Six-Year-Old Girl Dies After Brain Base Editing: Undisclosed Incident Exposes Regulatory Gaps

A personalized treatment that had carried the hopes of a family affected by a rare disease is becoming a grave case study in the clinical governance of gene editing. A joint investigation by Science and Retraction Watch found that a six-year-old girl with Snijders Blok–Campeau syndrome died within seven days after receiving an experimental brain base-editing therapy in Shanghai in March 2025; her death did not enter public view until more than a year later, following a media investigation.

The girl carried the p.R1025W variant in the CHD3 gene and had impairments in language, cognition, and motor development, but her condition was not immediately life-threatening. The research team designed two adeno-associated virus (AAV) vectors, each carrying part of a split adenine base editor, and delivered them into the cerebrospinal fluid through an intrathecal injection, with the aim of correcting a single DNA letter in neurons. This early-stage, open-label trial planned to enroll only one child. Its original primary outcome measure was therapy-related serious adverse events within 26 weeks after administration.

However, whether the safety data were sufficient to support moving into human testing has become the central controversy. According to information obtained by Le Monde, all four primates in the toxicology study developed liver injury, and one also developed kidney injury; the hospital ethics committee reportedly did not review the toxicology report before approving the treatment. The girl also developed kidney abnormalities after receiving the therapy, and the hospital ethics committee later determined that her death was clearly related to the therapy and involved thrombotic microangiopathy. These details currently come mainly from investigative reporting and related documents; the complete medical records and findings of the regulatory investigation have not been made public.

There is also a clear discrepancy between the public record and the actual outcome. Information for trial NCT06860672 on ClinicalTrials.gov has not been updated since March 6, 2025. When the investigation was published, the trial was still listed as recruiting and did not disclose that its sole participant had died. The report also stated that the girl’s family raised approximately US$860,000 to develop the therapy, but neither this funding nor the death appeared in the trial record or in a related paper published later.

An animal study published by the research team in Nature in February 2026 showed that the therapy could correct some genes, restore protein levels, and improve certain cognitive, motor, and autism-related behaviors in mice carrying the human CHD3 variant; primate experiments also showed widespread neuronal transduction and reassembly of the editor. On this basis, the paper supported the feasibility of clinical translation, while also acknowledging challenges including high vector doses, editor assembly efficiency, and off-target and bystander editing. Behavioral improvements in animal models cannot answer whether humans can tolerate the immune response and organ toxicity.

Seven external experts interviewed questioned whether the team had adequately explained the risks to the family, properly addressed the animal safety signals, and reasonably assessed the potential benefits. The family has requested an investigation and the retraction of the related paper. Nature said it had not been informed during submission and peer review that the human trial was being planned or conducted. It has now launched an investigation and added an editorial warning on July 29, alerting readers that the paper and some of its data had been called into question.

The incident also exposes a gray area in current systems governing “single-patient personalized therapies”: when development, funding, ethics review, and clinical implementation are highly concentrated, independent risk assessment and incident reporting become especially important. Shanghai’s local health authorities reportedly penalized Xinhua Hospital in September 2025 for inadequate oversight and for failing to register the research as a commercially funded project. Whether the fine and subsequent accountability measures can address the death of a child, the failure to disclose data promptly, and questions over whether informed consent was adequate remains subject to a complete and independent investigation.