Berotralstat for preventing recurrent attacks of hereditary angioedema

3.1 Hereditary angioedema is a rare genetic disorder. It affects approximately 1 per 10,000 to 50,000 people, and usually develops in the first 10 to 20 years of life. It is a relapsing condition that causes unpredictable and recurrent attacks of swelling. This is usually in the mouth, gut or airway, but it can affect multiple places in the body at once. It often leads to difficulty breathing and severe pain. The patient experts explained that acute attacks of hereditary angioedema are difficult to predict and can vary in severity from mild to life threatening. Attacks can significantly affect the quality of life of people with this condition, as well as that of their family members and carers. The patient and clinical experts explained that attacks can be triggered by anxiety and stress; for example, exams, surgery or dental treatment, as well as positive life events such as weddings and holidays. The clinical experts highlighted that usually attacks are treated as they happen. They advised that the aim of prophylactic treatment is to reduce the rate and severity of attacks and allow people to live an attack-free life. There are currently no effective licensed oral prophylactic treatments. Current oral long-term prophylactic treatment includes attenuated androgens, usually danazol. These are prescribed early in the treatment pathway but often have side effects and limited effectiveness. Also, access to androgens is often limited because of supply issues (see section 3.2). The clinical experts explained that long-term prophylactic treatment with injectable lanadelumab or C1 esterase inhibitors (C1‑INH) is only available in England for a very small number of people who have 2 or more clinically significant attacks per week as per NHS England's commissioning policy. The patient and clinical experts also highlighted that there are limited prophylactic treatment options for people with difficult intravenous access and needle phobia. The committee recognised that hereditary angioedema can be a severe and debilitating condition. It acknowledged the lack of effective prophylactic treatment options available to people with this condition. The committee concluded that there is an unmet need for effective treatment options for preventing recurrent attacks of hereditary angioedema.

3.2 The company originally positioned berotralstat for people with at least 2 angioedema attacks per month who have used androgens before, or if androgens are unsuitable. To align with its proposed positioning for berotralstat, in the economic model the company used data on the subgroup of patients in APeX‑2 (the main source of clinical evidence; see section 3.4 and section 3.5) who had at least 2 attacks per month and who had used androgens before. This population is narrower than that specified in the marketing authorisation and NICE scope. It is also narrower than the intention to treat population of APeX‑2 (n=80 in the intention to treat population compared with n=35 in the company's proposed positioning subgroup). The intention to treat population in APeX‑2 also includes patients who had fewer than 2 attacks per month, and those who had not used androgens before. The clinical experts explained that specifying 2 or more attacks a month is reasonable. This is because people having less frequent attacks may not want to take regular or preventative treatments to avoid attacks. However, they stated that supply of androgens in the NHS is inconsistent. They explained that access to androgens varies, is based on local arrangements, and people are unable to get them from local pharmacies. One expert highlighted that the Department of Health and Social Care's advice to clinicians is to not prescribe androgens to people who have not had them before. The committee heard that people under 18 cannot have androgens, but people under 18 are included in the marketing authorisation for berotralstat. It was concerned that positioning berotralstat after androgens may prevent some people from accessing berotralstat. Consultation comments highlighted that the term 'unsuitable for androgens' in the company's proposed positioning wording should apply to people under 18. During consultation, the company agreed that it is suboptimal to deny access to berotralstat because of androgen supply shortages and updated its proposed positioning wording to include 'unavailability of androgens'. However, the committee remained concerned that this may still inadvertently prevent some people from accessing berotralstat. So it based its decision making on a larger subgroup (see section 3.7), which included some people who had not had androgens before. Since the cost-effectiveness estimates for this subgroup are within the range normally considered cost effective, berotralstat is recommended whether or not people have had androgens before (see section 3.16).

3.3 The company submission compared prophylactic berotralstat with no prophylactic treatment. In both groups people had treatment if they had an attack. The company described this as standard care. Standard care treatments include C1-INHs, icatibant and conestat alfa. The ERG noted that this was narrower than the comparators specified in NICE's final scope for this appraisal. But the ERG's clinical expert agreed with the company's description of how hereditary angioedema is currently treated in the UK. The committee concluded that standard care is an appropriate comparator at the company's proposed positioning of berotralstat.

