3 Committee discussion
The appraisal committee considered evidence submitted by Aimmune Therapeutics UK Ltd, a review of this submission by the evidence review group (ERG), and responses from stakeholders. See the committee papers for full details of the evidence.
3.1 Peanut allergy is one of the most common food allergies in children, affecting between 0.5% and 2% of children in the UK. Severe reactions, notably anaphylaxis, can be life threatening, although deaths from peanut allergy-related anaphylaxis are very rare in the UK. Symptoms can have a rapid onset. They can include angioedema (facial swelling), respiratory symptoms (wheezing), conjunctivitis, oral allergy syndrome (lip or tongue swelling), rhinitis (blocked nose) and urticaria (blotchy red rash). It is not possible to predict the likelihood or severity of future allergic reactions based on previous reactions. Patient experts explained that peanut allergy affects 'all aspects of daily life' and 'can cause extreme anxiety' for children and young people with the allergy and their carers. It can have implications for shopping and preparing food, weaning infants, eating out, travelling, seasonal events, education, socialising and work. Parents may experience anxiety, in particular around the time their children start secondary school or leave home. The committee concluded that peanut allergy burdens patients and their carers.
3.2 The goal of treatment for peanut allergy is preventive, that is, to reduce the frequency and severity of allergic reactions and improve quality of life, to lessen anxiety and to normalise activities of daily living. The main preventive strategy for peanut allergy is strictly avoiding peanuts and being ready to respond to an emergency. Symptomatic treatment of mild reactions to accidental peanut exposure includes antihistamines, whereas severe anaphylactic reactions need emergency treatment including self-administered adrenaline. Clinical experts explained that strictly avoiding peanuts cannot be considered the only option, and active preventive treatments are needed to reduce the risks of accidental peanut exposure. They noted that most accidental exposures involve less than 300 mg of peanut protein. Tolerating this amount, equivalent to roughly 1.5 peanuts, would give 'bite protection' from small accidental exposures to peanut and would be a 'meaningful outcome'. Tolerating 1,000 mg of peanut protein is considered to be 'highly clinically significant'. The committee concluded that clinicians and people affected by peanut allergy would welcome a treatment option that would reduce the risks of accidental peanut exposure and improve quality of life of children with peanut allergy and their carers.
Lifetime treatment with Palforzia or regularly including peanuts in the diet is needed to maintain tolerance
3.3 Palforzia is an oral immunotherapy, consisting of peanut protein, as a defatted powder of Arachis hypogaea L. It is given through a 'structured dosing' approach, starting from a dose level as low as 0.5 mg, and gradually increasing. This allows people to build tolerance to peanut protein over time. Strict peanut avoidance and emergency preparedness continue during treatment. The experts explained that Palforzia allows for precise and reproducible dosing of 'minuscule amounts' of peanut protein, which is not possible with dietary peanut. However, according to the clinical experts, once people tolerate higher doses of peanut protein, they can start to include peanuts in their diet to maintain their tolerance instead of continuing treatment with Palforzia. This is not reflected in the marketing authorisation for Palforzia, which notes that daily maintenance is required to maintain the tolerance and clinical effects of Palforzia. Yet, it also states that no recommendation can be made about the duration of treatment beyond 24 months, and that the effect of stopping treatment on maintenance of clinical efficacy has not been evaluated. The clinical experts explained that people not regularly including peanuts in their diet after Palforzia treatment may lose tolerance and will need to return to strictly avoiding peanuts and being prepared for emergencies. They added that whether or not people keep peanuts in their diet after Palforzia treatment may be linked to taste aversion, motivation, adverse effects, restrictions around meals and exercise, and support received. The patient experts anticipated that people will be highly motivated to include peanuts in their diet after committing to Palforzia treatment, but noted that some may be averse to the taste of peanuts. Moreover, there may be psychological stress and anxiety associated with eating a food that people have diligently avoided and greatly feared for many years. The patient experts explained that people with peanut allergy and their carers may require additional psychological support during the transition. The committee concluded that people would need lifetime treatment with Palforzia or to regularly include peanuts in their diet to maintain tolerance to peanuts.
