A second undisclosed child death in a Chinese gene-editing trial has come to light — this time at HuidaGene Therapeutics, a Shanghai-based biotech, in a trial separate in every material respect from the Shanghai Jiao Tong University case reported weeks earlier. According to an investigation published by STAT News on August 5, 2026, a boy enrolled in HuidaGene’s HG302 trial for Duchenne muscular dystrophy (DMD) died of acute respiratory distress syndrome (ARDS) during a severe immune reaction after receiving a high dose of the company’s experimental CRISPR-based therapy — a death the company did not disclose for roughly twelve months, until STAT’s reporters pressed it directly.
What STAT’s investigation found
HG302 is an investigational gene-editing therapy built on HuidaGene’s proprietary CRISPR variant, which the company calls hfCas12Max, delivered intravenously via an adeno-associated virus (AAV) vector. The therapy is designed to edit the dystrophin gene so that a patient’s body resumes producing a shortened but partially functional form of dystrophin, the muscle-structural protein missing in DMD. According to STAT’s reporting, the trial enrolled four boys, aged four to eight, at Shanghai Children’s Medical Center. One of them died in August 2025 after high-dose systemic administration of the AAV vector triggered ARDS — a recognized, serious risk of high-dose systemic AAV gene therapy that has caused patient deaths in other DMD gene-therapy programs outside China as well.
STAT reports that the trial had no independent data safety monitoring board, meaning the dose-escalation decision that preceded the death was never reviewed by a body independent of the trial’s own investigators. The company disclosed the death only after a monthslong STAT investigation and repeated questions, more than a year after it occurred. In the interim, according to the reporting, HuidaGene’s public communications went quiet: its CEO had presented preliminary, inconclusive data at a 2025 conference, both the CEO and the company’s chief technology officer left in the summer of 2025, and a clinical trial registry listing was updated in February 2026 to mark the study “complete,” without disclosing the death or clarifying the status of the trial’s remaining participants.
How this differs from the Shanghai Jiao Tong case
This is a distinct case from the death CASRAI covered in Shanghai Jiao Tong Opens Investigation After Undisclosed Death in Gene-Editing Trial, published days earlier. The two incidents share a pattern — an undisclosed pediatric death in a Chinese gene-editing trial, surfaced by investigative reporting roughly a year after the fact — but they are not the same event, and conflating them understates how systemic the underlying gap is. The Shanghai Jiao Tong case involved a 6-year-old girl who died about a week after an in vivo base-editing procedure targeting a CHD3 mutation for Snijders Blok-Campeau syndrome, delivered via viral vector into spinal fluid at Xinhua Hospital, and surfaced through joint reporting by Science and Retraction Watch tied to a related Nature paper. The HuidaGene case involves a different institution (Shanghai Children’s Medical Center), a different sponsor (HuidaGene Therapeutics, not a university-affiliated academic lab), a different disease target (DMD, not SNIBCPS), a different editing modality (a CRISPR nuclease variant, not a base editor), a different route of administration (systemic intravenous AAV, not intrathecal), and a different cause of death (ARDS from an immune reaction to a high systemic dose, not the reaction described in the Shanghai Jiao Tong case). What connects them is not the trial itself but the regulatory pathway both used.
The regulatory gap: China’s investigator-initiated trial pathway
Both deaths occurred in trials run under China’s investigator-initiated trial (IIT) framework — a pathway, as STAT’s reporting describes it, that lets Chinese hospitals and their investigators start human studies without prior review or approval from national drug regulators. This differs sharply from China’s sponsor-led registration trial pathway, in which the National Medical Products Administration (NMPA) reviews and approves a trial before it can proceed, in the same way FDA review of an Investigational New Drug (IND) application gates a US sponsor-led trial. An investigator-initiated study is, by definition, one where the investigator — not a commercial or institutional sponsor — holds primary responsibility for the trial’s design and conduct; in the US, that structure still sits inside FDA’s IND framework and its associated oversight obligations. China’s IIT pathway, as used in both of these cases, removes the up-front regulatory gate entirely, leaving oversight to the hospital’s own ethics review and the investigator’s own judgment about safety monitoring, dose escalation, and disclosure.
