CADTH Reimbursement Review

Tezepelumab (Tezspire)

Sponsor: AstraZeneca Canada Inc.

Therapeutic area: Asthma

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Stakeholder Input

Clinical Review

Executive Summary

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

Introduction

Asthma is a chronic respiratory disorder characterized by reversible airway obstruction. Hallmarks of asthma include inflammation, bronchoconstriction, and airway remodelling, as well as hyper-responsive airways and mucous production.¹ Symptoms of asthma include wheezing, dyspnea, chest tightness, sputum production, and coughing, all of which can be exacerbated by exogenous influences such as allergens, upper respiratory tract infections, or environmental factors such as smoke or cold air.¹ Eosinophils are believed to be a major contributor to the inflammatory processes that are characteristic of the disease, according to the clinical expert consulted by CADTH for this review. An estimated 2.4 million Canadians 12 years or older suffer from asthma, representing 12% of all children and 8% of adults.²

The management of mild asthma is carried out using “relievers” such as short-acting beta2-agonists (SABAs) or rapid-acting beta2-agonists such as formoterol, combined with controllers such as inhaled corticosteroids (ICS) on an as-needed basis.¹ Alternatively, regular, daily treatment with low-dose ICS is used.¹ If regular low-dose ICS do not achieve good asthma control then treatment is typically escalated to the use of long-acting bronchodilators, most commonly long-acting beta2-agonists (LABAs), always in combination with ICS.¹ Oral corticosteroids (OCS) are used for acute exacerbations on a short-term basis in “bursts,” although some patients’ asthma can be severe enough to require OCS on an ongoing basis, according to the clinical expert consulted by CADTH for this review. According to the expert, the approach to managing asthma has evolved, such that patients are now routinely grouped into those who have type 2 inflammation and those who do not. Type 2 inflammation is mediated, in part, by cytokines such as interleukin (IL)-4, IL-5, and IL-13, and this explains why this phenotype may be more responsive to the biologics that target these cytokines, according to the clinical expert. Monoclonal antibodies are the newest entrants into the asthma treatment paradigm, beginning with an immunoglobin E (IgE) inhibitor (omalizumab) and, more recently, IL-5 inhibitors (mepolizumab, reslizumab, and benralizumab), an inhibitor of IL-4 and IL-13 (dupilumab), and now a thymic stromal lymphopoietin (TSLP) inhibitor, tezepelumab. According to the clinical expert consulted by CADTH for this review, none of the monoclonal antibodies are intended to be used in the first line but are reserved instead for those patients whose asthma is not well controlled with high doses of ICS plus a LABA.

As TSLP is a cytokine found near the beginning of the inflammatory cascade, it is therefore thought that blockading TSLP may have a broad effect on many mediators that play a role in the pathophysiology of various asthma phenotypes, including eosinophils, IgE, and a variety of ILs.³ Tezepelumab is administered by subcutaneous injection at a dose of 210 mg every 4 weeks. Its proposed indication is as an add-on maintenance treatment in adults and adolescents 12 years and older with severe asthma, and the sponsor’s reimbursement request is the same as the proposed indication. At the time of writing this report, tezepelumab was being reviewed by Health Canada under the regular review process, and it has been approved by the FDA.

The objective of this report was to perform a systematic review of the beneficial and harmful effects of tezepelumab injections as an add-on maintenance treatment in adults and adolescents 12 years and older with severe asthma.

Stakeholder Perspectives

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

Patient Input

Patient input for this review was received from Asthma Canada and the Lung Health Foundation (formally known as the Ontario Lung Association). Patients reported that living with asthma negatively affected their psychological and social well-being and results in poor quality of life. The patients indicated that their asthma affects their ability to complete daily activities, attend school or work, and participate in outdoor and/or physical activity, and interferes with social interactions. Patients also reported loss of productivity at school or work due to their asthma, leading to a decrease in performance or quality of work and/or schoolwork. Parents and caregivers expressed concern about accessing adequate and necessary medical care as severe exacerbations can cause loss of consciousness or hypoxia, in addition to urgent emergency department (ED) care to restore airway functions. Patients, parents, and caregivers noted that there was an unmet need for treatment options for severe asthma. Even with currently available treatment, 1 in 4 respondents indicated that they have poor symptom control. Patients, parents, and caregivers noted several barriers to accessing health care providers (e.g., respirologists and specialized asthma clinics) including travel time and cost, missed school or work, and the financial burden of prescription refills. Patients reported that the long-term use of OCS, while providing some degree of inflammation control after failing other options, is associated with notable side effects, which were reportedly a considerable source of distress. Patients and caregivers identified the following treatment priorities: the ability to control their day-to-day symptoms, the ability to control exacerbation, reduction of cost or coverage for current and upcoming treatments, and reduction in medication-associated side effects. Other key treatment outcomes highlighted by patients included improvement in quality of life, reduction in the number of medications required to take to maintain asthma control, and treatments with minimal side effects.

Clinician Input

Input From the Clinical Expert Consulted by CADTH

According to the clinical expert consulted by CADTH for this review, half of all patients with type 2 asthma remain poorly controlled with existing nonbiologic interventions. Although various measures, including improved adherence to therapy, can improve control in these patients, a subset of approximately 5% of patients remain poorly controlled no matter what the intervention. The minority of patients who have non–type 2 asthma are unlikely to benefit from current biologics.

According to the clinical expert, tezepelumab can treat patients with type 2 and non–type 2 asthma, and no other biologics are available that can reduce exacerbation frequency in patients with non–type 2 disease. The expert stated that tezepelumab could be used in the first line in patients with type 2 asthma or in patients who failed to improve on other biologics, as airway inflammation in type 2 asthma may be driven by factors other than those targeted by current biologics.

As the definition of severe asthma can vary, an indication for use in severe asthma is imprecise, according to the clinical expert, although there is no evidence of efficacy in patients who are OCS-dependent. The subset of patients who are OCS-dependent likely only represents a relatively small segment of patients with asthma in Canada.

According to the clinical expert consulted by CADTH, relevant outcomes to assess treatment response include improved health-related quality of life (HRQoL), decreased frequency of exacerbations, and improved asthma control, which would include improvement in or stabilization of forced expiratory volume in the first second (FEV₁), elimination of airflow reversibility to a bronchodilator, and reduction in symptoms. Once a patient meets the criteria for initiation, clear stopping rules will be difficult to develop, as asthma control can be affected by environmental factors. The decision to initiate administration of the drug should be limited to respirologists or allergists with experience using biologics, and, once started, tezepelumab could be maintained by a generalist.

Clinician Group Input

Input was received from 6 clinicians on behalf of the AllerGen Clinical Investigator Collaborative (CIC). The clinician group indicated that, while some treatments are effective for patients with severe eosinophilic type 2 (T2)-high asthma, no effective treatment options are available for those patients who have severe asthma that is not persistently T2-high asthma. Members of the CIC agreed that the use of tezepelumab in asthma should be restricted to patients with severe asthma, regardless of their eosinophilic asthma status. According to the CIC, severe exacerbation risk remains the single most important outcome to improve in severe asthma. The CIC suggested that tezepelumab should be discontinued if patients continue to experience severe exacerbations while on treatment. The only other reason cited for discontinuation by the CIC was side effects.

Drug Program Input

In response to a question from the drug plans regarding whether tezepelumab should be restricted based on phenotyping and/or biomarkers, the clinical expert consulted by CADTH stated that this is likely unnecessary as tezepelumab appears to have efficacy across all phenotypes. The drug plans asked about alignment with initiation criteria for other biologics, and the clinical expert confirmed that: 1) patients should have a documented diagnosis of asthma; 2) the patient should be inadequately controlled with high-dose ICS and 1 or more additional controllers; 3) the patient should have experienced 2 or more clinically significant asthma exacerbations in the past year; and 4) the patient should have completed a baseline assessment of asthma symptom control using a validated instrument. The clinical expert also confirmed that, in their opinion, the renewal criteria tezepelumab should be aligned with the renewal criteria for other biologics for severe asthma, and that there is no evidence to suggest tezepelumab should be combined with any other biologics. The clinical expert also noted that, with respect to prescribing criteria, tezepelumab should be restricted to respirologists and allergists for initiation of therapy; however, family physicians should be able to maintain treatment once initiated.

Clinical Evidence

Pivotal Studies and Protocol-Selected Studies

Description of Studies

Three multinational, sponsor-funded, double-blind randomized controlled trials (RCTs) were included in this systematic review. The NAVIGATOR study randomized 1,061 patients who were on medium- or high-dose ICS and who had 2 or more exacerbations in the past year at a 1:1 ratio to either tezepelumab or placebo over a treatment course of 52 weeks.⁴ The primary outcome was the annualized asthma exacerbation rate (AAER) and key secondary outcomes included the AAER in patients with baseline eosinophil counts of less than 300 cells/µL, change from baseline in pre-bronchodilator FEV₁, the Asthma Quality of Life Questionnaire Standardized for patients 12 years of age and older (AQLQ[S]12+), and the 6-item Asthma Control Questionnaire (ACQ-6). The SOURCE study randomized 150 patients with OCS-dependent asthma, 1:1, to either tezepelumab or placebo over a treatment course of 48 weeks.⁵ The primary outcome was the percent reduction in OCS dose while not losing asthma control, and key secondary outcomes included the AAER, time to first asthma exacerbation, rate of asthma exacerbation associated with ED visits, urgent-care visits or hospitalization, and patients who did not experience an asthma exacerbation over 48 weeks. The PATHWAY study was a phase II double-blind RCT that randomized 550 patients on medium- to high-dose ICS and at least 2 exacerbations (or 1 severe asthma exacerbation) in the past year, 1:1:1:1, to 3 different doses of tezepelumab, including the proposed dose in the draft product monograph, or placebo, over a treatment course of 52 weeks.⁶ Results are reported for the tezepelumab treatment group in the PATHWAY study that received the dose recommended in the draft product monograph (i.e., 210 mg subcutaneously every 4 weeks) only; results from the other tezepelumab arms are not reported in this review. The primary outcome was the AAER, and secondary outcomes included subgroups based on the primary outcome, change from baseline in FEV₁, and ACQ-6 score.

Across studies, the mean age of patients was between 49 and 53.5 years, and the majority were female, ranging between 59% and 68% of patients across studies. In the NAVIGATOR study, 62% of patients were White and 28% were Asian, while 84% of patients in the SOURCE study and 91% of patients in the PATHWAY study were White. In the NAVIGATOR study, 60% of patients had 2 exacerbations in the past 12 months and the remainder had more than 2, while in the PATHWAY study 78% of patients had 1 or 2 exacerbations and the remainder had 3 or more. In the SOURCE study, the protocol for which did not require more than 1 exacerbation in the past 12 months, 43% of patients had 1 exacerbation, 35% had 2, and 23% had more than 2 exacerbations. In the NAVIGATOR study, 75% of patients were on high-dose ICS and the remaining were on medium-dose ICS, while in the SOURCE study, all but 1 patient were on high-dose ICS. All patients in the SOURCE study were on OCS at baseline, while 9% were on OCS in the NAVIGATOR study and 8% were on OCS in the PATHWAY study.

Table 2: Summary of Key Results From Pivotal and Protocol-Selected Studies

AAER = annualized asthma exacerbation rate; ACQ-6 = 6-item Asthma Control Questionnaire; AE = adverse event; AQLQ(S)12+ = Asthma Quality of Life Questionnaire (Standardized) for patients 12 years of age and older; BD = bronchodilator; CFB = change from baseline; FEV₁ = forced expiratory volume in the first second; ICS = inhaled corticosteroids; LSM = least squares mean; NR = not reported; OCS = oral corticosteroids; OR = odds ratio; SAE = serious adverse events; SD = standard deviation; SE = standard error; Tez = tezepelumab; WDAE = withdrawal due to adverse event; vs. = versus.

^aModel: a negative binomial regression analysis with treatment, region, age group, and history of exacerbations as covariates. The logarithm of the time at risk is used as an offset variable. Annual exacerbation rates displayed are estimated marginal rates from the model. Absolute difference is the difference between the marginal rates. Confidence intervals for annual exacerbation rates and absolute differences are estimated via the delta method.

^bModel: a negative binomial regression analysis with treatment, region, and history of exacerbations as covariates. The logarithm of the time at risk is used as an offset variable. Annual exacerbation rates displayed are estimated marginal rates from the model. Confidence intervals for annual exacerbation rates and absolute differences are estimated via the delta method.

^cRate ratio, and 95% CI for the rate ratio were estimated from negative binomial regression with treatment group, and the stratification factors- baseline blood eosinophil count (≥ or < 250 cells/μL) and baseline ICS dose level (medium or high) as the covariates.

^dP values are not controlled for multiplicity or failed in the hierarchy and should be considered supportive in nature.

^eEstimate of the mean change from baseline at each week in tezepelumab is compared to placebo using a repeated measures analysis. Estimates are LSMs. The model with unstructured covariance structure is: change from baseline in FEV₁ = treatment group + region + age + baseline FEV₁ + visit + treatment × visit.

^fThe estimate of the cumulative odds ratio is obtained using a proportional odds model with treatment, region, and daily OCS dose at baseline as covariates

^gEstimate of the mean change from baseline at each week in tezepelumab is compared to the placebo using a repeated measures analysis. Estimates are least squares means. The model with Unstructured covariance structure is: Change from baseline in Rescue medication use weekly means = Treatment group + region + age + baseline rescue medication use + week + treatment * week.

^hEstimate of the mean change from baseline at each week in tezepelumab is compared to placebo using a repeated measures analysis. Estimates are LSMs. The model with unstructured covariance structure is: change from baseline in AQLQ(S) + 12 = treatment group + region + baseline AQLQ(S) + 12 + visit + treatment × visit .

ⁱEstimate of the mean change from baseline at each week in tezepelumab is compared to placebo using a repeated measures analysis. Estimates are LSMs. The model with unstructured covariance structure is: change from baseline in ACQ-6 = treatment group + region + baseline ACQ-6 + visit + treatment × visit.

^jSmall sample sizes for these outcomes in the PATHWAY study were due to an error in collecting electronic patient-reported outcome data.

Source: Clinical Study Report for NAVIGATOR,⁴ SOURCE,⁵ and PATHWAY.⁶

Efficacy Results

Mortality

Across all studies, only 1 death in the tezepelumab group and 2 deaths in the placebo group were reported. The 2 deaths in the placebo group were due to an unknown cause and heart failure, both in the NAVIGATOR study, and 1 patient in the tezepelumab group in the SOURCE study died due to cardiac arrest.

Acute Asthma Exacerbation

The AAER was the primary outcome of both the NAVIGATOR and PATHWAY studies. All results reported for the PATHWAY study included only the proposed Health Canada dosing. In the NAVIGATOR study the AAER over 52 weeks was 0.93 (95% confidence interval [CI], 0.80 to 1.07) with tezepelumab and 2.10 (95% CI, 1.84 to 2.39) with placebo, for a rate ratio of 0.44 (95% CI, 0.37 to 0.53; P < 0.001). In the PATHWAY study, the AAER over 52 weeks was 0.20 (95% CI, 0.13 to 0.30) with tezepelumab and 0.72 (95% CI, 0.59 to 0.88) with placebo, for a rate ratio of 0.29 (95% CI, 0.16 to 0.51; P < 0.001). In the SOURCE study, the rate ratio for AAER over 48 weeks was 0.69 (95% CI, 0.44 to 1.09).

The AAER associated with an ED visit or hospitalization was 0.06 (95% CI, 0.04 to 0.09) with tezepelumab and 0.28 (95% CI, 0.20 to 0.39) with placebo, for a rate ratio of 0.21 (95% CI, 0.12 to 0.37) in the NAVIGATOR study, and 0.16 (95% CI, 0.06 to 0.44) with tezepelumab and 0.28 (95% CI, 0.13 to 0.58) with placebo in the SOURCE study, for a rate ratio of 0.59 (95% CI, 0.19 to 1.82). The AAER associated with hospitalization was 0.03 (95% CI, 0.01 to 0.06) with tezepelumab and 0.19 (95% CI, 0.12 to 0.30) with placebo, for a rate ratio of 0.15 (95% CI, 0.07 to 0.33; P < 0.001) in the NAVIGATOR study, and 0.02 (95% CI, 0.00 to 0.07) with tezepelumab and 0.14 (95% CI, 0.08 to 0.22) with placebo, for a rate ratio of 0.14 (95% CI, 0.03 to 0.71) in the PATHWAY study.

Change in Pulmonary Function

The pre-bronchodilator FEV₁ increased in both the tezepelumab and placebo groups in the NAVIGATOR study, with a least squares mean (LSM) change from baseline to 52 weeks of 0.23 L (standard error [SE] = 0.018) with tezepelumab and 0.10 L (SE = 0.018) with placebo and a LSM difference between groups of 0.13 L (95% CI, 0.08 to 0.18; P < 0.001). In the SOURCE study, the change from baseline to week 48 was 0.21 L (SE = 0.046) with tezepelumab and −0.04 L (SE = 0.046) with placebo, for a LSM difference between groups of 0.26 litres (95% CI, 0.13 to 0.39), and in the PATHWAY study, the LSM change from baseline to week 52 was 0.076 L with tezepelumab and −0.056 L with placebo, for a LSM difference between groups of 0.132 (95% CI, 0.033 to 0.231).

Reduction in Oral Corticosteroids

Reduction in OCS use was the primary outcome of SOURCE. The cumulative odds ratio (OR) for patients having a reduction in OCS dose was 1.28 (95% CI, 0.69 to 2.35; P = −0.434). Tezepelumab therefore failed to demonstrate superiority over placebo for the primary outcome of this study.

Reduction in Use of Rescue Medication

Reduction in daily rescue medication was observed in both the tezepelumab and placebo groups in the NAVIGATOR study, an LSM change from baseline of −2.53 puffs (SE = 0.137) with tezepelumab and −2.36 puffs (0.137) with placebo, for an LSM difference between groups of −0.17 (95% CI, −0.55 to 0.21). Rescue medication use also declined in both groups in the SOURCE study, with an LSM change from baseline to 48 weeks of −0.85 puffs (SE = 0.280) with tezepelumab and −0.37 puffs (SE = 0.268) with placebo, for an LSM difference between groups of −0.47 (95% CI, −1.24 to 0.29).

