Beremagene geperpavec for treating skin wounds associated with dystrophic epidermolysis bullosa

Epidermolysis bullosa (EB) is a rare condition caused by gene mutations that mean certain skin anchoring proteins are not expressed correctly or are disrupted. This results in very fragile skin that blisters and breaks frequently, particularly in response to minor trauma or friction. There are various types of EB, but beremagene geperpavec is only indicated for dystrophic EB (DEB). DEB affects a part of the skin called the sublamina densa, below the lamina lucida. There are 2 types of DEB, defined by whether 1 or 2 copies of the collagen type 7 alpha 1 chain (COL7A1) gene are affected. In dominant DEB (DDEB), the mutation is only in 1 copy of the COL7A1 gene. Life expectancy is typically unaffected, and blistering is usually confined to areas exposed to friction. Recessive DEB (RDEB) is often more serious because both copies of the COL7A1 gene are affected. This form of the condition often leads to widespread blistering and more extensive skin damage, resulting in an increased risk of death from aggressive skin cancer. The clinical experts explained that wounds can also form internally, and can lead to widespread complications affecting body systems such as the eyes and gastrointestinal tract. Severe forms of DEB are likely to present from birth and so are often diagnosed in babies or children. The committee concluded that DEB can be a severe and life limiting condition, particularly in the more severe cases, that often affects young children.

The clinical experts explained that, in people with DEB, blistering may be caused by daily actions such as rubbing, scratching, or light bumps. A patient expert at the meeting explained that blisters were not limited to any specific area and are not self limiting. That is, they blister or a subsequent wound may continue to increase in size if not managed carefully. The clinical experts explained that there are 2 different types of wounds:

• recurrent wounds, which:

  • are generally small to medium sized

  • usually heal spontaneously within a few weeks

• chronic wounds, which:

  • tend to be larger and more painful

  • may cover an entire body region, such as the back or a limb

  • tend to remain open for at least 4 weeks, with some never fully healing.
    

The patient experts emphasised that DEB can have a debilitating impact on all aspects of life, particularly in more severe forms such as RDEB. It can cause constant blistering and pain, and substantial segments of skin may fall off. A patient expert explained that blistering and wounds can be so severe that people may be unable to walk, hold or touch anything. They explained that chronic wounds can consist of completely raw skin that can cover extremely large areas and are intensely painful. The patient experts also noted that with some wounds not healing for many months and increased risk of infection, this can lead to fatigue and sleep disruption, which are often exacerbated by persistent pain and itching. They also emphasised the large impact of non-skin-related symptoms. These include corneal abrasions, which can cause vision loss, and gastrointestinal and oral lesions that can lead to nutritional problems and may impair growth in children. A patient organisation emphasised the substantial psychological burden associated with the condition, driven by lifelong relentless pain, loss of independence, and stigma associated with the visible appearance of wounds. The patient expert also explained that wound management is highly burdensome, in their case taking them up to 7 hours per day. Some carers may need to leave work to provide full time care, which has financial implications. Education can also be disrupted by wound dressing and hospital visits. The clinical experts explained that high levels of inflammation from infection of chronic wounds can cause health conditions such as anaemia and osteoporosis. They emphasised that specialist care is delivered at 4 centres across the UK, so some people need to travel considerable distances to access treatment. The committee concluded that DEB has extensive and severe effects on the quality of life of people with the condition, their families and carers.

The clinical experts explained that there were no licensed disease-modifying treatments for DEB. Current care has 3 broad categories:

• wound management, which includes bathing to wash wounds, lancing and draining blisters and using non‑adhesive dressings and bandages to manage open wounds. NICE's highly specialised technologies guidance on birch bark extract for treating epidermolysis bullosa (from here referred to as HST28) recommends birch bark extract for partial thickness wounds (wounds that extend through multiple different layers of the skin surface) associated with DEB and junctional EB in people 6 months and over. The clinical experts explained that birch bark extract, if used, is always an add-on to standard wound management. Antibiotics are also used to treat infection

• surgery, which is used to manage complications of DEB such as fusion of fingers and squamous cell cancer

• pain management, which includes pharmacological and non‑pharmacological interventions to manage the pain and itch from DEB and from surgical and wound management procedures.

