The hospital, not the sponsor: where Chinese gene-therapy deaths are supposed to land first

Investigator-initiated trials quietly redistribute who owns a death. Under China's IIT rules, the hospital ethics committee, not the sponsor, is the first place a serious adverse event is supposed to land. The sponsor's job is to support; the institution's job is to record, report, and gate. When that split works, a fatality produces a paper trail within days. When it does not, the trail shows up a year later, in someone else's press release.

That is the pattern behind a 12-month silence around a boy's death in HuidaGene's HG302 Duchenne gene-editing trial at Shanghai Children's Medical Center. A Tech Times investigation, citing the company's own timeline, found that HuidaGene knew by January that it had enough information to submit its findings for peer review, and told no one else for another seven months. Wire versions have read the story as 'experts question HuidaGene.' The structural reading is sharper: in an IIT, the institution is supposed to be the first responder, and a 12-month gap is the institution failing the same way the sponsor did. Whether this reflects an institutional failure specific to this IIT structure or a broader systemic gap is a bounded inference from the documented timeline and IIT rules, not a proven causal link.

The mechanism may be portable: the documented IIT accountability gap in the HG302 case is one data point suggesting that any jurisdiction handing disclosure duty to a hospital ethics board without a public clock, a fine, or a registry lane creates conditions for a similar silence—but this is a generalization from a single case, not a proven systemic rule. The thing to watch is not the next HuidaGene statement. It is whether the hospital ethics board, the NMPA, or the trial registry files anything before the company does.