Health-Related Quality of Life

Mean AQLQ(S)12+ scores increased (improved) from baseline to 52 weeks in the NAVIGATOR study, both in the tezepelumab group, at 1.48 (SD = 1.26) and the placebo group, at 1.16 (SD = 1.17), with a difference between groups of 0.33 (95% CI, 0.20 to 0.47; P = 0.001). In the SOURCE study, the LSM change from baseline to week 48 was 0.94 (SD = 0.124) with tezepelumab and 0.58 (SD = 0.123) with placebo, for a difference between groups of 0.36 (95% CI, 0.01 to 0.70). In the PATHWAY study, the LSM change from baseline to week 52 was 1.17 (SE = not reported [NR]) with tezepelumab and 0.97 (SE = NR) with placebo for a difference between groups of 0.20 (95% CI, −0.09 to 0.48). Responders to the AQLQ(S)12+ were also reported, defined as those with a change from baseline of 0.5 or greater. In the NAVIGATOR study, 78% of tezepelumab patients and 72% of placebo patients were responders, for an OR of 1.36 (95% CI, 1.02 to 1.82). In the SOURCE study, 62% of patients in the tezepelumab group and 52% of the placebo group were responders, for an OR of 1.66 (95% CI, 0.81 to 3.43). In the PATHWAY study, 73% of patients treated with tezepelumab and 62% of those given a placebo were responders.

Symptoms

Symptoms were assessed using the ACQ-6. In the NAVIGATOR study, ACQ-6 scores decreased (improved) from baseline to week 52 in both the tezepelumab and placebo groups, for a difference versus placebo of 0.33 (95% CI, 0.20 to 0.47; P < 0.001). Responders to the ACQ-6 were also reported in the NAVIGATOR study, defined as those with a change from baseline of 0.5 or greater. In the NAVIGATOR study, 86% of tezepelumab patients and 77% of placebo patients were responders, for an OR of 1.99 (95% CI, 1.43 to 2.76). In the SOURCE study, the LSM change from baseline to 48 weeks for tezepelumab was −0.87 (SE = 0.125) and −0.51 (SE = 0.123) with placebo, for a difference between groups of −0.37 (95% CI, −0.71 to −0.02). In the PATHWAY study, the LSM change from baseline to week 52 was −1.20 (SE = NR) with tezepelumab and −0.91 (SE = NR) with placebo, for a difference between groups of −0.29 (95% CI, −0.56 to −0.01).

Harms Results

Adverse events (AEs) in the tezepelumab versus placebo groups occurred in 77% versus 80% of patients, respectively, in NAVIGATOR; 72% versus 86%, respectively, in the SOURCE study; and 66% of patients in each group in the PATHWAY study.

The most common AE was nasopharyngitis, occurring in 21% versus 21% of patients in the NAVIGATOR study, 15% versus 25% of patients in the SOURCE study, and 14% versus 12% of patients in the PATHWAY study in the tezepelumab versus placebo groups, respectively. Other common events (occurring in 10% or more of patients in any group of any study) were upper respiratory tract infections, headaches, and asthma.

Serious adverse events (SAEs) in the NAVIGATOR study for the tezepelumab versus placebo groups occurred in 9% versus 13% of patients respectively; in the SOURCE study they occurred in 15% versus 21% of patients, respectively, and in the PATHWAY study they occurred in 10% versus 13% of patients, respectively. The most common SAE was asthma.

Adverse events resulting in discontinuation of the study drug occurred in 2% versus 4% of patients in the NAVIGATOR study, 3% in each group in the SOURCE study, and 2% versus 1% of patients in the PATHWAY study in the tezepelumab versus placebo groups, respectively.

Notable harms in the CADTH systematic review protocol included infections. Severe infections occurred in the tezepelumab versus placebo groups in 9% versus 8% of patients in the NAVIGATOR study, respectively, and 5% versus 9% of patients in the SOURCE study, respectively. In the PATHWAY study, infections were reported as SAEs rather than severe infections, and these occurred in 1% versus 3% of patients in the tezepelumab versus placebo groups, respectively. No opportunistic infections and no helminth infections were reported across the studies. Injection-site reactions occurred infrequently across the studies. In the NAVIGATOR study, injection-site reactions occurred in 1.5% versus 0.9% of patients in the tezepelumab versus placebo groups, respectively; in the SOURCE study, none occurred with tezepelumab, and 1.3% of patients in the placebo group experienced these events. In the PATHWAY study, these events were reported by injection volume; at the 1 mL volume they occurred in 1.5% versus 2.9% of patients and at the 1.5 mL volume they occurred in 1.5% versus 1.4% of patients in the tezepelumab versus placebo groups, respectively. Hypersensitivity reactions reported as SAEs were infrequent; just 1 patient in each of the tezepelumab and placebo groups in the NAVIGATOR study and none in the other studies reported such a reaction.

Critical Appraisal

Although the NAVIGATOR and SOURCE trials accounted for multiplicity, early failure of the hierarchy in the SOURCE study meant that all of the P values for the key secondary outcomes should only be considered supportive and not suitable for drawing conclusions. Although the number of study withdrawals was generally low (less than 5%) across studies, additional data appeared to be missing for many of the continuous outcomes, including the patient-reported outcomes such as ACQ-6, AQLQ(S)12+, and EQ-5D 5-Levels questionnaire (EQ-5D-5L). Because the missing data also exceed the reported number of treatment withdrawals, it is unclear why the data were missing. In the SOURCE study, fewer patients at baseline had more than 2 asthma exacerbations in the past year in the tezepelumab group compared to the placebo group, and this may have biased the results in favour of tezepelumab if patients in the placebo group were more prone to having an asthma exacerbation.

With respect to external validity, the clinical expert consulted by CADTH for this review noted that 25% of patients in the NAVIGATOR study were on medium-dose ICS, suggesting that these patients may have been undertreated rather than having severe asthma. The clinical expert noted that they would not start a patient on a biologic for asthma until trying high-dose ICS. The lack of an active control, particularly another biologic, in any of the included trials is a limitation, as only indirect comparisons were available to assess the relative efficacy and harms of tezepelumab compared to other biologics.

Indirect Comparisons

Description of Studies

No head-to-head trials comparing the efficacy of tezepelumab with other biologics used to treat patients with severe uncontrolled asthma are currently available. The sponsor submitted 2 indirect treatment comparisons (ITCs), a network meta-analysis (NMA) and a matching adjusted indirect comparison (MAIC)–simulated treatment comparison (STC). Three additional ITCs^7-9 were identified in a systematic search of the literature performed by CADTH. Of the sponsor-submitted ITCs, both the NMA and MAIC-STC compared tezepelumab with dupilumab, mepolizumab, benralizumab, omalizumab, and reslizumab for uncontrolled moderate-to-severe asthma in adults and adolescents.^10,11 The 3 published ITCs identified by CADTH indirectly compared tezepelumab with dupilumab, benralizumab, mepolizumab, reslizumab, and omalizumab. Data on reslizumab are not reported in this review because reslizumab was not considered a relevant comparator in the CADTH systematic review protocol.

Efficacy Results

In the sponsor-submitted NMA, no differences were identified in terms of reduction of AAER, reduction of hospitalization due to AAER, FEV₁ improvement, symptom reduction (change of ACQ-6 score), and an OCS reduction of 50% or greater when comparing tezepelumab with dupilumab, mepolizumab, benralizumab, and omalizumab. The results of the sponsor-submitted MAIC-STC were aligned with that reported in the sponsor’s NMA. Findings from the 3 published ITCs were also aligned with the results reported in the sponsor’s NMA.

Harms Results

Safety outcomes (i.e., any AEs) were assessed in a published ITC by Ando et al. (2022)⁷ that compared tezepelumab with mepolizumab, benralizumab, and dupilumab. No difference in the risk of AEs was found in this ITC.

Critical Appraisal

Due to the considerable methodological limitations of the ITCs, such as heterogeneity across the included studies and the significantly reduced effective sample size after the match adjustment in the MAIC-STC, as well as the lack of subgroup analysis for the pure severe uncontrolled asthma group, the ITC results are subject to uncertainty. No definitive conclusion can be drawn on the comparative effectiveness and safety profile between tezepelumab and other relevant biologics as an add-on maintenance treatment of adults and adolescents 12 years and older with severe asthma.

Other Relevant Evidence

Description of Studies

The DESTINATION study¹² is a phase III, multi-centre, double-blind, randomized, placebo-control, parallel-group, long-term extension (LTE) study for patients who completed the NAVIGATOR⁴ or SOURCE⁵ trials. The DESTINATION study was designed to provide evidence of the long-term safety and tolerability of tezepelumab 210 mg administered every 4 weeks subcutaneously in adults and adolescents with severe, uncontrolled asthma for up to 2 continuous years, including 1 year of treatment in the predecessor NAVIGATOR and SOURCE parent studies. Adults (18 to 80 years old) and adolescents (12 to 17 years old) who had continued to receive the investigational product and attended the end-of-treatment visit in 1 of the parent studies were eligible for enrolment. A total of 951 patients were enrolled and randomized to the DESTINATION trial: 827 patients from the NAVIGATOR study and 124 from the SOURCE study. Patients previously randomized to 210 mg tezepelumab in either parent study were assigned to and remained on 210 mg tezepelumab administered every 4 weeks subcutaneously in the DESTINATION study (tezepelumab plus tezepelumab group). Patients previously randomized to the placebo arm in the parent studies were re-randomized in a 1:1 ratio to either 210 tezepelumab (placebo plus tezepelumab group) or matching placebo (placebo plus placebo group) administered every 4 weeks subcutaneously. Patients recruited from the SOURCE study were followed post-treatment for 12 weeks. Patients who enrolled from the NAVIGATOR trial who completed 100 weeks of tezepelumab treatment were eligible for either 12 weeks of follow-up or a 36-week extended follow-up. The primary outcome for the DESTINATION trial was to evaluate the long-term safety of tezepelumab in patients with severe asthma. The secondary outcome was to assess the effect of tezepelumab on the AAER over 104 weeks. This review of the DESTINATION study focused on the results from the tezepelumab plus tezepelumab and placebo plus placebo groups.

Efficacy Results

Asthma Exacerbations

Among patients enrolled in the LTE of the NAVIGATOR study, administration of tezepelumab plus tezepelumab resulted in a reduction in the rate of asthma exacerbation compared to placebo plus placebo (AAER = 0.50; 95% CI, 0.40 to 0.63). Similarly, treatment with tezepelumab plus tezepelumab reduced the rate of asthma exacerbations associated with hospitalization or ED visits compared with placebo plus placebo (AAER = 0.39; 95% CI, 0.22 to 0.69).

In patients enrolled in the LTE from the SOURCE study, the AAER for asthma exacerbations between tezepelumab plus tezepelumab and placebo plus placebo was 0.66 (95% CI, 0.37 to 1.19). For asthma exacerbations associated with hospitalization or ED visits, the AAER for tezepelumab plus tezepelumab versus placebo plus placebo was 0.27 (95% CI, 0.05 to 1.63).

Asthma Control

Improvement from baseline ACQ-6 scores over the LTE study period was observed in the tezepelumab plus tezepelumab group compared to the placebo plus placebo group in patients who were originally enrolled in the NAVIGATOR study (LSM difference = 0.31; 95% CI, −0.47 to −0.14). Similar trends in ACQ-6 scores were observed in patients originally enrolled in the SOURCE study, with the tezepelumab plus tezepelumab group seeing an improvement in ACQ-6 scores over the LTE study period compared to placebo plus placebo group (LSM difference = −0.74; 95% CI; −1.12 to −0.25).

Harms Results

Among patients who entered the DESTINATION study from the NAVIGATOR trial, 66.7% and 71.4% of patients in the tezepelumab plus tezepelumab and the placebo plus placebo groups reported at least 1 AE during the LTE study period, respectively. Among patients who remained on tezepelumab during the LTE period, AEs leading to discontinuation of the investigational product were reported by 4 patients (1%) and AEs leading to death were reported by 7 patients (1.7%). Among those who continued to received placebo in the LTE period, AEs leading to discontinuation of the investigational product were reported by 2 patients (1%) and AEs leading to death were reported by 1 patient (0.5%). Finally, SAEs during the LTE study period were reported in 35 (8.4%) and 22 (10.7%) of patients in the tezepelumab plus tezepelumab and placebo plus placebo groups, respectively. Notable harms of interest reported during the LTE study period included hypersensitivity (0.5% in both the tezepelumab plus tezepelumab group and placebo plus placebo group) and injection-site reactions (0.5% and 1.5% in the tezepelumab plus tezepelumab group and placebo plus placebo group, respectively).

Among patients who entered the DESTINATION study from the SOURCE trial, 71.1% and 68.8% of patients in the tezepelumab plus tezepelumab and placebo plus placebo groups reported at least 1 AE during the LTE study period, respectively. Among the tezepelumab plus tezepelumab group during the LTE period, no AEs led to discontinuation of the investigational product and 1 case (1.7%) involved an AE leading to death. In the placebo plus placebo group in the LTE period, there were no reported AEs leading to discontinuation of the investigational product or death. Finally, SAEs during the LTE study period were reported in 7 patients (11.7%) and 4 patients (12.5%) in the tezepelumab plus tezepelumab and the placebo plus placebo groups, respectively. No notable harms of interest were reported among patients enrolled in the DESTINATION study during the LTE study period.

Critical Appraisal

The DESTINATION trial provided additional data on the long-term efficacy of tezepelumab relative to placebo. Statistical hypothesis-testing was not part of the design. Blinding may have been compromised by accidental publishing of individual test results on the investigator’s portal (November 23, 2021) by the laboratory vendor before the primary database lock, which may have led to unblinding for investigators who may have viewed the data. There were several imbalances between treatment groups among those who enrolled from the SOURCE study. First, fewer patients in the placebo plus placebo group completed the treatment protocol. Second, a greater proportion of patients in the placebo plus placebo group reported use of additional controller medications at baseline. Although the direction of any bias is unclear, it is possible that the differential dropout rate between the 2 treatment groups may have introduced attrition bias in favour of the tezepelumab plus tezepelumab group. Likewise, while the direction of any bias is unclear, it is possible that the differential use of controller medication may have been a surrogate of disease severity and biased the results in favour of the tezepelumab plus tezepelumab group. Overall, the DESTINATION study population represented the population of patients with severe, uncontrolled asthma and severe, OCS-dependent asthma as derived from the parent NAVIGATOR and SOURCE studies, respectively. More than 90% of patient from the parent studies were enrolled in the DESTINATION trial. At LTE baseline, patient characteristics were similar to parent studies’ baseline. Completion of the LTE exceeded 96% across all treatment groups from the NAVIGATOR study. While completion of the LTE was lower among patients who entered the LTE from the SOURCE trial, completion of the LTE remained above 80%. Given that the patients enrolled in the LTE study were originally from the NAVIGATOR and SOURCE parent studies, and the eligibility criteria remained the same, it is reasonable to expect that the same limitations to generalizability are relevant to DESTINATION.

Conclusions

There is evidence that tezepelumab reduces the rate of asthma exacerbations in patients whose asthma remains uncontrolled despite the administration of medium- to high-dose ICS. This reduction in exacerbation risk appears to occur regardless of whether patients have type 2 or non–type 2 asthma. Additionally, tezepelumab appears to improve pulmonary function, as well as HRQoL and symptoms of asthma as measured by the AQLQ(S)12+ and ACQ-6. Data from an LTE suggest these benefits of tezepelumab on exacerbations and symptoms may continue through 2 years of treatment; however, these findings need to be confirmed in a study that formally compares tezepelumab to placebo over this time frame. There is no evidence that tezepelumab facilitates the reduction of OCS doses in patients with OCS-dependent asthma, or reduces exacerbations in these patients. With respect to harms, there are no obvious safety or tolerability issues associated with tezepelumab, and this conclusion takes into account data from an extension with at least 2 years of follow-up. Indirect evidence suggests that the efficacy and harms of tezepelumab are similar to those of other biologics used to treat asthma, although the degree of heterogeneity between the studies included in the indirect comparisons precludes drawing concrete conclusions on the comparative results.

Introduction

Disease Background

Asthma is a chronic respiratory disorder characterized by reversible airway obstruction. Hallmarks of asthma include inflammation, bronchoconstriction, and airway remodelling, as well as hyper-responsive airways and mucous production.¹ Symptoms of asthma include wheezing, dyspnea, chest tightness, sputum production, and coughing, and these symptoms can be exacerbated by exogenous influences such as allergens, upper respiratory tract infections, or environmental factors such as smoke or cold air.¹ An estimated 2.4 million Canadians 12 years or older, or 12% of all children and 8% of adults, suffer from asthma.²

According to the clinical expert consulted by CADTH for this review, there are several asthma phenotypes, 1 of which is associated with eosinophilic airway inflammation and an increased peripheral blood eosinophil count, and this may persist despite treatment with moderate- to high-dose ICS. Eosinophils promote airway inflammation and contribute to airway hyper-responsiveness and remodelling, among other functions, according to the clinical expert. The clinical expert noted that tissue eosinophilia is present in 40% to 60% of patients with asthma and the use of ICS typically reduces eosinophilic inflammation in these patients; however, a subset of patients (5% to 10% overall, and 50% of patients with severe asthma) continue to experience exacerbations despite treatment with high-dose ICS.

Standards of Therapy

Traditionally, the management of asthma is carried out using medications for the acute relief of exacerbations (colloquially known as “asthma attacks” and often referred to as “relievers” or “rescue medications”) and controllers, or maintenance drugs, which are used on a regular or chronic basis in an effort to prevent the onset of exacerbations.¹ According to the clinical expert consulted by CADTH for this review and based on the updated Global Initiative for Asthma guidelines, the pharmacologic management of asthma in Canada has recently evolved.¹ In step 1, patients begin using a low-dose ICS whenever a reliever medication is used. As symptoms persist, step 2 involves daily low-dose ICS or ICS plus formoterol on an as-needed basis. From there, patients may need to escalate to regular use of low-dose (step 3) or medium-dose (step 4) ICS plus a LABA.¹ Finally, step 5 involves the use of daily high-dose ICS plus a LABA, and if control of asthma is not achieved at that point, additional treatments are considered, such as low-dose OCS, inhaled tiotropium, and/or biologics.¹ According to the clinical expert consulted by CADTH for this review, other drugs that may be considered as add-on therapy include leukotriene receptor antagonists (LTRAs) and long-term therapy with macrolides, with the latter considered off-label. Nonpharmacologic therapies include asthma education, improvement of inhaler technique, allergen avoidance, and a written asthma action plan.¹ According to the clinical expert consulted by CADTH on this review, the treatment of comorbidities such as tobacco dependence, depression, and obstructive sleep apnea are also important in the management of asthma. With respect to harms associated with pharmacologic therapies, ICSs have short-term side effects such as oral candidiasis (“thrush”) and dysphonia; however, a number of concerning adverse effects, including osteoporosis, are associated with their long-term use, particularly at high doses.¹³ The use of systemic corticosteroids heightens the risk of harms, and their chronic use is to be avoided, according to the clinical expert.