The clinical experts explained that response rates to birch bark extract vary. They estimated that about 50% of people see a benefit in wound-closure rates. They emphasised that only about 10% of people with DDEB use birch bark extract in clinical practice. This is because they typically have smaller, fewer and less severe wounds compared with RDEB, many of which resolve spontaneously over time. The clinical expert also explained that birch bark extract usually only has a small benefit in severe RDEB. A patient expert emphasised that some people do not want to use birch bark extract, especially on large chronic wounds, because its oily consistency can hinder wound drying and may increase the risk of infection. The thickness of the preparation can also limit its practicality for larger wounds, because applying dressings evenly can be difficult and may leave residual product on the wound. So, many people rely on best supportive care alone. The committee noted that the summary of product characteristics permits beremagene geperpavec to be used on wound areas of up to 200 cm² per week. It recalled that people with DEB may have wounds distributed across their body, including some large chronic wounds. So, it thought that some people may wish to treat wounds over the maximum permitted surface area. The clinical experts at the meeting confirmed that beremagene geperpavec would likely be the preferred treatment for large, chronic wounds, because it directly addresses the underlying cause by replacing the missing COL7A1 gene. But birch bark extract would likely be used at the same time, to treat small and superficial wounds. The committee concluded that the relevant comparators were birch bark extract with best supportive care and best supportive care alone. It agreed that further real-world evidence would be useful to resolve the uncertainty about who uses birch bark extract in clinical practice. It also concluded that beremagene geperpavec and birch bark extract (with best supportive care) may also be used in combination in clinical practice.

Clinical-effectiveness data for beremagene geperpavec came from 2 clinical trials and an open-label extension study. Both trials used an intra-patient design in which different target wounds (a single wound used to assess outcomes such as wound closure and speed of wound healing) on the same person were treated. The trials were:

• GEM‑1/-2 (from here, referred to as GEM‑1), which was an open-label randomised controlled trial, consisting of 2 phases. Phase 1 enrolled 2 people aged 18 and over with 2 wounds of less than or equal to 10 cm². Phase 2 consisted of 3 parts and enrolled 10 people aged 2 years and over with at least 2 wounds of less than or equal to 50 cm². Wounds were treated with a range of doses between 8 × 10⁸ and 1.2 × 10⁹ plaque forming units (PFU) or placebo at a range of dosing frequencies. The primary outcome was the proportion of wounds with complete closure at weeks 8, 10 and 12. Time to, and duration of, wound closure was also collected

• GEM‑3, which was a double-blind, randomised controlled trial in 31 people aged over 6 months. The key outcomes were the proportion of wounds with complete wound closure at 6 months (the primary outcome) and at 3 months (considered a key secondary outcome). 2 wounds of similar size were treated, 1 with beremagene geperpavec and another with placebo. People had beremagene geperpavec at a dose according to their age and wound area up to a maximum dose of 1.2 × 10⁹ PFU for 6 months or until wound closure, after which they were followed up for 30 days

• an open-label extension study for GEM‑3, which enrolled 23 people with DEB who had not previously had beremagene geperpavec and 24 people from GEM‑3. People had weekly beremagene geperpavec up to the maximum weekly dose (1 × 10⁹ PFU) until wound closure. The key outcomes were safety and tolerability, health-related quality of life and durability of wound closure.

The GEM‑3 results suggested that more wounds closed when treated with beremagene geperpavec compared with placebo. Specifically:

• complete wound closure increased by 46% with beremagene geperpavec compared with placebo at 6 months (primary endpoint; beremagene geperpavec 67%, placebo 22%, 95% confidence interval [CI] 24 to 68; p=0.002)

• complete wound closure increased by 51% with beremagene geperpavec compared with placebo at 3 months (key secondary endpoint; beremagene geperpavec 71%, placebo 20%, 95% CI 29 to 73; p<0.001).

