Shanghai Jiao Tong University School of Medicine has established a task force to investigate the undisclosed death of a child who received an experimental gene therapy overseen by one of its researchers, and Nature has attached an editor’s note to a related paper, following joint reporting by Science and Retraction Watch published on July 23, 2026, with a follow-up on July 30, 2026. Both the university investigation and the journal’s review are open and unresolved as of this writing. Neither has produced a finding of misconduct, and Nature‘s notice is explicitly not a retraction.
What is confirmed
According to the Science/Retraction Watch reporting, a 6-year-old girl — referred to in the reporting by the pseudonym “Mei” to protect her identity — was the sole participant in an early-stage, single-patient (n-of-1) clinical trial of a base-editing gene therapy conducted in late March 2025 at Xinhua Hospital in Shanghai, affiliated with Shanghai Jiao Tong University School of Medicine’s Songjiang Research Institute. The therapy targeted a mutation in the CHD3 gene associated with Snijders Blok-Campeau syndrome (SNIBCPS), a rare neurodevelopmental condition. The trial used an experimental adenine base editor, delivered by viral vector, administered into the child’s spinal fluid.
The child died approximately seven days after the procedure. Science and Retraction Watch reported that clinicians and outside experts they consulted attributed the death to a severe immune reaction to the therapy. The reporting states that prior testing in monkeys had already shown signs of toxicity, including liver damage, before the human procedure went ahead, and that outside experts who reviewed the case for Science said the trial team downplayed those risks in describing the procedure to the child’s parents and appeared to have overlooked safety signals from the animal studies.
The neuroscientist who led the effort, Zilong Qiu of Shanghai Jiao Tong University School of Medicine, and colleagues subsequently published a paper in Nature in early 2026 — “In vivo base editing of Chd3 rescues behavioural abnormalities in mice” — reporting proof-of-concept preclinical work on the same base-editing approach in mice and non-human primates. According to the Retraction Watch account, the published paper did not disclose the child’s death, did not disclose the adverse findings in the monkey studies, and did not disclose that the child’s family had contributed more than $800,000 of their own and relatives’ savings toward developing the therapy. An earlier preprint version of related work reportedly thanked “families with SNIBCPS patients for their participation and support,” language that did not appear in the same form in the published paper.
The institutional and journal response
Shanghai Jiao Tong University School of Medicine announced, in the week of July 29, 2026, that it had formed what Retraction Watch described as a special task force to conduct a comprehensive investigation into the incident; the university has not published a timeline for completing it. Neither Qiu, the university, nor Xinhua Hospital had responded to requests for comment at the time the July 30 Retraction Watch article was published.
Nature posted an editor’s note to the paper on July 29, 2026, reading: “Readers are alerted that concerns have been raised regarding this article, including with the data presented. Further editorial action will be taken if appropriate as soon as the investigation into the concerns is complete and all parties have been given an opportunity to respond in full.”
Why the distinction between an editor’s note, an expression of concern, and a retraction matters
It is worth being precise about what has and has not happened, because the terms carry different meanings and different evidentiary thresholds under journal and publisher policy. An expression of concern is a formal editorial notice, typically issued under guidance from the Committee on Publication Ethics (COPE), used when a journal has well-founded concerns about a paper’s reliability or the conduct underlying it but the matter is not yet resolved — it is explicitly not a finding that the paper is wrong or that misconduct occurred. Nature‘s July 29 notice functions in that space but is worded and labelled by the journal as an editor’s note pending investigation, a step that signals concern without yet reaching the more formal designation. A retraction is a different and more severe action still, reserved for findings that a paper’s central conclusions cannot be relied upon; none has occurred here. Readers and institutions tracking this story should not treat the current notice as equivalent to either a retraction or a conclusion that research misconduct occurred — both would require the investigation to actually conclude and make findings.
The disclosure and oversight questions this raises
Independent of what the investigation ultimately finds, the reporting raises several disclosure questions that map directly onto standard research-administration practice:
- Adverse event and safety-signal disclosure. A death occurring within days of an investigational intervention is, under standard clinical trial oversight frameworks, the kind of event that triggers mandatory reporting obligations to institutional review bodies and, in many jurisdictions, to regulators — see how this is normally handled in Adverse Event Reporting to the IRB and the underlying definitions of an adverse event and a serious adverse event (SAE). Preclinical toxicity findings in animal models are also a recognized category of safety signal that ordinarily must inform both risk disclosure to participants and any subsequent publication.
- Financial disclosure. U.S. and international norms for clinical investigator financial disclosure — codified for FDA-regulated trials in 21 CFR Part 54 and, more broadly, in journal and funder conflict of interest disclosure requirements — exist precisely so that a funding relationship as direct as a patient family financing the treatment’s development is visible to reviewers, readers, and oversight bodies, not omitted from the published record.
- Informed consent for pediatric and first-in-human procedures. The reporting describes a consent process using simplified language for the child and raises questions, per outside experts quoted, about whether risk was adequately conveyed to the parents given the known preclinical toxicity signals. CASRAI’s informed consent form (ICF) entry outlines what a compliant consent process is expected to document.
- Single-patient trial registration and oversight. Single-patient, or n-of-1 trial, designs are an established and legitimate methodology, particularly for ultra-rare genetic conditions, but they still fall under the same clinical trial registration and ethical-review expectations as larger trials; the reporting has raised questions about whether that oversight was fully applied here.
What CASRAI is and is not asserting
CASRAI is not in a position to independently verify clinical or laboratory findings, and this article does not do so. Every specific claim above is attributed to the Science/Retraction Watch reporting and, where quoted directly, to Nature‘s own published notice. No named individual is accused here of research misconduct, fraud, or any other finding — that determination, if any, rests with Shanghai Jiao Tong University’s investigation and with Nature‘s own editorial process, both of which are ongoing. This entry will be a candidate for update if either body publishes a conclusion.
For institutions and research offices, the episode is a live illustration of why disclosure obligations — adverse events, preclinical safety signals, and funding relationships alike — are treated as non-negotiable in clinical research governance: they are what allow reviewers, journals, and future patients to make informed decisions, and their absence is what turned a scientific result into an investigation. See also CASRAI’s guide to respondent rights during a research misconduct investigation for how such investigations are normally expected to proceed, and the research integrity officer (RIO) entry for the institutional role typically responsible for coordinating a response like the one SJTU has now opened.