According to the clinical expert, the approach to managing asthma has also evolved, such that patients are now routinely grouped into those who have type 2 inflammation and those who do not. According to the clinical expert, type 2 inflammation is mediated, in part, by cytokines such as IL-4, IL-5, and IL-13, and this explains why the phenotype may be more responsive to biologics that target this cytokine. Monoclonal antibodies are the newest entrants into the asthma treatment paradigm, with an IgE inhibitor (omalizumab¹⁴) being the first drug approved, and, more recently, IL-5 inhibitors (mepolizumab,¹⁵ reslizumab,¹⁶ and benralizumab¹⁷), an IL-4 and IL-13 inhibitor (dupilumab¹⁸), and now a TSLP inhibitor, tezepelumab. According to the clinical expert, none of the monoclonal antibodies are intended to be used in the first line but are reserved instead for those patients whose asthma is not well controlled with high doses of ICS plus a LABA.

According to the clinical expert consulted by CADTH, the goals of asthma therapy are to maintain control of asthma, indicated by an absence of exacerbations, stable lung function, and improved symptoms. Improving these symptoms should improve HRQoL. The longer-term goal is to prevent airway remodelling, preventing future risk from severe exacerbations and, ultimately, reducing the risk of death. Reducing the risk of harms from pharmacologic therapies is also an important goal.

Drug

Tezepelumab is a TSLP inhibitor. As TSLP is a cytokine found at the beginning of the inflammatory cascade following epithelial signalling, blockading TSLP may have a broad effect on many mediators that play a role in the pathophysiology of various asthma phenotypes, including eosinophils, IgE, and a variety of interleukins.³ Tezepelumab is administered by subcutaneous injection at a dosage of 210 mg every 4 weeks.³ Its proposed indication is as an add-on maintenance treatment in adults and adolescents 12 years and older with severe asthma,³ and the sponsor’s reimbursement request is the same as the indication.¹⁹ It is currently being reviewed at Health Canada under the regular review process, and it has been approved by the FDA.²⁰

Table 3: Key Characteristics of Tezepelumab, Dupilumab, IL-5 Inhibitors, and Omalizumab

AD = atopic dermatitis; ICS = inhaled corticosteroid; IgE = immunoglobulin E; IL = interleukin; LABA = long-acting beta2-agonist; OCS = oral corticosteroids; TSLP = thymic stromal lymphopoietin.

^aHealth Canada–approved indication.

Sources: Product monographs for tezepelumab,³ dupilumab,¹⁸ mepolizumab,¹⁵ benralizumab,¹⁷ reslizumab,¹⁶ and omalizumab.¹⁴

Stakeholder Perspectives

Patient Group Input

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

Patient input for this review was received from Asthma Canada and the Lung Health Foundation (formally known as the Ontario Lung Association), both of which are registered charities. The information gathered from Asthma Canada was obtained from approximately 600 people living with asthma or caring for someone living with asthma via multiple online surveys and in-depth individual interviews. The information from Lung Health Foundation was obtained during phone interviews conducted in April 2021 with 3 female patients living with asthma. None of the patients, parents, or caregiver respondents had experience with tezepelumab.

Patients reported that living with asthma negatively affects their psychological and social well-being, results in poor quality of life that affects their ability to complete daily activities, attend school or work, and participate in outdoor and/or physical activity, and interferes with social interactions. Patients also reported loss of productivity at school or work due to symptoms, fatigue, and exacerbations, leading to a decrease in performance or quality of work or schoolwork. Parents and caregivers expressed concerns about accessing adequate and necessary medical care as severe exacerbations can cause loss of consciousness or hypoxia, in addition to urgent ED care to restore airway functions. The need for urgent medical attention was noted to be stressful when parents and caregivers try to navigate busy and overcrowded EDs.

Patients, parents, and caregivers noted that there was an unmet need for treatment options for severe asthma. Even with currently available treatment, 1 in 4 respondents indicated that they have poor symptom control. Patients, parents, and caregivers identified several barriers to accessing health care providers (e.g., respirologists and specialized asthma clinics), including travel and missed school or work. Moreover, a third of patients and caregivers reported skipping prescription refills for asthma medications due to the financial burden as many patients with severe asthma have a low income or are unable to work due to living with asthma or caring for someone with asthma. The patients, parents, and caregivers reported that the long-term use of OCS provided some degree of inflammation control after failing other options, but is associated with notable side effects, including weight gain, acne, excess facial hair, mood swings, high blood pressure, hyperactivity, high blood sugar, and increased infections, as well as osteopenia, osteoporosis, glaucoma, cataracts, and heart disease. The source of distress associated with side effects from current medications taken to manage asthma was further emphasized by the patients who were interviewed. Patients, parents, and caregivers reported that current treatments can be difficult to take due to the mode of administration and frequency of dosing.

Patients and caregivers identified the following treatment priorities: the ability to control their day-to-day symptoms; the ability to control an exacerbation; reduction of cost or coverage for current and upcoming treatments; and reduction in medication-associated side effects. Other key treatment outcomes highlighted by patients included improvement in quality of life, reduction in the number of medications required to take to maintain asthma control, and treatments with minimal side effects.

Clinician Input

Input From the Clinical Expert Consulted by CADTH

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

Unmet Needs

According to the clinical expert consulted by CADTH, patients with asthma can be subdivided into those predominantly driven by type 2 inflammation (T-helper 2 [Th2] cells and type 2 innate lymphoid cell cellular pathways), who present with allergic and/or eosinophilic asthma, and those who are non–type 2, and it is the latter group of patients who are difficult to control. Non–type 2 includes patients with obesity-associated asthma, pauci-granulocytic asthma, very late onset disease, and smooth-muscle hypertrophy–associated disease. Currently these patients are managed with ICS-LABA treatments and often require additional treatments such as chronic macrolide administration (not Health Canada–approved) and/or addition of a long-acting muscarinic antagonist (LAMA), according to the clinical expert. The expert indicated that patients who smoke often require smoking cessation interventions to help improve asthma control. The subset of patients with primary airway smooth-muscle hypertrophy may respond to bronchial thermoplasty. The clinical expert reported that these non–type 2 patients typically do not respond to current biologic therapies.

The clinical expert reported that corticosteroids play an important role in managing non–type 2 asthma and tend to target airway inflammation in a nonspecific manner, and their systemic side effects increase with increasing dose.

According to the clinical expert, approximately half of patients with type 2 asthma remain poorly controlled with existing nonbiologics. Although nonpharmacologic treatments such as environmental control, medication adherence, inhaler education, and self-management techniques can improve asthma control, approximately 5% of patients will remain poorly controlled despite otherwise good treatment, according to the clinical expert. The clinical expert indicated that these poorly controlled patients are at greater risk for severe asthma exacerbations and worsened quality of life. Although OCS can reduce the frequency of exacerbations, they carry a risk of severe long-term side effects. A recent position statement from Asthma Canada²¹ and a peer-reviewed retrospective cohort analysis recommend limiting regular OCS use and the frequency of short-term OCS “bursts.”²² Treatment of comorbidities of asthma is an integral part of therapy, as highlighted in a recent Canadian Thoracic Society position statement on the treatment of severe asthma.²³

Place in Therapy

The clinical expert indicated that tezepelumab can treat patients with type 2 asthma as well as those with non–type 2 disease. No other biologics are available that can effectively reduce exacerbation frequency in patients with non–type 2 disease, according to the clinical expert. The clinical expert noted that tezepelumab could be used either as a first-line biologic in patients with type 2 asthma or in patients who fail to improve on other biologics, as airway inflammation in type 2 asthma may be driven by factors that fall outside of those targeted by current biologics.

The clinical expert indicated that tezepelumab could drive a shift in how biologic medications are currently prescribed. Because it is effective in patients with type 2 or non–type 2 disease, the clinical expert anticipated that it could become the preferred biologic. Although good asthma care should include careful phenotyping, tezepelumab could simplify the assessment of patients for initiation of biologics as well as the assessment of maintenance, according to the clinical expert. The clinical expert added that, although there is no evidence that tezepelumab is effective in patients who use OCS on a chronic basis, other biologics with evidence of efficacy in this population are available, and regular OCS use is not common in patients with asthma in Canada.

Patient Population

The clinical expert indicated that most patients with asthma respond to therapy regardless of the underlying severity of the disease; however, the proposed Health Canada indication for tezepelumab is for add-on maintenance therapy in adults and adolescents 12 years and older with severe asthma. The clinical expert noted that this is not a precise indication as the definition of severe asthma can vary. In general, patients who remain poorly controlled on high-dose ICS-LABA or whose control worsens when they try to decrease the dose of ICS-LABA could be treated with tezepelumab, according to the clinical expert, who also noted that patients with type 2 or non–type 2 asthma would be eligible for treatment.

The clinical expert reported that patients with a clinical presentation suggestive of asthma require confirmation with spirometry or a peak expiratory flow (PEF) measurement showing reversibility of airflow obstruction. Alternatively, airway hyper-responsiveness, reactivity to cold air, exercise, or methacholine can also be used to confirm a diagnosis of asthma. The clinical expert reported that, when physicians rely on clinical judgment rather than these objective measures, there is evidence that asthma is over-diagnosed in about one-third of patients.²⁴ The clinical expert also noted that there is evidence that Canadian physicians under-diagnose asthma,²⁵ and more importantly, may under-diagnose the presence of severe asthma. Over-diagnosis of asthma can be prevented by restricting prescribing to a clearly defined group of physicians (respirologists and allergists) who then use standard practices to diagnose asthma and to assess asthma severity, according to the clinical expert.

The clinical expert highlighted that non–type 2 asthma is not a uniform disease. It is not clear which subtypes were chosen for the clinical trials but, for the time being, there is no clear method to define who is most likely to benefit. The default would be that all non–type 2 patients should be eligible if a treatment is funded.

Assessing Response to Treatment

The clinical expert indicated that outcomes of relevance to tezepelumab include improved HRQoL, decreased exacerbation frequency, improved asthma control, and decreased ED utilization. Biomarkers such as fractional exhaled nitric oxide (FeNO) as an indicator of airway inflammation or induced eosinophil counts are not used clinically in most Canadian sites due to access and lack of funding, according to the clinical expert, who reported that lung function can be measured in most Canadian sites and can be used to assess some airway responses (e.g., increased FEV₁ and/or stabilization with loss of reversibility). Airway hyper-responsiveness, as measured by methacholine challenge testing, is also expected to improve with tezepelumab, unlike the lack of effect seen with most other biologic medications. Airway hyper-responsiveness is not a typical outcome measured clinically but such services are available in many lung-function laboratories in Canada. The clinical expert noted that the most relevant clinical outcome is asthma control, including improvement or stabilization of FEV₁, elimination of airflow reversibility to bronchodilator, and reduction of night-time and daytime symptoms. A validated measure such as the ACQ-6 can be used to objectively assess improved control.

Discontinuing Treatment

The clinical expert indicated that, once the patient meets the criteria for initiation of tezepelumab, it is difficult to develop clear stopping rules other than drug intolerance. The clinical expert reported that the frequency of exacerbations is determined not just by medication use but also by environmental factors. For example, in Western Canada, the wildfire season could be expected to increase exacerbation frequency in a susceptible population, while in Central Canada, the pollen season could have a similar effect for type 2 asthma. Asthma control would have comparable environmental factors, according to the clinical expert, who noted that asthma control over a longer time frame than months would be needed to properly evaluate drug efficacy, and therefore any stopping rule would need exceptions for environmental factors.

Prescribing Conditions

The clinical expert indicated that tezepelumab should be started by either an allergist or a respirologist with experience using biologics. Ideally this should be restricted to asthma clinics but there is no accreditation process to support this designation, according to the clinical expert. The clinical expert noted that, once started, tezepelumab could be maintained by generalists. The expert noted that, in an era of virtual clinics, regional disparities in access to specialist care can be improved and should not prevent controlled access to drug initiation.

Clinician Group Input

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

Input was received by 6 clinicians on behalf of the AllerGen CIC, which is a group of clinical investigator sites that investigate potential new therapies for the management of asthma. Six sites from 5 provinces in Canada are involved in the CIC.

Unmet Needs

The clinician group reported that, while some treatments are effective for patients with T2-high asthma, there are no effective treatment options for those patients who have severe asthma that is not persistently T2-high asthma. Currently, the only treatment options for those patients who do not have T2-high asthma are conventional asthma therapies, according to the CIC. However, the CIC indicated that these patients are less responsive to ICS or OCS. The CIC reported that, although various treatments have been developed for these patients, limited impact on clinical outcomes has been reported.

Place in Therapy and Patient Population

Members of the CIC agreed that the use of tezepelumab in asthma should be restricted to patients with severe asthma. However, in contrast to other biologic therapies, the CIC reported that tezepelumab may be used in patients with both T2-high asthma and T2-low asthma. The CIC suggested that tezepelumab should be discontinued if patients continue to experience severe exacerbations while on treatment. The only other indication cited for discontinuation by the CIC was side effects.

Assessing Response to Treatment

According to the CIC, severe exacerbation risk remains the most important outcome to improve in severe asthma. The CIC reported that severe asthma exacerbation events are potentially life-threatening and have significant effects on both patients’ and their families’ lives and functioning. The CIC indicated that other outcomes that may be used to assess response to treatment include improved lung function, other measurements considered important in asthma control, and reduction in biomarkers associated with severe asthma, including blood eosinophil count, sputum eosinophil count, and exhaled nitric oxide levels.

Discontinuing Treatment

Drug intolerance would be the main reason to discontinue the drug; otherwise, once a patient has been initiated on tezepelumab, it is difficult to develop clear stopping rules. Response to a drug with respect to exacerbation frequency can be affected by external factors, such as seasonal wildfires and pollen levels, which vary in intensity across the country. To adequately assess response and account for these external factors, one would need to try the drug for more than simply months, and any stopping rules would need exceptions for those factors.

Prescribing Conditions

The CIC noted that, in general, severe asthma is managed in specialty practices in Canada. The CIC indicated that treatment with tezepelumab would be expected to be initiated by an expert in managing severe asthma. Likewise, the decision to continue or discontinue treatment with tezepelumab would be made by an expert in managing severe asthma.

Additional Considerations

The clinical expert noted that non–type 2 asthma is not a homogeneous disease. Because it is not clear what subtypes were chosen in the clinical trials there is no clear method for defining who among patients with type 2 asthma would benefit most from this drug or whether all patients with non–type 2 asthma should be eligible for it. The expert also wondered whether post-marketing follow-up could be requested to determine the optimal characteristics for a non–type 2 responder.

Drug Program Input

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

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

ACQ-6 = 6-item Asthma Control Questionnaire; OCS = oral corticosteroids; pCPA = pan-Canadian Pharmaceutical Alliance.

Clinical Evidence

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

Systematic Review (Pivotal and Protocol-Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of tezepelumab injection as an add-on maintenance treatment in adults and adolescents 12 years and older with severe asthma.

Methods

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

The literature search was performed by an information specialist using a peer-reviewed search strategy.

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

Table 5: Inclusion Criteria for the Systematic Review

ACQ-6 = 6-item Asthma Control Questionnaire; AE  =  adverse event; AQLQ(S) + 12 = Asthma Quality of Life Questionnaire for patients 12 years of age and older; ED = emergency department; FEV₁ = forced expiratory volume in the first second; HRQoL = health-related quality of life; ICS = inhaled corticosteroids; IgE = immunoglobulin E; IL-5 = interleukin-5; OCS = oral corticosteroids; PEF = peak expiratory flow; SABA = short-acting beta2-agonists; SAE = serious adverse event; SAMA = short-acting muscarinic antagonists; vs. = versus.

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy according to the PRESS Peer Review of Electronic Search Strategies reference.²⁶

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946–) via Ovid and Embase (1974–) via Ovid. All Ovid searches were run simultaneously as a multi-file search. Duplicates were removed using Ovid deduplication for multi-file searches, followed by manual deduplication in EndNote. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concept was tezepelumab. Clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

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

The initial search was completed on May 6, 2022. Regular alerts updated the search until the meeting of the CADTH Canadian Drug Expert Committee on September 28, 2022.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature reference.²⁷ Included in this search were the websites of regulatory agencies (FDA and European Medicines Agency). Google was used to search for additional internet-based materials. Appendix 1 provides more information on the grey literature search strategy.

Findings From the Literature

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

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Table 6: Details of Included Studies

AAER = annualized acute exacerbation rate; ACQ-6 = 6-item Asthma Control Questionnaire; AE = adverse event; AQLQ(S)12+ = Asthma Quality of Life Questionnaire Standardized for patients 12 years of age and older; ASD = asthma symptom diary; BD = bronchodilator; BMI = body mass index; CGI-C = Clinical Global Impression of Change; COPD = chronic obstructive pulmonary disease; DPI = dry powder inhaler; ECG = electrocardiogram; ED = emergency department; ePRO = electronic patient-reported outcome; EQ-5D-5L = EQ-5D 5-Levels questionnaire; FeNO = fraction of exhaled nitric oxide; FEV₁ = forced expiratory volume in the first second; FTP = fluticasone propionate; GINA = Global Initiative for Asthma; ICS = inhaled corticosteroids; IgE = immunoglobulin E; LABA = long-acting beta2-agonist; MDI = metered-dose inhaler; OCS = oral corticosteroids; PEF = peak expiratory flow; PGI-C = Patient Global Impression of Change; PGI-S = Patient Global Impression of Severity; PN = predicted normal; RCT = randomized controlled trial; SABA = short-acting beta2-agonist; SAE = serious adverse event; SC = subcutaneous; SCS = systemic corticosteroids; TB = tuberculosis.

Note: Three additional reports were included (FDA Clinical and Statistical Review, sponsor’s submission).

