Skip to content
ContentLora

    Tip: press / anywhere to search.

    Analysis

    How safe is in vivo gene editing? The 2026 debate

    In vivo gene editing delivers the editor straight into the body, and in 2026 the first such CRISPR therapy reached FDA review after strong Phase 3 results.[1][2] A 2025 clinical hold and patient death in another in vivo CRISPR program, plus liver deaths with an AAV gene therapy, keep safety at the centre of the debate.[3][4][5]

    Editor reviewedStrict sourcingUpdated Gene editing and gene therapyHealth and medicineLife sciences
    Show:

    In vivo editing promises one-time treatments given as an infusion, without removing and re-implanting cells.[6] The question for 2026 is whether its safety record is good enough for regulators to approve it widely. This page sets out the evidence. It is not medical advice.

    What the evidence shows

    • Lonvo-z, an in vivo CRISPR therapy for hereditary angioedema, cut attacks by 87% versus placebo in Phase 3.[1] The FDA accepted its application with Priority Review and set a target action date of 10 March 2027.[2]
    • CTX310 lowered LDL cholesterol by a mean 49% at its highest dose in Phase 1, with no treatment-related serious adverse events reported.[7]
    • In October 2025 the FDA placed holds on both nex-z Phase 3 trials after a patient had Grade 4 liver transaminase elevations and raised bilirubin.[3] The patient died on 5 November 2025. Intellia’s August 2026 quarterly report says the investigator attributed the death to septic shock secondary to a perforated duodenal ulcer, and that the clinical course also included acute liver injury.[4] The holds were lifted in January and March 2026 with enhanced liver monitoring.[8]
    • In July 2025 the FDA linked three deaths from acute liver failure to Sarepta’s AAV-based gene therapies.[5]
    • A 2025 review notes that in vivo approaches suffer from lower delivery efficiency, off-target effects and instability.[6] It adds that some gene-editing therapies have had serious or even fatal side effects.[9]
    • Even the approved ex vivo therapy, Casgevy, carries an off-target genome editing warning.[10]

    Reading the signals

    The efficacy evidence for in vivo CRISPR editing is now strong. A randomized Phase 3 trial is under FDA review, and that is a different level of evidence from the small dose-finding studies published in 2021.[1][11] The safety picture is less settled. The nex-z case involved a patient in a trial for amyloid heart disease whose reported cause of death was septic shock, but whose course also included acute liver injury after dosing. Separating the therapy’s role from other factors is hard from public information.[4][12] The FDA’s decision to let trials resume with liver monitoring suggests it judged the risk manageable, but not absent.[8]

    Edits are permanent. Off-target effects or late harms may only show up years after treatment, and trials of a few hundred people cannot rule out rare events. The AAV deaths show that delivery vehicles carry their own risks, separate from the editor.[5]

    How regulators are responding

    The FDA approved Casgevy’s pediatric expansion 53 days after filing under a national priority voucher pilot.[13] In February 2026 it published draft guidance on a “plausible mechanism” framework for individualized therapies that target genetic conditions with a known cause.[14] The 2025 personalized CPS1 base editor reported no serious adverse events in one infant.[15]

    Faster reviews and mechanism-based evidence standards move more of the safety burden to after approval. Long-term registries and follow-up will matter more as a result. WHO’s 2021 recommendations already called for registries of human genome-editing trials.[16]

    What may happen next

    The FDA decision on lonvo-z, due by 10 March 2027, is the key near-term test.[2] An approval would likely come with post-marketing safety requirements, though their scope is unknown. The resumed nex-z trials will be watched for further liver events.[8] Moderate confidence: more in vivo programs will reach pivotal trials in 2027, given BEAM-302’s pivotal cohort and Prime Medicine’s cleared Wilson disease and AATD trials.[17][18][19]

    Competing views

    Risks are manageable with monitoring

    Phase 3 and Phase 1 data show large effects with few treatment-related serious events, and the FDA let nex-z trials resume with added liver monitoring.[1][7][8]

    Safety signals call for caution

    A trial death after severe liver injury, AAV-linked liver deaths and the permanence of edits mean long-term safety is not yet established.[4][5][9]

    Regulation is adapting to the technology

    The FDA is speeding review of promising products and drafting new evidence standards for individualized editors, which shifts the burden to post-approval follow-up.[2][14][13]

    Questions readers ask

    Has anyone died in an in vivo CRISPR trial?

