Imlifidase for desensitisation treatment before kidney transplant in people with chronic kidney disease

3.1 Many people who are waiting for a deceased donor kidney are on dialysis. This filters waste products out of the blood. Both haemodialysis and peritoneal dialysis have a substantial effect on day-to-day life for someone with advanced chronic kidney disease. The patient expert explained that before both types of dialysis, the person needs to have surgery. People on dialysis have restricted fluid intake and diets, and may have very reduced energy levels. Also, people having haemodialysis need 2 or 3 sessions a week, each lasting 5 hours, so there is a substantial impact on time. They explained that it is often difficult for people on dialysis to make plans to see friends and family, or go on holidays, and that the time needed for haemodialysis can affect their ability to work full time. Long-term dialysis can also have a range of effects on physical and mental health, such as bone disease, heart disease, and a loss of hope. In some cases, people die while on the transplant waiting list. One of the patient groups highlighted that being on dialysis can feel like "sitting and waiting and feeling like everything's on hold". The patient expert explained that although people recognise that a kidney transplant is not without risk, and lifelong immunosuppression afterwards can have side effects (such as skin cancer risks with older regimens), a kidney transplant gives hope for a more normal life. The committee recognised that people who are on dialysis, especially for a long time while waiting for a kidney transplant, have reduced quality of life.

3.2 Some people who need a transplant have an immunological barrier to transplantation. They have antibodies to human leukocyte antigens (HLA), which is known as being sensitised. Exposure to tissue with different HLAs is the most common cause of sensitisation, and it can happen from transfusion of blood products, pregnancy or a previous transplant. People with a high level of sensitisation and no appropriate living donor can spend a long time on the waiting list for a deceased donor kidney. This is because they have antibodies against almost all donors' HLA (known as a positive crossmatch). In these circumstances, the donor kidney would be at very high risk of antibody-mediated rejection. One of the clinical experts explained that people who need a kidney transplant are encouraged to find a living donor if possible. This is because this creates the opportunity of either directed-donation transplant or transplant through the UK Living Kidney Sharing Scheme. If that is not possible, then people have dialysis until a suitable deceased donor is found through the national deceased donor allocation algorithm (UK Kidney Offering Scheme). NHS Blood and Transplant data reported in 2020 that the median wait for a deceased donor kidney was about 5 years for people who are highly sensitised, although a small number of people could wait more than 7 years. This is compared with a median waiting time of 1.5 to 2 years for people who are not sensitised at all. The UK Kidney Offering Scheme algorithm changed in 2019, with the aim of increasing access to transplants in the most sensitised population. Since 2019, the number of people in this group getting transplants has increased (see section 3.5). But the committee recognised that people who are highly sensitised still wait longer for a suitable donor kidney than those who are not. It recognised there is still an unmet need, and imlifidase offers the possibility of improving access to kidney transplantation.

3.3 While it is possible for a well-matched deceased donor kidney to become available for someone who is highly sensitised to HLA, it is unlikely. The likelihood of a favourable crossmatch may be measured by the calculated reaction frequency (CRF). That is, if someone waiting for a kidney had a CRF of 99%, this means 99% of the last 10,000 deceased donors would have HLA that would react with the blood serum of the person waiting for a kidney. In recent years, some centres have had success with antibody-incompatible transplants. Clinicians may 'delist' particular types of antigens from the individual's waiting list profile, because they believe those particular antibodies can be well managed to avoid antibody-mediated rejection. But the degree of risk-taking for incompatible transplants that centres are willing to take can vary. Delisting to increase the chances of finding a deceased donor match may not be possible for everyone who is highly sensitised. If these people do not have a suitable living donor available for a directed transplant or transplant through a kidney sharing scheme, then they have no other options but to continue waiting for a well-matched deceased donor kidney. If they wait too long, they may no longer be well enough to have a transplant.