Sources: Clinical Study Reports for NAVIGATOR,⁴ SOURCE,⁵ and PATHWAY.⁶

Description of Studies

NAVIGATOR

NAVIGATOR was a phase III, multinational, sponsor-funded double-blind RCT that compared tezepelumab to placebo in adults and adolescents with severe, uncontrolled asthma. The primary objective was to assess the effect of 210 mg tezepelumab subcutaneous every 4 weeks on asthma exacerbations in adult and adolescent patients with severe, uncontrolled asthma compared with placebo. The key secondary objectives were to assess the effect of 210 mg tezepelumab subcutaneous every 4 weeks on pulmonary function compared with placebo, on health status and health-related quality of life, asthma control, and asthma symptoms.

A total of 1,061 patients were randomized 1:1 to either tezepelumab 210 mg every 4 weeks or matched placebo. Patients were randomized from 297 centres in 18 countries, including Canada. Randomization was stratified by region and age (adult or adolescent).

The screening and run-in period for the NAVIGATOR study was 5 to 6 weeks. During this time patients were to undergo a variety of assessments, including pulmonary function; fill out and receive training on symptom diaries; and undergo a number of screening assessments, including a history and physical, electrocardiogram, and a review of concomitant medication. During the 7 days before randomization, patients had to meet at least 1 indicator of continued asthma symptoms, indicated by at least 2 days with a daytime or night-time asthma symptom score of 1 or more, or use of a SABA as a reliever on more than 2 days, or at least 1 night-time awakening due to asthma, and at baseline patients had to have an ACQ-6 score of 1.5 or greater to be randomized.

Figure 2: Study Design for NAVIGATOR

SC = subcutaneous; V = visit.

Source: Clinical Study Report for NAVIGATOR.⁴

SOURCE

The SOURCE study was a phase III, multinational, sponsor-funded, double-blind RCT that compared tezepelumab to placebo in patients with severe, refractory asthma who were on maintenance OCS and ICS and/or LABA, with or without additional asthma controllers. The primary objective of the study was to evaluate the effect of tezepelumab compared with placebo in reducing the prescribed OCS maintenance dose in patients with asthma requiring chronic treatment with maintenance OCS in addition to high-dose ICS plus a LABA. The key secondary objective was to evaluate the effect of tezepelumab compared with placebo on asthma exacerbations. A total of 150 patients were randomized 1:1 to either tezepelumab 210 mg subcutaneous every 4 weeks or matched placebo over a treatment course of 48 weeks. Patients were enrolled from 60 sites in 7 countries, and there does not appear to have been any Canadian sites. The run-in period in the SOURCE study was longer than in the NAVIGATOR and PATHWAY studies, as it included an 8-week OCS dose-optimization phase, in which the minimum dose at which patients were able to maintain asthma control was determined. The details of the dose-optimization phase are described in the Interventions section.

Figure 3: Study Design for SOURCE

SC = subcutaneous; V = visit; Wk(s) = week(s).

Source: Clinical Study Report for SOURCE.⁵

PATHWAY

The PATHWAY study was a phase II, multinational, sponsor-funded, double-blind RCT that compared 3 different dose levels tezepelumab to placebo in adults with inadequately controlled, severe asthma. The primary objective was to evaluate the effect of 3 dose levels of tezepelumab on asthma exacerbations in adult patients with inadequately controlled severe asthma. There were a number of secondary objectives, including to evaluate the effect of tezepelumab on asthma exacerbations, lung function, and asthma symptoms in the pre-specified subpopulations of asthma; the effect of tezepelumab on lung function, asthma symptoms, and other metrics related to asthma control and on other parameters of asthma exacerbations; the effect on tezepelumab on HRQoL; and assessment of the safety and tolerability of tezepelumab. A total of 552 patients were randomized 1:1:1:1 to tezepelumab 70 mg, tezepelumab 210 mg, tezepelumab 280 mg, or placebo subcutaneous every 4 weeks. This report will focus on the results for the tezepelumab 210 mg group, which is the recommended dose in the draft product monograph, compared to placebo. Results will not be presented for tezepelumab 70 mg and 280 mg because these doses are not recommended by Health Canada. Randomization was stratified by study site (Japanese or non-Japanese), eosinophil count, and ICS dose (250 cells/µL or greater and medium-dose ICS; 250 cells/µL or greater and high-dose ICS; less than 250 cells/µL and medium-dose ICS; and less than 250 cells/µL and high-dose ICS). Patients were randomized at 98 sites across 12 countries, and there does not appear to have been any Canadian sites.

The screening and run-in period for PATHWAY was up to 5 weeks and had a purpose similar to that of NAVIGATOR, with similar assessments performed.

Figure 4: Study Design for PATHWAY

EOS = eosinophils; ICS = inhaled corticosteroids; Q2W = every 2 weeks; Q4W = every 4 weeks; SC = subcutaneous.

Source: Clinical Study Report for PATHWAY.⁶

Populations

Inclusion and Exclusion Criteria

In the NAVIGATOR and SOURCE studies, enrolled patients were 12 to 80 years of age, with a documented diagnosis of asthma for at least 12 months before the study. The NAVIGATOR and SOURCE studies specified that patients had to have a morning pre-bronchodilator FEV₁ of less than 80% of predicted normal (< 90% for patients 12 to 17 years of age), and evidence of asthma, such as reversibility of 12% or greater or 200 mL or greater, either historically or during screening. In the NAVIGATOR study and PATHWAY, patients had to have 2 or more asthma exacerbations in the 12 months before screening (or in the PATHWAY study patients could have had 1 severe exacerbation resulting in hospitalization), while in the SOURCE study patients had to have 1 or more. An asthma exacerbation was defined by a worsening of asthma that required treatment with systemic corticosteroids for at least 3 consecutive days, an ED or urgent-care visit, or a hospitalization. For patients who were on OCS at baseline, a temporary increase for 3 or more days qualified as an exacerbation. In the NAVIGATOR and PATHWAY studies, patients had to have an ACQ-6 score of 1.5 or greater at screening. In the NAVIGATOR and PATHWAY study, patients had to have their asthma controlled by either medium- or high-dose ICS. In the SOURCE study, all patients had to have been receiving high-dose ICS and a LABA for at 3 months before the study, and had to have been using OCS for asthma for at least 6 months prior and received a stable dose of between 7.5 mg and 30 mg prednisone or prednisone equivalent, inclusive, for at least 1 month before the study.

Patients who had any other clinically relevant pulmonary disease, including chronic obstructive pulmonary disease, were excluded for all studies, as were patients who were current smokers, or those with a 10 pack-year history of smoking, and patients using vaping products.

Baseline Characteristics

Across studies, the median age of patients was between 49 and 53.5 years of age, and the majority were female, ranging between 59% and 68% of patients across studies. In the NAVIGATOR study, 62% of patients were White and 28% were Asian, while 84% of patients in the SOURCE study and 91% of patients in the PATHWAY study were White. In the NAVIGATOR study, 60% of patients had 2 exacerbations in the past 12 months, while the remainder had more than 2, while in the PATHWAY study 78% of patients had 1 or 2 exacerbations while the remainder had 3 or more. In the SOURCE study, which did not require more than 1 exacerbation in the past 12 months, 43% of patients had 1 exacerbation, 35% had 2, and 23% had more than 2 exacerbations. Mean eosinophil counts were higher in the NAVIGATOR study (340 cells/µL) and the PATHWAY study (372 cells/µL) than in the SOURCE study (242 cells/µL). The mean total IgE was also higher in the NAVIGATOR study than in the SOURCE study, at 565 IU/mL versus 299 IU/mL, respectively, and this was not reported in the PATHWAY study. In the NAVIGATOR study, 75% of patients were on high-dose ICS and the remaining were on medium-dose ICS, while in the SOURCE study, all but 1 patient were on high-dose ICS. According to the protocol, all patients in the SOURCE study were on OCS, while 9% in the NAVIGATOR study and 8% in the PATHWAY study were on OCS.

There were imbalances between the tezepelumab versus placebo groups within studies for mean baseline eosinophils in the NAVIGATOR (327 [SD = 293] versus 353 [SD = 488]), SOURCE (253 [SD = 203] versus 232 [[SD = 154]), and PATHWAY (365 [SD = 351] versus 380 [SD = 328]) studies. In the SOURCE study more patients in the tezepelumab group had 1 exacerbation in the past 12 months (46% versus 40%, respectively) and fewer in the placebo group had more than 2 (18% versus 28%, respectively). In the NAVIGATOR study, mean baseline IgE levels were lower with tezepelumab versus placebo at 516 (SD = 960) versus 614 (SD = 1,159). Reversibility of FEV₁ was larger with tezepelumab than with placebo, with a mean of 16.5% (SD = 14.9) versus 13.9 (SD = 14.6) and fewer tezepelumab patients had an FEV₁ reversibility of less than 12% (41% versus 54%, respectively).

Table 7: Summary of Baseline Characteristics

BD = bronchodilator; BMI = body mass index; DPI = dry power inhaler; FEIA = fluorescence enzyme immunoassay; FeNO = fractional exhaled nitric oxide; FEV₁ = forced expiratory volume in the first second; FTP = fluticasone propionate; ICS = inhaled corticosteroid; IgE = immunoglobulin E; NR = not reported; OCS = oral corticosteroids; PN = predicted normal; SD = standard deviation; Tez = tezepelumab.

Note: Medium-dose ICS defined as 500 mcq or greater FTP DPI-equivalent; high-dose ICS defined as greater than 500 mcg FTP DPI-equivalent.

Sources: Clinical Study Reports for NAVIGATOR,⁴ SOURCE,⁵ and PATHWAY.⁶

Interventions

Tezepelumab 210 mg or matching placebo was administered by subcutaneous injection every 4 weeks in all studies. The PATHWAY study was a dose-ranging study, and 2 other doses of tezepelumab were tested in the study: 70 mg every 4 weeks and 280 mg every 2 weeks, neither of which were of interest for this review as both were outside the recommended dose in the draft product monograph. In all studies, the study drug was administered by a qualified health care professional (a pharmacist or study nurse) at the site. After March 2020, due to the COVID-19 pandemic, the study drug could be administered in a patient’s home. The sponsor provided specific instructions as to drug administration, storage, and preparation. No changes were allowed to background medications except during the management of an asthma exacerbation. A SABA was to be held for at least 6 hours before scheduled spirometry, FeNO, or electrocardiogram, with the exception of any unscheduled visits due to asthma worsening. Both LABAs and LAMAs were also to be held, at least 12 hours for twice-daily dosing and 24 hours for once-daily dosing, and LTRAs were to be held for 24 hours before the assessment.

In the SOURCE study, patients had to be on OCS for at least 6 months before the study. There were a number of phases related to OCS use during the study, including an 8-week optimization phase before randomization, an induction phase that lasted the first 4 weeks after randomization, a reduction phase that lasted from weeks 4 to 40, and a maintenance phase that lasted from weeks 40 to 48. Details about each are provided in the following section.

During the OCS optimization phase of SOURCE, the minimum OCS dose while maintaining asthma control (optimized dose) was reached for all patients. The optimized OCS dose was kept stable for 2 weeks before randomization and was considered the baseline OCS dose. The baseline OCS dose was to be maintained at the same level from 2 weeks before randomization to the end of the induction phase. A reduction in the OCS dose commenced at this point, and continued at 4-week intervals until visit 15, in accordance with the OCS dose-titration schedule (reduction phase). During the reduction phase, a minimum stable OCS dose, or complete elimination of requirement for OCS, while maintaining asthma control, was reached for each patient. No adjustments should have been made to the OCS dose during the maintenance period until the end of treatment; however, exceptions may have occurred, such as in cases in which a patient had an exacerbation or for safety reasons.

In the SOURCE study, the magnitude of the OCS dose reduction depended on the dose the patient was taking. For patients on more than 10 mg to 30 mg of prednisone and/or prednisolone daily, their dose could be reduced by 5 mg per day, and for patients taking 7.5 mg to 10 mg daily, their dose could be reduced by 2.5 mg daily. Patients who were not considered to be candidates for OCS dose reduction in the optimization phase, or who were noncompliant with electronic diary completion and OCS, ICS-LABA, or other asthma-controller use, meaning they were unable to reduce their OCS dose, were considered screen failures. If a patient achieved asthma control at an OCS dose of less than 7.5 mg during the optimization phase, or asthma control was maintained after 3 consecutive OCS dose reductions, they were considered screen failures.

A set of criteria determined whether a patient was eligible to have an OCS dose reduced in the SOURCE study, and all criteria had to be met. These included pulmonary function (morning PEF of 80% or greater of mean morning measures as compared with the baseline mean), symptoms (no more than 2 nights with asthma-related awakenings [requiring rescue medications] over a 7-day period compared with baseline), rescue medication use (no more than 4 puffs per day above the baseline mean and fewer than 12 puffs per day on all days in the prior 14 days), exacerbations (none requiring increased systemic corticosteroids or hospitalization since the previous visit), investigator judgment (asthma control is sufficient to allow OCS dose reduction and no signs of adrenal insufficiency [at OCS dose reductions below 5 mg]). Whether to continue with OCS dose reduction was at the discretion of investigator even if the criteria had not been met, as long as the decision was documented and justified.

Patients across studies were allowed to take medications for conditions other than asthma, as deemed medically necessary. Some medications were prohibited within 4 to 12 weeks of the first study visit and for 4 weeks after the last dose of the study drug, largely because they had immunosuppressive or immunomodulatory effects. Exceptions were made for medical need, as judged by the investigator. Other monoclonal antibodies for asthma were not allowed within 4 months or 5 half-lives (whichever is longer) before randomization and for 4 weeks after the last dose of the study drug.

Outcomes

A list of efficacy end points identified in the CADTH review protocol that were assessed in the clinical trials included in this review is provided in Table 8. These end points are summarized in the following section. A detailed discussion and critical appraisal of the outcome measures are provided in Appendix 4.

Table 8: Summary of Outcomes of Interest Identified in the CADTH Review Protocol

AAER = annualized acute exacerbation rate; ACQ-6 = 6-item Asthma Control Questionnaire; AQLQ12+ = Asthma Quality of Life Questionnaire (Standardized) for patients 12 years of age and older; BD = bronchodilator; CFB = change from baseline; ED = emergency department; EQ-5D-5L = EQ-5D 5-Levels questionnaire; FEV₁ = forced expiratory volume in the first second; ICS = inhaled corticosteroids; OCS = oral corticosteroids.

Note: Numbers in square brackets, [ ], indicate ranking in the multiple testing hierarchy.

Sources: Clinical Study Reports for NAVIGATOR,⁴ SOURCE,⁵ and PATHWAY.⁶

Asthma Exacerbations

In the included studies, an asthma exacerbation (as recorded on the exacerbation electronic case report form page) was defined as a worsening of asthma that leads to any of the following:

• a temporary bolus or burst of systemic corticosteroids (or a temporary increase in stable OCS background dose) for at least 3 consecutive days to treat symptoms of asthma worsening; a single depot-injectable dose of corticosteroids will be considered equivalent to a 3-day bolus or burst of systemic corticosteroids

• an ED or urgent-care visit (defined as evaluation and treatment for less than 24 hours in an ED or urgent-care centre due to asthma that required systemic corticosteroids, as described previously)

• an inpatient hospitalization (defined as admission to an inpatient facility and/or evaluation and treatment in a health care facility for ≥ 24 hours) due to asthma.

The start of an exacerbation was defined as the start date of systemic corticosteroids, ED or urgent-care visits requiring systemic corticosteroids, or hospital admissions due to asthma, whichever occurs earlier. The end date was defined as the last day of systemic corticosteroids or ED, urgent-care facility, or hospital discharge, whichever occurs later.

Two or more exacerbations with the same start date and end date were counted as 1 exacerbation for the purposes of calculating the number and duration of exacerbations for a patient. In cases in which 1 or more exacerbations were recorded as starting or ending during another exacerbation, these were counted as 1 exacerbation, using the earliest exacerbation start date and the latest exacerbation stop date to calculate duration.

Additional systemic corticosteroid treatments, ED or urgent-care visits requiring use of systemic corticosteroids, or inpatient hospitalization due to asthma occurring during an exacerbation were not regarded as a new exacerbation. To be counted as a new exacerbation it must have been preceded by at least 7 days during which neither criterion is fulfilled. If the end date of the first exacerbation and the start date of the second exacerbation were less than 7 days apart, then these were counted as 1 exacerbation.

Asthma Quality of Life Questionnaire (Standardized) for Patients 12 Years of Age and Older

In the AQLQ(S) + 12 the patients are asked to recall their experiences during the previous 2 weeks and to score each of the 32 questions on a 7-point scale ranging from 7 (no impairment) to 1 (severe impairment).³⁵ The total score is calculated as the mean response to all questions. The 4 individual domain scores (assessing symptoms, activity limitations, emotional function, and environmental stimuli) are the means of the responses to the questions in each of the domains. No specific minimal important difference (MID) has been established for the AQLQ(S) + 12; however, due to similarities with the Asthma Quality of Life Questionnaire (Standardized), which has an MID of 0.5, a cut point of 0.5 points is considered to be clinically meaningful.^36-38 The AQLQ(S)12+ was assessed at weeks 4, 12, 24, 36, and 52 in the NAVIGATOR study, and weeks 4, 40, 44, and 48 in the SOURCE study. In PATHWAY study, the sponsor noted in the schedule of assessments that compliance with patient-reported outcome assessments were checked every 2 weeks.

Six-Item Asthma Control Questionnaire

The ACQ-6 questionnaire³⁹ includes questions on awakening at night by symptoms, limitations of normal daily activities, waking in the morning with symptoms, dyspnea, wheezing, and daily rescue medication. The questions of the ACQ-6 are measured on a 7-point scale scored from 0 (totally controlled) to 6 (severely uncontrolled). The ACQ-6 score computed as the unweighted mean of the responses to the 6 questions. If a response to any of the questions is missing, the ACQ-6 will be missing. A mean score of 0.75 or lower indicates well-controlled asthma, scores between 0.75 and up to 1.5 indicate partly controlled asthma, and a score higher than 1.5 indicates poorly controlled asthma.⁴⁰ Changes of at least 0.5 are considered clinically meaningful.^1,40-43 The ACQ-6 was to be completed at the beginning of each site visit by use of an electronic diary. This corresponded to week 2, 4, and then every 4 weeks until week 52 in the NAVIGATOR study, and every 4 weeks in the SOURCE study. In the PATHWAY study, the sponsor noted in the schedule of assessments that compliance with patient-reported outcome assessments was checked every 2 weeks.