3.4 Palforzia is used 'in conjunction with a peanut-avoidant diet', in line with its marketing authorisation. The company selected strict peanut avoidance as the only comparator for Palforzia. The committee agreed, concluding that strict peanut avoidance was the most appropriate comparator for Palforzia.
3.5 The committee recalled that 1 of the treatment goals is to reduce the risks associated with accidental peanut exposure (see section 3.2). Reactions to accidental exposures are uncommon, with 1 study suggesting an annual incidence of around 12%. Therefore, clinical trials use oral food challenges as a surrogate end point. In an oral food challenge, people are given increasing doses of precise amounts of peanut protein to assess their tolerance to the allergen. Their maximum tolerated dose, or 'tolerability threshold', is the highest dose tolerated with no more than mild symptoms. Having some tolerance to peanut protein can protect from risks of accidental exposure to peanut. The committee recalled that tolerating 1,000 mg peanut protein is considered clinically significant. The committee agreed that tolerance to peanut protein measured during oral food challenge is clinically relevant, and reflects an appropriate surrogate end point for assessing the efficacy of Palforzia.
3.6 Palforzia was studied in 2 randomised controlled trials conducted in North America and Europe, including the UK. All people had peanut allergy confirmed by peanut-specific immunoglobulin E (IgE) reactivity and skin prick test. All participants had an entry food challenge at screening to select people who had dose-limiting symptoms to peanut protein at a dose of 100 mg or less in PALISADE, and 300 mg or less in ARTEMIS. PALISADE enrolled people aged 4 to 55 years (n=496), and ARTEMIS enrolled people aged 4 to 17 years (n=175). People were randomised to receive Palforzia while avoiding peanuts, or placebo while avoiding peanuts. Treatment with Palforzia was given in 3 phases:
An 'escalation phase', which consisted of 5 dose levels (0.5 mg to 6 mg), all given on day 1 of treatment.
An 'up-dosing phase', which had 11 dose levels lasting 2 weeks each (3 mg to 300 mg).
A 'maintenance phase', in which people received 300 mg once daily for 24 to 28 weeks in PALISADE and 12 weeks in ARTEMIS.
Peanut desensitisation was measured in an exit food challenge at the end of maintenance treatment. The primary end point in ARTEMIS and PALISADE in Europe was the proportion of people who tolerated at least 1,000 mg peanut protein with no more than mild symptoms, whereas in PALISADE in North America, it was the proportion of people who tolerated at least 600 mg peanut protein. People completing maintenance treatment and tolerating at least 300 mg peanut protein in the food challenge could enrol in PALISADE follow-on, an open-label observational study, and continue the Palforzia maintenance dose for a further 28 to 56 weeks, after which there was an additional food challenge. The clinical experts explained that the baseline characteristics of people enrolled in PALISADE and ARTEMIS aligned with those of people with peanut allergy seen in NHS practice. They also explained that people in NHS practice would follow the same dosing schedule as in the trials. The committee concluded that PALISADE and ARTEMIS are generalisable to NHS practice.
3.7 Compared with placebo, Palforzia increased the proportion of people who could tolerate at least 1,000 mg peanut protein. Among people aged 4 to 17 years, the proportion of people tolerating at least 1,000 mg peanut protein was:
50.3% of people randomised to Palforzia versus 2.4% of people randomised to placebo in PALISADE (absolute treatment difference 48.7%; 95% confidence interval 38.0% to 57.7%; p<0.0001)
58.3% versus 2.3% of people, respectively, in ARTEMIS (absolute treatment difference 56.0%; 95% confidence interval 44.1% to 65.2%; p<0.0001).
