Dupilumab for treating severe asthma with type 2 inflammation

Standard care is the appropriate comparator in the updated population

3.7 The clinical trial populations included people with differing severity of asthma (defined by eosinophil level and the number of exacerbations in the previous year). These populations therefore included people who would be offered different treatment options in the NHS:

Clinical evidence

The evidence on clinical effectiveness is relevant to NHS clinical practice

3.8 The company's clinical evidence came from 3 randomised-controlled trials, DRI12544, QUEST and VENTURE. These compared dupilumab with placebo in people aged 12 years and over (except DRI12544, which only included people aged 18 years or over) with persistent asthma who had 1 or more exacerbations in the previous year. None of the trials had restrictions on blood eosinophils or FeNO. DRI12544 and QUEST included people with moderate-to-severe asthma not on maintenance oral corticosteroids. VENTURE included people with severe corticosteroid-dependent asthma (on maintenance corticosteroids). The 3 trials were conducted globally, and QUEST was the only trial that included people from the UK. The trial populations were based on use of moderate-to-high doses of inhaled corticosteroids. This was because they included people from countries like the US and Japan, where the clinical expert stated that there is reluctance to use high-dose inhaled corticosteroids. The committee concluded that there were some caveats, but that all 3 trials included were relevant to NHS clinical practice.

Dupilumab is more clinically effective than standard care in the clinical trial populations

3.9 All primary outcomes were reported for the intention-to-treat population in all 3 trials. In QUEST, the first coprimary outcome was annualised rate of severe exacerbations. There was a 47.7% (95% confidence interval [CI] 33.8% to 58.7%, p<0.0001) lower rate of severe exacerbations in the dupilumab group compared with placebo. Change from baseline in the forced expiratory volume in 1 second (FEV1) at 12 weeks was the second coprimary outcomes in QUEST and the primary outcome in DRI12544. There was an increase in FEV1 at 12 weeks when dupilumab was compared with placebo in DRI12544 (least squares [LS] mean difference 0.14 litre, 95% CI 0.08 to 0.19, p<0.0001) and QUEST (LS mean difference 0.20 litre, 95% CI 0.11 to 0.28, p<0.0001). In VENTURE, the primary outcome was the percentage reduction in oral corticosteroid dose from baseline. There was a greater reduction in oral corticosteroid use with dupilumab compared with placebo (LS mean difference 28 mg, 95% CI 16 to 41, p<0.0001) at 24 weeks. The proportion of people with treatment-related adverse events was similar within each trial between those having dupilumab and placebo. In DRI12544 and QUEST, the proportion of people with any treatment-related adverse events ranged from 74.7% to 84.1%. In VENTURE, a smaller proportion experienced any treatment-related adverse events (64.5% and 62.1% in the placebo and dupilumab arms respectively). The committee concluded that dupilumab was more clinically effective than standard care in the clinical trial populations and is a relatively safe treatment.

The clinical effectiveness estimates for dupilumab are uncertain in the company's updated population

3.10 The company's decision-problem subgroup analyses at the first appraisal committee meeting focused on the annualised rate of severe exacerbations for the posthoc population (that is, people with a blood eosinophil count of 150 cells per microlitre or more, FeNO of 25 parts per billion or more and 3 or more exacerbations in the previous year) from QUEST and VENTURE. Dupilumab reduced the rate of severe exacerbations when compared with placebo within this subpopulation in QUEST and VENTURE, although in small posthoc subgroups with 101 and 152 people respectively. There were improvements in the placebo groups for the primary outcomes of these trials. This was possibly because of regression to the mean and the placebo effect. The committee concluded that dupilumab is clinically effective and safe as an addition to standard care in people with a blood eosinophil count of at least 150 cells per microlitre or FeNO of 25 parts per billion or more and 3 or more exacerbations in the previous year who may or may not be taking maintenance oral corticosteroids. However, there were very limited evidence clinical efficacy data provided for the company's updated population because of small number of patients in the QUEST trial corresponding to each subgroup. The committee concluded that the clinical effectiveness of dupilumab in the company's updated base case was highly uncertain.