5-Level EQ-5D Questionnaire

The EQ-5D-5L^44-46 is a generic HRQoL instrument. The first 2 components of the EQ-5D-5L assess 5 domains: mobility, self-care, usual activities, pain and/or discomfort, and anxiety and/or depression, and each domain has 5 levels ranging from no problem to extreme problems. The next component of the EQ-5D-5L consists of a 20 cm Visual Analogue Scale (EQ VAS) that has end points labelled 0 and 100, with respective anchors of “worst imaginable health state” and “best imaginable health state.” Respondents are asked to draw a line from an anchor box to the point on the EQ VAS that best represents their health on that day. The EQ-5D-5L produces 3 types of data for each respondent, a profile indicating the extent of problems on each of the 5 dimensions represented by a 5-digit descriptor, a population-preference weighted health index score based on the descriptive system, and a self-reported assessment of health status based on the EQ VAS. The index score is calculated by applying a multi-attribute utility function to the descriptive system. Scores of 0 represent the health state “dead” and 1 represents “perfect health.” No MID was identified for the EQ-5D-5L that was specific to asthma; however, in the general population an MID of 0.056 has been established for index scores.⁴⁷ The EQ-5D-5L was administered every 2 weeks via electronic diaries in the NAVIGATOR and SOURCE studies.

Pulmonary Function

Spirometry was performed by the investigator or authorized designate, and followed guidelines set by the American Thoracic Society and European Respiratory Society. Patients were advised to avoid strenuous exertion for at least 30 minutes or large meals for the 2 hours before the assessment. Patients were to withhold SABAs at least 6 hours before their scheduled assessment and LABA-LAMAs for 12 hours prior (for twice-daily formulations) or 24 hours (for once-daily formulations). Administration of LTRAs was to be held for 24 hours, twice-daily theophyllines for at least 12 hours, and once-daily theophyllines at least 24 hours before scheduled assessment. In the NAVIGATOR study, spirometry was carried out at weeks 2 and 4 and then every 4 weeks until week 52. In the SOURCE study, spirometry was conducted at weeks 4, 12, 24, 40, and 48.

Statistical Analysis

The statistical analysis methods used in the NAVIGATOR, SOURCE, and PATHWAY trials are summarized in Table 9.

Primary Outcome(s) of the Studies

Power Calculation

In the NAVIGATOR study, the planned number of 530 patients were enrolled, with an overall type I error control at an alpha of 0.05 and type I error control for the primary outcome at an alpha of 0.01, resulting in an estimated 90% power for key secondary outcomes. For the primary outcome, with a placebo rate of 0.9, a shape parameter of 2.4, and a dropout rate of 10%, there was greater than 99% power to detect a rate reduction of 50% at a 2-sided significance level of 1%. For the primary outcome in patients with baseline eosinophils below 300 cells/µL, with a placebo rate of 0.6 and assuming half the patients will be in this group, there was 94% power to detect a rate reduction of 50% at a 2-sided significance of 5%. For the key secondary outcome, the nominal power was 95% or higher, assuming SDs of 400 mL for FEV₁ and 1.3 for all questionnaires.

For the SOURCE study, with 152 patients assigned to 2 treatment groups, it was estimated that a 2-sided 5% significance level would produce at least 90% power to reject the null hypothesis for the primary outcome, assuming an OR of 2.75 and the proportional odds assumption. Proportions assumed for placebo were 10% of patients for each of category 1 (90% to 100% reduction in OCS dose) and 2 (75% to < 90% reduction), 15% of patients for each of category 3 (50% to < 75% reduction) and category 4 (> 0% to < 50% reduction), and 50% for category 5 (no reduction or an increase). The minimum detectable OR that was still significant with these assumptions was 1.86. For the key secondary outcome, it was estimated that, with 76 patients per group, the study had greater than 80% power to reject the null hypothesis for rate ratios up to 0.39, using a 2-sided 5% significance level, and under the following assumptions: a placebo rate of 1.3 exacerbations per year in this study population, a conservative assumption on the dispersion parameter (2.4), and a uniform dropout rate of 10%.

For the PATHWAY study, the primary analysis was based on the intention-to-treat (ITT) population. A total of 124 patients per group were required to detect a 40% reduction in the AAER for each tezepelumab dose group compared to placebo. Assuming an AAER of 0.7 in the placebo group, a 2-sided significance of 0.1 was applied to the hypothesis of comparing a single tezepelumab dose group to placebo. An AAER of 0.7 in the placebo group was estimated based on internal and external studies with a similar patient population (before considering dropouts). The dispersion parameter was selected from a mepolizumab study in patients with severe eosinophilic asthma. The sample size was increased to accommodate an estimated 10% dropouts. The minimal detectable difference was approximately a 28% reduction in AAER. The minimum acceptable reduction in AAER for the study population was based upon the expected reduction of approximately 50% in a Th2-driven population, and a more modest reduction in a non-Th2 asthma population for which there were few data and no current competitors.

Statistical Test or Model

In the NAVIGATOR study, a hierarchical testing strategy was implemented to test for superiority of tezepelumab over placebo in each of the primary and key secondary end points, while controlling the overall type I error rate at 0.05 (2-sided). The primary end point in all patients was tested at a 2-sided significance level of 0.01. The primary analysis of the primary end point compared the AAER over 52 weeks between treatment groups using a negative binomial model. The response variable was the number of asthma exacerbations experienced by the patient over the study period. Treatment, region, age, and history of exacerbations were included as factors in the model. The logarithm of the time at risk for exacerbation in the study was used as an offset variable.

In the SOURCE study, a hierarchical testing strategy was also employed, and the order of outcomes in the hierarchy can be found in Table 8. The null hypothesis for the primary outcome was tested first, at a 2-sided 5% level of significance. If this was rejected, the key secondary outcome was tested at a 2-sided 5% significance level, and so on.

For the 5 ordered categories of the primary outcome, the cumulative OR of the percentage reduction category from baseline in daily OCS dose at 48 weeks while not losing asthma control, there are 4 possible cumulative odds for each treatment group, corresponding to the 4 different possible binary splits. These binary splits were defined as:

• category 1 versus categories (2, 3, 4, and 5)

• categories (1 and 2) versus categories (3, 4, and 5)

• categories (1, 2, and 3) versus categories (4 and 5)

• categories (1, 2, 3, and 4) versus category 5.

The ordered categories for OCS daily dose reduction were in turn defined as:

• category 1 = 90% to 100% reduction

• category 2 = 75% to lower than 90% reduction

• category 3 = 50% to lower than 75% reduction

• category 4 = 0% to lower than 50% reduction

• category 5 = no reduction or an increase.

Key secondary outcomes in the SOURCE study were tested in a similar manner to that used for the NAVIGATOR study.

In the PATHWAY study, there does not appear to have been a strategy for controlling for multiplicity. The primary and secondary outcomes were tested in a similar manner to those in the NAVIGATOR study. Analysis of the primary outcome, AAER, was conducted using a negative binomial regression model with treatment group, baseline eosinophils, and baseline ICS dose as covariates and the response variable in the model was the number of exacerbations experienced by a patient over the 52-week study period. The follow-up time was adjusted by the offset option in the model to adjust for different exposure times among patients during the events that occurred as the logarithm of number of days in the study.

Subgroup Analyses

Pre-planned subgroup analyses were conducted on the primary and key secondary outcomes in each of the included studies for the following subgroups identified in our review protocol: baseline eosinophil counts (various cut points), ICS dose (medium and/or high), exacerbations in the past year (various cut points), OCS use at baseline and allergy status (allergic or not allergic). In the NAVIGATOR study, the specific subgroup of patients with eosinophil counts lower than 300 cells/µL was included as part of the statistical hierarchy; however, no other subgroups across all the studies were adjusted for multiplicity, and no P values were reported for any analyses presented.

Sensitivity Analyses

In the NAVIGATOR study, sensitivity analyses were performed on the primary and key secondary end points, including analyses to explore the impact of missing data and early discontinuation from investigational product. Further analyses were also performed to explore the consistency of treatment effects across demographic and baseline subgroups. Tipping-point analyses assessed pre-specified degrees of improvement in the placebo group and degrees of worsening in the tezepelumab group after study withdrawal.

In the SOURCE study, sensitivity analyses were to be performed if the amount of missing data was sufficient to potentially affect interpretation, using multiple imputation-by-pattern mixture models, where imputation of missing OCS doses at each visit up to and including week 48 was performed in 2 steps. The non-monotone (intermediate) missing values were to be imputed first, assuming they were missing at random (MAR). A Markov chain Monte Carlo (MCMC) method was used to partially impute the data. The remaining missing values at each visit were to be imputed using a sequential regression method. At each iteration, missing values were to be imputed sequentially, 1 time point at a time. Specifically, the model for each visit focused on the OCS dose at the visit as the response variable and included the following predictors: the covariates incorporated in the primary statistical model, including treatment, region, and baseline OCS dose, and the OCS dose from baseline at all the prior visits. A total of 100 imputations were to be carried out with the seed of 518,009. These imputation processes were to be carried out twice depending on the assumed mechanism of missingness, MAR, or dropout reason–based multiple imputation, which assumes that the trajectories for patients in the tezepelumab group who drop out for a treatment-related reason will follow the pattern of placebo observations based on an assumption of missing not at random, whereas remaining patients who dropped out were to be imputed assuming they were MAR.

A single imputation method, average dose, was also used; where a patient who withdrew at any point after their baseline assessment, the final OCS dose was to be imputed to the average daily dose that a patient was taking in the 14 days before their discontinuation from study drug.

Additionally, the actual percentage reduction in daily OCS dose was compared between groups using a van Elteren test. Patients who received systemic corticosteroids for other (non-asthma) reasons were summarized. All sensitivity analyses were also repeated using on-treatment data only. The final OCS dose for those patients discontinuing the study drug was to be the actual dose received from 1 dose level higher when asthma stability was verified, at the time of discontinuation of study drug.

In the PATHWAY study, sensitivity analyses were to be carried out using a Poisson regression model to assess robustness of the distributional assumptions. Correction for potential over-dispersion was to be made by the Pearson chi-square method.

Secondary Outcomes of the Studies

In the NAVIGATOR study, the main analysis of key secondary outcomes analyzed patients according to the treatment they were randomized to and included results from patients who received other controller or rescue therapies, as well as those who discontinued therapy early. Missing data from study discontinuations were modelled based on what was observed during the study using direct likelihood approaches, which are valid under the assumption that data are MAR. The change from baseline was compared to placebo using a mixed method for repeated measures model. The response variable in the model was change from baseline at each scheduled post-randomization visit, and treatment, visit, region, age (adolescent or adult) and treatment-by-visit interaction will be included as factors in the model. Baseline value of the corresponding outcome will also be included as a continuous linear covariate. Key secondary outcomes in the SOURCE and PATHWAY studies were tested in a similar manner to those in the NAVIGATOR study, although no testing hierarchy was planned for the PATHWAY study.

Table 9: Statistical Analysis of Efficacy End Points

AAER = annual asthma exacerbation rate; ACQ-6 = 6-item Asthma Control Questionnaire; AQLQ(S)12+ = Asthma Quality of Life Questionnaire (Standardized) for patients 12 years of age and older; CFB = change from baseline; EQ-5D-5L = EQ-5D 5-Levels questionnaire; FEV₁ = forced expiratory volume in the first second; ICS = inhaled corticosteroids; MCMC = Markov chain Monte Carlo; MMRM = mixed model for repeated measures; NA = not applicable.

Sources: Clinical Study Report for NAVIGATOR,⁴ SOURCE,⁵ and PATHWAY.⁶

Analysis Populations

In the NAVIGATOR and SOURCE studies, efficacy analyses were performed on the full analysis set (FAS), which consisted of all randomized patients who received any study drug. The safety analysis set included all patients who received at least 1 dose of the study drug.

In the PATHWAY study, the ITT population was defined as patients who were randomized and received any study drug, and patients were analyzed according to their randomized treatment group. The as-treated population included any patients who received the study drug, and patients were analyzed according to the treatment they actually received. There was also a per-protocol population, which included patients who did not have significant protocol violations and received at least 80% of the intended doses of investigational product.

Results

Patient Disposition

Study withdrawals were generally below 10% in each group in each study, with the exception of the PATHWAY study, from which 11% of tezepelumab patients and 6% of placebo patients withdrew. This difference in study withdrawals between groups in the PATHWAY study was also the largest difference between groups across all the studies. The most common reason for withdrawal was “withdrawal by patient” and “other,” which occurred in 5% and 3% of tezepelumab versus placebo patients, respectively.

Exposure to Study Treatments

In the NAVIGATOR study, the mean duration of exposure with tezepelumab was 352.3 days (SD = 59.1) and in the placebo group mean duration of exposure was 342.2 (SD = 74.9) days. In the SOURCE study, the mean duration of exposure was 318.7 days (SD = 59.8) in the tezepelumab group and 325.8 days (SD = 51.7) in the placebo group. In the PATHWAY study, the mean duration of exposure was not reported; however, 70% of patients in each of the tezepelumab and placebo groups received all 26 planned doses.

Adherence to the study drug was monitored at clinic visits, at which patients had drugs administered and adherence to background therapy was monitored by electronic diaries. Days with missing diary entries were assumed to indicate nonadherence. Adherence to study drug was 99% across both groups in the NAVIGATOR study, and adherence to background medications for asthma was 80% in the tezepelumab group and 78% in the placebo group. In the SOURCE study, adherence to the study drug was 97% across both groups, and adherence to OCS during the optimization phase was 97% and during the down-titration period was 90% with tezepelumab and 93% with placebo. Adherence to background asthma therapy was 92% with tezepelumab and 93% in the placebo group. Adherence was not reported in the PATHWAY study, although it was monitored, according to the sponsor.

Table 10: Patient Disposition

ITT = intention-to-treat; NR = not reported; Tez = tezepelumab.

^aThis patient withdrew from the study as they were unable to continue due to an adverse event (preferred term: invasive breast carcinoma).

^bOther reasons included patient’s decision, AE, failure to meet eligibility criteria, visit out of window, sponsor decision, lack of available investigational product, pregnancy, use of prohibited medications, and missed dose.

Sources: Clinical Study Report for NAVIGATOR,⁴ SOURCE,⁵ and PATHWAY.⁶

Efficacy

Only those efficacy outcomes and analyses of subgroups identified in the review protocol are reported in the following section. Appendix 3 provides detailed efficacy data.

Mortality

Across all studies, there was only 1 death in the tezepelumab group and 2 deaths in the placebo group. The 2 deaths in the placebo group were due to an unknown cause and heart failure, both in the NAVIGATOR study, and 1 patient in the tezepelumab group died in the SOURCE study, due to cardiac arrest.

Asthma Exacerbations

The AAER was the primary outcome of both the NAVIGATOR and PATHWAY studies. In the NAVIGATOR study, the AAER over 52 weeks was 0.93 (95% CI, 0.80 to 1.07) with tezepelumab and 2.10 (95% CI, 1.84 to 2.39) with placebo, for a rate ratio of 0.44 (95% CI, 0.37 to 0.53; P < 0.001). Various sensitivity analyses were conducted to account for missing data, and results were consistent with that of the primary analysis, according to the sponsor. In the PATHWAY study, the AAER over 52 weeks was 0.20 (95% CI, 0.13 to 0.30) with tezepelumab and 0.72 (95% CI, 0.59 to 0.88) with placebo, for a rate ratio of 0.29 (95% CI, 0.16 to 0.51). In the SOURCE study, the AAER over 48 weeks was 0.69 (95% CI, 0.44 to 1.09) (Table 11).

The AAER associated with an ED visit or hospitalization was 0.06 (95% CI, 0.04 to 0.09) with tezepelumab and 0.28 (95% CI, 0.20 to 0.39) with placebo, for a rate ratio of 0.21 (95% CI, 0.12 to 0.37) in the NAVIGATOR study, and was 0.16 (95% CI, 0.06 to 0.44) with tezepelumab and 0.28 (95% CI, 0.13 to 0.58) with placebo in the SOURCE study, for a rate ratio of 0.59 (95% CI, 0.19 to 1.82) (Table 11).

The AAER associated with a hospitalization was 0.03 (95% CI, 0.01 to 0.06) with tezepelumab and 0.19 (95% CI, 0.12 to 0.30) with placebo, for a rate ratio of 0.15 (95% CI, 0.07 to 0.33; P < 0.001) in the NAVIGATOR study, and 0.02 (95% CI, 0.00 to 0.07) with tezepelumab and 0.14 (95% CI, 0.08 to 0.22) with placebo in the PATHWAY study, for a rate ratio of 0.14 (95% CI, 0.03 to 0.71) (Table 11).

Subgroup analyses were performed for a number of subgroups relevant to our protocol and those reported for the primary outcome of each study are reported in Table 25. In the NAVIGATOR study, the AAER in patients with baseline eosinophil counts below 300 cells/µL was part of the statistical hierarchy and was therefore controlled for multiplicity, with an AAER for tezepelumab versus placebo of 1.02 versus 1.73, and a rate ratio of 0.59 (95% CI, 0.46 to 0.75; P < 0.001). None of the other subgroup analyses reported were adjusted for multiplicity and no P values were reported.

Pulmonary Function

The pre-bronchodilator FEV₁ increased in both the tezepelumab and placebo groups in the NAVIGATOR study, with a LSM change from baseline to 52 weeks of 0.23 L (SE = 0.018) with tezepelumab and 0.10 L (0.018) with placebo for an LSM difference between groups of 0.13 L (95% CI, 0.08 to 0.18; P < 0.001). In the SOURCE study, the change from baseline to week 48 was 0.21 L (SE = 0.046) with tezepelumab and −0.04 L (SE = 0.046) with placebo, for an LSM difference between groups of 0.26 L (95% CI, 0.13, 0.39), and in the PATHWAY study, the LSM change from baseline to week 52 was 0.076 L (SE = NR) with tezepelumab and −0.056 L (NR) with placebo, for an LSM difference between groups of 0.132 L (95% CI, 0.033 to 0.231) (Table 11).

