Fosdenopterin for treating molybdenum cofactor deficiency type A

Molybdenum cofactor deficiency (MoCD) type A type is a rare genetic condition that can appear shortly after birth, with symptoms appearing within 28 days of birth for most people. The condition is caused by defects to the gene that makes molybdenum cofactor. Without molybdenum cofactor, an enzyme called sulfite oxidase does not function properly to process sulfites. This causes the build-up of sulfites to toxic levels in the body and in particular the brain, leading to seizures and severe brain abnormality. Without treatment, people who have the condition have a median survival of around 4 years.

The company classified 2 types of MoCD type A: early and late onset. It noted that early-onset MoCD type A occurs within the first month of life, and can cause more severe with symptoms such as seizures and feeding difficulty. Late-onset MoCD type A normally occurs within the first 2 years of life, and usual symptoms include developmental delay and uncontrolled muscle movements. The clinical expert noted that the aforementioned use of early and late onset is misleading, and that classifying the condition based on a strict cut-off of 28 days from birth may not be appropriate (see section 3.9). They highlighted that late-onset MoCD type A is uncommon. They also explained that clinicians prefer the terms typical and atypical MoCD type A because these capture the nature of the condition more precisely. They added that typical MoCD type A symptoms usually present within 28 days of life, but noted that some people may have sustained brain injury before birth. The clinical expert cited a recently published consensus guidelines for MoCD type A, which describe how fetal seizures may be noticed during late pregnancy as 'increased hiccupping'. Other symptoms of typical MoCD type A include acute encephalopathy (brain disorder), feeding difficulties and irritability. They explained that the usual symptoms of atypical MoCD type A include loss of muscle control and movement disorder, but that seizures and ongoing brain damage may also occur.

The clinical expert highlighted that they were aware that around 70 people have or have had MoCD type A in the UK over about 30 years. They also said that that the rate of MoCD type A has been increasing slightly in more recent years. The committee acknowledged the clinical expert's opinion and concluded that describing MoCD type A as typical and atypical was a reasonable approach. It thought that it was not appropriate to apply a strict cut-off of 28 days from birth for people with typical MoCD type A.

Patient experts noted in their submissions that MoCD type A markedly affects the daily life of people with the condition and their carers. They highlighted that it was hard having to adjust their lives to becoming full-time carers and learning how to give the various treatments needed throughout the day. The patient experts noted that, often, people would need to take numerous medicines to control their symptoms (including for seizures). They highlighted that, while some people may not have seizures, others can have 10 to 60 seizures per day. They also noted that some children would not be able to have food by mouth, so would need have nasogastric tubes fitted to have food. The committee heard that:

• sometimes people with MoCD type A may not need antiseizure medicines or nasogastric feeding

• some children with MoCD type A:

  • may be non-verbal

  • may not have full control of their bowels and may need a nappy beyond the expected age

• children with MoCD type A would need:

  • help with dressing

  • cleaning of central lines needed for some treatments (for example, fosdenopterin).

There are no treatment options for MoCD type A aimed at targeting the underlying condition. Standard of care is aimed at managing symptoms such as seizures and loss of muscle control. The company proposed that fosdenopterin would be started when people are suspected of having MoCD type A (a presumptive diagnosis). Then, a genetic test would be done to confirm if they have the condition. People confirmed as having MoCD type A would continue having fosdenopterin and people with a negative genetic test would stop treatment. The clinical expert noted that people diagnosed early and started on fosdenopterin treatment early could be prevented from the severe neurodisability and brain damage associated with MoCD type A. They explained that the treatment restores normal biochemistry and, if used early enough, can lead to better outcomes. They added that not every person necessarily needs antiseizure treatment. But brain damage can occur before treatment, so long-term symptoms may occur regardless of fosdenopterin treatment. The clinical expert highlighted that speed of diagnosis is dependent on familiarity with MoCD type A in a particular hospital. If a hospital is familiar with the condition, biochemical tests can be done within 24 hours to make an initial decision. This means that starting treatment the next day is possible. Then, genetic confirmatory test can then done afterwards. If the hospital is not familiar with MoCD type A, tests to diagnose it may be the second- or third-line tests done. This can delay diagnosis. The committee concluded that fosdenopterin provides a treatment option for people with MoCD type A and speed of diagnosis is an important factor for outcomes.

