It was an unsettling sense of deja vu for researchers and bioethicists when French authorities reported in January that one man died and five others were seriously injured after being administered an investigational drug in a Phase I clinical trial.
The clinical trial community had followed a seemingly similar Phase I disaster in London, England, a decade earlier when a study drug critically injured six men, resulting in swollen heads, unconsciousness, and extreme pain. In the 2016 incident, one study volunteer died and five others suffered from serious neurological problems, according to a report by ANSM, the French national drug safety agency.
“My initial thought was that history had repeated itself,” says Stephen Senn, PhD, head of the Competence Center for Methodology and Statistics General Management at the Luxembourg Institute of Health.
“However, although the outcome — six affected and one severely so — was extremely similar, the background to what happened was very different,” Senn says. “If anything, this trial is more worrying.”
The 2006 trial was later seen to have been conducted rashly, giving too many volunteers the study drug at once when the trial began. The 2016 study had taken a more cautious approach.
“In the case of [the 2006 trial], it is easy to see how with more care the problems could have been avoided,” Senn says.
In both cases, the volunteers’ adverse reactions occurred quickly and virulently, and there were no early indications of how it could have happened. Both disasters also raise questions about informed consent in Phase I trials and whether volunteers can truly appreciate the risks of first-in-human investigational drug studies.
The most recent incident involves a study drug referred to as BIA 10-2474. The Phase I trial was conducted by Biotrial, a clinical research organization, on behalf of Bial, a Portuguese pharmaceutical company. The new molecule under development was an FAAH enzyme inhibitor for treating pain, and the trial had been approved by French regulatory authorities, according to a February 2016 news release by Bial.
The 2006 case involved a humanized agonistic anti-CD28 monoclonal antibody, called TGN1412, which was studied by TeGenero AG of Wurzburg, Germany. (For more information, see the May 2006 issue of IRB Advisor.)
“Clearly, there is a comparison to be made with TGN1412 in that both ended in disaster, but it’s difficult to say more yet,” says Noel Snell, an honorary senior lecturer at the National Heart & Lung Institute in Great Britain.
“TGN1412 was a biological, and, as I understand it, BIA 10-2474 is a low-molecular-weight compound, a standard enzyme inhibitor,” Snell explains.
In the 2006 British trial, six young volunteers were injected with the first doses of TGN1412 about the same time. All six began to exhibit serious symptoms, eventually requiring organ support and weeks of hospitalization. Critics said the trial should have begun with just one person receiving the first dose of the study drug and then proceeded slowly after monitoring the first volunteer’s reaction to the new agent, according to a December 2006 report by the Pharmaceutical Journal.
The report concluded that doses of study drugs should be calculated according to a minimal anticipated biological effect for new classes of medicines.
A recent ANSM report outlined the steps taken to study the BIA 10-2474 drug’s effect on volunteers, beginning with the first cohort of subjects on July 7, 2015. In that cohort, two volunteers were given the treatment (one placebo and one study drug) and 24 hours later, five more volunteers were given the study drug and one the placebo. There were no adverse events. The study proceeded, solely escalating the dose with second, third, and fourth cohorts in August, and fifth, sixth, and seventh cohorts in September and an eighth cohort in October.
After the initial single administration phases, the trial began a study of the interaction with food, followed by a multiple-dose stage. There were multiple, ascending doses, from 2.5 mg to 50 mg, and each dose involved a cohort of eight volunteers, with six receiving the product and two a placebo. For cohorts one through four, there were no serious adverse events, the ANSM report says.
Then, what had been a fairly routine Phase I trial changed: On Jan. 6, 19 days after the fourth cohort, six volunteers received the study drug at a dose of 50 mg. Three days after receiving the first dose of the study drug, one of the volunteers had an adverse reaction and was hospitalized. Within a few days, the remaining five volunteers who received the study drug also were hospitalized, French officials say.
“The study’s design was fine, and it was cautious,” says Arthur Caplan, PhD, director of the division of medical ethics at New York University School of Medicine in New York City.
In using a careful initial approach, it appears that researchers of investigational drugs have learned at least one lesson from the 2006 disaster, Caplan adds.
The Royal Statistical Society in Great Britain urged French investigators to include independent statistician members in their research into what went wrong. The society also called for greater transparency in the study’s regimen and information sharing.
The very fact that the French trial gave only one patient the study drug at initiation was an improvement over the TGN1412 trial, notes Senn, who commented about the tragedy in a Jan. 22, 2016, statement by the Royal Statistical Society.
“Here, it seems that a cautious dose-escalation was involved, with single doses first and then moving to multiple doses, a reduction in the unit dose,” Senn adds. “Furthermore, the problem arose with the last multiple dose cohort and, that, after an intended 10 days. The volunteer who died had only received five doses at the time he was hospitalized.”
IRBs and bioethicists might point to Bial and Biotrial’s decision to dose all subjects simultaneously after the first cohort as one factor in the disaster. “Bial and Biotrial seem to have ignored the advice to not dose all subjects simultaneously — although, of course, this has always been standard practice until the TGN1412 problems,” Snell says. “Although this may reduce risks — depending on how long it takes for adverse reactions to develop — it only reduces the number of subjects potentially at hazard because the first one still has a problem.”
IRBs and clinical research sites could use the two Phase I trial disasters as a cautionary tale, suggesting that greater attention needs to be paid to informed consent, bioethicists suggest.
“Informed consent should give an idea of the overall risk to subjects in Phase I studies, which is incredibly low over the years, and also potential specific risks from the agent being studied,” Snell says. “This will be highly dependent on the pre-clinical tox [finding], which we know is a poor predictor of human side effects, but I’d like to see the tox package for this drug.”
But anecdotal evidence suggests that Phase I study volunteers might ignore the risks and focus on other things, including the money they receive for participating: “It’s debatable how ‘informed’ the consent process is,” Snell says. “When the TGN1412 disaster was all over the media, applications to participate in Phase I studies actually increased.”
As the British Phase I trial disaster showed, healthy volunteers for these first-in-human drugs often are seduced by the money and don’t consider the risks, Caplan says.
“People who sign up for Phase I studies are younger and pretty much think of themselves as invulnerable,” he said.
Senn would recommend that informed consent documents in Phase I studies include a separate document that explains the risk assessment that was made before the trial began.
There is one way that the research community could enhance participant safety in Phase I trials, and that would be to create a systematic registry that is used to prevent people from signing up for more than one Phase I trial at a time, Caplan says.
“People can shop around and lie about their participation,” he adds.
“Sadly, the lesson is there is danger in every Phase I study, but it’s just not common,” Caplan says. “Two disasters in 10 years is like the airlines — a good safety record.”
IRBs and clinical research sites simply need to make certain Phase I studies have a plan for dealing with any adverse event or catastrophe that might occur, including prompt reporting, transparency, shutting down the trial quickly in the event of a problem, and having a failure plan, he suggests.
From an IRB and investigators’ perspective, the take-home message is to never get comfortable in any Phase I trial. “You still must proceed cautiously and wait and see what happens,” Caplan explains. “That’s the best chance of minimizing risk.”