3.4 Imlifidase is an enzyme that breaks down a major class of human antibodies (immunoglobulin G). This includes the antibodies that a person already has against potential donor kidneys. If imlifidase is given immediately before a transplant, it can change a positive crossmatch to a negative one. This allows a brief window for a transplant to be done without rapid rejection. It is considered innovative by some clinical experts. Because the treatment has a transient effect, antibody levels in the body rise after transplant. Some people who had imlifidase in the trials also had a more intensive regimen of immunosuppression drugs after transplant than is currently used in the NHS for transplants without imlifidase. The committee was aware that some people who might have imlifidase might need more intensive immunosuppression regimens. But it was also aware of the impact staying on dialysis can have on health and quality of life. The committee concluded that some people who are highly sensitised may need more intense immunosuppression after having a transplant with imlifidase.

3.5 The deceased donor UK Kidney Offering Scheme was updated in 2019 (see section 3.2). This allowed kidneys to be donated from people who had died by circulatory death, in addition to those who had died by brainstem death, to be included in the UK Kidney Offering Scheme. It also increased the priority level of people who were previously harder to find a match for, or who have waited over 7 years for a transplant. People who joined the waiting list before the change, and who are highly sensitised and would have been unlikely to have a transplant, may no longer be in this population, because transplant rates have increased with the increased prioritisation. The company used data provided by NHS Blood and Transplant and clinical expert input to define its proposed eligible population for imlifidase. According to the company definition, people must have the following criteria to be eligible for imlifidase:

3.6 When a deceased donor kidney becomes available, it is allocated to an eligible person through the UK Kidney Offering Scheme. Various factors are considered to account for the suitability, urgency and need of a person who could have the donor kidney. The committee considered the impact this might have on the organ's cold ischaemic time (that is, the length of time between a kidney being removed from a donor and being transplanted). The clinical experts explained that the average cold ischaemic time varies across the transplant centres in the UK but is around 12 to 16 hours. It also varies for donations after brain stem death and for donations after circulatory death. The committee understood that going beyond a 12‑hour cold ischaemic time with kidneys after circulatory death may present a greater risk of delayed graft failure, and therefore they need to be transplanted within a shorter time window. Other factors can increase cold ischaemic time, including transporting the kidney and the number of crossmatch tests needed. The clinical experts agreed that an increased cold ischaemic time is likely to have some negative effects on transplant outcomes. But the company noted that recent data from NHS Blood and Transplant suggested that many transplants with a cold ischaemic time of more than 24 hours are still being carried out successfully. The clinical experts explained that procedures can be put in place to mitigate the impact of extended cold ischaemic time and prevent delays to transplantation. These could include pre-donation blood samples and virtual crossmatch testing. Already in NHS practice, a second person from the waiting list is lined up ready to have a transplant as a back-up, in case the first person matched cannot have the transplant. So, for imlifidase, if a negative crossmatch was not reached in time, the donor kidney could be used for someone else. The committee noted that the centres in the clinical trial were not based in the UK and might have been well placed for short cold ischaemic times, by providing high numbers of transplants and donors close by. The NHS England commissioning lead explained that a clinically led national multidisciplinary team would be needed to develop the pathways and protocols for using imlifidase if it was recommended. The committee agreed that these protocols could allow people needing imlifidase to be treated in a specialist centre with experience of transplantation in people who are highly sensitised. It concluded that protocols should be developed in a small number of specialist centres to mitigate the impact of increased cold ischaemic time on donor kidneys when using imlifidase.

3.7 Adding a second imlifidase infusion could potentially increase the cold ischaemic time because an additional crossmatch test would be needed. Clinical feedback at consultation suggested that adding a second dose could add an extra 8 to 10 hours to the transplant process. Only a small number of people in the clinical trials who had imlifidase needed a second imlifidase infusion (the exact proportions are confidential). The clinical experts advised that 1 infusion would be sufficient in most situations. The committee agreed that giving only 1 infusion of imlifidase would allow for safe transplantation within acceptable cold ischaemic time thresholds. The committee accepted that this would be a pragmatic option given the challenges of the increased cold ischaemic time, and would allow for most people who are highly sensitised to have imlifidase. It concluded that there should be a limit of only 1 imlifidase infusion for people who are highly sensitised.