    Yes. A participant in Intellia's Phase 3 MAGNITUDE trial of nex-z who had Grade 4 liver enzyme elevations died in November 2025, according to trade press reports. The FDA placed both nex-z Phase 3 trials on hold and lifted the holds in early 2026 with enhanced liver monitoring.[3][4][8]

    What safety warnings apply to the approved CRISPR therapy?

    Casgevy, an ex vivo therapy, carries warnings for neutrophil engraftment failure, delayed platelet engraftment, hypersensitivity reactions and off-target genome editing risk.[10]

    Why is delivery a safety issue?

    Viral vectors such as AAV can keep the editor active for a long time, which may raise off-target risk, and AAV gene therapies have been linked to fatal liver failure. Lipid nanoparticles mostly reach the liver, so other tissues are hard to target.[20][5][21]

    Sources

    Each numbered claim is a statement we checked against the sources listed with it. Status shows how well established it is.

    1. [1]

      In the Phase 3 HAELO trial, a single infusion of lonvo-z reduced hereditary angioedema attacks by 87% versus placebo over weeks 5 to 28 (mean monthly attack rate 0.26 versus 2.10), and 62% of treated patients were attack-free and therapy-free versus 11% on placebo. confirmedas of 2026-10-10

    2. [2]

      On 8 September 2026 Intellia said the FDA had accepted its lonvo-z application with Priority Review and a target action date of 10 March 2027; the company says lonvo-z would be the world's first in vivo CRISPR-based therapy if approved. confirmedas of 2026-10-10

    3. [3]

      On 29 October 2025 the FDA placed clinical holds on Intellia's Phase 3 MAGNITUDE and MAGNITUDE-2 trials of nex-z after a patient dosed in MAGNITUDE had Grade 4 liver transaminase elevations and increased bilirubin. confirmedas of 2026-10-10

    4. [4]

      Intellia's August 2026 quarterly report states that the MAGNITUDE participant who had Grade 4 liver transaminase elevations and increased bilirubin after a nex-z dose died on 5 November 2025, and that the principal investigator reported the cause as septic shock secondary to a perforated duodenal ulcer, with a clinical course that also included acute liver injury treated with corticosteroids and an autopsy report supporting the clinical diagnoses. confirmedas of 2026-08-06

    5. [5]

      In July 2025 the FDA said three deaths appeared to result from acute liver failure in people treated with AAV-vector gene therapies from Sarepta, requested that Sarepta suspend Elevidys distribution, and placed some of its trials on clinical hold. confirmedas of 2026-10-10

    6. [6]

      Early CRISPR therapies edited cells outside the body (ex vivo); newer approaches deliver the editor into the patient (in vivo), which a 2025 review notes faces lower delivery efficiency, off-target effects and instability. confirmedas of 2026-10-10

    7. [7]

      In November 2025 CRISPR Therapeutics reported Phase 1 data, published in NEJM, for CTX310, an LNP-delivered CRISPR therapy that edits the ANGPTL3 gene in liver cells; at the highest dose mean ANGPTL3 fell 73%, triglycerides 55% and LDL cholesterol 49%, with no treatment-related serious adverse events reported. confirmedas of 2026-10-10

    8. [8]

      The FDA lifted the clinical hold on Intellia's MAGNITUDE-2 trial in January 2026 and the hold on MAGNITUDE in March 2026, after the company agreed to study modifications including enhanced monitoring of liver laboratory tests; enrollment then advanced in both Phase 3 trials. confirmedas of 2026-08-06

    9. [9]

      A 2025 review of CRISPR therapeutics notes that some therapies have low engraftment or editing rates, do not work, or have serious or even fatal side effects. confirmedas of 2026-10-10