The marketing authorisation for beremagene geperpavec included all people with DEB with mutation in the COL7A1 gene. The committee recalled this included people with RDEB and DDEB. But the EAG advised that only 1 person GEM‑1, 1 person in GEM‑3 and 1 extra person in the open-label extension study had DDEB. The company considered that the lack of data in DDEB was a consequence of the rarity of the condition. But it expected beremagene geperpavec to be effective in both subtypes because it replaces the missing functional COL7A1 gene. The company provided evidence of response in the DDEB subtype from both the US and the open-label extension study to support this. The committee recalled the difference in severity and treatment outcomes for the different DEB subtypes (see section 3.1). The EAG was concerned that the treatment effect from the company's clinical trials may not be generalisable to people with DDEB. It noted biological rationale for differences in treatment effect by subtype, given that DDEB is associated with less severe collagen abnormalities and retention of a functional COL7A1 gene. The clinical experts at the meeting supported this, stating that people with DDEB generally have a smaller wound burden, with mostly recurrent wounds that close within a few weeks with best supportive care. They explained that only a small proportion of people with DDEB were likely to have beremagene geperpavec in clinical practice, and this population would be determined by the specialist centres. This would mostly be adults with recurrent wounds that have become chronic, or wounds that are particularly painful or in areas affecting functionality and that may not heal without further treatment. But most people with RDEB were likely to have beremagene geperpavec in clinical practice. The committee noted that the company had not provided analyses considering the impact of beremagene geperpavec by subtype. It noted that the population with DDEB who would have beremagene geperpavec in clinical practice was uncertain because it had not been defined by the company. It agreed that the following should be submitted at consultation:

• the proportion of people having beremagene geperpavec for RDEB and DDEB in clinical practice

• the criteria for identifying people with DDEB for treatment, including any wound-burden thresholds for treatment

• clarification of which people with DDEB are represented by the RDEB data in the GEM studies.

There were no clinical trials directly comparing beremagene geperpavec with birch bark extract. The committee noted that clinical-effectiveness data for birch bark extract was available from the EASE trial. This was a phase 3 randomised double-blind trial that included people with both subtypes of DEB as well as junctional EB (in which blistering occurs in the basement membrane of the skin). A single target wound of between 10 cm² and 50 cm² was randomised to have either birch bark extract gel or a control gel for 90 days. After this, people could have birch bark extract for up to 24 months. The company stated that an indirect treatment comparison (ITC) between GEM‑3 and EASE was not feasible. This was because there were differences between trials in:

• the overall trial design, frequency of placebo use and permitted concomitant treatments

• the characteristics of the people included, such as:

  • median age

  • wound size

  • DEB subtype proportions

• trial endpoint definitions, including:

  • complete wound closure: GEM‑1 and GEM‑3 used stricter closure criteria and only counted wound closure at the end of the assessment period, but EASE counted closure at any time during the assessment period

  • time to wound closure.

The company was concerned that the EAG's ITC only captured the rate of wound closure, because time to closure and duration of closure were not assessed in GEM‑3 (see section 3.6). It reiterated that beremagene geperpavec was expected to keep wounds closed for longer than birch bark extract or best supportive care because it addresses the underlying condition (see section 3.5). So, by excluding benefits related to the duration of wound closure, the EAG's ITC was likely to substantially underestimate the overall treatment effect of beremagene geperpavec. The company was also concerned about the use of the EASE data, in which not all open wound area was treated. This introduced a disconnect with the PLS used to estimate vial usage, which assumed that the entire open wound area was treated (see section 3.14). The committee also noted that the median wound size in EASE was considerably larger than in GEM‑3, which may have introduced uncertainty into the results. It considered that EASE included wound sizes that may be different to those treated in clinical practice (see section 3.5). It noted this likely favoured beremagene geperpavec because GEM‑3 had smaller wounds to close. But the committee noted that the company's approach did not use data from GEM‑3, which it decided was the most robust source of clinical-effectiveness data. It agreed that excluding duration of wound closure was a limitation of the EAG's approach, meaning the ITC likely underestimated the full treatment effect of beremagene geperpavec. The committee noted the limited clinical evidence and that the ITC was an anchored comparison of available phase 3 data. It noted that the different methods of assessing wound closure between trials (with stricter closure thresholds requiring confirmation in the GEM trials) may also have resulted in an underestimate of the relative treatment effect of beremagene geperpavec with the EAG's approach. But it agreed that the EAG's ITC did provide a comparison of EASE and GEM‑3. Despite the limitations, the committee concluded that it was a useful comparison of birch bark extract and beremagene geperpavec.