3.8 Principle 7 of the principles that guide the development of NICE guidance and standards states that recommendations should be based on population benefits and value for money. As stated in NICE's guide to the methods of technology appraisal, 'the reference-case perspective on outcomes aims to maximise health gain from available healthcare resources.' The committee understood that any donor kidney used with imlifidase could have been used for someone else who is likely to incur lower costs, and have better outcomes and equal related savings from avoiding dialysis. The clinical experts had a wide range of views on which costs and benefits should be included. The company felt a utilitarian analysis at the population level would not capture the benefit of increased equity of access to transplants. It considered that allocation of deceased donor kidneys already relies on a trade-off between equitable access and providing best quality matching. The committee recognised the equity issues of people who are highly sensitised and agreed that these should be taken into account. The company had not provided much evidence of the differences in outcomes for people who are highly sensitised who had imlifidase, compared with those who are not highly sensitised and who would no longer get a transplant. The committee recognised these populations were different. But it was aware that both groups would be competing for the same resource. The company had not explored the potential consequences of this. The committee noted that principle 9 of the principles that guide the development of NICE guidance and standards aims to reduce health inequalities, which emphasises that NICE guidance should support strategies that improve population health as a whole. It considered that people who would be eligible for imlifidase are likely to have been on the waiting list for a long time and they would likely still be disadvantaged using the new Kidney Offering Scheme algorithm. The committee concluded that kidneys are a scarce resource, but decisions should consider opportunity costs as well as equity of access for people who are highly sensitised.

3.9 Evidence for the clinical effectiveness of imlifidase originally came from 4 non-UK based, uncontrolled, open-label studies. The primary outcomes reported on safety and ability to achieve a crossmatch conversion after treatment with imlifidase. For this reason, they had short follow-up times that ranged between 64 days and 180 days. This meant that longer-term outcomes to assess the success of transplant were not estimated. The clinical experts agreed that the trial outcomes were too short for this clinical context (with potential graft loss at 5, 10 and 15 years). The company had acknowledged that longer-term data was needed and provided further clinical evidence for imlifidase from the trials originally included. The ERG had requested the company provide clinical evidence for 3 populations. These included:

3.10 In the company's model, antibody-mediated rejection had been captured using data from its clinical trials. There was a high rate of antibody-mediated rejection (40%) in the company's original clinical data. There was no comparator arm in the trials nor a matched population. So, it was also not clear whether the 40% antibody-mediated rejection was a consequence of a very unwell population in the imlifidase trials, or a consequence of people having had imlifidase in the trials. Clinical experts explained that in clinical practice they would normally expect only 10% of people to have antibody-mediated rejection after an average transplant, based on UK experience. The committee noted that the antibody-mediated rejection rates were still high in the company's newly defined population. The exact rates cannot be reported because they are commercial in confidence. The clinical experts explained that it is difficult to establish exact rates because reasons will vary depending on individual characteristics. But a 30% to 50% antibody-mediated rejection rate for people who are highly sensitised in the first month after transplant would be plausible. At consultation, the company stated the antibody-mediated reaction rates in its trials were in line with what is expected in clinical practice for HLA-incompatible kidney transplants. In the company's trial data, no antibody-mediated rejection events were reported after the first year after transplant. So, the company had only included the events and related costs of antibody-mediated rejection events in the first 2 cycles of its model. The committee was concerned that antibody-mediated rejection had not been fully accounted for in the company's model. The model also did not differentiate between a graft needing intensive immunosuppression therapy and one that was more successful. Antibody-mediated rejection can be chronic and difficult to treat, with changes in immunosuppression regimens, biopsies and limited graft survival. Feedback at consultation clarified that an antibody-mediated rejection rate of 40% was consistent with HLA antibody-incompatible transplantation. The committee concluded that some antibody-mediated rejection is to be expected in people who are highly sensitised, but their quality of life will be improved while the transplant is working.

