aHorizontal separation lines indicate placebo and different drug classes.

^bPresented models are suggested base case results, on balance based on model fit parameters and 95% CrI.

^cSignificant difference based on 95% CrI: for binomial outcomes, OR = 1 is the line of no effect; for continuous outcomes, difference = 0 is the line of no effect.

Source: Full network meta-analysis report submitted by the sponsor.¹⁰²

Table 29: Summary of Outcomes at 12 to 16 Weeks —bDMARD Purely Naive Network (2 of 2)^a