NICE's final scope for this evaluation included established clinical management without fosdenopterin as the relevant comparator. The committee understood that this largely constituted antiseizure medicines. The company explained that it had included the 10 most common antiseizure medicines used in its clinical trials as the relevant comparators (standard care). The committee understood that fodenopterin would be used in addition to current standard treatment, including antiseizure medicines, and that fosdenopterin is a life-long treatment. The committee recalled that sometimes some people may not need antiseizure medicines (see section 3.2). It asked the clinical expert if seizures are a common occurrence in this population. The clinical expert explained that most people with MoCD type A need treatment for seizures, but that seizures may not necessarily present immediately. The committee concluded that the company's proposed description of standard care was appropriate.

The clinical evidence submitted by the company came from 3 fosdenopterin trials (MCD‑501 [4 people], MCD‑201 [8 people], and MCD‑202 [3 people]). These were pooled and compared with a natural history study (MCD 502 [37 people]). The trials included people with MoCD type A. The company provided results for both the full analysis set (FAS, that is, all people having or not having treatment) and the genotype-matched analysis set (GMAS, that is people having treatment genotype-matched to people not having treatment). The key outcome measure reported by the company was overall survival. For the FAS, the median overall survival was not reached for the fosdenopterin group but was 50.7 months for people having standard care. The company reported better survival probability for people having fosdenopterin compared with standard care (hazard ratio [HR] 5.5, 95% confidence interval [CI] 1.44 to 21.04). For the GMAS, the median overall survival for fosdenopterin was not reached and was 47.8 months for people having standard care (HR 7.1, 95% CI not reported). The company submitted additional results on feeding status. These suggested that, compared with placebo, fosdenopterin increased the time before non-oral feeding was needed (75 months, 95% CI 0.2 to not estimable compared with 10.5 months, 95% CI 4.9 to 53.6). The company also reported results of seizure control. It reported that the odds ratio of seizures being absent or resolved compared with being controlled or present was 1.216 (95% CI, 0.337 to 4.387) for fosdenopterin compared with standard care in the FAS. For the GMAS, the odds ratio was 1.461 (95% CI, 0.368 to 5.808). The committee noted that the impact of fosdenopterin on seizures was highly uncertain because the confidence intervals were wide and they crossed 1. The EAG explained that the primary issue related to the small trial population. It was also concerned about the high number of censoring events in the overall survival results, which could have made interpreting the trial results more difficult. The committee noted that very few people in the clinical trials (fosdenopterin: 1 of 15, standard care: 4 of 37) had late-onset MoCD type A. It also noted that there was limited evidence for the use of fosdenopterin in this population (see section 3.9). The committee concluded that the size and nature of the trials the results were uncertain but suggested that fosdenopterin could improve survival outcomes compared with standard care.

The EAG noted that the fosdenopterin trials were all single-arm trials. This increased uncertainty, given that the population included in the clinical trial was already small. It also noted difficulty understanding the flow of people during the trials. It was unclear whether people were presumptively diagnosed with MoCD type A and started on treatment with fosdenopterin before having a confirmatory genetic test (see section 3.13). The EAG highlighted that the company used results from 2 different data cuts, July 2019 and October 2021, for its economic and clinical analyses. The company explained that it did not have full access to the individual patient level data because the clinical trials were done by a different company. The EAG noted that the population addressed in the company's model was in babies, children and young people with MoCD type A. But it was concerned that 1 person in the comparator arm was diagnosed at month 484 (40 years). This meant that, before their diagnosis, they had the same risk of death and ability to feed orally as the general population, which could have biased the results. The committee had concerns about how generalisable the fosdenopterin trials were to UK clinical practice. It asked the clinical expert if they considered the trials were generalisable. The clinical expert noted that the trials were not reflective of the UK MoCD type A population. This was because, in their experience, about 90% of people with the condition are from South Asian ethnic backgrounds. But, in the trials, around 70% of people who had fosdenopterin were from White ethnic backgrounds. But the clinical expert said they had not seen any evidence to suggest that treatment efficacy would differ between ethnic groups. The committee understood that MoCD type A is linked with a high incidence of consanguinity rather than with being from a South Asian ethnic background (see section 3.19). The clinical expert explained that the number of people who had seizures (fosdenopterin 71%, standard care 92%) and feeding difficulties (fosdenopterin 64%, standard care 84%) at baseline was similar to that in UK clinical practice. The committee concluded that the generalisability of the trials was uncertain but, in the absence of further data, the trials were sufficient for decision making.