3.11 In its revised model, the company used NHS Blood and Transplant data to estimate the proportion of people who were not having imlifidase who had dialysis. It originally adjusted the proportions so that everyone would have had dialysis for at least 2 years. The ERG agreed that NHS Blood and Transplant data was an appropriate source to inform this distribution, but it did not agree that everyone would be having dialysis. Based on clinical opinion, it considered that there may be a small number of people who could otherwise meet the eligibility criteria but might not be able to have imlifidase if it assumed everyone had to have had dialysis for at least 2 years. The ERG therefore assumed that 5% of people in its base case would not have dialysis before imlifidase. The company later agreed that people who had not previously had dialysis would also be eligible for imlifidase (see section 3.5). It accepted that being on dialysis should not be a requirement but considered that a 5% proportion was too high. Based on clinical feedback it suggested it was unlikely that people who did not have dialysis would stay on the kidney waiting list for longer than 6 months. The committee recognised that there was some uncertainty around applying the estimate. But it concluded that some people would not be having dialysis before having a transplant with imlifidase.

3.12 The company's original submission assumed that 100% of people who had imlifidase would go on to have a kidney transplant. However, this was not the case in its clinical trials. For its base case, the ERG used the trial data from everyone who had imlifidase. Two out of 54 people did not get the full dose of imlifidase before transplant, so 96.3% had a transplant in the imlifidase arm of the model. The ERG also considered a scenario taking into account the 1 person (out of 52) who did not have a negative flow cytometry crossmatch (the outcome of the trial) but who had a negative virtual crossmatch after imlifidase and had a transplant anyway. In the ERG's scenario, the proportion of people having a transplant in the imlifidase arm was informed by those who had a full dose, multiplied by those who had a negative crossmatch. So, 94.4% had a transplant in the imlifidase arm in this scenario. The committee considered both the ERG base case and scenario plausible and took these into account for decision making. The company updated its base case in line with the ERG preference that 96.3% of people having imlifidase will have a transplant after treatment. The committee accepted this change but recognised that there was still some uncertainty around the appropriate value, based on the small number of people there is data for. It concluded that not everyone who has imlifidase goes on to have a kidney transplant, but the exact proportion is uncertain.

3.13 To extrapolate 6 months of post-transplant data from its trials, the company used the iBox predictive model for kidney graft survival. This was developed using data from a general transplant population in France, rather than a population consisting only of people who are highly sensitised. The iBox model was run with the company's trial data based on its original target population and using a Weibull distribution to extrapolate this to project long-term graft survival with imlifidase. Although the ERG considered iBox to be a high-quality predictive model, it was aware that iBox is a proprietary model that is not owned by the company. It had been unable to check how various factors were weighted, and the statistical power is unknown. The committee had originally considered the iBox projection and extrapolation to be too optimistic. It was concerned that the projection of trial data done through the iBox model was not a good long-term fit. This was because the 10‑year graft survival rates looked similar but seemed to improve for the company's highly sensitised population in relative terms at 20 years. This would suggest that people who are highly sensitised do relatively better over time, or the iBox general population (including people who are not highly sensitised) does relatively worse over time. This is implausible without evidence to support it. The committee considered that:

3.14 To estimate overall survival in people who had a functioning graft, the company base case extrapolated overall survival from all people who had imlifidase and a transplant (the 'all imlifidase' population) in the company trials. The company also provided a scenario analysis using data from the 'unlikely to have a transplant' population in its trials. Although the ERG considered using the 'all imlifidase' group reasonable, it considered it to make the cost-effectiveness results highly uncertain. The ERG noted the overall survival data was too uncertain to produce reasonable long-term estimates to be used for modelling a lifetime horizon in the population considered for this evaluation. The ERG considered that, in the absence of better data, either extrapolation using the 'all imlifidase' or the 'unlikely to be transplanted' data could be reasonably used to inform overall survival with a functioning graft. The ERG explored using the 'unlikely to be transplanted' overall survival data in a scenario analysis. The committee noted this substantially increased the ICER. It considered the 'unlikely to be transplanted' population to be more appropriate to extrapolate overall survival with a functioning graft. This would be the population that would be most likely to have imlifidase in clinical practice (see section 3.8). It concluded the extrapolations for overall survival with a functioning graft should be taken from the 'unlikely to be transplanted' population.