These findings were supported by the key secondary outcomes of tolerating at least 600 mg or at least 300 mg peanut protein. The committee concluded that Palforzia improved tolerance to peanut protein in people aged 4 to 17 years with peanut allergy compared with placebo.
There is no direct clinical trial evidence that Palforzia reduces the frequency and severity of reactions to accidental peanut exposure
3.8 The committee recalled that reactions to accidental peanut exposure are uncommon (see section 3.5). It noted that these reactions, including severe anaphylactic reactions needing treatment with adrenaline, infrequently occurred in the Palforzia trials. It also noted that there were no proven differences in this outcome between the Palforzia and placebo groups. The company considers the exact rates to be confidential so they cannot be reported here. The committee appreciated that avoiding anaphylaxis from accidental exposure to peanuts is a goal of treatment with Palforzia, but also that treatment with Palforzia increases the risk of anaphylactic reactions as an adverse event of treatment (see section 3.9). A clinical expert noted that although there is no direct evidence that Palforzia 'prevents anaphylaxis' following accidental exposure, there is evidence from clinical trials that the maximum severity of symptoms at the exit oral food challenge was lower with Palforzia than with placebo. The committee concluded that there is no direct clinical trial evidence that Palforzia reduces the frequency and severity of reactions to accidental peanut exposure.
The committee would prefer a meta-analysis, but this would not have a meaningful impact on the results
3.10 The company explained that it attempted a meta-analysis of Palforzia clinical trial data but did not consider it robust because of differences in study design between PALISADE and ARTEMIS. These differences included trial location (North America and Europe versus Europe only), age of participants (4 to 55 years versus 4 to 17 years), severity of peanut allergy (reactions to 100 mg or less versus 300 mg or less in the entry food challenge test), and duration of maintenance treatment (24 to 28 weeks versus 12 weeks). The ERG explained that despite these differences, the company could have done a network meta-analysis, a simple meta-analysis, or used individual participant data. The ERG also noted that the company did not consider a phase 2 study, the ARC001 trial. The company explained that it excluded this study because it had only 55 participants and was done solely in the US. The ERG accepted that including this study would not add much insight. The committee agreed with the ERG that despite differences between trials, there were several ways in which the company could have done a meta-analysis of trial data. It noted that meta-analysis would have allowed it to assess heterogeneity and uncertainty in the treatment effect. The committee concluded that it would have preferred to have seen the results of a meta-analysis, but accepted this would be unlikely to have any meaningful impact on the results presented.
3.11 The company used a Markov state transition model to estimate the cost effectiveness of Palforzia plus avoiding peanuts compared with avoiding peanuts only. The model structure had 5 phases: escalation, up‑dosing, maintenance, extension and extrapolation. After maintenance, health states were based on the amount of peanut protein tolerated in an oral food challenge. Simulated people who tolerated 300 mg or more peanut protein could also move to an 'include peanuts in diet' health state. From all health states, people could move to 'spontaneous tolerance' with the lifetime probability of 5%, or 'death' health states. The committee concluded that the model structure and health states were reasonable, and suitable for decision making.
3.12 The model assumed that peanut allergy does not increase the risk of death compared with that of the general population. The company modelled the risk of death from UK life tables for the general population. The company assumed that Palforzia had no effect on the risk of dying. All gains in quality-adjusted life years (QALYs) were from improvements in quality of life for patients and their carers. The committee recalled that deaths from peanut allergy are very rare in the UK. However, the committee considered it possible that Palforzia could decrease that risk and noted that the assumption of no survival gain may be conservative (see section 3.22). The committee concluded that it is reasonable to assume no survival benefit of Palforzia in the model.
