A Shanghai Trial Edited a Child's Brain, Then Buried the Death

Neuroscientist Zilong Qiu's team at Shanghai Jiao Tong University designed a CRISPR-derived base editor to convert the mutant adenine toward the healthy letter without cutting both DNA strands. The editor was too large for one delivery vehicle, so the instructions rode two AAV9 vectors into cerebrospinal fluid on March 24, 2025, at Xinhua Hospital. Both vectors had to reach the same neurons. The family had raised on the order of $860,000 to help fund the personalized build. Seven days later she was dead of thrombotic microangiopathy, a clotting catastrophe already known to trail high-dose AAV work. The hospital ethics board judged the link "definitely related."

Signals ignored

Monkey toxicology had already drawn blood on the page. All four treated animals developed moderate-to-severe liver injury; one high-dose animal also showed kidney damage. CRISPR Medicine News, summarizing the Science report, notes that the hospital ethics committee approved the single-patient trial before it reviewed the final toxicology report. Independent experts questioned whether enough neurons could be edited to matter, and whether the risks were explained with the honesty a nonfatal baseline demands.

The trial ran as investigator-initiated research inside the hospital. That route did not require prior review by China's National Medical Products Administration. Months later, local health authorities fined the hospital for oversight and registration failures and left the lead researcher unsanctioned. He Jiankui's 2018 embryo-editing scandal was supposed to have closed this kind of gap. The gap reopened as paperwork.

The Nature-shaped hole

Neither the team nor the hospital disclosed the death when it happened. When associated preclinical research appeared in Nature on February 18, 2026, the girl's case was absent. The family later asked that related trial write-ups reflect what the therapy actually did. A scientific paper is a public instrument. Publishing the animal story while the human endpoint stayed offstage is how a field launders ambition into prestige.

The Chinese Society of Gene and Cell Therapy answered the disclosure with the language of reform: stronger preclinical evidence, tighter ethics governance, timely reporting of serious adverse events, and China's newer State Council Order No. 818 for investigator-initiated research. Those sentences are necessary. They arrive after a child is gone.

Gene editing will keep pressing toward the brain because that is where some of the cruelest single-letter diseases live. The Shanghai case warns about sequence. Ambition is not the indictment. When the indication is nonfatal, when primate livers are already failing, and when an ethics board signs before the tox file is closed, the first human should not be a fundraising milestone. And when that human dies, the literature does not get to pretend the experiment never left the animal house.