The company's model after technical engagement consisted of a state transition model with 7 health states to model the disease course of DEB. Beremagene geperpavec was compared with best supportive care and birch bark extract. Both beremagene geperpavec and birch bark extract were assumed to be used as add-ons to best supportive care. The company model after technical engagement aligned with that used in HST28. The model was based on the body-surface-area percentage (BSAP) score, which is a way of measuring DEB disease severity. Health states 1 to 6 were defined by taking the BSAP score range reported for EASE and dividing it up equally between 6 states. The higher BSAP score health states represented more severe disease and a seventh health state represented death. The committee noted that BSAP only describes the amount of the skin surface covered in wounds. It does not capture the impact or location of wounds, and does not describe other aspects of the condition, such as damage to the gastrointestinal tract. In each of the 13-week cycles, people in the modelled cohort could move between any of the health states and to the absorbing death state. Movement between the health states was determined by transition probabilities (see section 3.14). Transition probabilities, risk of death and drug use were calculated according to age (0 to 6 years, 6 to 18 years and 18 years and over) and aggregated for the whole cohort. The model also included an increased risk of death for people with RDEB based on rates from Petrof et al. (2022). Beremagene geperpavec was assumed to be used for life but 1% of people were modelled to stop treatment every year (see section 3.17). The committee concluded that the company's model after technical engagement was appropriate for decision making.

The company derived the baseline distribution in each health state from Bruckner et al. (2020) using the percentage of body surface area covered by wounds in the RDEB and DDEB population. Bruckner et al. defined 3 severities, based on the body surface area covered by open wounds: mild (BSAP less than 10%), moderate (BSAP 10% to 30%) and severe (BSAP over 30%). The EAG explained that these definitions did not fully align with those in the company's model. So, it preferred to use the baseline distribution by health state from EASE. The committee noted that this aligned with the EAG's preferred source of clinical-effectiveness evidence for birch bark extract in the ITC (see section 3.10). So, to best align with the available clinical evidence, it agreed that the EAG's approach using the EASE baseline distributions by health state was appropriate for decision making.

The company assumed that 1% of people would stop treatment with beremagene geperpavec every year. Because no one stopped treatment in GEM‑1 and GEM‑3, this was based on the stopping rate used in HST28. The company applied the same stopping rate of 1% annually for birch bark extract across all cycles, except for the first cycle in which a stopping rate of 8.3% was applied. The company assumed that people stopping beremagene geperpavec or birch bark extract would transition to health states based on the distribution observed in the best supportive care arm. The committee was concerned that the long-term stopping rate of 1% for beremagene geperpavec was low and unsupported by evidence. It noted that varying this assumption had a large impact on the cost-effectiveness results. It agreed that scenarios assuming alternative long-term stopping rates should be provided at consultation as well as further rationale and data for these scenarios. The committee noted that people could also stop treatment because they had chronic wounds that did not respond to treatment, with stopping rates for beremagene geperpavec estimated in the SEE (see section 3.15). The committee recalled the uncertainty associated with the SEE estimates (see section 3.15). It noted that GEM‑3 may provide evidence on short‑term response rates with beremagene geperpavec (67% of wounds closed at 6 months and 71% at 3 months). It thought that the company should provide further details at consultation about the modelling of responders and non‑responders and consider whether the available trial data could be used to inform the modelling. The committee concluded that additional justification for the stopping assumptions with beremagene geperpavec should be presented at consultation. This should include why people stop beremagene geperpavec, long-term annual stopping rates and initial stopping because of non-response to treatment, as well as an exploration of alternative data sources.

The company emphasised the lack of trial data to inform mortality estimates for people with DEB. Because of this, it used mortality estimates from Petrof et al. (2022) for people with RDEB. Different survival estimates for severe and non-severe RDEB were applied to the modelled health states based on the BSAP affected. Long-term mortality was extrapolated using a Gompertz curve in the company's base case after technical engagement. By reducing wound burden (that is, moving people to health states associated with a lower risk of death), the company modelled a mortality benefit for beremagene geperpavec compared with the comparators. The committee noted that 37% of the modelled population had DDEB and 63% had RDEB. It noted that general population mortality was assumed for people with DDEB, making subgroup distribution important for overall life years. The committee recalled that the proportion of people with each subtype who would have beremagene geperpavec in clinical practice was highly uncertain (see section 3.9). It noted that the EAG had explored a scenario in which the excess mortality associated with RDEB was applied to the entire modelled population, which effectively assumed that everyone had RDEB. But the committee found this assumption unrealistic, because some people with DDEB would also be expected to have beremagene geperpavec in clinical practice. So, the committee concluded that, during consultation, the impact of varying the proportions of people with RDEB and DDEB should be examined. These analyses should be done alongside an assessment of any limitations or loss of robustness associated with modelling these subgroups separately.