The model does not reflect that in the NHS some people may need an oral food challenge before starting Palforzia
3.13 The company based its model on results from PALISADE; ARTEMIS was used in a scenario analysis. In PALISADE, everyone had an oral food challenge before they entered the trial to confirm they were sensitive to less than 100 mg peanut protein. However, the company and the ERG models excluded an oral food challenge before starting Palforzia. The committee noted that the company's trial results for people with confirmed severe peanut allergy may not apply to an unscreened NHS population, some of whom might have developed tolerance spontaneously, or be sensitive to higher doses of peanut protein than people in the trial. The British Society of Allergy and Clinical Immunology 'strongly recommended' doing a food challenge before starting Palforzia in the NHS to confirm whether a peanut allergy is still present, and to determine its severity. The clinical experts explained that re-testing for peanut allergy with a skin prick test and/or IgE testing already forms part of routine care to check for tolerance that sometimes develops spontaneously. It is also done to remind teenagers of their allergy – older children may not have experienced a reaction to peanuts since they were much younger and may not remember being allergic. The clinical experts noted that up to half of people may need a food challenge to determine whether treatment with Palforzia is suitable for them. They also noted that people with recent severe reactions to small amounts of peanut protein from accidental exposure would not need a food challenge. The committee acknowledged that the model did not include a food challenge before starting treatment with Palforzia. The committee concluded that the model does not reflect the likely need for oral food challenge before starting Palforzia in the NHS.
The model does not reflect that most people would not need an oral food challenge after Palforzia treatment in NHS practice
3.14 The company and ERG models included a single oral food challenge for people taking Palforzia, at around 2 years, and none for people avoiding peanuts only. The clinical experts confirmed that people who avoid peanuts are not usually offered food challenges in NHS practice. They also confirmed that if a food challenge were needed after Palforzia, they would offer one only after 1 to 2 years of treatment. They explained that fewer than 20% of people would need a food challenge to determine tolerance after treatment with Palforzia. This is because taking maintenance doses of Palforzia is a de facto food challenge, and if people take 300 mg Palforzia every day, they will be able to tolerate the same amount of peanut protein. Therefore, most people could start including peanuts in their diets without the oral food challenge. The committee concluded that the model does not reflect that most people would not need an oral food challenge after Palforzia treatment in NHS practice.
Few people are likely to continue treatment with Palforzia lifelong in NHS practice, and the model overestimates this
stay on Palforzia lifelong and continue avoiding peanuts
start including peanuts in their diet permanently
start including peanuts in their diet, but then switch back to avoiding peanuts.
The company estimated the likelihood of being in each of these 3 groups based on a consensus clinical opinion. The company considers these estimates confidential so they cannot be reported here. The ERG acknowledged that the company used reasonable methods to reach these estimates but highlighted that they are not evidence based and are uncertain. The patient experts explained that people who have committed to 2 years' treatment with Palforzia will be highly motivated to maintain tolerance, start dietary peanuts and adhere to a diet that includes peanuts. The clinical experts agreed that most people would be happy to start including peanuts in their diet, although some may then stop. The clinical experts explained that only people with borderline tolerance to 300 mg peanut protein may need to continue treatment with Palforzia. They expected this would be fewer than 5% of people who had completed maintenance treatment with Palforzia. The committee noted that this proportion was lower than the company assumed in its model. The committee concluded that fewer people than modelled are likely to continue treatment with Palforzia lifelong in NHS practice.
3.16 The company conducted a de novo utility study to estimate quality of life in children and young people with peanut allergy using the EQ‑5D‑Y (the youth EQ‑5D), and for their carers using EQ‑5D. The company pooled data from various sources and populations to obtain utility estimates, including:
38 online surveys completed by young people aged 12 to 17 with peanut allergy themselves ('self-reported utilities'); none of whom had prior treatment with Palforzia
62 online surveys completed by carers on behalf of their children aged 4 to 11 with peanut allergy ('proxy-reported utilities'); none of the children had prior treatment with Palforzia
50 interviews with carers reporting utilities on behalf of their children aged 4 to 11 with peanut allergy; none of the children had prior treatment with Palforzia
7 'Palforzia surveys': surveys among people who have had prior treatment with Palforzia, of which 2 were self-reported by young people aged 12 to 17, and 5 were proxy-reported by parents of children aged 4 to 11 with peanut allergy.