3.4 The clinical-effectiveness evidence for berotralstat is from APeX‑2. This is a 3-part, phase 3, randomised, double-blind, placebo-controlled trial in people 12 years or older with type 1 or type 2 hereditary angioedema. Part 1 of APeX‑2 compared berotralstat 150 mg (n=40) with placebo (n=40) over a follow-up period of 6 months. People had standard care if they had an attack during the trial period in both the berotralstat and placebo arms (see section 3.3). The placebo arm of APeX‑2 informed the clinical evidence for the standard care arm used in the economic model. Berotralstat 110 mg was also included in APeX‑2 but was not considered relevant to this appraisal because this dose will not be licensed or marketed in the UK. The committee was aware of the small sample size of the trial, particularly for the trial data relevant to the company's proposed positioning (see section 3.2). However, it acknowledged that doing a robust trial in hereditary angioedema is difficult because of the rarity of the disease.

3.5 Results from the intention to treat population of APeX‑2 show a statistically significant reduction in mean monthly attack rates of 44% with berotralstat compared with placebo. Similar results were observed for the subgroup of patients from APeX‑2 who had 2 or more attacks at baseline. The patient experts explained that a prophylactic treatment that reduces attack rate could potentially be life changing for people with this condition. However, they explained that although the reduction in attack rate is a clinically important outcome for people with hereditary angioedema, the reduction in attack severity would be equally important. They noted that if a treatment did not reduce attack rate, but reduced attack severity, they would still value the option to have that treatment. They further highlighted that the hospitalisation of people with hereditary angioedema is often because of attack severity rather than attack rate. The company and the ERG stated that the location of the attack (such as the limbs or the airway) and duration of attack were used as a proxy for attack severity in the model. The company explained that the direct measure of attack severity in the trial was subjective and did not show that berotralstat reduced attack severity more than placebo. This was despite the proxy measures showing that berotralstat reduced severity. So this subjective measure was not considered credible enough to be included as an outcome in the analysis. The committee recognised that it is important to consider evidence on attack severity as well as attack rate when assessing the clinical effectiveness of berotralstat. The committee was dissatisfied that direct data on severity from APeX‑2 was not presented and applied in the model but accepted that there are limitations with this data. It concluded that the clinical evidence suggests berotralstat is more effective than placebo in reducing attack rate, but its effect on attack severity is not known.

3.6 The company submitted a cohort-level Markov model with 2 health states: alive and dead. The alive health state was split into 2 substates: attack free or attack. The time spent in each of these substates was determined by treatment-specific attack rates from APeX‑2. The model used percentage reductions from baseline attack rates in the berotralstat and placebo arms of APeX‑2, applied to the baseline attack rates specified in the model. People in the attack substate incur the costs of an acute attack and lower health benefits compared with those in the attack-free substate. The ERG advised that the model structure is generally acceptable and similar to a previous appraisal in this disease area (see NICE's technology appraisal guidance on lanadelumab). In the berotralstat arm of the model, the company applied a treatment continuation rule. This rule states that people can only continue taking berotralstat if they have a reduction in attack rate of at least 50% compared with baseline by 3 months. However, the committee noted that there was no continuation rule in APeX‑2 or the marketing authorisation. It was concerned with the choice of a 50% or more reduction in attack rate from baseline as the cut-off point to continue treatment beyond 3 months. The company stated that this 50% or more cut-off point was based on people having 2 or more attacks per month. It noted that applying this cut-off point results in a reduction of at least 1 attack per month. The patient experts explained that if people had fewer attacks but did not reach the threshold of a 50% reduction, they would likely want to continue treatment anyway. Also, even if the number of attacks did not decrease, but the severity did, they would consider it beneficial to continue treatment. The committee noted the importance of the patient experts' comments, and was concerned that it would be difficult to implement the continuation rule in clinical practice. In response to consultation, the company highlighted that a continuation rule is already being used in clinical practice through the early access to medicines scheme for berotralstat. It noted that continuation rules are in place for other treatments for hereditary angioedema through NHS England's commissioning policy on C1‑INHs. Comments from consultees, including patient organisations, also supported applying the continuation rule in clinical practice. They suggested that clear guidance on stopping if the response to berotralstat is not good enough should be provided and that a 50% reduction in attacks is reasonable. Clinical advice to the ERG also supported the continuation rule. However, the ERG highlighted that the early access to medicines scheme for berotralstat and NHS England's commissioning policy for use of C1‑INHs do not strictly define the percentage reduction in attack rate. The committee was satisfied that it was appropriate to include a continuation rule in the economic model. It concluded that the model structure is acceptable for decision making.