The committee noted that the licence for beremagene geperpavec limited the maximum dose permitted by age to:

• a total of 1 ml (2 syringes) to treat 100 cm² of open wounds in children 3 years old and under

• a total of 2 ml (4 syringes) to treat 200 cm² of open wounds in people over 3 years old.

The company assumed that everyone having beremagene geperpavec would have it at home after their first dose. At the first administration, the company assumed a one-off cost for 30 minutes of nurse time, which would be used to train people on self-administration. Costs for homecare delivery were then included for weekly doses. NHS England comments submitted at technical engagement suggested that the company's current administration costs were considerably underestimated, because:

• initial estimates using standard NHS England homecare delivery charges for aseptic preparation and delivery (which may be underestimated because of compounding of gene therapies) were considerably above the company's modelled costs

• based on current experience with beremagene geperpavec, people would likely need more training and clinical time than modelled for initial training in self-administration.

The summary of product characteristics for beremagene geperpavec states that 1 vial of 1-ml extractable volume of suspension provides a total of 2.5 ml when mixed into gel before administration. This constitutes 4 syringes of 0.5 ml when mixed into gel before administration, covering a maximum open wound area of 50 cm². The company assumed that the remaining 0.5 ml of beremagene geperpavec would be wasted because of potential remaining pockets of air during the preparation phase. So, 1 vial provides the maximum licensed dose per week for people aged over 3 years (see section 3.20). The company recalled that children 3 years old and under and those with wound areas under 60 cm² would have lower doses (see section 3.20). So, the company considered that there was potential for vials to be shared among these people. In its base case, the company assumed that vial sharing would occur for 80% of people having beremagene geperpavec. It supported this assumption with clinical expert feedback and UK data showing the plausibility of vial sharing for oncology drugs. It also provided beremagene geperpavec specific data from France that suggested current vial sharing levels of 35%. But the EAG questioned the plausibility of vial sharing in NHS clinical practice. It noted that the current stability data stated that a syringe kept under refrigeration should be used within 7 days if mixed under aseptic conditions and within 24 hours otherwise. It considered that the company had assumed that everyone having beremagene geperpavec would have treatment at home after their first dose (see section 3.21). It agreed that, with home administration, vial sharing was unlikely to be possible because mixed syringes would need to be transported to different people to allow use within the stability timeframe. The company emphasised a proposed licence variation based on new data that suggested beremagene geperpavec could be stored in a domestic freezer for 42 days. It considered this would enable vial optimisation for a high proportion of people because they could keep a long-term supply at home. But the committee recalled NHS England input that stated that most people would be on the maximum dose of beremagene geperpavec, which was 1 whole vial (see section 3.20). It noted that the company's and EAG's base cases both assumed vial wastage of 20% (0.5 ml), and considered a company scenario that assumed reduced wastage. The committee recalled the uncertainty around the feasibility of home administration and its preference to assume that all people would have beremagene geperpavec in specialist centres (see section 3.21). It agreed that, without further clarity on the feasibility of home administration, the potential for vial sharing was highly uncertain. It stated that the company should quantify the impact of vial sharing and drug wastage assumptions when outlining its proposed homecare administration model. Also, the potential implications of any future licence variation in relation to storage requirements should be clearly reflected in both the anticipated benefits and associated risks. The committee concluded that the company should submit additional information at consultation to clarify the feasibility of vial sharing and wastage in clinical practice.

In the company's base case after technical engagement, birch bark extract tube usage was estimated using the PLS. The EAG noted that the model predicted extremely high birch bark extract usage with an average 927 tubes per person in year 1, reducing to 516 tubes by year 10. The patient expert explained that, because birch bark extract is administered at each dressing change, higher usage was expected compared with beremagene geperpavec, which is administered weekly. The clinical expert also supported the plausibility of a modest reduction in use over time, because some wounds close with birch bark extract treatment (see section 3.5). But the EAG was concerned that the company's predicted tube usage was considerably higher than the mean 29.67 tubes used per month in EASE. It preferred to use this value in its base case, with no reduction over time. The committee agreed that the exact number of birch bark extract tubes used in clinical practice was uncertain. But it found the EAG's assumption to be more appropriate because it was based on direct trial evidence and aligned with its preferred source of clinical-effectiveness data for birch bark extract. The committee concluded that the birch bark extract usage rate from EASE was preferred for decision making.