The company pooled data from all these populations and sources (n=157) to inform its cost-effectiveness modelling. The ERG preferred to use only self-reported data from young people who had not had prior treatment with Palforzia (n=38). The ERG was concerned that carers may project some of the negative impact on their own quality of life when reporting proxy utility values for children. This risks double counting the negative impact on carers, because the company included carer disutility separately in its model (see section 3.17). Furthermore, the ERG explained that the company's survey among people who have had Palforzia, or their carers, may have been biased, because they might have not remembered what their (or their children's) quality of life was before or during the treatment. The committee noted the large difference between the utility values estimated by the company and the ERG. It agreed with the ERG that utility values from the Palforzia survey were not plausible. It also noted that the difference between the company and ERG approaches seemed to be driven mainly by differences in methods of data collection, rather than whether utilities were self- or proxy-reported. The committee agreed it was unclear which utility values better reflected quality of life of people with peanut allergy in NHS practice. The committee concluded that the utility values were uncertain and agreed to consider both the ERG-preferred approach and the pooled sample from all treatment-naive people, that is, excluding those from the Palforzia survey (n=150).
3.17 The patient and clinical experts explained that carers often 'carry the mental load' of peanut allergy day to day. The model included carer disutility until people with peanut allergy reached 18 years of age. The utility values for carers were collected in the company's de novo utility study, using EQ‑5D (see section 3.16). The company assumed that on average, there was more than 1 carer per child (the company considers the exact number to be confidential so it cannot be reported here). The ERG agreed it was reasonable to include carer quality of life in the model, but highlighted that the number of carers per child is uncertain. The committee noted that the number of carers per child assumed by the company was in line with other technology appraisals that included carers' quality of life. The committee concluded that including carers' quality of life in the model is appropriate, and that the methods used by the company to capture this were acceptable.
It is reasonable to assume the utility gains for children with peanut allergy and their carers can be realised without oral food challenge
3.18 The company and ERG assumed that treatment with Palforzia during escalation and up‑dosing has a negative impact on quality of life of children with peanut allergy and their carers. However, once people start maintenance treatment, there is a small gain in quality of life compared with baseline. Any further gains are then realised only after the oral food challenge, and are linked to the tolerance level reached, with higher benefit related to higher tolerance levels. The ERG explained this is because utility values associated with a change in tolerance level are realised only after the results of a food challenge become known. Both the ERG and company models also assumed that people who avoid peanuts but do not take Palforzia do not have any oral food challenge and therefore have no gains in quality of life. That is, the model assumed they have the same baseline quality of life throughout the model time horizon. The clinical experts explained that people on Palforzia could have quality-of-life benefits without an oral food challenge because they know they can tolerate a 300 mg dose of peanut protein, and can start including peanuts in their diet. The committee recalled that tolerating 300 mg peanut protein gives 'bite protection' (see section 3.2). It also recalled that taking maintenance doses of Palforzia is a de facto food challenge and people could start including peanuts in their diet after reaching tolerance to 300 mg peanut protein. The committee concluded that utility gains for children with peanut allergy and their carers related to Palforzia treatment can be realised without an oral food challenge.
3.21 The committee noted that company and ERG base-case models differed only in their approach to estimating utility values (see section 3.16). In a deterministic pairwise analysis, the company base-case incremental cost-effectiveness ratio (ICER) was £23,745 per QALY gained compared with avoiding peanuts only. The ERG deterministic base case was £36,565 per QALY gained. Corresponding probabilistic ICERs were £25,940 and £39,716 per QALY gained, respectively, compared with avoiding peanuts only. However, the committee noted that neither the ERG nor company base cases fully captured its preferences that:
up to 50% of people may need an oral food challenge before starting Palforzia in NHS practice to confirm peanut allergy (see section 3.13)
fewer than 20% of people would need an oral food challenge after Palforzia treatment to determine if they can start introducing peanuts into the diet, instead of continuing Palforzia (see section 3.14)
fewer than 5% of people would be expected to continue treatment with Palforzia lifelong in NHS practice (see section 3.15)
utility values for people with peanut allergy are highly uncertain; therefore, both the ERG-preferred utilities from only adolescent self-reported, treatment-naive sample (n=38) and pooled utilities from all people naive to treatment (n=150) should be considered (see section 3.16)
Palforzia can have a positive impact on quality of life of people with peanut allergy and their carers without an oral food challenge (see section 3.18)
all costs related to setting up oral immunotherapy treatment in NHS practice should be included (see section 3.19).
