Etranacogene dezaparvovec for treating moderately severe or severe haemophilia B

Haemophilia B is an X‑linked, congenital bleeding condition characterised by a deficiency of coagulation factor IX (FIX). It mainly affects men, but can affect women in rare cases. The severity of haemophilia B generally correlates with the level of FIX in the blood and is defined as either severe (FIX level below 1%), moderate (FIX level 1% to 5%) or mild (FIX level 5% to less than 40%). Moderately severe haemophilia B does not have a standard definition but is generally considered to be when the FIX level is 1% to 2%. The main symptom of haemophilia is prolonged bleeding but other complications include bleeding into joints and muscles without having had an injury. Patient experts explained that bleeds are not only physically painful but can also have a substantial psychological impact on people with the condition and their family. They often have anxiety or worry about their condition, causing great mental distress. The patient experts explained that FIX prophylaxis treatment for moderately severe or severe haemophilia B (see section 3.2) often requires self-infusion or infusion by carers as often as every 3 to 4 days, which is a substantial treatment burden. They added that this makes planning difficult, especially when travelling, and impairs the ability to be spontaneous. Because of the heavy treatment burden, 1 patient expert described a one-off treatment with etranacogene dezaparvovec, with potential to stop the need for regular FIX prophylaxis, as life-changing. Stakeholders noted that despite having regular prophylaxis treatment, people with severe haemophilia B still have painful bleeds, some of which require lengthy recovery periods, hospital visits and potential hospital stays. Over time these bleeds will lead to joint damage, pain and disability. A one-off treatment with etranacogene dezaparvovec could lead to fewer hospital visits, better joint health and lower rates of disability over time. The committee concluded that moderately severe or severe haemophilia B substantially affects health-related quality of life.

The clinical management of haemophilia B usually involves long-term FIX prophylaxis treatment and on-demand treatment with FIX concentrates. FIX prophylaxis treatment involves regular administration of standard half-life FIX concentrates (every 3 to 4 days) or extended half-life FIX concentrates (every 1 to 2 weeks) to prevent bleeding. On-demand treatment includes administration of FIX concentrates at the time of a bleeding episode or before events that may present a higher risk of bleeding. The company noted that, despite being eligible for regular FIX prophylaxis treatment, a small number of people with the condition opt to only have on-demand treatment because of personal choice or clinical challenges. The company proposed that etranacogene dezaparvovec would mainly replace FIX prophylaxis treatment but could also replace on-demand only treatment. The comparators included in the company submission were 4 FIX prophylaxis treatments available on the NHS:

• nonacog alfa (BeneFIX, standard half-life)

• eftrenonacog alfa (Alprolix, extended half-life)

• albutrepenonacog alfa (Idelvion, extended half-life) and

• nonacog beta pegol (Refixia, extended half-life).
  
  The committee concluded that FIX prophylaxis treatment was the most appropriate comparator.

The primary clinical-effectiveness evidence was from the HOPE‑B trial. HOPE‑B is an ongoing phase 3, open-label, single-dose, single-arm multinational trial evaluating etranacogene dezaparvovec in adult males with moderately severe or severe haemophilia B who had routine FIX prophylaxis treatment (n=54). HOPE‑B included a lead-in period (minimum 6 months) in which people had FIX prophylaxis treatment. After the lead-in period, people had etranacogene dezaparvovec. Because there was no control arm, outcomes assessed during the lead-in period were compared with outcomes in the post-treatment follow-up period. The company submission presented data up to 24 months after treatment with etranacogene dezaparvovec. At consultation, the company provided additional data for up to 36 months after treatment.

The HOPE‑B primary outcome is annualised bleeding rate (ABR). Several bleeding outcomes were reported, including various types of bleeds: all bleeds, joint bleeds, spontaneous bleeds and bleeds that needed FIX treatment. At 7 to 24 months after etranacogene dezaparvovec, results showed that:

• the adjusted ABR for all bleeding episodes decreased from 4.19 to 1.51, a reduction of 64% (p=0.0002)

• the adjusted annualised spontaneous bleeding rate decreased from 1.52 to 0.38, a reduction of 75% (p=0.0005)

• the adjusted annualised joint bleeding rate decreased from 2.35 to 0.46, a reduction of 80% (p<0.0001)

• the adjusted ABR for bleeds that needed FIX treatment decreased from 3.65 to 0.99, a reduction of 73% (p=0.0001).
  