The summary of product characteristics permits beremagene geperpavec to be used on wound areas of up to 200 cm² per week. The committee acknowledged the clinical experts' advice that some people would use birch bark extract at the same time, to treat small and superficial wounds (see section 3.5). At technical engagement, the company stated that modelling the combination of beremagene geperpavec and birch bark extract would require a substantial restructuring of the existing model. It also emphasised a lack of data on combination use, even though there are countries where both treatments are commercially available.. So, it decided that incorporating such a scenario would introduce significant additional uncertainty. The EAG supported this. It explained that the lack of robust evidence would mean that estimating clinical effectiveness would rely on applying a PLS approach, which it considered methodologically flawed and prone to bias. The committee recognised that the exclusion of combination use from the company's model was a key uncertainty and was associated with additional costs that could be substantial. It thought that the company could, in theory, have included birch bark extract in the PLS for use on untreated wounds in people having beremagene geperpavec. But it decided any attempt to model this scenario would be highly uncertain because of:

• the complexity of modelling combination treatment

• the lack of data on combination treatment's effectiveness and

• uncertainty about which treatment would be used for which type of wound.

The committee recalled that DEB is associated with both skin-related and non-skin-related symptoms (see section 3.1). The company assumed that reducing wound burden lowers the prevalence of complications, and so beremagene geperpavec improves outcomes by keeping people in milder health states. Because of this, it considered costs for clinical manifestations in the model. Rates of developing clinical manifestations that are reversible (anaemia, infections and osteoporosis) and irreversible (complete or partial mitten deformity of foot, complete mitten deformity of hand, and squamous cell cancer) were modelled by health state. This was based on rates reported in people with RDEB in Eng et al. (2021). The EAG advised that the relationship between wound burden and reduction of clinical manifestations was highly uncertain. It noted the limited supporting evidence, particularly given that beremagene geperpavec is a localised treatment. It also emphasised that the company's modelling of irreversible clinical manifestations lacked face validity. This included a higher risk of squamous cell cancer for people using birch bark extract than those having best supportive care alone. So, the EAG's base case excluded costs for clinical manifestations. The clinical experts at the committee meeting stated there was no evidence that beremagene geperpavec has a systemic effect on the body. But they noted that improved wound healing could reduce inflammation-related complications such as anaemia, osteoporosis, and gastrointestinal issues linked to malnutrition, particularly when treatment is started early. The patient experts emphasised that improvements in nutrition and sleep may reduce corneal abrasions, which often occur during nighttime wakings triggered by itching or pain. Better wound closure may also allow time and resources to be redirected to other affected areas (such as the mouth, oesophagus and eyes), supporting additional healing. They also explained that beremagene geperpavec may help maintain functionality of hands and feet by reducing scarring after surgery for mitten deformities, the benefits of which can be short-lived because of the high rate of wound reopenings. The committee thought it was plausible that there would be some benefits for beremagene geperpavec beyond the direct impact on the skin. But the extent of this impact was highly uncertain and not supported by evidence. So, the committee concluded that the model should exclude the cost of clinical manifestations. But it noted that some benefits of using beremagene geperpavec, including functional benefits after surgery, were likely not captured in the model.

The company included costs for concomitant medications used as best supportive care based on prevalence data from Eng et al. (2021). This estimated concomitant medication usage by severity. The EAG noted that statistically significant differences between mild and severe cases were only reported in Eng et al. for opiates, gabapentin and MediHoney. All other medications showed no significant association with severity. Given this, the EAG excluded costs for concomitant medications in its base case. The committee agreed that this exclusion was appropriate for decision making.