The committee therefore considered a number of scenario analyses done by the ERG. It noted that assuming a food challenge before starting treatment with Palforzia had a minimal impact on the ERG base-case ICER, increasing it to £37,059 per QALY gained. Although it did not see a scenario assuming lower use of oral food challenge after treatment with Palforzia, it agreed it would also be likely to have a minimal impact on the ICERs. The committee noted that the choice of utility values and the proportion of people continuing treatment lifelong had the biggest impact on the cost-effectiveness estimates. Using utility values from the 150 treatment-naive people in the sample decreased the ERG base-case ICER to £27,735 per QALY gained. Assuming nobody continues Palforzia lifelong decreased the ERG base-case ICER to below £20,000 per QALY gained. The committee noted it did not see any scenario analyses assuming 5% people would continue Palforzia lifelong, but agreed that ICERs would slightly increase compared with scenarios assuming nobody continues Palforzia lifelong. When using utilities from the full sample of 150 treatment-naive people, the ICERs would be lower than £20,000 per QALY gained. The committee therefore agreed that the most plausible ICERs, when excluding the costs of setting up oral immunotherapy treatment in NHS practice, would be around £20,000 per QALY gained. The committee further noted it did not see any scenario analyses including all costs related to setting up oral immunotherapy treatment in NHS practice, in line with its preferred assumptions, and that it may be difficult to estimate the impact on the cost-effectiveness estimates. However, the committee agreed there may be some benefits of Palforzia not captured in the company and ERG models (see section 3.22). It also recalled that peanut allergy is a burden for children and their carers (see section 3.1). Therefore, the committee concluded that Palforzia is likely to be a cost-effective use of NHS resources, and therefore recommended it for routine NHS use.
3.22 The committee agreed with the clinical experts and the company that Palforzia is a 'potential step change' in treating peanut allergy and as such is innovative. It further noted some potential benefits that may not have been captured in the modelling:
The company's model assumed no survival benefit from Palforzia (see section 3.12), so if even 1 death from anaphylaxis is prevented by Palforzia, this would be an additional benefit for people with peanut allergy and the NHS.
Even if people do not continue to benefit from treatment for their whole life, for example if they return to avoiding peanuts later in life, they may still have benefited from treatment and peanut tolerance during important years as a young adult, when they are growing up, gaining independence and travelling.
The committee concluded that Palforzia is innovative and may have some benefits that were not adequately captured in the model, and took this into account in its decision making.
3.23 The committee recalled that Palforzia has a marketing authorisation for treatment of peanut allergy in people aged 4 to 17 and those who turned 18 while on treatment (see section 2.1). This means most adults with peanut allergy will have no access to treatment. However, the committee noted that its remit only allows it to appraise a technology within its marketing authorisation.
3.24 The committee recalled that treatment with Palforzia is resource intensive because patients must attend a clinic prepared to treat anaphylaxis for initial dose escalation and the first dose of each new up‑dosing level, in line with its marketing authorisation (see section 3.19). The patient and clinical experts explained that there may be unequal access to specialist allergy clinics in England because of geographic location or socioeconomic status. The committee considered these issues but noted they are related to implementing guidance in NHS practice and therefore outside of its remit. The committee concluded there are no equalities issues that can be addressed in the guidance.