  The committee noted that at 7 to 24 months post-treatment, 27 out of 54 people had bleeds (average of 2.7 bleeds per person). It noted that the average number of bleeds after treatment was not substantially different from the lead-in period (average of 3.4 bleeds per person). A clinical expert explained that this may be because people may need a period of relearning in the first couple of years after having etranacogene dezaparvovec, to differentiate between joint pains and bleeds. However, a scan would be needed to confirm whether the pain was because of a bleed. Therefore, it was possible that people in the trial recorded bleeds in their patient diary when they were actually experiencing joint pain. A key secondary outcome was the change in FIX level between baseline and the post-treatment period. At 24 months post-treatment, the least squares mean increase in endogenous FIX from baseline was 34.13 IU/dl (p<0.001). At consultation, the company provided additional follow-up data for 36 months of treatment that continued to show a reduction in bleeding rates (the exact results are confidential so cannot be reported here). The committee concluded that bleeding rates were lower after etranacogene dezaparvovec than during the lead-in period.

The EAG highlighted that the company did not report participants' FIX levels during the lead-in period, but instead estimated baseline FIX levels based on their historical haemophilia B severity. This approach meant it was not possible to compare FIX levels during routine prophylaxis treatment in the lead-in period with FIX levels after treatment with etranacogene dezaparvovec. The company said it used this approach because FIX levels would vary depending on the type, brand, dose and frequency of FIX prophylaxis treatment that people were having. It also noted that FIX levels fluctuate after FIX prophylaxis treatment so it would be challenging to identify a representative measurement. It added that a benefit of etranacogene dezaparvovec would be more stable FIX levels because of endogenous production (that is, the body producing its own FIX). The company believed that using a historical estimate of baseline FIX levels instead of actual measurements better represents endogenous FIX production in the lead-in period and leads to a fairer comparison between treatments. The EAG considered that it was not possible to determine how etranacogene dezaparvovec affects FIX levels without comparing FIX levels during FIX prophylaxis with levels after etranacogene dezaparvovec. It added that understanding the change in FIX levels after treatment with etranacogene dezaparvovec would corroborate the other clinical outcomes, and show how etranacogene dezaparvovec reduces bleeds (for example, by increasing FIX levels). A clinical expert said that FIX levels tend to fluctuate after FIX prophylaxis treatment, and that the risk of bleeds is particularly high when FIX levels are low. They added that etranacogene dezaparvovec treatment stabilises FIX levels and so reduces the risk of bleeding from low FIX levels. The committee considered that a representative measure of people's actual FIX levels during the lead-in period would be useful. But it understood the company's rationale and accepted the company's approach for reporting change in FIX levels. The committee concluded that etranacogene dezaparvovec produces a clinically meaningful increase in endogenous FIX levels.

The EAG noted the clinical benefits reported in HOPE‑B may have been overestimated. It suggested that reduced physical activity during the COVID‑19 pandemic may have meant there were fewer bleeding episodes needing on-demand FIX replacement. The EAG also noted that after the lead-in period, the trial protocol prohibited prophylactic FIX replacement for FIX levels of 5% or more but investigating clinicians could give FIX replacement at their discretion. The EAG considered that clinicians may be less likely to give ad hoc FIX replacement within the trial than in routine practice, to adhere as closely as possible to the preferred study procedures. It considered it plausible that use of FIX replacement would be higher in clinical practice than in HOPE‑B. The company highlighted that FIX replacement use remained substantially reduced up to 24 months post-treatment. It suggested that if the COVID‑19 pandemic had lowered activity levels, increased activity after the pandemic would have led to more bleeds and increased use of FIX replacement, which was not the case. It also highlighted that the reduction in annualised spontaneous bleeding rates (not related to trauma or activity) from the lead-in period to 7 to 18 months post-treatment, was maintained at 24 months post-treatment. The patient experts shared their experience that activity levels actually increased during the COVID‑19 pandemic. Clinical experts added that there was no noticeable difference in reported bleeding episodes during the pandemic. The clinical experts also noted that decisions about giving FIX replacement would be based on normal clinical practice and not influenced by a trial setting. The committee considered it plausible that physical activity (or its intensity) may have increased during the COVID‑19 pandemic and recalled that bleeding episodes did not noticeably change during the pandemic. The committee concluded that the COVID‑19 pandemic and trial protocol did not have a substantial impact, if any, on the magnitude of clinical benefits reported in HOPE‑B.