Reduction in Oral Corticosteroid Use

Reduction in OCS use while still maintaining asthma control was the primary outcome of SOURCE. The cumulative OR for patients having a reduction in OCS dose was 1.28 (95% CI, 0.69 to 2.35; P = 0.434) (Table 11). Tezepelumab therefore failed to demonstrate superiority over placebo for the primary outcome of this study. There were 54% of patients in the tezepelumab group and 46% of placebo patients who achieved a dose reduction of 90% or more. Other thresholds for dose reduction for tezepelumab versus placebo included 75% or greater to less than 90% (7% versus 5%, respectively), 50% or greater to less than 75% (14% versus 18% respectively), and greater than 0 to less than 50% (19% versus 18%, respectively).

This outcome was not assessed in the NAVIGATOR or PATHWAY study.

Reduction in Rescue Medication

Reduction in daily rescue medication was observed in both the tezepelumab and placebo groups in the NAVIGATOR study, with an LSM change from baseline of −2.53 puffs (SE = 0.137) with tezepelumab and −2.36 puffs (0.137) with placebo, for an LSM difference between groups of −0.17 (95% CI, −0.55 to 0.21) (Table 11). Rescue medication use also declined in both groups in the SOURCE study, with an LSM change from baseline to 48 weeks of −0.85 puffs (SE = 0.280) with tezepelumab and −0.37 puffs (SE = 0.268) with placebo, for a LSM difference between groups of −0.47 (95% CI, −1.24 to 0.29).

This outcome was not assessed in the PATHWAY study.

Health-Related Quality of Life

Health-related quality of life was assessed using the AQLQ(S)12+ and the EQ-5D-5L instruments. Mean AQLQ(S)12+ scores increased (improved) from baseline in the NAVIGATOR study, both in the tezepelumab 1.48 (SD = 1.26) and placebo 1.17 (SD = 1.16) groups, with a difference between groups of 0.33 (95% CI, 0.20 to 0.47; P = 0.001) (Table 11). In the SOURCE study, the LSM change from baseline was 0.94 (SE = 0.124) with tezepelumab and 0.58 (SE = 0.123) with placebo for a difference between groups of 0.36 (95% CI, 0.01 to 0.70). In the PATHWAY study, the LSM change from baseline to week 52 was 1.17 (SE = NR) with tezepelumab and 0.97 (SE = NR) with placebo for a difference between groups of 0.20 (−0.09 to 0.48). Responders to the AQLQ(S)12+ were also reported, defined as those with a change from baseline of 0.5 or greater. In the NAVIGATOR study 78% of tezepelumab patients and 72% of placebo patients were responders, for an OR of 1.36 (95% CI, 1.02 to 1.82). In the SOURCE study, 62% of tezepelumab patients and 52% placebo were responders, for an OR of 1.66 (95% CI, 0.81 to 3.43). In the PATHWAY study, 73% of tezepelumab patients and 62% of placebo patients were responders.

In the NAVIGATOR study, on the EQ VAS, the LSM change from baseline was 14.64 (SE = 0.708) and 11.86 (SE = 0.712) with placebo, for a LSM difference between groups of 2.78 (95% CI, 0.81, 4.75) (Table 11). In the SOURCE study, the LSM change from baseline to week 48 was 9.21 (SE = 2.209) with tezepelumab and 2.00 (SE = 2.226) with placebo, for an LSM difference between groups of 7.21 (95% CI, 1.01 to 13.41). In the PATHWAY study, the LSM change from baseline was 11.7 (SE = 19.1) with tezepelumab and 10.5 (SE = 17.4) with placebo. No between-group differences were reported.

The EQ-5D-5L health index scores were also reported. In the NAVIGATOR study, the LSM change from baseline for tezepelumab was 0.11 (SE = 0.008) and 0.07 (SE = 0.009) for placebo, for a LSM difference between groups of 0.03 (95% CI, 0.01 to 0.06). In the SOURCE study, the LSM change from baseline with tezepelumab was 0.07 (SE = 0.026) and 0.0 (SE = 0.027) for placebo, for a LSM difference between groups of 0.07 (95% CI, −0.01 to 0.14). In the PATHWAY study, the LSM change from baseline to week 52 was 0.0858 (SE = 0.1994) with tezepelumab and 0.0719 (SE = 0.1824) with placebo.

Asthma Symptoms

Symptoms were assessed as a key secondary outcome using the ACQ-6. In the NAVIGATOR study, ACQ-6 scores decreased (improved) from baseline in both the tezepelumab and placebo groups, for a difference versus placebo of 0.33 (95% CI, 0.20 to 0.47; P < 0.001) (Table 11). Responders to the ACQ-6 were also reported, defined as those with a change from baseline of 0.5 or greater; 86% of tezepelumab patients and 77% of placebo patients were responders, for an OR of 1.99 (95% CI, 1.43 to 2.76). In the SOURCE study, the LSM change from baseline to 48 weeks was −0.87 (SE = 0.125) for tezepelumab and −0.51 (SE = 0.123) with placebo, for a difference between groups of −0.37 (95% CI, −0.71 to −0.02). In the PATHWAY study, the LSM change from baseline to week 52 was −1.20 with tezepelumab and −0.91 with placebo, for a difference between groups of −0.29 (95% CI, −0.56 to −0.01).

Reduction in ICS Dose

This outcome was not investigated.

Table 11: Efficacy, FAS Population (NAVIGATOR, SOURCE) and ITT Population (PATHWAY)

AAER = annualized asthma exacerbation rate; ACQ-6 = 6-item Asthma Control Questionnaire; AQLQ12+ = Asthma Quality of Life Questionnaire (Standardized) for patients 12 years of age and older; BD = bronchodilator; CFB = change from baseline; CI = confidence interval; ED = emergency department; EQ-5D-5L = EQ-5D 5-Levels questionnaire; FEV₁ = forced expiratory volume in the first second; ICS = inhaled corticosteroid; LSM = least squares mean; NR = not reported; OCS = oral corticosteroids; OR = odds ratio; SD = standard deviation; SE = standard error; Tez = Tezepelumab; vs. = versus.

^aModel: a negative binomial regression analysis with treatment, region, and history of exacerbations as covariates. The logarithm of the time at risk is used as an offset variable. Annual exacerbation rates displayed are estimated marginal rates from the model. All CIs for annual exacerbation rates and absolute differences are estimated via the delta method.

^bThe rate ratio, and 95% CI for the rate ratio were estimated from negative binomial regression with treatment group, with stratification factors of baseline blood eosinophil count (≥ or < 250 cells/μL) and baseline ICS dose level (medium or high) as the covariates.

^cModel: a negative binomial regression analysis with treatment, region, age group, and history of exacerbations as covariates. The logarithm of the time at risk is used as an offset variable. Annual exacerbation rates displayed are estimated marginal rates from the model. Absolute difference is the difference between the marginal rates. All CIs for annual exacerbation rates and absolute differences are estimated via the delta method.

^dP values are not controlled for multiplicity or failed in the hierarchy and should be considered supportive in nature.

^eEstimate of the mean change from baseline at each week in tezepelumab is compared to placebo using a repeated measures analysis. Estimates are least squares means. The model with unstructured covariance structure is: change from baseline in FEV₁ = treatment group + region + age + baseline FEV₁ + visit + treatment × visit.

^fThe estimate of the cumulative odds ratio is obtained using a proportional odds model with treatment, region, and daily OCS dose at baseline as covariates.

^gEstimate of the mean change from baseline at each week in Tezepelumab is compared to the Placebo using a repeated measures analysis. Estimates are least squares means. The model with Unstructured covariance structure is: Change from baseline in Rescue medication use weekly means = Treatment group + region + age + baseline rescue medication use + week + treatment * week.

^hEstimate of the mean change from baseline at each week in tezepelumab is compared to placebo using a repeated measures analysis. Estimates are least squares means. The model with unstructured covariance structure is: change from baseline in AQLQ(S)12+/ACQ-6 = treatment group + region + baseline AQLQ(S)12+/ACQ-6 + visit + treatment × visit.

ⁱEstimate of the mean change from baseline at each week in Tezepelumab is compared to the Placebo using a repeated measures analysis. Estimates are least squares means. The model with Unstructured covariance structure is: Change from baseline in VAS/index score = Treatment group + region + age + baseline VAS/index score + visit + treatment * visit.

^jSample sizes for these outcomes in the PATHWAY study were low due to an error in collection of ePRO data.

^kThe estimate of the odds ratio is obtained using a generalized estimating equation model for repeated measures binary data with an unstructured covariance structure and treatment, region, age, visit, visit × treatment, and baseline ACQ-6 and/or AQLQ(S)12+ score as covariates.

Sources: Clinical Study Report for NAVIGATOR,⁴ SOURCE,⁵ and PATHWAY.⁶

Harms

Only those harms identified in the review protocol are reported here. Table 12 provides detailed harms data.

Adverse Events

Adverse events in the tezepelumab versus placebo groups occurred in 77% versus 80% of patients, respectively, in the NAVIGATOR study; 72% versus 86%, respectively, in the SOURCE study; and 66% of patients in each group in the PATHWAY study (Table 12).

The most common AE was nasopharyngitis, occurring in 21% versus 21% of patients in the NAVIGATOR study, 15% versus 25% of patients in the SOURCE study, and 14% versus 12% of patients in the PATHWAY study in the tezepelumab versus placebo groups, respectively. Other common events (occurring in 10% or more of patients in any group of any study) were upper respiratory tract infections, headaches, and asthma.

Serious Adverse Events

Serious adverse events in the NAVIGATOR study, tezepelumab versus placebo, occurred in 9% versus 13% of patients. In the SOURCE study they occurred in 15% versus 21% of patients, respectively, and in the PATHWAY study they occurred in 10% versus 13% of patients, respectively (Table 12). The most common SAE was asthma.

Table 12: Summary of Harms, Safety Analysis Set

AE = adverse event; NA = not assessed; SAE = serious adverse event; Tez = tezepelumab; URTI = upper respiratory tract infection.

Sources: Clinical Study Report for NAVIGATOR,⁴ SOURCE,⁵ and PATHWAY.⁶

Withdrawals due to Adverse Events

Adverse events resulting in discontinuation of the study drug occurred in 2% versus 4% of patients in the tezepelumab versus placebo groups, respectively, in the NAVIGATOR study, 3% in each group in the SOURCE study, and 2% versus 1% of patients, respectively, in the PATHWAY study (Table 12).

Notable Harms

Notable harms in the CADTH systematic review protocol included infections. Severe infections occurred in the tezepelumab versus placebo groups in 9% versus 8% of patients in the NAVIGATOR study, respectively, and 5% versus 9% of patients in the SOURCE study, respectively (Table 12). In the PATHWAY study, infections were reported as SAEs rather than severe infections, and these occurred in 1% versus 3% of patients in the tezepelumab versus placebo groups, respectively. No opportunistic infections and no helminth infections were reported across the studies.

Injection-site reactions occurred infrequently across the studies. In the NAVIGATOR study, injection-site reactions occurred in 1.5% versus 0.9% of patients in the tezepelumab versus placebo groups, respectively; in the SOURCE study, there were none with tezepelumab group, and 1.3% of patients in the placebo group experienced these events. In the PATHWAY study, these events were reported by injection volume, and at the 1 mL volume they occurred in 1.5% versus 2.9% of patients in the tezepelumab versus placebo groups, respectively, and at the 1.5 mL volume they occurred in 1.5% versus 1.4% of patients, respectively. Hypersensitivity reactions reported as SAEs were infrequent, at 1 patient in each of the tezepelumab and placebo groups in the NAVIGATOR study, and none in the other studies.

Critical Appraisal

Internal Validity

All 3 studies appeared to take appropriate steps for the randomization to maintain allocation concealment (use of an interactive voice- or web-response system), and to maintain blinding (matching placebo injections). The phase III studies (NAVIGATOR and SOURCE) used a hierarchical testing procedure to control for multiplicity. The lack of multiplicity control in the PATHWAY study means that all reported P values should be considered supportive and not used to draw conclusions. The phase II dose-ranging design is an additional limitation when interpreting data from the PATHWAY study, as this design was not appropriate for determining superiority. Given the statistical testing could not reject the null hypothesis for the primary outcome in the SOURCE study, failure of the statistical hierarchy means that any P values less than the pre-specified 2-sided 5% significance level reported for outcomes after the primary outcome should also be considered supportive of efficacy. Failure of statistical comparison on the primary outcome and the impact on the hierarchy for the secondary outcome limits any conclusions that can be drawn from the SOURCE trial. The FDA review of tezepelumab focused on the results of the NAVIGATOR study because of these limitations with the other studies.

Withdrawals from the study were generally below 10% across the studies, and the only study for which this was not the case, PATHWAY, had numerically more withdrawals in the tezepelumab group compared to the placebo group (11% versus 6%, respectively). This difference in withdrawals could potentially bias results in favour of tezepelumab for outcomes such as exacerbations, as patients no longer in the study are no longer at risk of an exacerbation. Additionally, as the number of withdrawals increases there is an increasing risk that baseline characteristics will no longer be balanced between groups, and that these imbalances may bias results either for or against the study drug. The sponsor performed a variety of sensitivity analyses in an attempt to account for missing data. Despite steps to account for missing data, more than 10% of data were missing from end-of-study assessments for continuous outcomes such as FEV₁, ACQ-6, AQLQ(s)12+, and EQ-5D-5L. For example, in the NAVIGATOR study, for the week-52 assessment of FEV₁, the reported sample was 471 patients (of 528 at baseline) with tezepelumab and 453 of 531 patients in the placebo group, which were 11% and 14% less, respectively, than the baseline FAS population. Similar but slightly smaller amounts of missing data were seen for the ACQ-6 (8% tezepelumab and 11% placebo) and the AQLQ(S)12+ (10% and 12%, respectively). Larger amounts of data were missing for the EQ-5D-5L (15% and 17% in the tezepelumab and placebo groups, respectively, in the NAVIGATOR study). All of these numbers are higher than the reported number of patients who withdrew from study treatment (7% tezepelumab and 11% placebo), and higher than the number of patients reported to have withdrawn from the study (3% and 4%, respectively). It is unclear why data were missing for patients who were not only in the study but remaining on treatment, and this should be considered a limitation of the findings for these outcomes. Similar amounts of missing data were seen for these outcomes in the SOURCE study. In the PATHWAY study, there was a larger amount of missing data for the week-52 assessments for the ACQ-6 and AQLQ(S)12+, and this was attributed to an error in the administration of the electronic patient-reported outcome instruments. This resulted in 60% to 70% of the FAS missing from the week-52 analysis, meaning that these data are unlikely to be reliable. The sponsor reported data for these outcomes for the last reliable time point, week 48, and those results were consistent with data reported at week 52.

There were pre-specified subgroups in each of the included studies; however, only 1 of these in the NAVIGATOR study (AAER in patients with eosinophil counts < 300 cells/µL) were adjusted for multiple statistical comparisons, as it was part of the hierarchical testing procedure. No P values were therefore reported for the remaining subgroups across all the included studies, and no P values were reported for interactions. As a result, data for pre-specified subgroups were reported, and only limited conclusions can be drawn from these data. Additionally, for this subgroup in the NAVIGATOR trial (AAER in patients with eosinophils < 300 cells/µL at baseline) it was not clear from the Clinical Study Report whether this outcome was a co-primary outcome, and the FDA rejected it as a co-primary outcome in their review.

The primary and key secondary outcomes assessed in the included trials are all well-established, well-validated outcomes used for asthma. Minimal important differences have been established for the ACQ-6 and AQLQ(S)12+, and these are reviewed in detail in Appendix 4. Although not part of the statistical hierarchy, data were reported for ACQ-6 and AQLQ(S)12+ “responders,” and a response was defined using the MID for each of these instruments.

Noteworthy differences between groups in baseline characteristics were observed in the SOURCE study, in which fewer patients in the tezepelumab than in the placebo group had more than 2 exacerbations in the past year (18% versus 28% of patients, respectively). Given the importance of asthma exacerbations as a marker of asthma control, and the fact that AAER was a key secondary outcome in the SOURCE study, this difference may be important. If patients in the tezepelumab group were less prone to having multiple exacerbations the results may have been biased in favour of tezepelumab for the AAER outcome. Although a lower exacerbation rate with tezepelumab may also indicate their asthma was better controlled than in the placebo population, baseline ACQ-6 scores, another measure of asthma control, were similar between groups.

External Validity

The primary and key secondary outcomes were all relevant outcomes for assessing patients with asthma, notably exacerbation rate, pulmonary function (FEV₁), HRQoL (AQLQ and EQ-5D-5L) and symptoms (ACQ-6). Reduction in OCS dose without losing asthma control was the primary outcome of the SOURCE study, and this was an appropriate outcome given the primary objective of the study. However, failure to demonstrate superiority of tezepelumab over placebo at this early stage of the hierarchy meant that all testing of all subsequent outcomes yielded P values that were “supportive.”

Although the clinical expert consulted by CADTH for this review noted that they would not consider putting a patient on tezepelumab, or any biologic, until they trialled them on high-dose ICS, 25% of patients in the NAVIGATOR study were on a medium-dose ICS at baseline. This, and the relatively large number of patients who demonstrated a high degree of airway reversibility suggests that this population may have been undertreated upon entering the trial, and this may affect the generalizability of the findings. The clinical expert also noted that the proposed indication for tezepelumab is for patients with severe asthma, and the expert would not consider patients who are taking a medium dose of ICS to have severe asthma.

None of the included trials compared tezepelumab to any of the other biologics approved for management of asthma, the IgE inhibitor omalizumab, the IL-5 inhibitors, or the IL4 and IL-13 inhibitor (dupilumab). As omalizumab and the IL-5 inhibitors have been approved for several years in patients with eosinophilic asthma or those with allergic asthma, it is plausible that a trial involving these patient populations that included 1 or more of these drugs as comparators could have been designed. These direct comparisons would have been valuable given the unique mechanism of tezepelumab. The lack of an active-controlled trial is a limitation of this review.

The clinical expert consulted by CADTH for this review noted that adherence to therapy is a major issue in the management of asthma. It is unknown what patient adherence to therapy was like before entering the NAVIGATOR or PATHWAY study. For example, all that was known is that patients were poorly controlled on their current asthma regimens, and it is possible that poor adherence was a major contributor to this poor asthma control. It is also possible that patients, when observed more closely in a clinical trial, may have increased their adherence to their background therapy. It is not known whether this may have biased results in favour of tezepelumab or placebo, or would have affected both groups equally.

