At least a dozen people have died in clinical trials of gene therapies in the United States, Europe and Russia over the past three decades [7], which averages out to fewer than one a year [8]. That is a numerator without a denominator: nothing in the record says how many patients were dosed to produce it, so two deaths in China cannot be read as evidence that Chinese trials carry more risk. What separates these two cases is who did the finding. One death was disclosed by the company running the trial and the other by reporters, not by a regulator [27].
Gene therapies can cause severe immune reactions that kill [6]. The objection researchers are raising here is narrower, and concerns a decision taken with harm data already in hand [9]. Fyodor Urnov, who studies molecular therapeutics at the University of California, Berkeley, says the trial should not have gone ahead once the therapy showed evidence of harm in animals [21], and that the research and medical team committed several "egregious wrongs" [18]. Every animal in the toxicology cohort was affected [24]. Philippe Campeau of the University of Montreal, who helped characterize Snijders Blok-Campeau syndrome, says the girl's symptoms were mild [23] - the other side of a risk-benefit calculation that the syndrome's fuller presentation, including impaired intellectual development, might otherwise support [16]. Zilong Qiu, who developed the treatment, also corresponded directly with the family [19]. Neither he nor the clinician Yongguo Yu answered Nature's questions about the toxicology studies [22], and Shanghai Jiao Tong University did not answer Nature's questions about its investigation [5].
Investigator-initiated trials let researchers test therapies quickly because the drug regulator is not supervising them [11]. That is also why the only parties positioned to read the monkey toxicology results before the girl was dosed were the people who had developed the therapy and the hospital administering it [20]. Jiayou Shi, a lawyer at the National Research Center for Civil and Commercial Law at Renmin University of China, expects gene-therapy regulations to be strengthened in response [12]. Joy Zhang, a sociologist of science at the University of Kent, puts a halt to gene-editing research or stricter funding requirements on the table [13], and says the damage to the credibility of Chinese institutions is already done [10].
The thing this does not tell you is cause. Nature's account places the girl's death after she received the therapy and states no cause of death for either child [25]. It does not say when HuidaGene learned of the boy's death, so the interval before the 5 August announcement is not in the record [26]. No partner, licensee or investor outside China appears in the account at all [28], which means the case that outside collaborators are newly exposed rests on rules not yet written rather than on anything documented so far. The measurable question is whether the May increase in oversight [11] puts an independent reader of animal data between a developer and a patient before the next first-in-human dose.