3.7 To align with its proposed positioning for berotralstat, the company's model inputs were based on data from a subgroup of APeX‑2 with a small number of patients (n=35, 17 patients who had berotralstat and 18 who had standard care; see section 3.2). The ERG highlighted its concerns with using clinical evidence for attack rate reductions based on a small sample size. It suggested that analysis using the intention to treat population would provide this evidence for a larger number of people. This would also reduce uncertainty in the cost-effectiveness analysis. In response to technical engagement, the company considered using the intention to treat population from APeX‑2 to inform its economic model. But because this included people who would not have berotralstat in UK clinical practice, it suggested that this would undermine the cost-effectiveness evidence used for decision making. Instead, it provided a scenario analysis using clinical evidence from a larger subgroup (n=57) of people with at least 2 attacks per month who may not have used androgens before. The ERG agreed with using this larger subgroup because it included more people than the company's proposed positioning subgroup. At the second meeting, the ERG also highlighted that the company's updated positioning (see section 3.2) may include more people who have not had androgens before. This made using data from the larger subgroup more relevant. The committee recalled its concerns about the company's positioning (see section 3.2). It concluded that the larger subgroup including those who may not have used androgens before is more appropriate for decision making.

3.8 The company's original model used observed data from APeX‑2 to inform treatment-specific baseline attack rates. It used the monthly percentage reduction in attack rates from baseline to 12 months for the berotralstat arm, and to 6 months for the standard care arm. To extrapolate the long-term percentage reduction in attack rate in each treatment arm beyond the specified periods, it used the last observed percentage reduction carried forward over the remaining time horizon of the model. The ERG raised several concerns with the company's original base-case analysis:

3.9 The company's model took treatment-arm specific costs for managing acute attacks from APeX‑2. This resulted in the estimated costs per attack being lower in the berotralstat arm than in the standard care arm. This was because of a reduced need for multiple administrations of treatments to manage acute attacks. However, the ERG's clinical expert suggested that there was no plausible reason for berotralstat to consistently affect the cost of treating attacks. Because of the small sample size of the company's proposed positioning subgroup (see section 3.2), the ERG advised that it would be more appropriate to use equal acute attack treatment costs between the berotralstat and standard care arms, based on the intention to treat population. In response to technical engagement, the company highlighted that use of acute treatments in the berotralstat and standard care arms of APeX‑2 was consistent between its proposed positioning subgroup, the intention to treat population and the larger subgroup. Clinical advice to the company suggested that a reduced need for multiple treatments for acute attacks in the berotralstat arm was because of reduced attack severity. During technical engagement, the clinical experts highlighted that prophylactic treatment would likely reduce both the rate and severity of attacks, resulting in lower costs per acute attack overall. They explained that the number of people who need a second dose of treatment to manage acute attacks would reduce if berotralstat reduces attack severity. The committee considered that alternative published data sources may provide information about the use of treatments for acute attacks. However, it concluded that treatment-arm specific costs for managing acute attacks taken directly from APeX‑2 were appropriate for decision making.

3.10 The company used utility values from Nordenfelt et al. (2014), a Swedish registry study that included EQ‑5D‑5L values for both the attack-free and attack substates. The ERG highlighted that EQ‑5D data was collected in APeX‑2. It considered that this should have been explored further, particularly in the APeX‑2 intention to treat population given the small sample size of the company's proposed positioning subgroup (see section 3.2) and the continuation rule (see section 3.6). During technical engagement, the company explained that using the EQ‑5D data from APeX‑2 resulted in implausible utility values for the attack-free health state because they were higher than those of the general UK population. The clinical experts explained that the effect of an attack on quality of life is more likely to be influenced by personal factors and severity of attacks, rather than prior treatment with androgens or attack rate. They advised that quality of life is better for those in the berotralstat arm compared with the standard care arm when attack free. The ERG also highlighted that the utility values from Nordenfelt et al. were based on a larger sample size and that the attack utility data was collected systematically. In contrast, in APeX‑2, the quality-of-life data collection may not have coincided with an attack. The committee was concerned that using utility values directly from APeX‑2 may not adequately capture the effect of attacks on health‑related quality of life and does not reflect the effect of attack severity. But it noted that the latter was likely to apply to the utility values from Nordenfelt et al. too. The ERG explained that the duration of attack, which is used as a proxy for attack severity, is captured in the quality-adjusted life year (QALY) calculation. But it noted that there was not much difference in the duration of attack between the different treatment arms in APeX‑2. The committee concluded that analysis using utility values that reflect attack severity as well as attack rate reduction would have been preferable.