The model structure is appropriate for decision making

3.11 The company submitted a 4‑state Markov model comparing dupilumab with standard care in people with severe asthma and type 2 inflammation. The model consisted of 4 live health states: uncontrolled asthma; controlled asthma; moderate exacerbation; and severe exacerbation. In addition, the model included states for asthma-related deaths and death from other causes. Response to treatment was defined as a 50% or greater reduction in the annual exacerbation rate, which was assessed at 52 weeks. People whose asthma responded continued on dupilumab and those whose did not transferred to standard care. The company derived the efficacy and clinical parameters in the model from the QUEST clinical trial. The committee concluded that the model structure was appropriate for decision making.

The evidence for the company's updated population is limited and effectiveness estimates are based on assumptions

3.12 The committee noted that the company split the updated population into 3 subgroups: the adolescents who are aged 12 to 17, the adults who are not eligible for biologicals (blood eosinophil count 150 to 299) and adults who previously received biologicals but did not respond (blood eosinophil count of 300 and more). The committee noted that the clinical effectiveness evidence available for each of the subpopulations was limited because the number of patients included in the QUEST trial corresponding to each subgroup was small. The trial included 2 patients corresponding to the adolescent subgroup and 14 patients corresponding to the subgroup of adults who are not eligible to biologicals. The QUEST protocol excluded patients who had been on biologicals but 1 patient in the trial was included who had previously received a biological. The ERG noted that the estimates of transition probabilities for the company's updated population were highly uncertain, due to small sample sizes. The company assumed that clinical effectiveness was the same for each subgroup based on trial estimates for the overall company's updated population. The committee particularly considered the assumption of equal efficacy of dupilumab regardless of whether people had received prior biological therapy. The company provided clinical expert opinion that switching from other biologicals (interleukin-5 inhibitors: mepolizumab, reslizumab, benralizumab) to dupilumab (interleukin-4/13 inhibitor) was acceptable because the mechanisms of action were different enough. The committee was concerned about this assumption and considered that although it is plausible that people who did not respond to other biologicals could respond to dupilumab, assuming that the response rate would be as good as in people not eligible for other biologicals was optimistic. The committee considered this to be a key area of uncertainty and noted that it would have liked to see exploration of scenarios with a range of alternative response rates for the group of adults who did not respond to biological therapy. The committee concluded that the effectiveness estimates in the company's updated population were highly uncertain.

The company's updated base case does not include a multiplier for long-term severe exacerbation rates

3.13 The committee noted that asthma-related mortality often drives cost effectiveness in asthma models. The annual severe exacerbation rate (2.39 exacerbations per year) in the placebo arm of the QUEST trial was lower than observed in clinical practice in the year before trial enrolment (4.46 exacerbations per year).The company's model after technical engagement used the exacerbation rates from QUEST and VENTURE in the first year of the model and increased the number of severe exacerbations in subsequent years for both dupilumab and standard care by applying a multiplier. The ERG considered the trial to be the best source of exacerbation data and did not include an exacerbation multiplier in its base case model which resulted in higher incremental cost effectiveness ratios (ICERs). The company provided evidence on severe exacerbation rates from 3 severe asthma cohorts: WATCH (Wessex Asthma Cohort of Difficult Asthma), U-BIOPRED (Unbiased BIOmarkers in PREDiction of respiratory disease outcomes) and the Sanofi Real World Evidence (RWE) study. It also accepted the committee and ERG's concerns about the uncertainty of using a multiplier. The exacerbation rates in the company's updated base case model were taken from the QUEST trial for the duration of the model without an exacerbation multiplier. The committee concluded that the updated base case without the exacerbation multiplier was appropriate.

Real world evidence is best source of data to inform the setting of treating severe exacerbations

3.14 The company assigned different mortality rates to severe exacerbations treated in hospital emergency care, inpatients and general practice. In the QUEST trial 6.7% of severe exacerbations were treated in hospital (3.0% in emergency care, 3.7% in inpatients and 93.3% in general practice). The company originally based resource use associated with severe exacerbations on the UK Difficult Asthma Registry registry data (O'Neill 2015) with 26.5% (7.8% in emergency care, 18.7% in inpatients) and 74.0% in general practice) as a better estimate or resource use in the NHS. The ERG base-case model used the QUEST data for the setting of severe exacerbations. The clinical expert explained that the number of patients treated in hospital in clinical practice is likely to be higher than that seen in the trial because patients in trials are well monitored on optimised treatment, more motivated and have better adherence to treatment. The committee requested further exploration of different sources of data for the setting of treating exacerbations, to inform the model. The company submitted data on the setting of treating severe exacerbation rates from 3 different sources (WATCH, U-BIOPRED and the Sanofi RWE study). The Sanofi RWE study was based on case notes from severe asthma centres in the NHS, in which the definition of severe exacerbation was established to match the QUEST trial definition and was used in the company's updated model. The ERG considered the Sanofi RWE study to be of reasonable quality which produced consistent results with other sources. The committee concluded that the Sanofi RWE study on the setting of severe exacerbations was appropriate for use in the company's updated base case.

