Read our COVID-19 research and news.

Health minister Marisol Touraine of France.

Health minister Marisol Touraine of France.

Philippe Grangeaud/Flickr (CC BY-ND 2.0)

French company bungled clinical trial that led to a death and illness, report says

Why one man died and four others fell ill during a drug safety study in France last month is still very unclear. But a preliminary inspection report lashes out at Biotrial, the company that conducted the study, for how it responded after the first volunteer in the clinical trial was hospitalized. Three major errors by Biotrial put other volunteers at risk, says the report, published yesterday by France's General Inspectorate of Social Affairs (IGAS).

Biotrial was testing a compound called BIA 10-2474, developed by Portuguese pharma company Bial as a candidate drug for a range of diseases. The study started in July 2015, and initial administrations of the drug produced no severe side effects; things went wrong in a group of eight people who entered the study on 6 January and received multiple, high doses. Six of them were scheduled to receive 10 daily doses of BIA 10-2474, two others a daily placebo. On 10 January, a subject identified as "volunteer 2508" complained of headaches and blurry vision; he was taken to the hospital in the early evening and stayed there overnight.

The next morning at 8 a.m., Biotrial staff proceeded to give the seven remaining volunteers their daily dose of BIA 10-2474 or placebo without first finding out how volunteer 2508 was doing. Biotrial apparently assumed his condition wasn't serious and that he was recovering. "We expected to see him come back," one staffer told the investigators. "The hospital didn't call us," said another. But at 9 a.m., a Biotrial doctor was told by the hospital that the patient's condition had worsened and that he had been sent to get an MRI. The patient was declared brain-dead later that day, and the study was halted. Four others who had received their doses that day fell ill and subsequently needed hospitalization.

The fact that the remaining volunteers weren't given any information about the status of volunteer 2508 that Monday morning is the second major shortcoming, the report says. As a result, they were unable to reconsider their participation in the study. This violated what Biotrial had promised in an informed consent form: "You will be informed about any new significant information that could affect your willingness to continue the trial."

The third big mistake was not reporting the disaster to the French National Agency for Medicines and Health Products Safety (ANSM) until 14 January, 3 days after Biotrial stopped the study. That made it impossible for ANSM to take measures to protect volunteers who took part earlier or those enrolled in other studies of the same class of molecules.

Because BIA 10-2474 works on the body's endocannibinoid system, it would also have been better to exclude volunteers who regularly use cannabis, the report says, to reduce the risk of any interactions. Biotrial's exclusion criteria for use of psychoactive substances were inconsistent and not precise enough, the report says. According to the company, however, volunteer 2508 had declared that he didn't use cannabis; a urine test taken before the trial started had come back negative. The report says more study is needed on this issue.

How the other four patients are doing isn't entirely clear. At a press conference where she presented the report yesterday, French health minister Marisol Touraine said they are improving and have all gone home, but that it's too soon to say whether they can fully recover. Last month, a doctor who treated them said that three might have irreversible brain damage.

The study of BIA 10-2474 had been going on for half a year. Between July 2015 and October 2015, Biotrial tested single doses of the drug on a total of 64 volunteers, with doses increasing from 0.25 to 100 milligrams. After that, 12 volunteers took part in a study that compared a single 40-milligram dose taken on an empty stomach with the same dose combined with a fatty lunch. In November 2015, a series of multiple-dose trials began, each enrolling eight subjects. The first group received 2.5 milligrams daily, the second 5 milligrams, and so on; the accident happened in the fifth group, which received 50 milligrams daily, the highest dose.

The 90 people who received BIA 10-2474 earlier have all been contacted; so far, 75 have had a checkup and an MRI, Touraine said yesterday, and all seem fine.

IGAS will not present a full report on the matter until March. ANSM is doing its own investigation, as is the public prosecutor's office in France.

Responding to the report, Touraine said that from now on any hospitalization during a clinical trial should be regarded as "new fact" that needs to be brought to the attention of health authorities immediately. Such events should "lead to an immediate suspension of the trial until the safety of volunteers is guaranteed," she said. "Volunteers must be clearly informed about the suspension of the study and its reasons."