The clinical expert consulted by CADTH noted that excluding current smokers from the included trials is a generalizability issue, although it is not an uncommon practice in asthma trials. Otherwise, the demographics of the patients included in the clinical trials generally reflected those patients that would be expected to be treated with tezepelumab, in the clinical expert’s opinion.

Indirect Evidence

Objectives and Methods for the Summary of Indirect Evidence

Due to the lack of direct evidence comparing tezepelumab with other existing therapies as add-on maintenance treatment in adults and adolescents 12 years and older with severe asthma, the sponsor submitted 2 ITCs.^10,11 In addition, CADTH conducted an independent literature search for published ITCs that compared tezepelumab with other relevant comparators for the treatment of patients with in adults and adolescents 12 years and older with severe asthma. A focused literature search for ITCs dealing with tezepelumab and asthma was run in MEDLINE All (1946–) on May 9, 2022. No limits were applied. Three additional ITCs^7-9 comparing these treatments were identified. The objective of this section is to summarize and critically appraise the indirect evidence from the 2 sponsor-submitted ITCs and the 3 ITCs identified in the CADTH literature search. To align with the CADTH systematic review protocol, only information pertaining to the criteria outlined in Table 5 are presented in this section.

Description of Sponsor-Submitted Indirect Comparisons

The sponsor submitted 2 ITC reports.^10,11 One was an NMA¹⁰ and the other was an MAIC-STC.¹¹ The objectives of the ITCs were to identify, evaluate, and synthesize the evidence on the clinical efficacy and safety of tezepelumab compared with other biologics in patients with moderate-to-severe uncontrolled asthma. The MAIC-STC analyses were performed to account for key areas of heterogeneity between trials that could not be addressed by NMA methods.

Systematic Literature Review of the Indirect Treatment Comparisons

A systematic literature search strategy was developed based on the population, intervention, comparator, outcome, and study design (PICOS) criteria presented in Table 13 to identify relevant studies investigating the efficacy and safety of tezepelumab with other existing treatments. Supplemental grey literature searches were also performed for key clinical conferences in January 2021. A targeted grey literature search of the US National Institutes of Health ClinicalTrials.gov and searches of health technology assessment websites were also completed to identify additional studies of interest. Hand-searches of the bibliographies of relevant systematic reviewers, meta-analysis, and NMAs (published within the past 3 years) identified from the database search were performed to identify additional relevant studies.

Following screening and a feasibility assessment, a total of 36 studies^4,5,32,49-80 were included in the sponsor-submitted ITCs.

Network Meta-Analysis Feasibility Assessment

An NMA feasibility assessment was conducted by assessing heterogeneity across all included studies to ensure they were clinically and methodologically similar for comparison. The available study characteristics, inclusion and exclusion criteria, and potential treatment-effect modifiers were assessed. A list of potential treatment-effect modifiers is presented in Table 14. It was determined that a NMA was feasible, but MAIC and STC methods would be used to better adjust for differences between studies.

Methods of the Sponsor-Submitted Network Meta-Analysis

Based on the findings from the feasibility assessment, a full network Bayesian NMA¹⁰ was performed. Various subgroup analysis and sensitivity analyses were conducted. Methods for the NMA analysis are briefly summarized in Table 15.

Table 13: Study Selection Criteria and Methods for Indirect Treatment Comparison

AQLQ = Asthma Quality of Life Questionnaire; EMA = European Medicines Agency; ED = emergency department; FeNO = fractional exhaled nitric oxide; GINA = Global Initiative for Asthma; ICS = inhaled corticosteroids; IgE = immunoglobulin E; IL = interleukin; LABA = long-acting beta2-agonist; PICOS = population, intervention, comparator, outcome, and study design; RCT = randomized controlled trial; SGRQ = St. George's Respiratory Questionnaire.

^aThis criterion was relaxed after initial screening to permit inclusion of studies in which at least 75% of patients reported LABA use (plus at least medium-dose ICS), despite not requiring use of a LABA or other controllers as a part of their inclusion criteria.

^bLatest version of Japan prescribing information (as of March 2020).

^cSystematic reviews, meta-analyses, network meta-analyses, and bibliographies of these records were reviewed and cross-referenced with the included study lists to ensure that no primary studies were missed.

^dSearch captured all languages, but non-English citations were excluded during full-text screening.

Sources: Sponsor-submitted indirect treatment comparisons.^10,11

Table 14: Categorized List of Potential Treatment-Effect Modifiers

ACQ = Asthma Control Questionnaire; FEV₁ = forced expiratory volume in the first second; ICS = inhaled corticosteroid; OCS = oral corticosteroids.

Sources: Sponsor-submitted indirect treatment comparisons.^10,11

Table 15: Indirect Treatment Comparison Analysis Methods

AAER = annualized asthma exacerbation rate; ACQ = Asthma Control Questionnaire; FE = fixed-effect; FEV₁ = forced expiratory volume in the first second; OCS = oral corticosteroids; RE = random-effect.

Source: Sponsor’s network meta-analysis.¹⁰

Statistical Approach

Bayesian NMAs were performed for each of the 5 outcomes using ITC methodology based on the National Institute for Health and Care Excellence (NICE) Decision Support Unit (DSU) Technical Support Document.⁸¹ Model selection was based on model fit statistics, total number of studies, and studies per connection, and plausibility of the results. The model fit statistics assessed were the deviance information criterion, total residual deviance, and the random-effects standard deviation. When model fit was unclear, both fixed-effect and random-effect model results were presented.

The NMAs were performed using R statistical software. The MCMC sampling process used to construct the posterior distribution for all model parameters involved discarding a burn-in of 40,000 or more iterations before accepting a subsequent 40,000 or more iterations for NMA estimation. Three MCMC chains were run in parallel. Monte Carlo errors, trace plots, and Gelman–Rubin diagnostic criteria (ideally close to 1, with values > 1.05 indicating nonconvergence) were generated and reviewed to assess model convergence.

Primary Analyses and Summary Measures of Treatment Effect by Outcome

All NMAs were conducted using the ITT populations of relevant RCTs. Changes in the AAER and changes in the AAER leading to hospitalization were modelled using a Poisson likelihood and a log link, with corresponding rate ratios and 95% credible intervals (CrIs) calculated. For changes from baseline in the Asthma Control Questionnaire (ACQ) score and pre-bronchodilator FEV₁, normal likelihood functions were used. The relative treatment effect was defined as the mean difference (95% CrI) in change from baseline between treatments. For change from baseline in OCS dose by predefined reduction categories, treatment effects were modelled using an ordinal model (probit scale) considering each category reported in the included RCTs (no reduction or increase in use, 1% to 49% reduction, 50% to 74% reduction, 75% to 89% reduction, or 90% to 100% reduction). The proportion of patients in each of the predefined reduction categories for each treatment were calculated. Additionally, analyses were conducted for each predefined reduction category, where risk ratios and 95% CrIs were generated. All RCTs included in this analysis assessed OCS reduction between week 20 and week 28. However, OCS reduction in the SOURCE study⁵ was assessed at week 48.

Assessment of Heterogeneity and Inconsistency

The available study characteristics, inclusion and exclusion criteria, potential treatment-effect modifiers, and outcome definitions were assessed to ensure similarity. Additionally, clinical experts were consulted to further investigate the potential impact of heterogeneity on effect estimates. Because no direct evidence connects any of the relevant biologics, consistency could not be assessed.

Subgroup and Sensitivity analysis

Clinically relevant subpopulation and/or subgroup analysis (e.g., baseline eosinophil count, number of exacerbations at baseline) were performed.

Sensitivity NMA analyses were also conducted by removing some studies with different characteristics (such as those studies that used best standard care as a comparator and not a true placebo group).

Results of Sponsor-Submitted Network Meta-Analysis

Summary of Included Studies

A total of 36 trials^4,5,32,49-80 were included in the NMA. An overview of the included trials is presented in Table 26. Overall, 3 trials for tezepelumab, 6 for benralizumab, 3 for dupilumab, 3 for mepolizumab, 16 for omalizumab, and 5 for reslizumab were included. For the purpose of this review, NMA results related to reslizumab are not summarized because reslizumab was not identified as a relevant comparator in the protocol for the CADTH systematic review.

Assessment of Risk of Bias of Included Trials

It was reported that the methodological quality of all included studies was assessed using the NICE quality appraisal checklist of quantitative intervention studies.⁴⁸ As the technical report for the NMA states that the assessed trials were considered to have adequate internal validity and generalizability to the target population, sensitivity analyses based on the results of the quality assessment of individual studies were not conducted.⁸²

Network Meta-Analysis Results

The key findings of the sponsor-submitted NMA are presented in Table 16 and Table 17.

Reduction in Annualized Asthma Exacerbation Rate

Primary Analyses

A visual representation of the evidence network for reduction in the AAER is provided Figure 5. The common comparator to all trials is placebo, with a single multi-dose study creating 1 closed loop.

Figure 5: Evidence Network for Annualized Asthma Exacerbation Rate

AAER = annualized asthma exacerbation rate; B = benralizumab; BSC = best standard care; D = dupilumab; M = mepolizumab; O = omalizumab; OAT = optimized asthma therapy; PBO = placebo; Q2W = every 2 weeks; Q4W = every 4 weeks; R = reslizumab; t = tezepelumab.

Source: Sponsor-submitted Indirect treatment comparison.¹⁰

Sixteen relevant studies reporting AAER were included, resulting in a network of 10,092 patients. Based on the number of studies included and the model fit statistics, the random-effects model was considered the best fit. In terms of the AAER, tezepelumab was not favoured over other biologics (Table 16).

Subgroup Analyses

For patients with an eosinophil count of 150 cells/μL or higher, tezepelumab showed a lower AAER compared with omalizumab (relative risk [RR] = 0.63; 95% Crl, 0.43 to 0.94) and benralizumab (RR = 0.63; 95%Crl, 0.49 to 0.82). There appeared to be no differences between tezepelumab when compared with mepolizumab and dupilumab.

For patients with an eosinophil count of 300 cells/μL or higher and lower than 300 cells/µL, 3 or more exacerbations in past 12 months, and allergic asthma, tezepelumab was not favoured over the other biologics.

For eosinophil counts below 150 cells/μL, tezepelumab had a lower AAER compared with dupilumab 300 mg (RR = 0.48; 95% Crl, 0.28 to 0.84).

Sensitivity Analyses

Overall, the results of the sensitivity analyses were aligned with the primary analyses of reduction in AAER.

Reduction in AAER Leading to Hospitalizations

Primary Analyses

A visual representation of the evidence network for reduction of AAER leading to hospitalizations is provided in Figure 6.

Figure 6: Evidence Network for AAERs Leading to Hospitalizations

AAER = annualized asthma exacerbation rate; B = benralizumab; BSC = best standard care; D = dupilumab; M = mepolizumab; O = omalizumab; OAT = optimized asthma therapy; PBO = placebo; Q2W = every 2 weeks; Q4W = every 4 weeks; R = reslizumab; t = tezepelumab.

Source: Sponsor-submitted indirect treatment comparison.¹⁰

Eleven relevant studies reporting AAER leading to hospitalizations were included, which resulted in a network of 6,965 patients. Unlike AAER, only pooled data were available for the 2 dosages (200 mg and 300 mg every 2 weeks) of dupilumab. All trials except for 1, the EXALT (2011) study for omalizumab, reported the rate of exacerbations resulting in hospitalization or emergency room/department visit. The EXALT (2011) study reported the rate of exacerbations resulting in hospitalization. Based on the number of studies included and the adequate model fit statistics, the random-effects model was considered the best fit. The NMA found that tezepelumab was not favoured over other biologics for AAER leading to hospitalizations (Table 16).

Subgroup Analysis

No subgroup analyses were conducted for AAER leading to hospitalizations due to the lack of available data.

Sensitivity Analysis

The results of the sensitivity analysis for AAER leading to hospitalizations were consistent with the primary analysis.

Change From Baseline in Asthma Control Questionnaire Score

Primary Analysis

A visual representation of the evidence network for change in baseline ACQ score is provided in Figure 7.

Figure 7: Evidence Network for Change From Baseline in ACQ Score

ACQ = Asthma Control Questionnaire; B = benralizumab; BSC = best standard care; D = dupilumab; M = mepolizumab; O = omalizumab; PBO = placebo; Q2W = every 2 weeks; Q4W = every 4 weeks; R = reslizumab; t = tezepelumab.

Source: Sponsor-submitted Indirect treatment comparison.¹⁰

A total of 19 relevant studies reporting ACQ score were included, resulting in a network of 8,791 patients. Based on the number of studies included and the adequate model fit statistics, the random-effects model was considered the best fit. The NMA indicated that tezepelumab was not favoured over other biologics for the change in ACQ (Table 17).

Subgroup Analysis

For patients with an eosinophil count of 150 cells/μL of higher, compared with benralizumab, tezepelumab demonstrated increased improvement in ACQ score (mean difference = −0.23; 95% CI, −0.4 to −0.04). Tezepelumab was not favoured over other biologics in any other subgroup analysis.

Sensitivity Analysis

The results of the sensitivity analysis for ACQ score were consistent with the primary analysis.

Change From Baseline in Pre-Bronchodilator FEV₁

Primary Analysis

A visual representation of the evidence network for change from baseline in pre-bronchodilator FEV₁ is provided in Figure 8.

Figure 8: Evidence Network for Change From Baseline in Pre-Bronchodilator FEV₁

B = benralizumab; D = dupilumab; FEV₁ = forced expiratory volume in 1 second; M = mepolizumab; O = omalizumab; PBO = placebo; Q2W = every 2 weeks; Q4W = every 4 weeks; Q8W = every 8 weeks; R = reslizumab; t = tezepelumab.

Source: Sponsor-submitted indirect treatment comparison.¹⁰

A total of 19 relevant studies reporting pre-bronchodilator FEV₁ were included, which resulted in a network consisting of 9,046 patients. Based on the number of studies included and the adequate model fit statistics, the random-effects model was considered the best fit. The NMA indicated that, in terms of FEV₁, tezepelumab was not favoured over other biologics studied (Table 17).

Subgroup Analysis

In terms of FEV₁, the results of all subgroup analyses were consistent with the primary analysis, meaning tezepelumab was not favoured over other biologics.

Sensitivity Analysis

In terms of FEV₁, the results of all sensitivity analyses were consistent with the primary analysis, meaning tezepelumab was not favoured over other biologics.

Change From Baseline in the OCS Dose Reduction of 50% or Greater

Primary Analysis

A visual representation of the evidence network for change from baseline in the OCS dose by predefined reduction categories is provided in Figure 9.

Figure 9: Evidence Network for Change From Baseline in the OCS Dose by Predefined Reduction Categories

B = benralizumab; D = dupilumab; M = mepolizumab; PBO = placebo; Q2W = every 2 weeks; Q4W = every 4 weeks; Q8W = every 8 weeks; t = tezepelumab.

Source: Sponsor-submitted indirect treatment comparison.¹⁰

Oral Corticosteroid Dose Reduction of 50% or Greater

Based on the number of single study connections in the network and the model fit statistics, the fixed-effects model was considered the best fit. The NMA indicated that tezepelumab was favoured to dupilumab 300 mg (OR = 0.36; 95% Crl, 0.14 to 0.93) for the outcome of OCS dose reduction of 50% or greater. No differences between tezepelumab and other biologics were identified (Table 17).

Subgroup Analysis

For patients with an eosinophil count of 300 cells/μL or greater, compared with mepolizumab, tezepelumab was favoured for the outcome of an OCS reduction of 50% or greater (OR = 11.07; 95% Crl, 1.17 to 312.70). Tezepelumab was not favoured over other biologics in this subgroup. Tezepelumab was not favoured over other biologics for the remaining subgroup analyses.

Sensitivity Analysis

Data of the sensitivity analysis for an OCS reduction of 50% or greater were consistent with that observed in primary analysis.⁸²

Description of MAIC-STC

The MAIC-STC approach used individual patient-level data (IPD) from the tezepelumab trials, and weighted the trial population to match average baseline characteristics reported for the comparator trials. The IPD from the NAVIGATOR⁴ and SOURCE⁵ studies was used for tezepelumab for all MAIC and STC analyses. For comparators, the most relevant study or studies were selected based on a comparison of the study characteristics (e.g., follow-up duration), inclusion and exclusion criteria, availability of key potential treatment-effect modifiers, and reporting of outcomes. For some comparators, multiple studies were pooled as no clear rationale was identified for selecting 1 study over the other. Potential treatment-effect modifiers were identified based on use in previous published ITCs, clinical relevance, and input from the external experts.

Methods of Sponsor-Submitted MAIC-STC

MAIC-STC Analysis Methods

Anchored MAIC and STC analyses were performed using methods outlined by the NICE DSU Technical Support Document.⁸¹ The IPD for tezepelumab from the NAVIGATOR and SOURCE studies were used in the analyses, and aggregate data from relevant published trials were used for the comparators.

MAIC Analysis

Matching patients (i.e., IPD to match the inclusion and exclusion criteria of the comparator trial): First, the inclusion and exclusion criteria were aligned between the 2 studies being compared. Patients from the NAVIGATOR⁴ and SOURCE⁵ studies were removed from the IPD if they did not satisfy the eligibility criteria used in the relevant comparator RCT. The estimated treatment effect therefore represents that for tezepelumab in the populations studied in the comparator trials. Bucher ITCs, matched but unadjusted for any differences in potential treatment effect modifiers, were conducted and presented alongside MAIC results.

Adjustment for differences in treatment effect modifiers: After completing the matching phase of the anchored MAIC, patients remaining in the NAVIGATOR study IPD were re-weighted using the ranked treatment-effect modifiers (Table 14 provides a pre-specified list) that were available from RCTs of both treatments so that both their means and SDs matched those reported in the comparator trial. Inverse propensity score weighting, in which patients in the tezepelumab trial were weighted by the inverse odds of being in that trial versus the comparator trial, was used. The propensity score model was estimated using the generalized method of moments based on the aggregate data and IPD. Matching and adjusting were successful if the relevant treatment effect modifiers were similar between the treatment arms of the studies. Treatment outcomes were then compared between balanced trial populations.

Estimating indirect treatment effects: After matching patients from relevant comparator studies, a weighted estimate of the outcome was derived using anchored MAIC adjustment weights. The anchored MAIC was first performed using all of the ranked treatment-effect modifiers, data permitting. That is, the effect sample size decreases as the number of matching variables increases. Statistical significance testing was defined using a 2-tailed P value of less than 0.05, and all comparisons between groups were reported with point estimates (e.g., ratio ratios) and 95% CIs.

