France’s National Agency for Medicines and Health Products Safety (ANSM) released a report this week outlining how its Temporary Specialist Scientific Committee (TSSC) evaluated what went wrong with a first-in-man clinical trial in Rennes, France that resulted in the death of one healthy volunteer and the hospitalization of five others in January.
ANSM launched its investigation into the Phase I trial in January, which the contract research organization (CRO) running the trial, known as Biotrial, said was “conducted in full compliance with the international regulations and Biotrial’s procedures were followed at every stage throughout the trial.”
The CRO has since said that it “cannot, alone, be the only name associated with the painful events which led to the death of the volunteer...The accident happened during the clinical trial for the molecule BIA 10-2474, performed for the Portuguese pharmaceutical company, Bial.”
The compound in question is what is known as a "reversible" inhibitor of FAAH (fatty acid amide hydrolase), an anandamide-degrading enzyme (hydrolase), one of the main mediators of what is known as the endocannabinoid system, ANSM said.
The members of the TSSC, after having studied data from literature and the documents provided to them (BIA 10-2474 preclinical data, data on the trial conducted by Biotrial), met in a one-day plenary session on 15 February.
According to a report of the committee’s minutes, Bial planned to develop the compound as an analgesic, though clinical development of several similar compounds “was abandoned after Phase 2 clinical trials due to insufficient effectiveness (analgesic especially) without any specific toxicity being noted in humans or animals.”
In terms of toxicology, TSSC says the BIA 10-2474 animal studies “were carried out properly in accordance with current standards (those of the ICH especially),” and “no toxicity, especially neurological (central or peripheral) comparable to that observed in the accident in Rennes, [France] appears to have been demonstrated in animals, despite the use of 4 different species and high doses administered over long periods.”
Also, the committee noted that “the fact that the toxicology studies, although conducted on four animal species with doses up to 650 times the dose absorbed by the hospitalised volunteers, do not apparently show any lesions or picture likely to predict such toxicity.”
However, the committee calls on Bial to clarify:
- The reasons for using four different species for the toxicology studies
- The circumstances of death by bronchopulmonary disease in dogs,
- The circumstances of death during studies on primates at high doses
- The results of any microscopic examinations of the brains of deceased primates
- The reasons for down-titration in the 13-week study in dogs
- The reasons for the apparent lack of preclinical pharmacology studies for confirming, before transfer to humans, the analgesic effect of BIA 10-2474, especially compared to benchmark analgesics.
As far as the way in which Biotrial handled the running of the trial, the committee highlighted that “the trial was not immediately suspended whereas one of the volunteers had been hospitalised for a sudden-onset event.”
In addition, volunteer selection “did not apparently take neuropsychological assessment (clinical interview with cognitive assessments and tests) into account, whereas the ‘somatic’ explorations appeared to be exhaustive.”
But as far as the symptoms observed in the hospitalized patients, ANSM says, “The entire picture, both clinical and radiological, was therefore completely unusual [bold is ANSM’s], with no relatedness to a known disease or toxicity.”
The seriousness of the result of this trial warrants new best practices with regard to first-in-man trials moving forward, the committee said. Some of those practices include:
- Demonstration of pharmacological activity, comparative whenever possible, should be a requirement in the future before in‐human administration or even before continuing toxicology studies can be envisaged. Preclinical pharmacology studies should be conducted as early as possible, on an adequate dose range (dose‐effect curves) and should be designed so as to be reasonably predictive of real‐life, future therapeutic efficacy.
- A neuropsychological assessment with clinical interview and cognitive tests should be a compulsory part of assessment during volunteer screening and inclusion in a Phase 1 trial for drugs with "central nervous system" tropism.
- Detailed and well‐supported arguments for the choice of maximum dose to be tested in volunteers with respect to the presumed effective dose should be provided. For example, in this case, it appears unjustified to plan to test a dose (100 mg) 80 times higher than that presumed to induce complete and prolonged FAAH inhibition (claimed mechanism of action of the drug tested).
- A large‐scale consensus process should cover Phase I dose‐escalation strategies to establish recommendations for more reasonable and careful practices than those applied.
- Pharmacokinetic parameter variability and extremes, and not only the mean, should be taken into account for setting the next dose level.
The TSCC said that at its 24 March meeting it will put forth more comprehensive recommendations that it would like to see applied at the international level.