3.11 The company's original model included a caregiver disutility based on a time trade off study that reflected how anxiety and the need to provide care affect caregivers' health-related quality of life. This was applied in the model for all the time spent caring for a person with an attack in the alive health state. The ERG explained that applying a single carer disutility for every attack and for every person may be too simplistic. It noted that it is unlikely that all attacks will affect carers to the same extent. It also had concerns with how large the carer disutility is, but this figure is considered confidential by the company and cannot be reported here. It suggested that this was too large when compared with the range identified in the NICE's decision support unit review of other technology appraisals (0.01 to 0.173 per year). Following technical engagement, the company revised its base case by applying carer disutility to 52% of attacks based on a burden of illness study. The patient experts explained the effect hereditary angioedema attacks have on carers, and the level of anxiety associated with caring for a family member with hereditary angioedema. The committee heard that, despite a reduction in attack rate, the level of anxiety remains, although often to a lesser extent for both patients and carers. The committee was aware that NICE's guide to the methods of technology appraisal states that the perspective on outcomes should be 'all direct health effects, whether for patients or, when relevant, carers'. However, it noted that although many diseases and conditions may adversely affect carers, few technology appraisals model this. For example, carer disutility was not included in a previous appraisal in this disease area (see NICE's technology appraisal guidance on lanadelumab). It considered that there was no clear evidence to suggest that the utility gains for carers associated with berotralstat use would be substantially greater than those with displaced treatments. It concluded that it was not appropriate to include health-related quality of life effects for carers in the base case.

3.12 The committee considered the cost effectiveness for berotralstat compared with standard care using its preferred assumptions, that is:

3.13 The committee considered the advice about life-extending treatments for people with a short life expectancy in NICE's guide to the methods of technology appraisal. It noted that berotralstat is a long-term prophylactic treatment and that the company did not make a case for berotralstat to be considered a life-extending treatment. The committee concluded that berotralstat does not meet the criteria to be considered a life-extending treatment at the end of life.

3.14 The committee considered berotralstat to be innovative because it would be the first licensed oral prophylactic treatment option for people with recurrent attacks of hereditary angioedema. This would mean people would have access to medicine that is more convenient than injectables. The patient and clinical experts explained the importance of reducing attack rate and people being attack free. They highlighted the potential for berotralstat to improve unpredictable and recurrent attacks of swelling and overall quality of life of people with this condition. The committee noted that berotralstat was granted early access to medicines scheme status. This gives people with life threatening or seriously debilitating conditions access to medicines that do not yet have a marketing authorisation or when there is a clear unmet medical need. The committee concluded that berotralstat is an innovative prophylactic treatment for recurrent attacks of hereditary angioedema, but all relevant benefits are reflected in the cost-effectiveness estimates.

3.15 No equalities issues were raised during scoping and technical engagement. The committee considered the implications of the company's positioning for berotralstat (see section 3.2), including any equality considerations. No additional equality issues were raised. The committee concluded that there were no equalities issues relevant to the recommendation.

3.16 Berotralstat is clinically effective at reducing attack rate compared with placebo. The committee took into account all commercial discounts for berotralstat and standard care treatments and agreed that the most plausible ICER was within the range NICE normally considers to be a cost-effective use of NHS resources. So, it concluded that berotralstat is recommended for preventing recurrent attacks of hereditary angioedema for people 12 years and older. But it is recommended only if they have at least 2 attacks per month, and it is stopped if the number of attacks per month does not reduce by at least 50% after 3 months.