Primary analyses and summary measures of treatment effect by outcome: Exacerbation rates were modelled using a negative binomial generalized linear model (GLM) with anchored MAIC adjustment weights, using the number of person-years at risk. Rate ratios and 95% CIs were calculated. For changes from baseline in ACQ score and pre-bronchodilator FEV₁, estimates were derived by fitting a GLM with anchored MAIC adjustment weights. The treatment effect between tezepelumab and the comparator was then derived as the difference between the predicted mean from the adjusted NAVIGATOR trial data and the estimated mean (based on summary-level data) from the comparator study. For the ACQ and FEV₁, the relative treatment effect was defined as the mean difference in change from baseline between treatments. For change from baseline in OCS dose, a binomial GLM with anchored MAIC adjustment weights was used, and ORs with 95% CIs were calculated.

Performance assessment: The performance and suitability of each anchored MAIC model were assessed based on the following criteria: the effective sample size, distribution, and characteristics (i.e., zero or extreme values) of patient weights, and balance between patient populations on potential treatment-effect modifiers. All MAIC scenarios were conducted first with all potential treatment-effect modifiers, and then modifiers were subsequently dropped 1-by-one, starting with those ranked last, until a single modifier remained.

STC Analysis

Primary analyses and summary measures of treatment-effect by outcome: The AAER and AAER leading to hospitalizations were modelled using negative binomial regression, where the number of person-years at risk was used. Rate ratios and 95% CIs were calculated. For change from baseline in ACQ score and pre-bronchodilator FEV₁, linear regression was used. The relative treatment effect was defined as the mean difference in change from baseline between treatments. For reduction in the OCS dose by 50% or greater, logistic regression was used. Odds ratios and 95% CIs were calculated.

General Model Selection Considerations

Regression models were fit using IPD from the NAVIGATOR and SOURCE studies, incorporating interaction terms between treatment and clinically relevant treatment-effect modifiers. The statistical performance of each regression model was assessed based on the criteria of model convergence and model fit statistics (e.g., Akaike information criterion). When models failed to converge, variables were either recategorized (i.e., by collapsing categorical variables into broader categories) or removed in order of least importance.

Results from MAIC-STC

Eleven studies^{4,32,49-51,55-60,67,68,77,78} were included in the MAICs and STCs.

A summary of MAIC and STC results from the primary analyses is presented in Table 16 and Table 17 (results are shown for the scenario that adjusted for the maximum number of variables, scenario A). Overall, MAIC-STC results for all of the outcomes evaluated did not demonstrate that tezepelumab was favoured against comparator treatments.

Critical Appraisal of the Sponsor-Submitted Network Meta-Analysis and MAIC-STC

Overall, the ITCs were conducted according to accepted methodological guidance. A systematic literature review was used as the basis for each ITC method. A pre-specified analysis plan for conducting NMAs and MAICs-STCs was used to guide the analyses. This included identification of potential sources of heterogeneity between trials and key clinical subpopulations that were then analyzed to further validate findings from the primary results. The ITCs were performed according to well-established methods outlined by the NICE DSU technical support documents.^81,83 Geometry of the evidence networks was provided for each outcome analysis. Distribution and characteristics (i.e., zero or extreme values) of patient weights were assessed in each MAIC-STC.

The key limitation of the NMA was the heterogeneity across the included studies in terms of the eligibility criteria, study designs, patient characteristics, comparators, and outcome assessment. Several trials, including the tezepelumab trials, did not select patients based on peripheral eosinophil levels. As a result, patients were included in the analyses with wide ranges of baseline eosinophil counts, and the analyses likely included patients who would not actually be eligible for mepolizumab or benralizumab, based on their respective trial inclusion criteria and approved indications. Some trials included adult patients only (≥ 18 years) and some trials included those aged ≥ 12 years, although patients younger than 18 years accounted for a small percentage of the populations. As well, some of the included studies enrolled patients with moderate-to-severe uncontrolled asthma. The proposed tezepelumab indication and requested reimbursement criteria are for adults and adolescents 12 years and older with severe asthma. No subgroup analysis was conducted for severe asthma only. Doses of ICE also varied. For example, mepolizumab studies used higher ICS doses, whereas ICS doses for the benralizumab studies were medium or high. There were some variations in terms of the outcome evaluation time points across trials and differences in the study design (e.g., randomization versus partial randomization, blind versus open label, and parallel versus cross over). Differences were also observed for potential treatment-effect modifiers, exacerbations in the past 12 months, and number of OCS users in the included studies. The differences in these potential treatment-effect modifiers were described as clinically important by the clinical expert consulted by CADTH and may affect the validity of NMA results. Indeed, the technical report for the NMA pre-specified numerous potential sources of heterogeneity and effect modifiers, and acknowledged that key assumptions, including homogeneity, may have been violated given the differences between the included studies. No meta-regression analyses to adjust for factors that may bias comparisons were conducted, but random-effects models were used where possible and the impact of the various sources of heterogeneity was explored using subgroup and sensitivity analyses where data were available.

Appropriate MAIC and STC methods were used to try to account for the heterogeneity. All analyses were anchored through placebo, except for some outcomes comparing tezepelumab to omalizumab, in which an assumption was made that the best supportive care arm in the EXALT trial of omalizumab and the placebo arm of the NAVIGATOR trial could be considered a common comparator. It is difficult to determine to what extent this may have affected comparative estimates, particularly when considering that many (perhaps all) patients in other trials who received placebo would have been receiving standard-of-care treatment (i.e., ICS with or without a LABA) in the background.

The ITC technical report outlined the pre-specified steps involved in identifying and clinically validating the important effect modifiers and matching factors for the ITCs. The clinical expert consulted by CADTH agreed that the list of potential effect modifiers included most of the key factors to consider. Given the availability of only aggregate data for the comparator trials, limited or a lack of reporting on certain covariates, and narrow criteria used in certain trials, many of the identified effect modifiers could not be included in the models, likely leading to residual heterogeneity between populations. The MAIC method reduces sample sizes when the matching process includes many variables, as was the case in the provided MAICs, which further compromised the already constrained precision of estimates on comparative efficacy. Reductions in effective sample size (i.e., > 90%) were observed in some MAIC analyses. In many instances randomization was not preserved and could result in biased estimates of the treatment effect. Differences in trial design, such as timing of outcome evaluation and duration of follow-up, cannot be accounted for using either method. These factors likely led to the observed similarity of the results between the NMA and MAIC or STC, unrealistic treatment-effect differences for some comparisons and outcomes, and the generally wide CIs that indicated limited precision and lack of confidence in the validity of the results. Other limitations included the exclusion of non-English publications. As there were no direct comparisons between treatments, the assessment of inconsistence was not applicable. Furthermore, safety outcomes and quality of life outcomes (such as the AQLQ[S]12+ questionnaire) were not assessed in the sponsor-submitted ITCs.

Description of Indirect Treatment Comparisons identified by a CADTH Literature Search

A total of 116 citations were identified in the literature search. Following screening of titles and abstracts, 113 citations were excluded and 3 potentially relevant reports from the electronic search were retrieved for full-text review. All of the 3 ITCs^7-9 were considered relevant to this review.

All 3 published ITCs used Bayesian NMA methods to compare the available biologics, including tezepelumab, for moderate-to-severe asthma. The results of NMAs were similar to the sponsor-provided NMA, indicating that tezepelumab was not clearly favoured over the other biologics for any of the outcomes evaluated. The NMAs had the same limitations as the sponsor-provided analysis, primarily the high degree of heterogeneity across the studies.

Other Relevant Evidence

This section includes submitted LTE studies and additional relevant studies included in the sponsor’s submission to CADTH that were considered to address important gaps in the evidence included in the systematic review.

Long-Term Extension Studies

One LTE study, DESTINATION,¹² provided evidence of the long-term safety and tolerability of tezepelumab 210 mg administered every 4 weeks subcutaneously in adults and adolescents with severe, uncontrolled asthma for up to 2 continuous years, including 1 year of treatment in the predecessor NAVIGATOR⁴ and SOURCE⁸⁴ parent studies. Data presented are based on a clinical database lock date of December 9, 2021.

Methods

The DESTINATION trial is a phase III, multi-centre, double-blind, randomized, placebo-controlled, parallel-group LTE study for patients who completed the NAVIGATOR⁴ or SOURCE study.⁸⁴ The primary objective of the DESTINATION trial was to evaluate the long-term safety and tolerability of tezepelumab over 140 weeks, inclusive of the treatment period of either parent study. Adults (18 to 80 years old) and adolescents (12 to 17 years old) who had continued to receive the investigational product and attended the end-of-treatment visit in 1 of the parent studies were eligible for enrolment. A total of 951 patients were enrolled and randomized at 182 centres in 18 countries into the DESTINATION study: 827 patients from the NAVIGATOR study and 124 from the SOURCE study. The study design and sequence of treatment periods are illustrated in Figure 10.

Patients previously randomized to 210 mg tezepelumab in either parent study were assigned to remain on 210 mg tezepelumab administered every 4 weeks subcutaneously in the DESTINATION study (tezepelumab plus tezepelumab group; n = 415 from NAVIGATOR, n = 60 from SOURCE). Patients previously randomized to the placebo arm in the parent studies were re-randomized at a 1:1 ratio to either tezepelumab 210 mg (placebo plus tezepelumab group; n = 206 from NAVIGATOR, n = 32 from SOURCE) or placebo (placebo plus placebo group; n = 205 from NAVIGATOR; n = 32 from SOURCE) administered every 4 weeks subcutaneously. Randomization was stratified by parent study. Patients recruited from the SOURCE study were followed up post-treatment for 12 weeks. Patients who were enrolled from the NAVIGATOR study who completed 100 weeks of tezepelumab treatment were eligible for either 12 weeks of follow-up or a 36-week extended follow-up. This review of the DESTINATION study focused on the results from the tezepelumab plus tezepelumab and placebo plus placebo groups.

The study design of DESTINATION is depicted in Figure 10.

Figure 10: DESTINATION Study Design

EOT = end of treatment; FU = follow-up; LTE = long-term extension; R = randomization; V = visit.

Source: Clinical Study Report for DESTINATION.¹²

The screening and/or randomization visit of the DESTINATION study was the same as the end-of-treatment visit from the parent NAVIGATOR (week 52) or SOURCE (week 48) studies.

Populations

Baseline patient and disease characteristics were as determined at baseline at time of enrolment in the parent study. Among patients who entered the DESTINATION study from the NAVIGATOR study, the mean age was 49 years (range = 12 to 80 years), and the majority were White (67%) and female (63%), and had baseline eosinophil counts of fewer than 300 cells/μL (58.4%) and FeNO levels below 25 ppb. Of patients who entered the LTE from the SOURCE study, the average age was 53.4 years (range = 22 to 76), and the majority were White (84%) and female (62.7%), with baseline eosinophil counts below 300 cells/μL (65.3%) and FeNO levels below 25 ppb. Overall, patients’ baseline characteristics were similar between the parent and LTE studies. The initial imbalances between groups in mean baseline eosinophil counts within the parent studies were no longer notable at LTE baseline.

Interventions

Investigational Products

The treatment protocol employed in the DESTINATION trial was similar to those described in the parent studies

The first dose of the investigational product was administered the same day as the end-of-treatment visit at week 52 and week 48 from the parent NAVIGATOR and SOURCE studies, respectively. The treatment-period duration was 52 weeks for patients who previously completed the NAVIGATOR study and 56 weeks for those who previously completed the SOURCE study. The last dose of the investigational product was administered at week 100 and the end-of-treatment visit was conducted at week 104. No investigational products were administered at week 104 or during the 12-week safety follow-up period.

Background Asthma Medication

In DESTINATION, tezepelumab was administered as an add-on maintenance therapy to ICS treatment plus at least 1 additional asthma-controller mediation with or without OCS, as described in the NAVIGATOR and SOURCE studies. Patients who were maintained on OCS, LABA, LAMA, theophylline, or LTRA in addition to ICS were allowed to continue treatment with these medications through the LTE. Starting from visit 1, investigators were also encouraged to consider stepping down the background medications when asthma symptoms were well controlled with stable lung function for at least 3 months in line with Global Initiative for Asthma guidelines.¹

Concomitant Therapy

Investigators were permitted to prescribe concomitant medications or treatment to provide adequate supportive care as necessary, except for those medications identified as excluded in the prohibited medication list previously described for the NAVIGATOR and SOURCE studies.

Patients were permitted to use inhaled short-acting bronchodilators as needed as reliever or rescue medication due to worsening asthma. Any medications or vaccines, including over-the-counter or prescription medicines, vitamins and/or herbal supplements that the patients received at the time of enrolment or during the study were documented, along with the reason for use and dates of administration.

Outcomes

Primary Outcome

The primary outcome in DESTINATION was the exposure-adjusted incidence rates of AEs and SAEs over 104 weeks. For each treatment group, exposure-adjusted incidence rates were defined as the number of patients reporting AEs divided by the total exposure duration for that treatment group. For individual patients, exposure duration (days) was determined using the start and end dates of the applicable analysis period.

Secondary Outcome

The secondary objective of the trial was to assess the long-term effects of tezepelumab on asthma exacerbations which was measured as the AAER over 104 weeks where baseline was week 0 in the parent study.

Exploratory Outcomes

Exploratory outcomes included the long-term effects of tezepelumab on pulmonary function, asthma control, health status of patients with airway obstructive disease, lung function, other end points associated with asthma exacerbations; maintaining asthma control; and maintaining OCS dose at 5 mg or less.

Harms

All AEs and SAEs (new and ongoing from the parent study) were collected from the time patients signed the informed consent to the end of the treatment and follow-up periods. In addition, the following AEs of special interest were documented: anaphylactic reactions, immune complex disease (type III hypersensitivity reactions), hypersensitivity reactions, malignancy, helminth infections, severe infections, injection-site reactions, opportunistic infections, Guillain-Barré syndrome, and adrenal crisis.

Statistical Analysis

Sample-Size Determination

The sample size for the DESTINATION study was not based on statistical consideration but determined by the number of patients who completed the double-blind treatment periods in either of the parent studies and met all eligibility to enter the LTE study.

Statistical Methods

No formal hypothesis-testing was conducted in the DESTINATION study. Analyses of the primary (safety) and secondary (efficacy) end points and the exploratory end points were performed on data from the primary database lock (December 9, 2021), which was conducted after the last patient completed week 104 (October 26, 2021). Presentation of all results were stratified by parent study. Data were not aggregated across the 2 parent studies.

Analysis Population

Safety analyses were performed using the safety analysis set consisting of all patients who were randomized and received at least 1 dose of the investigational product in either of the parent studies, regardless of protocol adherence and continued participation in either the studies, and regardless of enrolment in the LTE study.

Efficacy analyses were performed using the FAS consisting of all patients who were randomized and received at least 1 dose of the investigational product in either parent study, irrespective of their protocol adherence and continued participation in either the studies or their enrolment in the LTE study. Efficacy analyses were also performed using the FAS-LTE set consisting of patients who were randomized and received at least 1 dose of the investigational product in the DESTINATION study. Patients were assigned to their randomized treatment according to ITT principles.

Patient Disposition

Patient disposition is summarized in Table 18.

Patient rollover into the DESTINATION study was 94.6% and 90.5% of patients who completed the NAVIGATOR and SOURCE studies, respectively. Of the 951 adults and adolescents from the parent studies enrolled in the DESTINATION study, 827 patients (including 72 adolescents) were from the NAVIGATOR study and 124 were from the SOURCE study. A total of 950 patients received treatment in the DESTINATION study. Overall, patient disposition was similar across treatment groups among patients with NAVIGATOR as the parent study. For patients with SOURCE as the parent study, the overall patient disposition in the LTE was similar across the tezepelumab plus tezepelumab, and placebo plus tezepelumab groups with slightly fewer patients who completed both treatment and study for the placebo plus placebo groups.

Exposure to Study Treatments

Extent of Exposure

Durations of exposure to tezepelumab by parent study are summarized in Table 19. The overall total exposure to tezepelumab was 1,283.27 patient-years, consisting of 575.62 patient-years from treatment of 602 patients in the parent study period and 708.94 patient-years from treatment of 712 patients in the LTE study period.

Table 18: Patient Disposition for DESTINATION (LTE Analysis Set)

LTE = long-term extension; Pbo + Pbo = placebo parent plus placebo LTE; Pbo + Tez = placebo parent plus tezepelumab LTE; Tez + Tez tezepelumab parent plus tezepelumab LTE.

^aIncluded patients who completed treatment and the LTE period, and patients who discontinued treatment but completed study assessments during LTE period.

Source: Clinical Study Report for DESTINATION.¹²

Table 19: Extent of Exposure for DESTINATION (LTE Analysis Set)

LTE = long-term extension; Pbo + Pbo = placebo parent plus placebo LTE; Pbo + Tez = placebo parent plus tezepelumab LTE; SD = standard deviation; Tez + Tez = tezepelumab parent plus tezepelumab LTE

Source: Clinical Study Report for DESTINATION.¹²

Treatment Adherence

Treatment adherence to the investigational product was consistently high (> 97%) during the parent and LTE study periods.

Efficacy

Asthma Exacerbations

Asthma exacerbations over 104 weeks are summarized in Table 20.

Among patients enrolled in the LTE from the NAVIGATOR study, tezepelumab plus tezepelumab resulted in a reduction in the rate of asthma exacerbation compared to placebo plus placebo (AAER = 0.50; 95% CI, 0.40 to 0.63). Similarly, treatment with tezepelumab plus tezepelumab reduced the rate of asthma exacerbations associated with hospitalization or ED visits compared with placebo plus placebo (AAER = 0.39; 95% CI, 0.22 to 0.69).

In patients enrolled in the LTE from the SOURCE study, the AAER for asthma exacerbations between tezepelumab plus tezepelumab and placebo plus placebo was 0.66 (95% CI, 0.37 to 1.19). For asthma exacerbations associated with hospitalization or ED visits, the AAER for tezepelumab plus tezepelumab versus placebo plus placebo was 0.27 (95% CI, 0.05 to 1.63).

Asthma Control

The ACQ-6 change scores are summarized in Table 21.

Table 20: Asthma Exacerbation for DESTINATION (FAS-LTE)