Mortality estimates are uncertain and probably overestimated in the company economic model

3.15 The ERG explained that the original company model (using the confidential exacerbation multiplier) predicted 20% mortality over 10 years in the standard care arm. The committee questioned the clinical plausibility of this estimate because it seemed high compared with the approximate 1,300 asthma-related deaths a year in the UK. The higher death rate was a result of interaction between the exacerbation multiplier (see section 3.13) and using registry data to inform the setting of treating exacerbations (see section 3.15). The committee noted that the model did not offer plausible estimates, and requested that any additional analyses presented by the company include 10‑year mortality rates for dupilumab and standard care and show the flow of patients though different health states in the model for the purposes of model validation. Following the first committee meeting, the company conducted a literature search for UK asthma-related mortality data, but no further publications were retrieved. The committee noted that when the exacerbation multiplier was removed (in the updated company model, see section 3.14), 10-year mortality with standard care was reduced to 18%. The ERG considered that the mortality was probably still overestimated, but the plausibility of model survival projections was difficult to judge without UK data available. The ERG explained that the model predicted a mean age of deaths of 70 years with standard care (73 years with dupilumab), compared to an estimated life expectancy of 80 years with standard care in the benralizumab appraisal TA565. The committee was concerned that mortality could be overestimated because asthma-related mortality was one of the drivers of the model. The committee also noted that alternative methods had been used in the benralizumab appraisal, to adjust for high mortality. The committee concluded that the mortality rates were uncertain, and that alternative scenarios could be tested to explore the impact of the mortality on the ICER.

The company's updated base-case ICER is £35,968 per QALY gained for dupilumab compared with standard care in the proposed population

3.16 The company's base-case deterministic ICER for dupilumab compared with standard care is £35,968 per quality-adjusted life year (QALY) gained in the company's updated population. This included the confidential discount for dupilumab. The committee concluded that dupilumab does not represent a cost-effective use of resources, so could not be recommended for treating severe asthma with type 2 inflammation.

Additional benefits in people with severe asthma and type 2 inflammation, and nasal polyps or atopic dermatitis, may not have been adequately captured

3.17 The committee recognised that there is an unmet need for people with severe asthma caused by type 2 inflammation. The committee also heard that dupilumab is effective in people with comorbidities (such as nasal polyps and atopic dermatitis). It concluded that these additional benefits of dupilumab had not been captured in the QALY calculation.

There are limited data available on dupilumab for young people

3.18 Dupilumab is licensed in people aged 12 years and over. The company provided a subgroup analyses for the subgroup of people aged 12-17 years. The committee noted that the QUEST trial only included 2 patients aged under 18 years meeting the criteria of the updated base case population. There is an unmet need in this population with uncontrolled severe asthma with type 2 inflammation. Current NICE recommended biologicals are licensed in adults for eosinophilic asthma only. Mepolizumab is currently the only other biological that is licensed for treating children aged 6 years or over for severe refractory eosinophilic asthma. However NICE's technology appraisal guidance on mepolizumab recommends it for use in adults. The committee concluded that there are limited data available for dupilumab in young people, and acknowledged this during decision making.

Dupilumab is not recommended for treating severe asthma with type 2 inflammation

3.19 The committee acknowledged that dupilumab is effective for preventing exacerbations in people with severe asthma with type 2 inflammation compared with standard care. However, the cost-effectiveness estimates for dupilumab compared with standard care were higher than what NICE considers a cost-effective use of NHS resources. The committee identified several uncertainties in the modelling assumptions, particularly about mortality estimates and response rates in adults who did not respond to biological therapy. These uncertainties resulted in uncertainty about the true ICER. Therefore, the committee was unable to recommend dupilumab as a cost-effective treatment for use in the NHS for treating severe asthma with type 2 inflammation.