The company submitted a 2‑state survival model capturing people alive or dead. It did so because it considered improvement in overall survival to be the primary benefit of treatment with fosdenopterin. It also expected treatment to stabilise seizure, reduce the need for nasogastric feeding and to stabilise mobility. But the company noted that, because of a lack of data, it could not capture these additional benefits in distinct health states. The model applied a lifetime horizon of 100 years, a cycle length of 4 weeks, and discounts costs and quality-adjusted life years (QALYs) at 3.5%. Utilities from Dravet syndrome (a severe form of epilepsy that starts in childhood) was applied in the model as a proxy for MoCD type A utilities (see section 3.10). The committee concluded that the company's model was sufficient for decision making but highlighted that there was substantial uncertainty associated with it because it:

• did not capture all the relevant outcomes for MoCD type A (see section 3.8)

• used a proxy condition as a source of utility values (see section 3.10).

The EAG raised concerns about the simplicity of the company's model. It noted that the model was based on overall survival and did not capture all the relevant outcomes for MoCD type A, including developmental status, mobility, feeding status and seizure. So, the impact of fosdenopterin on these outcomes could not be tested in the model. The EAG explained that it had requested further data from the company at clarification including data for people's ability to feed orally. It used this data to make changes to the model. The EAG preferred to assume that:

• people having fosdenopterin and feeding orally would have similar outcomes to the general population

• people unable to feed orally would have similar outcomes to the standard-care population.

In its base case, the company considered fosdenopterin for everyone, that is, both the early- and late-onset populations (see section 3.1). It defined early-onset MoCD type A as being when symptoms start within 28 days of birth. The EAG noted that, in the included trials, most people in the fosdenopterin (14 of 15) and standard-care groups (33 of 37) had early-onset MoCD type A. It explained that including the 1 person with late-onset MoCD type A (1 of 15) who had fosdenopterin did not affect the model results. The EAG highlighted that more late-onset data was needed to be able to model this group. So, its base case only included the early-onset population. During technical engagement, the company submitted a case study (Lund et al, 2024) that discussed treatment of 2 people with late-onset MoCD type A who had treatment with fosdenopterin. The company noted that people with late-onset MoCD type A are less likely to have sustained brain damage and are more likely to respond to treatment. The EAG highlighted that the case study may provide supportive evidence for using fosdenopterin in late-onset MoCD type A. But it thought that, the company's model still did not include further data to support this. The committee asked the company whether it could describe the outcomes of the person with late-onset MoCD type A who had fosdenopterin in the clinical trials. The company explained to the committee that this person had developmental delay in their second year of life, which led to diagnosis and treatment with fosdenopterin. The clinical expert explained that there was limited data to provide evidence for this group. They added that they are aware of 1 person with late-onset MoCD type A who started treatment on day 55. But they had no evidence to quantify the treatment benefit for this person. The committee recalled the EAG's concern that 1 person in the standard-care group had been diagnosed at age 40 years. Before that, they could have had the same risk of survival as the general population (see section 3.6). It was concerned that limited evidence had been presented to fully support the efficacy of fosdenopterin in the late-onset population defined by the company (see section 3.5). The committee concluded that the model population should reflect the efficacy data, that is, the early-onset population.

The company did not collect quality-of-life data in any of its clinical trials. It used data from a study of quality of life in people with Dravet syndrome (Lagae et al. 2018) to generate proxy sex- and age-adjusted utility values for its modelling. Dravet syndrome is a severe form of epilepsy that typically starts during childhood. The company explained that it did this because:

• no data on quality of life for MoCD type A was available

• Dravet syndrome represented a proxy seizure-based condition that was relevant and provided data for children of similar age range.

