Sometime in February 2026, the public registry entry for HuidaGene Therapeutics’ HG302 study was updated to read “complete.” The study had enrolled four boys with Duchenne muscular dystrophy, all between four and eight years old, at Shanghai Children’s Medical Center. The entry did not mention that one of the four had been dead since the previous August.
The boy died in August 2025 of acute respiratory distress syndrome, brought on by a severe immune reaction to a high dose of the adeno-associated virus vector carrying HuidaGene’s gene-editing therapy, STAT reported. He was the fourth child dosed and had received the highest dose given in the trial to that point. The company’s chief executive, Alvin Luk, presented early data from the study at a scientific conference around that time and did not disclose the death. HuidaGene did not acknowledge it publicly until August 5, 2026, and only after reporters kept asking. For most of the year in between, the official record said the study had finished cleanly.
HG302 ran without an independent data monitoring committee. Such a board exists to halt a study when a patient is harmed, before the next patient is dosed. No board outside HuidaGene reviewed the dose escalation before it reached the fourth child, and none flagged the death when it came. The registry did the opposite of flagging it.
The HuidaGene death is not the only one. It is the third to surface in the same corner of China’s trial system in about a year. In May 2025, a six-year-old girl with the rare neurodevelopmental disorder Snijders-Blok-Campeau syndrome received the first gene-editing therapy directed at the human brain, at Xinhua Hospital in Shanghai, and died seven days later of thrombotic microangiopathy. The hospital’s own ethics board judged the death “definitely related” to the treatment, according to Science, and the case was never made public until reporters found it. On August 9, 2026, RiboX Therapeutics reported that a patient with systemic sclerosis had died after receiving its experimental in vivo CAR-T therapy; the patient, the company said, had significant pre-existing conditions. Two children and one adult, three companies, one shared feature: all three were investigator-initiated trials, opened by hospitals without the central regulatory sign-off a national drug agency would demand.
Follow the incentive and the pattern stops looking like coincidence. Investigator-initiated trials are faster and cheaper than the regulator-supervised kind: a company can generate early human data in a hospital-run study in Shanghai for a fraction of what a monitored U.S. trial costs, then carry that data back toward an American filing. The volume has moved with the money. China overtook the United States in annual clinical-trial registrations around 2020 and has widened the gap since, listing roughly 7,100 trials to America’s 6,000 by 2024, according to World Health Organization registry data.
The receipts reached Congress this month. Reps. John Moolenaar of Michigan and Ben Cline of Virginia wrote to Acting FDA Commissioner Kyle Diamantas asking the agency to stop accepting Chinese clinical data in drug applications unless FDA inspectors had audited the trial site within the previous year, and to review the products already approved in the United States on the strength of China-generated results, BioSpace reported. “Accepting Chinese clinical data poses real risks for patients,” the lawmakers wrote. The offshoring of early-stage trials to China, they added, “risks rewarding a system that has shown it is willing to treat children’s deaths as an acceptable cost of faster, cheaper research.”
The request lands on an agency that has spent years conceding it cannot inspect what it already oversees. The FDA inspected just 6 percent of foreign drug-manufacturing facilities in 2022, down from 37 percent in 2019, ProPublica found. Its reach into foreign trial sites is thinner still.
The Government Accountability Office counted 46 FDA inspections of clinical-trial sites in China across more than a decade, from 2012 through early 2023. Four years ago, experts told the BMJ that FDA oversight of clinical trials was “grossly inadequate.” When lawmakers say the agency takes foreign data on faith, they are describing the documented status quo, not a hypothetical: it accepts the results and rarely sets foot where they were produced.
Congress has opened a second front. In late June 2026, the House Select Committee on the Chinese Communist Party sent letters to five drugmakers, AbbVie, Bristol Myers Squibb, Eli Lilly, Merck, and Pfizer, demanding records by July 17 on where their China trials run, how consent was obtained, and how patient data is protected, Healio reported. The committee said Merck alone has sponsored or collaborated on 224 clinical studies in China since 2005, including 31 in Xinjiang and 40 at facilities with ties to the Chinese military, raising two questions: whether proprietary American biotech data could flow to the Chinese military through trials at military hospitals, and whether trials in Xinjiang met any real standard for voluntary, informed consent. Each letter noted there was no evidence the company had broken the law. At the time of Healio’s report, only Merck had responded publicly, affirming its commitment to good-clinical-practice standards.
The documents establish four things: three patients died, at least one death was hidden for a year while the registry called the study complete, that study had no independent monitoring board, and the FDA accepts data from this system while inspecting almost none of it. They do not establish that the cost-and-speed model caused any single death. A patient with advanced disease can die in a well-run study, and a child can suffer a fatal immune reaction under careful monitoring. What the records show is a structure that stripped out the checks meant to catch harm early, and a disclosure habit that kept the harm off the books until outsiders forced it into view. The pattern is too steep to wave off as three unrelated tragedies.
What remains is a question of who acts. As of this writing, Acting Commissioner Diamantas has not announced any change to how the FDA treats foreign trial data, and the letter to him sits unanswered. Four of the five drugmakers have said nothing publicly since the July 17 deadline passed. No safety board has convened to review the HG302 dose escalation, because there was never one to convene. The records that would settle the questions, the trial protocols, the inspection files, the consent forms from Shanghai and Xinjiang, remain with the agency and the companies that generated them. They have not been produced.
Sources
- STAT – Once again, child dies in gene-editing trial in China, rekindling debate on transparency and safety
- Science – Exclusive: Death of girl in Chinese gene-editing trial was never made public
- BioSpace – Lawmakers urge FDA to enact new China policies after 3 gene therapy deaths
- Endpoints News – Third death revealed in China’s popular but opaque trials
- Healio – US House committee opens investigations into five drugmakers over clinical trials in China
- House Select Committee on the CCP – Investigation into clinical trials at Chinese military sites
- ProPublica – FDA foreign facility inspections
- BMJ – FDA oversight of clinical trials is “grossly inadequate,” say experts