Login to Your Account

In wake of a fatal French clinical trial, EMA publishes revised phase I trial guideline


By Nuala Moran
Staff Writer

LONDON – A new guideline for the conduct of phase I trials in Europe will come into effect in January 2018, two years after one man died and four others suffered neurological damage in the phase I study of BIA 10-2474.

The EMA published the revised guideline last week, following a public consultation launched in May 2016. The guideline sets out new approaches to mitigating and managing the risks to trial participants, beginning with the calculation of the starting dose, to subsequent dose escalations and the criteria for the maximum dose.

In addition, the guidance stresses the need to conduct rolling reviews to assess any safety implications of emerging data, and outlines what action should be taken before progressing to the next dose level.

The EMA launched the consultation in response to the fiasco that occurred at a Biotrial Research SAS clinical research facility in Rennes, France, last January. The study was testing BIA-10-2474, an oral fatty acid amide hydrolase (FAAH) inhibitor, which the owner, Bial Portela & CA, of Porto, Portugal, was developing as a treatment for neuropathic pain. (See BioWorld Today, Jan. 20, 2016.)

While the EMA sets down the guideline for moving to first-in-human studies, it is not involved in the oversight of clinical trials, which is the responsibility of national regulatory agencies. In the case of BIA-10-2474 that was the French regulator L'Agence nationale de sécurité du médicament et des produits de santé (ANSM).

Two French investigations – one carried out by an independent scientific committee set up especially by ANSM and the second by the social affairs directorate IGAS (Inspection générale des affaires sociales), which oversees health – made a series of recommendations on the authorization and conduct of phase I studies.

However, both investigations concluded the trial protocol was approved in line with the current guideline, and given that, the European Commission and the Heads of Medicines Agencies, a body representing all of Europe's national regulators, looked to the EMA to improve practices at a European level.

Alongside aiming to reflect the lessons learned from the Bial study, the updated guidance is intended to encompass the evolution in trial design that has occurred in the decade since the current guideline was published in 2007 – itself as a response to the disastrous Tegenero trial of TGN-1412 in London, in which six healthy volunteers suffered life-threatening cytokine storms. (See BioWorld Today, May 26, 2006.)

In the 10 years since, first-in-human trials have become increasingly complex, including both single and multiple ascending-dose phases, food interaction studies and assessment of the investigational compounds effects in subjects of different ages, for example.

One difficulty facing the EMA in revising the phase I guideline was that while 10 other FAAH inhibitors have been tested in a number of phase I and phase II studies with no reports of serious adverse events, it is still not known what caused the neurological damage in the BIA-10-2774 study.

The trial of BIA-10-2474 began in July 2015, and 84 volunteers had received the drug in the single ascending-dose, interaction with food, and multiple ascending-dose arms of the study before the serious adverse events happened during administration of the highest dose of 50 mg in the multiple ascending-dose arm.

The eight volunteers in the 50-mg cohort started to receive the drug on Wednesday Jan. 6. It was on the fifth day of receiving the single dose, on Sunday, Jan. 10, that the volunteer who died became unwell and was taken to Rennes University hospital.

The doctor administering to the sick man informed a Biotrial doctor at 8:40 p.m. on Sunday that the volunteer had been hospitalized. Even so, the other volunteers were given their sixth 50-mg dose of BIA-10-2474 on the morning of Monday, Jan. 11. (See BioWorld Today, Feb. 9, 2016.)

The safety and tolerability profile of BIA-10-2474, which acts via the endocannabinoid system, was favorable up to 20 mg, with no alerts or signals in any of the safety parameters that would have led trial investigators to anticipate such serious harm as occurred at the 50-mg dose.

Bial said there was no reason to modify the dose-escalation plan approved by the authorities in the trial protocol.

However, both of the French investigations questioned whether it was appropriate to stick to the approved dose escalation, noting it was not reviewed in the light of pharmacokinetic data from volunteers in previous cohorts, which showed FAAH inhibition was complete at the 5-mg dose.

That is reflected in EMA's revised guideline, which says pharmacokinetic, pharmacodynamic and safety data for each cohort should be reviewed and planned doses adapted accordingly. The review also should consider whether other aspects of the protocol need to change to ensure safety.

Precautions to be taken when dosing individual subjects within a cohort also are more prescriptive in the revised guideline. Rather than just allowing an adequate period of time after administration to the first subject in a cohort to observe any reactions and adverse events, there should be a defined review for the "sentinel" subject before the dosing of any more subjects.

Inadequate animal models

The harm caused to healthy volunteers in the BIA-10-2474 trial underlines the limitations of animal models. In preclinical testing, at doses of up to 100 mg in four different species, BIA-10-2474 did not cause brain lesions of the type observed in the injured volunteers.

In addition to not highlighting potential toxicity, the inhibition of FAAH in animal studies did not read across to the dose required to inhibit the enzyme in humans.

Moreover, the dose-response curve goes from non-existent to complete over a very narrow range. That led the French investigation to criticize the (albeit approved) dose escalation in the BIA-10-2474 trial for the sharp jump between cohort four, which received 20 mg, and the harmed cohort five, dosed at 50 mg.

The French scientific committee concluded the most likely explanation of the toxicity was a gradual accumulation of the drug in the brain, related to specific features of BIA-10-2474, and which caused – unspecified – off-target effects. (See BioWorld Today, June 1, 2016.)

The revised guideline reflects that, saying there should be better integration of preclinical pharmacology and toxicology data into an overall risk assessment for first-in-human and early stage trials.

That will require sponsors to include the extrapolation and verification of the assumptions made in translating preclinical data to humans.

In addition, the guideline notes qualitative and quantitative differences may exist in biological responses to a product in animals compared to humans. Given that, sponsors should use in vitro human cell systems or human-derived material to understand those differences and improve the understanding of the relevance of the animal models, the EMA said.

The inputs the EMA received in response to the public consultation will be published next month.