    10. [10]

      Casgevy's prescribing information carries warnings for neutrophil engraftment failure, delayed platelet engraftment, hypersensitivity reactions and off-target genome editing risk. confirmedas of 2026-10-10

    11. [11]

      In a 2021 NEJM study, NTLA-2001, a lipid nanoparticle carrying Cas9 mRNA and a guide RNA targeting the TTR gene, given by infusion, lowered blood TTR protein by a mean 87% at the 0.3 mg/kg dose, with mainly mild adverse events. confirmedas of 2026-10-10

    12. [12]

      Nex-z (nexiguran ziclumeran) is Intellia's CRISPR/Cas9 therapy targeting the TTR gene for transthyretin (ATTR) amyloidosis with cardiomyopathy and hereditary ATTR with polyneuropathy, developed with Regeneron. confirmedas of 2026-10-10

    13. [13]

      On 1 July 2026 the FDA expanded Casgevy's approval to patients aged 2 years and older with sickle cell disease or transfusion-dependent beta-thalassemia, 53 days after filing, under the Commissioner's National Priority Voucher pilot program. confirmedas of 2026-10-10

    14. [14]

      On 25 February 2026 the FDA published draft guidance on a "plausible mechanism" framework for generating evidence of effectiveness and safety for individualized therapies that target specific genetic conditions with a known biological cause; comments closed on 27 April 2026. confirmedas of 2026-10-10

    15. [15]

      The case was reported in NEJM in May 2025; CPS1 deficiency has an estimated 50% mortality in early infancy, and in the seven weeks after the first infusion the infant tolerated more dietary protein and half the starting dose of a nitrogen-scavenger medicine, with no serious adverse events. confirmedas of 2026-10-10

    16. [16]

      In July 2021 the World Health Organization issued its first global recommendations on governing human genome editing, including steps toward a trial registry and a confidential channel for reporting unethical or illegal research. confirmedas of 2026-10-10

    17. [17]

      In July 2026 Beam dosed the first patient in the global pivotal cohort of its BEAM-302 trial in AATD-associated lung disease. confirmedas of 2026-10-10

    18. [18]

      New Zealand's Medsafe cleared Prime Medicine's trial application for the Wilson disease prime editor PM577a in June 2026, the company's first clinical authorization for an in vivo prime-editing therapy, and the FDA cleared its US investigational new drug application in July 2026, establishing a global open-label Phase 1/2 study with initial clinical data anticipated in 2027. confirmedas of 2026-08-06

    19. [19]

      On 24 September 2026 Prime Medicine said the FDA had cleared its investigational new drug application for PM647, an in vivo prime editor designed to correct the E342K (Pi*Z) mutation in SERPINA1; the global single-arm Phase 1/2 study will start with adults who have lung-only AATD and then add a cohort with significant liver disease, with initial clinical data expected in 2027. confirmedas of 2026-09-24

    20. [20]

      Adeno-associated virus (AAV) vectors carry a single-stranded DNA genome of about 4.7 kilobases, which limits cargo size, and their long-lasting expression of Cas9 may raise off-target risk. confirmedas of 2026-10-10

    21. [21]

      Lipid nanoparticles can carry Cas9 mRNA and guide RNA into tissues, but reliably targeting tissues other than the liver remains a challenge. confirmedas of 2026-10-10

    Revision history (2)
    1. Page created.
    2. Updated the nex-z death and hold claims to the company filing versions (reported cause, January and March 2026 lifts) and the Prime Medicine program claims.

    Created Oct 10, 2026. Last reviewed by an editor on Oct 10, 2026. Next scheduled review: Jan 10, 2027.

    Cite this page

    "How safe is in vivo gene editing? The 2026 debate." ContentLora, updated Oct 10, 2026. https://contentlora.com/analysis/in-vivo-editing-safety-debate

    Spotted an error? Suggest a correction or emailcorrections@contentlora.com.

    Keep exploring

    ☰ Contents