Because of confidential commercial arrangements for beremagene geperpavec and the comparators, the exact cost-effectiveness estimates are confidential and cannot be reported here. The company's base-case incremental cost-effectiveness ratio (ICER) for beremagene geperpavec was dominant compared with birch bark extract (lower overall costs and higher QALYs). Compared with best supportive care it was within the threshold normally considered a cost-effective use of NHS resources. The committee recalled that clinical-effectiveness data in the company's model came from the PLS. But the EAG preferred to use the EASE transitions with the relative risk from the ITC applied for beremagene geperpavec (see section 3.16). The committee noted that changing this assumption had a large impact on this ICER. This meant that the EAG's base case was considerably over the threshold considered an acceptable use of NHS resources when comparing beremagene geperpavec with best supportive care or birch bark extract.

The committee noted that some potential benefits of beremagene geperpavec may not have been included in company's model, including:

• some children with DEB are born with damage to the skin, the severity and persistence of which result in limiting functions including development of mitten deformities which limits the ability to walk (see section 3.2). The clinical experts explained that early wound closure with beremagene geperpavec has the potential to reduce severe scarring and improve the likelihood of maintenance of hand function and walking in children with severe skin damage in infancy and childhood

• beremagene geperpavec may be part icularly beneficial when used with or after surgery to treat mitten deformities or squamous cell cancer. The clinical experts stressed that reducing post-surgical scarring would help maintain functionality of hands and feet and increase the longevity of the benefits. This is especially true for mitten deformities, because the high rates of wound reopening mean the benefits are short term (see section 3.2). The committee also recalled that wounds are most likely to reopen around points of friction such as the knees, elbows, wrists and ankles, which can limit mobility (see section 3.2). The committee noted that, by promoting wound closure at these points, people may retain mobility and hand function

• treatment effect was modelled to be constant across age groups. But the treatment effect of beremagene geperpavec was likely larger in babies and young children because they have remaining stem cells for wound closure, less scarring, fewer chronic wounds and less fragile skin (see section 3.2). But this benefit is partly offset in older people with harder-to-treat wounds, so the size of any uncaptured benefit was uncertain

• the committee's preferred source for clinical-effectiveness data for beremagene geperpavec (the EAG's ITC) did not directly capture the duration of wound closure. This was because this data was not collected in GEM‑3 (see section 3.11). Unlike birch bark extract, beremagene geperpavec directly targets the underlying cause of DEB. So, some incremental benefit in the duration of wound closure is likely

• wound management can be highly burdensome for people with DEB and their carers, often requiring many hours of care each day (see section 3.3). So, by reducing the area of open wounds, beremagene geperpavec may reduce the time and resources needed for daily wound care. The committee noted that both the company's and the EAG's base cases included cost savings associated with reduced bandage use. But it agreed that the potential quality‑of‑life benefits associated with reduced wound-care burden were unlikely to have been fully captured.

NICE's technology appraisal and highly specialised technologies guidance manual notes that, above a most plausible ICER of £25,000 per QALY gained, judgements about the acceptability of a technology as an effective use of NHS resources will take into account the degree of certainty around the ICER. The committee will be more cautious about recommending a technology if it is less certain about the ICERs presented. But it will also take into account other aspects including uncaptured health benefits. The committee noted the high level of uncertainty, specifically about:

• the proportion of people with DDEB who would have beremagene geperpavec in clinical practice compared with people with RDEB, and how these people are reflected in the clinical evidence

• the clinical effectiveness of beremagene geperpavec compared with birch bark extract and best supportive care

• how the model captured the natural history of the condition

• the proportion of people stopping treatment with beremagene geperpavec in the model

• vial usage for beremagene geperpavec including home administration, vial sharing and wastage

• the impact of using beremagene geperpavec and birch bark extract on different wounds at the same time.

For the model assumptions, the committee preferred to:

• use the baseline distribution by health state from EASE (see section 3.13)

• use the EAG's approach to relative effectiveness (the transitions from EASE for birch bark extract and best supportive care, with the relative risk from the EAG's ITC applied to the birch bark extract transitions for beremagene geperpavec; see section 3.16)

• use the EASE data to inform the tube usage for birch bark extract (see section 3.23)

• apply the costs from NHS England for administration by 2 nurses in a day care unit for all beremagene geperpavec administrations (that is, assume no home administration for beremagene geperpavec; see section 3.21)

• exclude the costs for clinical manifestations of DEB (see section 3.25)

• exclude the costs of concomitant medications (see section 3.26)

• add a severity weighting of 1.2 to the QALYs (see section 3.27).

