Editing genes inside the body: what the first human results show

Gene editing changes the DNA that a living thing carries. In vivo gene editing does this by delivering the editing tools directly into the body, so the change happens inside a person's own cells. The older route is ex vivo editing, where cells are taken out, edited in a laboratory and returned. In vivo treatment matters because many diseases affect organs that are hard to reach with cell treatments, and because a single well-targeted dose could in principle replace years of repeated medicine. Whether that holds is still being tested in a small number of people.

How the tools work

The two human studies below use CRISPR gene editing and base editing. In both, a short guide RNA matches a chosen spot in the DNA and leads an enzyme to it. According to MedlinePlus, the Cas9 enzyme cuts the DNA at the targeted location, and researchers then use the cell's own repair machinery to add or delete pieces of genetic material, or to replace a segment with a customised sequence.

Delivery is the second problem. Editing tools are large and fragile, so they need a carrier. In VERVE-102, the editor is sent as messenger RNA, together with a guide RNA, packed into fatty particles called lipid nanoparticles that are designed to reach the liver (NEJM, VERVE-102).

What is approved

No in vivo gene editing treatment appears on the US Food and Drug Administration's list of approved cellular and gene therapy products, last updated on 17 September 2026. The list gives product names and manufacturers only, so this check covers names, not mechanisms, and only the US.

The closest approved product is Casgevy. On 8 December 2023 the FDA approved it for sickle cell disease in patients aged 12 and over with recurrent vaso-occlusive crises. The agency called it the first FDA-approved therapy using CRISPR/Cas9, and said the patient's blood stem cells are modified by genome editing (FDA). That is a cell-based route, not direct delivery into the body.

The first in vivo editing treatment could be approved soon. On 8 September 2026 Intellia Therapeutics said the FDA had accepted its licence application for lonvoguran ziclumeran (lonvo-z) for hereditary angioedema, with priority review and a target decision date of 10 March 2027 (company press release). A target date is not a decision, and the outcome is not known. For what the difference means, see our guide on how a drug approval differs from a trial result.

What has been shown in people

Two peer-reviewed trials give the main human evidence. Both were published in the New England Journal of Medicine in 2026. We read their abstracts; the figures below come from those abstracts unless we say otherwise. Both trials were funded by the companies developing the treatments.

HAELO (phase 3, hereditary angioedema). This double-blind trial tested lonvo-z, an investigational in vivo CRISPR treatment, in 80 people aged 16 or over with hereditary angioedema caused by C1 inhibitor deficiency (NEJM, HAELO). Fifty-two were assigned a single 50 mg intravenous infusion of lonvo-z and 28 were assigned placebo. From week 5 to week 28, the average monthly attack rate was 0.26 in the lonvo-z group and 2.10 in the placebo group, an 87% relative reduction. Median follow-up at the data cut-off of 10 February 2026 was 7.5 months.

On safety, the paper reports adverse events in 92% of the lonvo-z group and 86% of the placebo group up to week 28. Among events that were more common with lonvo-z, the most frequent included infusion-related reactions, headache, fatigue, back pain and upper respiratory tract infection. No serious adverse events, or events of grade 3 or higher, were reported in the lonvo-z group.

A further figure comes only from the company. According to an Intellia press release, 62% of treated patients were entirely attack-free and therapy-free over the six-month evaluation period, against 11% on placebo. We have not checked this against the full paper.

VERVE-102 (phase 1, PCSK9 and cholesterol). This open-label study gave a single intravenous infusion of an adenine base editor to adults with heterozygous familial hypercholesterolaemia or premature coronary artery disease (NEJM, VERVE-102). The editor is designed to switch off the PCSK9 gene in the liver, which affects LDL cholesterol. Thirty-five people across six dose levels, from 0.3 to 1.0 mg of total RNA per kg of body weight, had at least 28 days of follow-up. A TCTMD report gives their mean age as 52.

Average PCSK9 levels fell by 51% at the lowest dose and 88% at the highest. LDL cholesterol fell by 9% at the lowest dose and 62% at the highest, an absolute fall of 78 mg per decilitre. Neither the abstract nor the TCTMD report gives the time point for these figures. Fifteen people had at least one year of follow-up, and the authors say the reductions appeared to last, though we have not seen one-year figures.

No dose-limiting toxic effects occurred. The paper reports mild to moderate infusion-related reactions and temporary rises in the liver enzyme alanine aminotransferase. One participant with gastro-oesophageal reflux disease had aspiration pneumonitis. TCTMD adds that about 20% of participants had grade 1 or 2 infusion-related reactions, 31% had a treatment-related adverse event, and that the pneumonitis was grade 3 and judged unrelated to the treatment.

These results show the editor is biologically active. Lower PCSK9 and LDL are laboratory measures, though, and the study did not test whether heart attacks or strokes fall.

A safety event in another trial

In vivo editing has already had a serious setback. Intellia's nexiguran ziclumeran (nex-z), a CRISPR treatment designed to switch off the TTR gene in people with transthyretin amyloidosis, is in two phase 3 trials, MAGNITUDE and MAGNITUDE-2. According to the company, on 29 October 2025 the FDA placed both trials on clinical hold after a patient dosed with nex-z in MAGNITUDE had grade 4 rises in liver enzymes and raised bilirubin (company press release). On 6 November 2025 the company said that patient had died, and that the treating physician described it as a case with complicating comorbidities that was still being evaluated (company press release).

According to Intellia, the FDA lifted the hold on MAGNITUDE-2 in January 2026 (company press release) and the hold on MAGNITUDE on 2 March 2026. The agreed measures include closer monitoring of liver blood tests, guidance on short-term steroid treatment if liver enzymes rise soon after dosing, and excluding patients with certain liver abnormalities. These are company statements; we have not seen the FDA's own correspondence.

Limits and open problems

Delivery. Both published trials and the nex-z programme target the liver. The sources we checked do not show reliable delivery to muscle, brain, lung or blood-forming cells in people.

Liver safety. The VERVE-102 trial saw temporary liver enzyme rises, and the nex-z hold followed a severe liver event. How often such events happen, and in whom, will need larger and longer studies.

Off-target edits. The TCTMD report, citing the researchers, says that safety, "including potential risks of off-target gene editing, will have to be studied over several years." It reports that preclinical tests found no off-target problems, but an outside commentator raised the possibility that genes other than PCSK9 could be edited.

Durability and irreversibility. One-year VERVE-102 data cover 15 people, and TCTMD reports that participants are expected to be followed for up to 15 years. A treatment that changes DNA may be permanent. MedlinePlus notes that most genome editing changes are limited to body cells, which are not passed to children, while edits to egg or sperm cells could be passed on.

Cost and access. The sources we checked do not address the price of in vivo treatments or how they would be made at scale.

What to watch next

• The FDA decision on lonvo-z, with a target date of 10 March 2027.
• Results from the MAGNITUDE and MAGNITUDE-2 trials of nex-z, now that dosing has restarted under extra liver monitoring.
• Longer follow-up for VERVE-102, and enrolment in its trial, which TCTMD reports is planned for up to 85 people.
• Human data from delivery methods that reach tissues other than the liver.
• Longer-term measures of off-target changes and immune responses.

This article is general information, not medical advice.