After the First Dose: Britain’s CRISPR Rules Have No Route for a Second

In a UK gene-therapy centre, the conversation is beginning to change. The first wave of CRISPR-based treatments is moving out of early trials and into specialist NHS pathways. For many patients, the immediate question is no longer whether editing can work, but what happens if it works only partly, fades, or must be delivered by another route. The second edit is no longer a thought experiment. It is a regulatory, immunological and ethical problem that Britain has not yet answered.

The one-shot assumption

Much of the UK’s approval machinery is built around a single administration. A marketing authorisation specifies a dose, a schedule and a patient group. A NICE evaluation weighs the cost of that intervention against the health gains it is expected to produce. A clinical trial protocol usually defines one administration, followed by a period of follow-up. Consent forms are written for that first intervention, not for an open-ended series of them.

That assumption is understandable. Gene editing has been presented, rightly, as a potentially permanent correction. But permanence is not the same as completeness. A patient may receive an edit that reduces disease burden without removing it. A delivery vector may fail to reach enough cells. An immune response may blunt the effect of a first dose. In each case, the clinical answer may be another edit, a revised edit, or a different delivery route. The regulatory answer is far less clear.

Immunology sets the first limit

The most immediate obstacle is biological. Repeat dosing with an adeno-associated virus vector is difficult because the first exposure often generates neutralising antibodies against the capsid. A second infusion may be cleared before it reaches target tissue. For CRISPR components delivered as proteins, pre-existing and treatment-induced immunity to Cas9 or other editing proteins can pose similar problems. Lipid nanoparticles can be redosed more readily, but repeated administration may still provoke innate immune activation.

These are not abstract concerns for a 2026 clinic. They shape trial design, patient selection and the honesty of consent. A patient who agrees to a first dose should be told whether a second is likely to be possible, and under what conditions. At present, that information is often uncertain because the data have not been collected in a systematic way. UK trials have tended to focus on safety and initial efficacy, not on the immunology of re-administration.

A regulatory gap with three doors

For an unlicensed CRISPR therapy in a clinical trial, a repeat dose would usually require a substantial amendment to the protocol, fresh ethical review and MHRA approval. That route exists, but it is designed for changes to a study, not for a planned sequence of treatments. For a licensed product, repeat administration may require a variation to the marketing authorisation, supported by new clinical evidence. NICE would then need to consider whether its original appraisal still holds. For a clinician considering use outside the licensed indication, funding and liability become immediate questions.

The UK is not alone in facing this gap. But Britain’s separation of trial regulation, licensing and NHS commissioning makes the sequel particularly awkward. A therapy can be approved for one dose, commissioned for one dose and evaluated for one dose, while the clinical need for a second falls between those categories. Commissioners will reasonably ask whether a second administration is a new treatment or part of the first. The answer will determine who pays, who approves and who carries the risk.

Consent is not a lifetime subscription

There is also an ethical question that precedes the financial one. Consent to a first CRISPR treatment cannot automatically cover a second. The risks are cumulative, the patient’s circumstances may have changed, and the evidence base for re-dosing may be thin. A second edit may carry different off-target risks, especially if it uses a different delivery system or a different editing tool. Re-consent should be the default, not an afterthought.

The UK’s long-term follow-up arrangements for gene therapies are already stretched. Repeat treatment would place further demands on registries and clinical teams. It would also raise questions about how patients are told that a first edit has not been curative. That conversation needs a protocol of its own, separate from the consent form for the first procedure.

Who pays for the second cut?

The NHS budget adds another layer. NICE’s cost-effectiveness models are usually built around a single intervention with a durable effect. If a product needs repeat dosing, the original model may overstate its value or understate its cost. Managed access arrangements could accommodate uncertainty, but they are not designed for sequential edits. A hospital trust may find itself funding a second administration from its own budget while national commissioners argue that the first dose was the commissioned treatment.

This is not an argument against repeat CRISPR treatment. It is an argument for planning. The UK has an opportunity to build repeat dosing into trial protocols, collect immunogenicity data and develop a clear regulatory route before the requests become routine. That would mean guidance from the MHRA and the Health Research Authority, a NICE process for reassessment, and commissioning rules that recognise the possibility of a sequel.

The rules should arrive before the requests

Britain’s CRISPR sector has spent years arguing that gene editing is a serious clinical option, not a science-fiction promise. The second edit is part of that seriousness. A one-shot model may be biologically correct for some therapies, but it cannot be the default for all. As first-generation treatments reach patients, the country needs to decide how it will handle the second cut. The alternative is a system that can approve a cure but cannot authorise its follow-up.

The time to write those rules is before the second dose is requested. In 2026, that moment is close enough to plan for, and too close to ignore.