Neither case reported here involved an independent Data Safety Monitoring Board (DSMB) reviewing safety data as the trial progressed. In a conventional sponsor-led trial with a documented Data Safety Monitoring Plan, a dose-escalation decision following a prior serious adverse event would ordinarily be reviewed by a body independent of the investigators before proceeding. STAT’s reporting states that did not happen in the HuidaGene trial. That absence, not any single design flaw in the gene-editing approach itself, is the through-line between these two cases.
Why safety events in IITs go under-reported
Sponsor-led registration trials carry layered reporting obligations: a regulator has already reviewed the protocol before enrollment starts, a sponsor typically maintains an independent DSMB or equivalent safety oversight body, and serious adverse events generally trigger defined reporting timelines to the regulator and, in many jurisdictions, to a public trial registry. Clinical trial transparency norms — prospective registration, timely results reporting, disclosure of adverse events — were built around that sponsor-led structure. An IIT that never passes through a sponsor’s regulatory-facing infrastructure can fall outside all of it: there is no regulator who was notified at the outset and is therefore positioned to ask why a death wasn’t reported, and a registry entry can be quietly updated to “complete” without the update itself triggering any external review. In both cases reported this month, the deaths came to light only because investigative journalists asked direct questions — not because a regulatory or institutional reporting mechanism surfaced them.
What this means for research administrators and IRBs
For institutions and ethics boards reviewing collaborations, co-enrollment, or data-sharing arrangements that touch investigator-initiated trials conducted overseas, these two cases are a concrete argument for treating the sponsor-vs-investigator-initiated distinction as a real risk factor in due diligence, not a technicality. Questions worth asking before an institution attaches its name to, cites results from, or shares data with an overseas IIT: Is there an independent DSMB or equivalent safety-review body, and does it actually convene before dose-escalation decisions? What is the trial’s registration and results-reporting status, and has it been updated to reflect adverse events as they occur rather than only at closeout? And where a partner institution or co-investigator’s own paper or dataset draws on the trial’s results, has that fact been disclosed to the journal, funder, or IRB reviewing the associated work — the gap that also drove the parallel Nature editor’s-note review in the Shanghai Jiao Tong case? Neither of these deaths reflects a flaw specific to gene-editing therapy as a modality; both reflect what happens when a trial proceeds without the independent safety-review layer that sponsor-led oversight normally provides.
Frequently asked questions
Is this the same case as the Shanghai Jiao Tong University gene-editing death?
No. They are two distinct deaths, in two distinct trials, at two distinct institutions, involving two different diseases and two different gene-editing techniques, disclosed within weeks of each other by different investigative outlets. See CASRAI’s separate coverage of the Shanghai Jiao Tong case for that trial’s specifics.
What is an investigator-initiated trial (IIT), and why does China’s version lack regulatory pre-approval?
An investigator-initiated study is a trial designed and led by the investigator rather than a commercial or institutional sponsor. As STAT’s reporting describes it, China’s IIT pathway lets hospitals and their investigators begin such studies without the prior review a sponsor-led registration trial would need from the National Medical Products Administration — leaving safety oversight to the hospital’s own ethics process rather than a national regulator.
Was there independent safety monitoring in the HuidaGene trial?
STAT’s reporting states the trial had no independent data safety monitoring board, so the dose-escalation decision that preceded the fatal adverse event was not reviewed by a body independent of the trial’s own investigators.
Sources
Primary source: STAT News, “Once again, child dies in gene-editing trial in China, rekindling debate on transparency and safety,” August 5, 2026. The death was also reported by Bloomberg and other outlets in the days following STAT’s report. This article was last checked against the cited source on August 7, 2026; details in STAT’s subscriber-only reporting beyond what is described here have not been independently verified by CASRAI.