In its base case, the company used its calculated Dravet syndrome utility values for the standard-care arm (see section 3.10). For the fosdenopterin arm, it assumed that, after treatment for 1 year, people would have the same quality of life as the general population. It did so because it expected people having early treatment with fosdenopterin to have comparable long-term utilities with the general population. The EAG disagreed with this approach. It noted that no evidence was submitted by the company to support its assumption. Also, it was concerned that the utility values for people having fosdenopterin were overestimated. The EAG preferred to differentiate quality of life based on ability to feed orally (see section 3.8). It assumed that people having fosdenopterin:

• who are able to feed orally would have utilities midway between the standard-care population and general population

• who are unable to feed orally would have the same utilities as the standard-care population.

The company applied a caregiver disutility (-0.14), in its base case. It also assumed that people having standard care would need 1.8 carers throughout their life to give care corresponding to 14.8 hours per day. It took these estimates from NICE's technology appraisal guidance on fingolimod for the treatment of highly active relapsing–remitting multiple sclerosis and on fenfluramine for treating seizures associated with Dravet syndrome. For people having fosdenopterin, the company assumed that they would only need 1 carer until age 5 years, and none after that. The EAG noted that the company did not fully capture care in places such as special needs school. It highlighted that this should be included within the scope for personal social services. The EAG also noted that fosdenopterin is given intravenously through a catheter every day. It added that some people having treatment would be unable to feed orally beyond age 5 years, so would likely need care after this age. The company responded that people who have treatment early would have care needs similar to those of the general population. It added that any care would fall within regular parenting hours. The clinical expert explained to the committee that people who can feed orally may need less support but would likely need some support beyond age 5 years. The committee understood that the carers of some people with MoCD type A thought that their children would always need care. In some of these children, diagnosis was as early as 3 days after birth, and the children had started having treatment by day 4. The committee was aware that the EAG's base case assumed a differentiated care burden for people having fosdenopterin based on their ability to feed orally. People feeding orally were assumed to need less care (1 carer) until age 18 years and none after that. But people unable to feed orally were assumed to need care similar to that of the standard-care population (1.8 carers). The committee concluded that the EAG's approach of differentiating carer needs by feeding status was plausible.

Fosdenopterin can be used for people suspected of having MoCD type A before the condition is confirmed by genetic testing. The company's model assumed that people in the fosdenopterin and standard-care groups had already had their condition confirmed. The EAG noted that the company's model did not capture the cost of presumptive treatment, that is, from the time of starting treatment but before a genetic confirmation of the condition. The company explained to the committee that the presumptive treatment period is not substantial because fosdenopterin is a life-long treatment. It noted that it had done some analysis to explore the impact of presumptive treatment costs and that this increased the incremental cost-effectiveness ratio (ICER) by about 1%. The EAG noted that it was unclear from the company's submission what proportion of people having treatment presumptively would later have a negative genetic test result. It did a scenario analysis exploring that 10% to 95% of people would have a negative test after presumptively starting treatment. The EAG analysis assumed people would have presumptive treatment for either 7 or 28 days. The committee noted that the EAG's analysis represented the maximum fosdenopterin acquisition cost that the NHS would be expected to cover. It asked the clinical expert what proportion of people who start treatment would later have a negative test. The expert did not have a precise figure but estimated that around 50% was reasonable. They explained that confirmation of MoCD type A could often be within 2 weeks but there could be some delays. The committee recalled that the speed of genetic testing can be affected by a hospital's familiarity with MoCD type A (see section 3.3). So, it preferred to assume that people would have presumptive treatment for at least 28 days to account for differences in familiarity of the condition. The committee concluded that the cost of presumptive treatment should have been captured in the company's model for 28 days, using the most reliable data if available, or the clinical expert's estimates (50%).

Fosdenopterin vials (9.5 mg) need to be used within 4 hours once open. The dose used by people is individualised based on their weight. This means any unused vials would need to be disposed of. The EAG noted that, because of the lack of options in vial sizes and the storage needs, a substantial amount of the treatment would be wasted throughout a person's lifetime. It estimated that in the first 5 years of a person's life, around 37% of fosdenopterin would be wasted. The company responded that it did not have any plans to introduce smaller vial sizes. The committee was aware that the model already captured the cost of treatment wastage. It concluded that having smaller vial sizes would have reduced wastage and markedly improved the cost-effectiveness estimates.