The committee noted that there was considerable uncertainty surrounding the cost effectiveness of beremagene geperpavec for wounds associated with DEB. It agreed that during consultation the company should:

• provide details on the population who would have beremagene geperpavec in clinical practice, including:

  • the proportion of people with RDEB and DDEB having beremagene geperpavec in clinical practice and scenarios exploring any alternative plausible distributions

  • the criteria for identifying people with DDEB for treatment, including any wound-burden thresholds for treatment

  • the differences between baseline characteristics of people with DDEB in clinical practice and the GEM populations (see section 3.9)

  • an assessment of any limitations or loss of robustness associated with modelling subgroups separately (see section 3.19)

• attempt to address the limitations associated with the committee's preferred approach, which is to use the EAG's ITC for the source of clinical-effectiveness data (see section 3.16)

• make substantial corrections to the PLS if it remains the company's preferred source of clinical-effectiveness data, including:

  • using the PLS directly to generate cost-effectiveness outcomes, eliminating the need for a separate Markov model

  • allow for increases in BSAP affected by wounds from baseline

  • exploring alternative sources for wound dynamic data to ensure clinically plausible outputs, including addressing the committee's and EAG's face validity concerns (see section 3.18)

• provide data on beremagene geperpavec by age and ensure the model accurately captures the natural history of the condition over time

• provide additional justification for stopping assumptions with beremagene geperpavec, including:

  • further details on the modelling of responders and non‑responders and exploring use of the available trial data to model response rates

  • scenarios assuming alternative long-term stopping rates (see section 3.17)

• explore alternative approaches to modelling vial usage for beremagene geperpavec, including a scenario in which everyone is assumed to have the maximum dose with no reductions over time (see section 3.22)

• provide a breakdown of the costs associated with home delivery and confirm with NHS England how these would be funded, including the impact on vial sharing and wastage and any licence variation (see section 3.23)

• consider if vial sharing is possible when beremagene geperpavec is administered at home, and outline the rationale if assuming any vial sharing in this setting (see section 3.23)

• update the calculations informing the severity modifier to assess the impact of any new assumptions (see section 3.27).

Having concluded that beremagene geperpavec could not be recommended for routine use in the NHS, the committee then considered if it could be recommended for use during a managed access period for treating skin wounds associated with DEB. The committee noted that the company had not submitted a managed access proposal. It also noted that GEM‑3 has completed, but there is an ongoing post-authorisation study that may provide further data. The committee agreed that further real-world evidence data collection may help address some of the high uncertainty surrounding the outputs of the PLS. These include beremagene geperpavec vial usage in the NHS and concurrent use of birch bark extract and beremagene geperpavec (see section 3.20 and section 3.24). The committee concluded that a managed access proposal would be welcomed at consultation.

The committee considered that some people with skin wounds associated with DEB may meet the criteria for disability because of limited mobility from progressive scarring leading to hand or foot abnormalities. Disability is protected under the Equality Act 2010. The company emphasised that families caring for children with DEB often face substantial financial and practical burdens including high care costs, challenges accessing health and social care, and broader financial strain. This may be particularly pronounced in some ethnic minority groups. Stakeholders and a patient expert emphasised that restricting beremagene geperpavec to outpatient treatment at specialist centres could exacerbate inequalities. This is because the need to travel regularly for treatment may increase time and cost burdens, particularly for people living far from centres, on low incomes, or facing language barriers. The committee considered these issues. But, because its recommendation applies to the whole population in the license for beremagene geperpavec, it agreed these were not potential equalities issues that could be addressed by this evaluation. The committee concluded that all equalities issues for beremagene geperpavec had been considered in its decision making.

The committee noted the important uncertainties in the cost-effectiveness evidence. This meant it was not possible to reliably estimate the cost effectiveness of beremagene geperpavec. The most likely cost-effectiveness estimates were substantially above the threshold considered a cost-effective use of NHS resources. So, beremagene geperpavec should not be used. The committee concluded that the company should provide additional information for consideration at the next evaluation committee meeting to:

• reduce the very high levels of uncertainty in the analysis

• enable plausible model outputs to be produced and

• allow a more accurate estimate of the cost-effectiveness of beremagene geperpavec.