First Child to Have Brain Gene-Edited Died a Week Later. The Trial Was Never Registered.

She was six, and she was the first child ever to have her brain cells edited with a base-editing CRISPR therapy. A week after the procedure, she was dead. For the next sixteen months, almost no one outside the lab knew, because the procedure was never registered as a clinical trial, and no rule in any country forced the silence to end.

Mei was born with a mutation in the CHD3 gene, the cause of Snijders Blok-Campeau syndrome, a rare neurodevelopmental disorder that can mean intellectual disability, a head circumference that often runs larger than average, elevated risk of seizures, and, in some children, congenital heart defects. There is no approved treatment. The family, identified in joint reporting by Science Magazine and Retraction Watch only as "Mei," found Zilong Qiu's work through a private WeChat group for parents of children with autism and developmental disorders, then helped fund the attempt by paying more than $800,000, per the joint investigation.

The procedure used a base editor, a CRISPR-adjacent tool that swaps a single DNA letter without cutting both strands of the helix, packaged inside two adeno-associated viruses and delivered via lumbar puncture into the cerebrospinal fluid. The viruses were designed to cross into brain cells and rewrite the CHD3 mutation in place. Hong Kong Free Press and the original Retraction Watch report describe the dose as hundreds of trillions of viral particles, an order of magnitude with no published precedent in a living human brain. Zilong Qiu is a neuroscientist at Shanghai Jiao Tong University School of Medicine.

Mei died in late March 2025, roughly a week after the injection. The cause of death has not been publicly detailed; the joint investigation reports a clinical course consistent with a serious adverse event tied to the procedure, including rapid neurological decline. What is on the public record is the next sixteen months. No registry entry was filed before, during, or after the procedure. China's Clinical Trial Registration system, the U.S. ClinicalTrials.gov database, and the WHO's International Clinical Trials Registry Platform have no record of a CHD3 base-editing trial under Zilong Qiu's name. The only formal record of the work is a 2026 Nature paper describing base editing of the CHD3 locus in mice, not in patients.

That gap is what made the silence possible. Trial registration, the kind a sponsor files on ClinicalTrials.gov or ChiCTR before the first patient is dosed, is what activates almost every downstream safeguard in modern clinical research: independent ethics review with a published protocol, a data and safety monitoring board that can pause the study, a serious adverse event reporting clock that forces a public disclosure within days, and a results database where the outcome must be posted within a year. None of those mechanisms touched this case, because none of them was triggered.

The minimum architecture for the next generation of these trials is a short list of rules that registration would activate. Before the first dose, the protocol, vector design, and intended delivery route must be filed in a public registry, with the registry entry locked before recruitment begins. After any serious adverse event, defined to include any death within 30 days of an investigational gene-editing procedure regardless of suspected causality, the event must be reported to the registry within 48 hours, with the public record updated at the same time. An independent data and safety monitoring board, with no financial relationship to the sponsor or the institution, must hold the authority to pause the trial and the obligation to publish its reasoning. Long-term outcomes for every dosed patient must be posted to a public registry, including deaths, in a standardized format, within 12 months.

On 2026-07-27, AFP reported that Shanghai Jiao Tong University School of Medicine had opened a formal investigation into the concealment and into Zilong Qiu's conduct. The university's official statement, published in Chinese and relayed in English through AFP and Gizmodo, asserts a commitment to research integrity and a commitment to opposing any research that violates research ethics. The statement does not name a timeline, an external panel, or a commitment to publish findings.

Three things will tell the field whether the SHTMU investigation becomes the model or another internal review: whether the panel is internal or external, whether its scope covers the absence of registration as well as the concealment, and whether the conclusions are public, with a published report and a posted registry correction. The first institutional review of a first-in-human brain gene-editing death is the moment when the rules either get written or get deferred to the next case.