Because HOPE‑B was a single-arm trial, the company did indirect treatment comparisons to compare the clinical effectiveness of etranacogene dezaparvovec with FIX prophylaxis. The company used the inverse probability of treatment weight method for the comparison with Idelvion because participant-level data was available for both treatments. A matching-adjusted indirect comparison method was used for the comparisons with Alprolix, Refixia and BeneFIX because only summary data was available. The indirect treatment comparisons suggested statistically significant improvements in bleeding outcomes for etranacogene dezaparvovec compared with each of the comparators. The results are considered confidential by the company so cannot be reported here. The EAG believed that the indirect treatment comparisons used the best available methods, but the different methods used in the studies seriously undermined the results of comparisons. The EAG noted that the comparator studies differed from HOPE‑B in several important ways, principally relating to analysis populations, outcome definitions and background care. The committee understood the EAG's concerns but acknowledged these limitations related to the quality of the studies rather than the methods used to do the indirect treatment comparisons. The committee concluded that the magnitude of improvement in bleeding outcomes for etranacogene dezaparvovec compared with FIX prophylaxis treatments was uncertain, and it would take this into account in its decision making.

The company presented a cohort-based Markov model. The modelled cohort moved through 4 health states which were based on bleeding events. These were 'no bleed', 'non-joint bleed', 'joint bleed' and 'death', with everyone starting in the 'no bleed' state. Bleeding rates from HOPE‑B and the company's indirect treatment comparisons (see section 3.7) were used to calculate transition probabilities between the health states. The committee concluded that the company's model structure was appropriate for decision making.

The company modelled treatment with etranacogene dezaparvovec followed by FIX prophylaxis treatment after etranacogene dezaparvovec failure (see section 3.10). The company's model comparison was:

• etranacogene dezaparvovec followed by FIX prophylaxis treatment after etranacogene dezaparvovec failure, compared with

• FIX prophylaxis treatment.
  
  The company presented both a fully incremental analysis (which included 8 treatment combinations) and a pairwise analysis (which included 4 comparisons). The clinical experts explained that in clinical practice, the choice of FIX prophylaxis treatment is based on a variety of factors. This includes the dosing schedule, FIX activity levels, bleeding patterns, mechanism of action and availability of each treatment. A clinical expert said that the most frequently prescribed treatments in clinical practice are extended half-life treatments. This is primarily because they need less-frequent dosing than standard half-life treatments (see section 3.2). At consultation, the company provided an updated pairwise comparison using a basket of comparators weighted by NHS market share. The comparison was:

• etranacogene dezaparvovec followed by a basket of FIX prophylaxis treatments after etranacogene dezaparvovec failure, compared with

• a basket of FIX prophylaxis treatments.
  
  The EAG was concerned that a binary decision between etranacogene dezaparvovec and a basket of alternative treatment options could create misleading conclusions about the cost effectiveness of each treatment strategy and result in the incorrect application of an incremental analysis. The committee noted that some of the alternative treatment options were standard half-life FIX treatments. It was aware that there is declining use of standard half-life treatments in clinical practice. The committee considered that the updated pairwise comparison better reflected treatment choices in clinical practice for moderately severe or severe haemophilia B and reflected the most commonly used comparator according to market share. It concluded using a basket of FIX prophylaxis treatments weighted by NHS market share was the most appropriate comparator for the economic modelling.

The company's economic model included a predicted failure rate of etranacogene dezaparvovec based on extrapolations of observed data from Shah et al. (2022; see section 3.11). The company base case defined treatment failure (that is, the FIX level at which FIX prophylaxis treatment would restart) as a FIX level below 2%, based on advice from 8 NHS clinicians. Once etranacogene dezaparvovec failed, it was assumed that people resumed treatment with 1 of the 4 FIX prophylaxis treatments. The EAG consulted with an NHS clinician who advised that a FIX level of 2% to 5% would be considered as a 'trough' (a minimum level when people are routinely having FIX prophylaxis treatment). They also advised that this level may be too low for people to engage safely in some routine activities such as certain sports. The EAG also noted that people in HOPE‑B only stopped FIX prophylaxis treatment when FIX levels were more than 5%. The EAG's base case therefore considered that FIX prophylaxis treatment was more likely to be reintroduced when FIX levels dropped below 5% rather than 2%. The clinical experts explained that restarting FIX prophylaxis treatment is based on many factors, including bleeding symptoms, FIX level and personal preference. One clinical expert said an appropriate definition of treatment failure would be a FIX level between 2% and 3% and another said below 3%. The committee considered that the definition of treatment failure, and when people need to restart FIX prophylaxis treatment, would vary based on several factors, including bleeding symptoms and activity levels. At the first committee meeting, the committee requested that the company should model restarting FIX prophylaxis treatment at a FIX level of 3%. At consultation, the company provided this scenario analysis. The committee concluded that restarting FIX prophylaxis treatment at a FIX level of 3% was appropriate.