NICE's health technology evaluations manual specifies that a most plausible ICER of below £100,000 per QALY gained for a highly specialised technology is normally considered an effective use of NHS resources. For a most plausible ICER above £100,000 per QALY gained, judgements about the acceptability of the highly specialised technology as an effective use of NHS resources must take account of the size of the incremental therapeutic improvement. This is revealed through the number of additional QALYs gained and by applying a 'QALY weight'. The committee noted that a weight of between 1 and 3 in equal increments can be applied when the QALY gain is between 10 and 30 QALYs. It noted that some of the company's and EAG's analyses showed QALY gains within this range. The committee understood that carer QALYs are not normally included in the QALY weighting calculation and that the company's and EAG's analyses also did not include this. The committee recalled that there were multiple uncertainties with the clinical data and the economic modelling. But it considered the high unmet need for people with MoCD type A. Taking into account its preferred assumption (see section 3.17), it concluded that a QALY weight of 1.005 should be applied.

The company and EAG differed on 3 further assumptions, which the committee was aware had a small to negligible impact on the cost-effectiveness estimates. The company assumed that people having fosdenopterin would only also have 1.0 antiseizure medicine. The EAG preferred to use an estimate of 2.2, which was the lower end estimate used in the Dravet syndrome systematic literature review submitted by the company. The clinical expert explained that MoCD type A is not characterised solely by seizures, but estimated that people with MoCD type A would have 2 to 3 antiseizure medicines, on average. The dose of fosdenopterin is calculated based on a person's weight and the company used weight estimates from 3 different sources combined. The EAG made adjustments to smoothen the weight data, which the company accepted during technical engagement. To reflect the improvement in developmental status after having treatment, the EAG also preferred to assume people having fosdenopterin would be in the twenty-fifth percentile of the weight distribution,. The company's preferred percentile is considered confidential and cannot be reported here. The company assumed that everyone with MoCD type A would need a review by a metabolic physician twice yearly. The EAG's clinical expert noted that such reviews would likely be specific to taking fosdenopterin so should be excluded from the standard-care group. The committee considered these assumptions and had the following conclusions:

• 2.2 antiseizure medicines should be applied based on the systematic literature review results and the clinical expert's opinion

• a twenty-fifth percentile weight distribution should be used for modelling weight to capture the improvement in having treatment with fosdenopterin

• a metabolic physician review should only be included for modelling fosdenopterin.

The exact cost-effectiveness estimates cannot be reported here because there are confidential discounts for fosdenopterin. Both the company's and EAG's base-case ICERs were substantially above the range that NICE normally considers an acceptable use of NHS resources, even when the QALY weighting was applied. The committee noted that there were several uncertainties, including with the model's ability to capture outcomes relevant to MoCD type A. The committee noted the uncertainty around the cost-effectiveness estimates. It thought that this could have been reduced by adapting the model to capture the expected correlation of people's outcomes (brain damage, seizure status and ability to feed orally) with economic outcomes such as utility values and carer burden (see section 3.2, section 3.8, section 3.11 and section 3.12)

Having concluded that fosdenopterin could not be recommended for routine use, the committee then considered whether it could be recommended with managed access for treating MoCD type A. It noted that the company had not submitted a managed access proposal, so it could not make a recommendation for managed access at this stage.

The clinical expert explained that almost everyone in the UK with MoCD type A are from minority ethnic backgrounds. They noted that there is a high incidence of MoCD type A in people from South Asian ethnic background (see section 3.6). The committee understood that MoCD type A is linked with high incidence of consanguinity rather than with a South Asian ethnic background. The committee noted that issues related to differences in prevalence cannot be addressed in a technology evaluation. In its submission, the company also noted that fosdenopterin could reduce carer burden and that caring may be disproportionately done by women. The committee considered this issue and recalled that it had taken carer burden into account in its decision making (see section 3.12). The patient experts' submission noted that families on low income may not have access to the medical-grade freezer needed to store fosdenopterin. The committee considered this issue but noted that its recommendation does not restrict access to treatment for some people over others. So, it agreed that this was not a potential equalities issue. The committee noted that it was not presented with evidence to support the improved impact of fosdenopterin for any of the groups highlighted.

The committee concluded that, with its preferred assumptions and QALY weighting applied, the cost-effectiveness estimates for fosdenopterin were substantially above the range considered an acceptable use of NHS resources for a highly specialised technology. So, it did not recommend fosdenopterin for treating MoCD type A.