Most medicines are paid for repeatedly, and their value is judged against what the NHS spends on them each year. A gene therapy is paid for once and is expected to work for a lifetime. That mismatch is why the appraisal of a single treatment now sets the terms on which every later one will be argued.

What NICE decided

In February 2025, NICE recommended exagamglogene autotemcel, sold as Casgevy, for severe sickle cell disease. The recommendation carries conditions. It applies to people aged 5 years and over with recurrent vaso-occlusive crises and a suitable genotype, when a stem cell transplant is appropriate but no matched related donor is available, and only for people who have had at least two such crises per year over the previous two years.

It is not recommended for routine use. NICE's guidance states that some of the most likely cost-effectiveness estimates are higher than what it normally considers an acceptable use of NHS resources, even after accounting for the treatment's potential effect on health inequalities. It is instead recommended during a managed access period, under an agreement between NICE, NHS England and the manufacturer.

The committee gave reasons for accepting that uncertainty. These included the health inequalities experienced by people with sickle cell disease, the innovative nature of the technology, and benefits to carers' quality of life that the standard measure does not capture.

The agreement is between NICE, NHS England and the manufacturer, and its conditions govern use. NICE states that the treatment can only be used if those conditions are followed, and that more evidence will be collected during the period to address the uncertainties, after which the guidance will be reviewed and updated.

The clinical position is narrow by design. Standard care for sickle cell disease includes hydroxycarbamide, blood transfusions and iron chelation to remove excess iron. People well enough can receive a stem cell transplant where a suitable donor exists. Where transplant is clinically suitable but no matched related donor is available, this treatment is a possible cure.

In sickle cell disease, damaged red blood cells can block blood flow to parts of the body, depriving them of oxygen and causing severe pain. That event is a vaso-occlusive crisis, and it is the outcome the eligibility criteria count.

The thresholds that decide

NICE does not operate a single cut-off. Its manual sets a range of £25,000 to £35,000 per quality-adjusted life year gained for most technologies, and £100,000 per quality-adjusted life year for highly specialised technologies. As the estimate rises through that range, the committee requires an increasingly strong case before recommending.

A quality-adjusted life year combines length and quality of life into one unit. For highly specialised technologies, NICE may apply a weight of between 1 and 3 to those units where there is compelling evidence of significant gain, using equal increments across a range of 10 to 30 quality-adjusted life years gained.

Severity is handled separately. The committee considers the future health lost by people living with a condition under standard care, expressed as an absolute and a proportional shortfall, and may weight the units more heavily for more severe conditions. For highly specialised technologies the severity is taken as already reflected in the selection of technologies for evaluation.

These are the numbers a committee works with, and they explain the shape of this decision. An appraisal that lands above the usual range is not automatically refused. It is refused unless the committee can identify a reason to accept it, and the reasons available to it are specified in advance rather than chosen case by case.

Why this appraisal was hard

The clinical evidence carried its own limits. NICE's committee noted that the treatment was not compared with anything else in the trial, that the number of people treated was small, and that how well it works in the long term was unclear.

The economic model was uncertain on several counts, including its structure, the survival and quality-of-life outcomes used for both arms, how long the treatment's effect lasts, how often people withdraw before infusion, the characteristics of people starting treatment, and the frequency of complications.

Those uncertainties compound rather than add. The duration of the treatment effect is the largest single driver of value in a one-time therapy, because the benefit is counted over a lifetime. If that duration is unknown, the estimate is not merely imprecise; it is sensitive to an assumption the trial was not long enough to test. The withdrawal rate does similar work at the other end, because a patient who leaves the pathway before infusion receives none of the benefit.

Those are the ordinary reasons an appraisal reaches a conditional answer rather than a yes or a no. The list price reported for the treatment is £1.65 million. NHS England secured a reduced price through a reimbursement agreement, and the commercial terms are confidential, so the figure the committee actually used is not public.

What managed access defers

Managed access is a deferral, not a resolution. Treatment is funded while more evidence is collected, and NICE reviews the guidance afterwards. The guidance for this treatment was reviewed on 30 October 2026 to widen the eligible population to people aged 5 years and over, following a licence extension.

The same structure appears elsewhere in this category. NICE's guidance for onasemnogene abeparvovec in spinal muscular atrophy, and its later guidance for the presymptomatic form of the condition, both operate with confidential discount arrangements rather than published prices.

The second of those carries its own example of how the category moves. The presymptomatic guidance partially updates the earlier one, extending the treatment to babies aged 12 months and under before symptoms appear, and it was issued in April 2023.

Gene therapies in this category are priced and appraised one at a time, and each negotiation establishes terms that the next one inherits. The same manufacturer and NHS England reached a comparable reimbursement agreement for transfusion-dependent beta thalassaemia the previous August, under the same fund. The pattern is visible in the surrounding guidance: conditional funding, confidential discounts and scheduled reviews appear across the class rather than in this appraisal alone.

What is not yet known is whether the evidence collected during a managed access period will be sufficient to convert a conditional recommendation into a routine one. That depends on long-term follow-up that does not yet exist, and on whether the price revisions built into the agreement bring the estimate inside the threshold.

Summary

NICE has accepted a one-time cure at a price it has not published, conditionally and temporarily, because the long-term evidence does not yet exist.

  • The recommendation is conditional. Exagamglogene autotemcel is available during a managed access period, not for routine use.
  • Stated uncertainty was accepted. NICE cited treatment effect duration, trial size, absence of a comparator and health inequalities in reaching that decision.
  • The thresholds are published. £25,000 to £35,000 per quality-adjusted life year normally, and £100,000 for highly specialised technologies.
  • The price is not published. The £1.65 million list price was reduced through a confidential commercial agreement.
  • The review is scheduled but the evidence is not in. Guidance was reviewed on 30 October 2026 to widen eligibility, while the long-term data are still being collected.

Whether deferred evidence closes the cost-effectiveness gap is not yet known.

Research direction and editorial oversight: Juno Uni Hera. AI-assisted research and drafting; human review is ongoing.

Sources