The EAG noted that there was a difference between the company's deterministic and probabilistic cost-effectiveness results. The EAG considered the divergence was due to the way health state utilities were modelled. In the company's model, utility values were inputted as independent beta distributions. The EAG considered that modelling best practice should include a baseline and incremental utility to provide to provide a structural correlation between health state utilities. The company clarified it had already included the structural correlation between the health state utilities in its model. Its probabilistic utility values were determined by the lowest value of either the minimum function of the value of the beta distribution of the comparator or etranacogene dezaparvovec. It explained that stakeholders had agreed that quality of life with etranacogene dezaparvovec is superior to that with FIX prophylaxis and on-demand replacement. The EAG considered the company's approach could bias the health state utility estimate of FIX prophylaxis treatment which will always be below that of etranacogene dezaparvovec. The company accepted that the deterministic ICERs should be used for decision making. The committee concluded that the deterministic cost-effectiveness results were the most appropriate.

The committee considered the results of its preferred analysis using:

• treatment failure defined as a FIX level of 3% (see section 3.10)

• a basket of FIX prophylaxis treatments weighted by NHS market share (see section 3.9)

• updated Shah et al. (2022) analysis to include the person who had a partial treatment dose (see section 3.11)

• deterministic cost-effectiveness results (see section 3.13).
  
  Because etranacogene dezaparvovec and the comparators have confidential commercial arrangements, the exact ICERs are confidential and cannot be reported here. Etranacogene dezaparvovec followed by a basket of FIX prophylaxis treatments after etranacogene dezaparvovec failure had lower costs and higher quality-adjusted life-years (QALYs) than the comparator (a basket of FIX prophylaxis treatments, weighted by NHS market share).
  
  The committee considered a scenario analysis by the EAG which truncated the durability at each year over the time horizon of the model. This showed that although etranacogene dezaparvovec followed by a basket of FIX prophylaxis treatments was cost effective in the company's base case, the treatment effect would need to persist for a considerable duration beyond the trial data for etranacogene dezaparvovec to remain cost effective (the exact details are confidential and cannot be reported here).

NICE's manual for health technology evaluations notes that judgements about the acceptability of a technology as an effective use of NHS resources will take into account the degree of certainty around the incremental cost-effectiveness ratio (ICER). The committee will be more cautious about recommending a technology if it is less certain about the ICERs presented.

The committee noted concerns around the high level of uncertainty, specifically:

• the results of the indirect treatment comparisons (see section 3.7)

• the Shah et al. (2022) durability extrapolation including the lack of long-term data and methodological challenges in the extrapolation
  
  Because of the high level of uncertainty in the clinical and economic evidence, the committee was concerned about the limited available trial data. It was concerned, because of the short trial follow up, that there was insufficient data to reliably conclude on the likelihood of the treatment effect being long enough to achieve cost effectiveness (see section 3.14).
  
  The committee concluded that it could not recommend etranacogene dezaparvovec for routine use in the NHS. This was because although several scenarios varying treatment effect duration resulted in ICERs within the range considered a cost-effective use of NHS resources, the true duration of treatment effect was still highly uncertain and the consequences of decision error would have a significant impact on the NHS. The committee considered that the limited data informing the durability assumptions (see section 3.11) and underpinning the analyses was so uncertain that the plausible ICERs could increase far beyond the acceptable cost-effectiveness range, if the treatment effect waned earlier than the lifetime effect predicted by the company.

The committee noted that haemophilia B is rare in women and HOPE‑B did not include women. It was aware of clinical advice received by the EAG that the few women who have severe and moderately severe haemophilia B would be affected in the same way as men with the condition. At consultation, the company clarified that its summary of product characteristics states that etranacogene dezaparvovec is not recommended in women of childbearing potential and should not be used during pregnancy or during breastfeeding. Stakeholders also commented that HOPE‑B excluded people with uncontrolled HIV and hepatitis, but that this should not exclude them from any recommendation. The committee noted that the summary of product characteristics for etranacogene dezaparvovec does not exclude people with HIV and historical hepatitis B and C. But, that the safety and efficacy of etranacogene dezaparvovec has not been established in people with acute or uncontrolled chronic hepatitis infections. The committee concluded that any recommendation would not exclude people with HIV or hepatitis, and would not need to differentiate between men and women.

The committee also noted benefits of etranacogene dezaparvovec that were not included in the economic model. The company noted that etranacogene dezaparvovec has the potential to reduce long-term joint damage because it reduces bleeding episodes which are associated with joint damage. The company also noted that etranacogene dezaparvovec might lower mortality, which could lead to higher QALY benefit. The committee noted that the effect of etranacogene dezaparvovec on mortality has not been shown. The committee agreed to take the possibility of uncaptured benefit into its decision-making but concluded that this did not have a material effect because of the committee's considerable concerns about the uncertainty in the long-term durability estimates.