3.15 NICE's guide to the methods of technology appraisal notes that above a most plausible ICER of £20,000 per quality-adjusted life year (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. The exact ICERs are confidential, but the committee preferred the ERG's assumptions; in the ERG's base case:

3.16 Consultation feedback noted that changes to the Kidney Offering Scheme would improve access but never completely resolve inequity of access for people who are highly sensitised (see section 3.8). In particular, feedback stated that in people who have had previous blood transfusions, blood type and pregnancy are some of the risk factors that can increase the chances of developing an HLA sensitisation. For these reasons, some people from Black, Asian or minority ethnic family backgrounds, and people who have been pregnant, would be more likely to be highly sensitised. People with these protected characteristics may wait longer to have a transplant and might have difficulty accessing a matched kidney without imlifidase, compared with people who are not highly sensitised. This could have negative outcomes for people from these groups. The company stated that imlifidase could allow transplantation for people who are highly sensitised, regardless of the cause of their sensitisation. The committee was mindful of its responsibilities for people with protected characteristics under the Equality Act 2010 (see principle 9 of the principles that guide the development of NICE guidance and standards). The committee recognised that everyone who is highly sensitised and is offered a kidney under the Kidney Offering Scheme and meets the company's defined population would be considered eligible for imlifidase. It concluded that people with these protected characteristics have an increased chance of becoming very highly sensitised, and this should be taken into account in its decision making.

3.17 The committee considered whether imlifidase was innovative. It considered that imlifidase has the potential to provide a step-change to current treatment. But it was mindful of ensuring all costs and benefits were captured. The company had said that introducing imlifidase could allow people who would previously have been unlikely to get a transplant to go on to have a successful transplant, thereby improving equity of access for certain groups (see section 3.16). For this reason, the company suggested that imlifidase was innovative because it provided substantial benefits that may not be captured by measuring health gains directly. The committee agreed that imlifidase is a novel treatment because of its mechanism of action and that it could provide a brief window for a transplant to happen without rapid rejection. But it noted the challenges of introducing the technology, relating to increased cold ischaemic times and the issues around factoring in a second imlifidase infusion if it was needed (see section 3.6). The committee acknowledged that these factors must be taken into account in understanding whether a technology provides a step-change in treatment. It considered that implementation should be carefully considered. It was guided by the clinical experts that it was important to limit the number of centres providing imlifidase to minimise the impact on cold ischaemic time (see section 3.6). It considered that protocols were needed to support this when using imlifidase. The committee concluded that imlifidase could provide a step-change in treatment but that to allow this, implementation should be done using robust NHS systems.

3.18 The committee considered whether a managed access agreement would be appropriate. It considered that managed access is not appropriate to explore uncertainty around patient eligibility or the treatment pathway. It noted that a principle of managed access is that the entire eligible population should have access to treatment. It also noted that there are ethical issues with making a managed access recommendation when there are a finite number of donor kidneys. The committee considered it would be unlikely that a managed access recommendation for imlifidase aligned with the principles of resolving uncertainty through data collection, and considered whether the whole patient population could get access. It considered that the ongoing studies are unlikely to provide meaningful additional data for decision making. Collecting additional data in clinical practice would have ethical implications, which could add extra time to access to treatment. It concluded that a managed access agreement is not appropriate.

3.19 The conditional marketing authorisation specifies that imlifidase should be reserved for people unlikely to have a transplant under the available kidney allocation system, including prioritisation programmes for people who are highly sensitised. The committee understood that it can be very difficult for some people who are highly sensitised to have an appropriately matched kidney transplant. It recognised that the changes to the UK Kidney Offering Scheme in 2019 had improved access to transplants for people who are highly sensitised, but that there is still an unmet need for this population. The committee preferred the ICERs based on the ERG analyses over the company's analysis. But these were also associated with a high level of uncertainty related to integration into the existing treatment pathway and long-term clinical effectiveness. It considered that kidneys are a scarce resource that need to be allocated fairly. The committee recalled that cold ischaemic time could be mitigated by allowing only 1 dose of imlifidase (see section 3.7). Protocols would need to be developed to allow imlifidase to be used in a small number of specialist centres with experience of transplantation in people who are highly sensitised (see section 3.6). The committee recognised that the company had attempted to address the uncertainties in its cost-effectiveness analyses. It recognised that imlifidase could help improve equity of access to kidneys for people who are highly sensitised (see section 3.16). It noted the protocols that would need to be developed by the NHS when using imlifidase would be clinically led, and should take into account the following criteria: