Pharmacodynamics and pharmacokinetics

In terms of animal pharmacodynamics, the trial protocol reports that the biological activity of BIA 10-2474 was tested in models of predictive efficacy in treating pain. "BIA 10-2474 produced analgesic/anti-inflammatory activity in the mouse Formalin-Paw and Tail-Flick tests in a time-and dose-dependent manner. BIA 10-2474 also markedly potentiated the antinociceptive effects of exogenous anandamide in the mouse Formalin-Paw and Tail-Flick tests". In other words, BIA 10-2474 worked as a pain-killer in mice using two different tests (the "analgesic/anti-inflammatory" effect); and, in mice that had been given a dose of the neurotransmitter anandamide, BIA 10-2474 also improved its effects in numbing pain ("antinociceptive" effect).^[8] The expert committee of the ANSM took the view that this was insufficient basis for commencing human trials, and that further evidence of BIA 10-2474 as an analgesic was warranted.^[15] The committee's final report noted that in fact the original formalin paw study report presented additional data for the gabapentin comparator which was omitted in the Investigators Brochure.^[6] The final ANSM report stated that "doses used in these tests differ greatly (from 0.3 to 10 mg/Kg), without it being possible to trace a dose‐effect curve or to estimate an effective dose 50 (which is a surprising shortcoming)".^[6]

In terms of animal pharmacokinetics, in rats and dogs given a radio-labelled dose of BIA 10–2474, the drug was detectable in the blood for up to a day afterwards (oral or i.v.). Oral bioavailability was not reported. Terminal half-life (persistence in the blood) of the BIA 10–2474 in rats was 45 hours (oral) or 4 hours (i.v.) and in dogs it was 104 hours (oral) or 52 hours (i.v.). The authors made no prediction of a likely half-life in humans. Around two-thirds of the total dose was eliminated in the urine, about one-fifth in the faeces and the remainder was heavily metabolized in all species studied (rat, mouse, dog, monkey). Metabolism of BIA 10-2474 was essentially complete by 72 hours. Main metabolites were not described. The studies used total detectable radioactivity to calculate half-life and did not assess what proportion of this was due to metabolites.^[8]

The preliminary report by the French Inspector General for Social Affairs (IGAS), for the French Ministry of Health, remarked on the difference between the rat oral vs. I.V. half life, and, that in light of the adverse events seen in humans on repeated dosing, this may indicate "a mechanism of accumulation".^[21] Separately, accumulation of BIA 10-2474 given orally in humans was supported by pharmacokinetic data from the trial itself, reviewed by the ANSM expert committee. The molecule showed non-linear pharmacokinetics at doses between 40 and 100 mg, suggesting elimination pathways had become saturated, leading to accumulation.^[15]

The final report by the ANSM Committee noted the drug had a very steep dose-effect curve in humans "going from absence of to almost complete inhibition" over a narrow concentration range.^[6] The report also found, based on clinical data from the Rennes trial, that complete inhibition by BIA 10-2474 was very long-acting, and persisted well beyond the point where it was no longer detectable in blood of the trial subjects.

Safety pharmacology and toxicology

The protocol presents a summary of safety pharmacology studies in two species (rat, dog) and repeated dose toxicity studies in four species (13-week sub-chronic studies in mouse, dog and monkey; a 26-week chronic study in the rat). Of note, few adverse events were observed in any of the studies, with the oral No Observed Adverse Effect Level (NOAEL) varying between 10 mg/kg/day in rats to 75 mg/kg/day in monkeys. The authors suggest that these were the maximum doses tested in these studies, though it is not clear. The authors also report no effects of significance in the animal models used for the CNS safety pharmacology studies, which studied a dose of up to 300 mg/kg/day. The protocol nominates a human NOAEL of 100 mg as a human equivalent dose to the 26-week rat NOAEL, though with no description how this was calculated. The summary presented however includes no assessment of the relevance of the animal species selected for study (that is, in terms of physiological and genetic similarities with humans and the mechanism of action of the study drug).^[8] Early reports that the molecule had been studied in chimpanzees turned out to be incorrect.^[4][22][23]

At a hearing convened by the ANSM in March 2016, Bial clarified that the apparently extensive animal toxicology studies (and number of species) was due to a delay in commencement of clinical development, thus some studies had already been completed that would not have been required for a Phase I study. The ANSM Committee found no evidence that the studies had been performed because the company had doubts as to the tolerance of the molecule.^[6]

Notably absent from the protocol were calculations of receptor occupancy; predictions of in vivo ligand binding saturation levels; measures of target affinity or assessment of non-target binding interactions as suggested by the European guidance for Phase I studies (depending on whether BIA 10-2474 could be considered to require 'special consideration' as outlined in the guideline).^[8][24] On these issues, the French regulator's expert committee pointed out, based on the company's IC50 data, that complete FAAH inhibition should have been achieved with a dose of 1.25 mg in humans. In fact, the trial tested doses up to 80 times more (100 mg BIA 10–2474) than should have been required.^[15] The ANSM committee also requested data on non-target enzyme affinity from the manufacturer Bial.

Although the protocol summary reports no animal deaths during the studies, the ANSM expert committee reported that in fact several monkeys died or had to be euthanised during the dose escalation studies and that an explanation from Bial was pending. Also, two animals had to be euthanised in the 13-week dog study due to lung lesions - both from the top dose group.^[15][25] None of these animal deaths were described in the trial protocol.^[8]

It is not clear whether Bial disclosed these adverse animal findings either to Biotrial or to the ANSM in their application for clinical trial authorisation.^[26]

The findings presented in the trial protocol provide no explanation for the type and severity of events that would be later observed in Rennes.^[15] The preliminary report from the IGAS investigation, released by the French Minister for Health in February 2016, found there was no legal requirement in France for the trial sponsor to disclose all pre-clinical data to the ANSM. The Minister described this as an opportunity for improvement.^[27] Nevertheless, the IGAS report commented that the investigation thus far had not found any reason to fault the ANSM's authorisation of the trial based on the preclinical data.^[21]

The final report by the ANSM Committee concluded regarding the preclinical studies that "no aspects of the data that the TSSC has studied constituted a signal likely to contraindicate administration in humans." The report went on however to criticise Bial's Investigator's Brochure: "the brochure contains many mistakes, inaccuracies, figure inversions or incorrect translation of source documents, making understanding difficult in several aspects. This is highly surprising given the regulatory importance of this document."^[6]

Overview

In 2015 Biotrial, a contract research organization, initiated a first-in-human trial of BIA 10–2474 in healthy volunteers, with secondary endpoints to investigate neuropathic pain.^[6] The study was approved by the French regulatory authority, the ANSM, on 26 June 2015, and by the Brest regional ethics committee on 3 July 2015.^[3] The trial protocol was leaked by Le Figaro^[20] before a different version was released by ANSM.^[8][18] The manufacturer Bial refused the regulator's request to release the Investigator's brochure and the product dossier (Investigational Medicinal Product Dossier), citing French law on trade secrets.^[18]

Trial details

The study was entitled "A double-blind, randomised, placebo-controlled, combined single and multiple ascending dose study including food interaction, to investigate the safety, tolerability, pharmacokinetic and pharmacodynamic profile of BIA 10-2474, in healthy volunteers". The trial commenced on 9 July 2015 at a single centre in the city of Rennes, and commenced recruitment of 128 healthy volunteers, both men and women aged 18 to 55. Participants of the study were to receive €1,900 and, in turn, asked to stay at Biotrial's facility for two weeks during which time they would take the drug for ten days and undergo tests.^[2][28] The study drug was presented as capsules in three different strengths (0.25, 2.5 and 10 mg). The protocol describes four distinct parts to the study:

• a single dose ascending part
• a cross-over part to evaluate fed vs. fasting conditions (could be either single or multiple dose)
• a multiple dose ascending part
• a pharmacodynamics part to evaluate the effect of BIA 10-2474 vs placebo with different challenge agents^[8]

The protocol specifies the first three parts would be double-blind but the pharmacodynamics part would be open label. According to the protocol, dose levels and numbers of groups could be increased, or were not yet defined, and would depend on what was observed with initial dosing, (an approach known as adaptive trial design). Thus many details of the key multi-dose part are not included in the trial protocol.^[8] The absence of these details was criticised by scientists and in the media,^[29][30] before the ANSM published further details of the dosing.^[31] The Royal Statistical Society was particularly critical, stating that it had "clear statistical reservations about the trial's study design", and that the protocol lacked features such as a risk assessment, recommended in the wake of the TGN1412 incident to prevent severe adverse events occurring in these circumstances.^[30]

Starting dose and subsequent doses

For the single dose part of the study, the protocol describes eight groups of eight volunteers (3:1 randomised) who were to receive single doses of BIA 10–2474 at 0.25, 1.25, 2.5, 5.0, 10, 20, 40 and 100 mg, with the possibility of additional groups to be added if no maximum tolerated dose was reached.^[8] In describing the rationale for the starting dose, the authors of the protocol conclude that:

No target organ was identified during toxicology studies and few adverse clinical findings were observed at the highest dose tested. For the single ascending dose part [of the clinical trial], a starting dose of 0.25 mg was judged to be safe for a first-in-human administration.[8]

The protocol provides for the first two subjects of the first single dose cohort to receive the first dose as a sentinel dose, that is, either 0.25 mg BIA 10-2474 or a placebo on the first day, then wait for 24h before treating the other 5:1 subjects. In the event no safety concerns had emerged in the foregoing, all other single- and multi-dose groups were to be dosed with a 10-minute interval between recipients.^[8]

For the multiple ascending dose part of the study, the protocol planned four groups of eight volunteers (3:1 randomised) who were to receive a single dose orally, once daily for 10 days at different dose levels. However, the protocol defines no doses for these groups, stating that this will be based on the outcome of the single dose portion of the trial. Further dose groups up to a maximum of eight groups could be added depending on whether adverse events were observed. The authors note that nonetheless, the starting dose will not exceed 33% of the maximum tolerated dose (MTD) identified in the single dose groups (or 33% of the maximum administered dose if the MTD is not reached).^[8]

Further details of the study as it was actually conducted have been published by the French agency (ANSM). The dose groups for the single dose part were as described in the protocol, with no additional groups undertaken. In the crossover part, a single 40 mg dose was given to a group of 12 subjects. In the multiple dose ascending part, the doses were 2.5, 5.0, 10, 20 and 50 mg BIA 10–2474, each to be given once per day for 10 days to groups of 8 volunteers (3:1 randomised). The severe adverse events were observed in the 50 mg dose group.^[4][31]

The ANSM expert committee reported^[15] that complete FAAH inhibition should have been achieved by a dose of 1.25 mg remarking that:

It appears unjustified to plan to test a dose (100 mg) 80 times higher than that presumed to induce complete and prolonged FAAH inhibition.[15]

Further, the ANSM committee remarked that the gap between the 20 mg and 50 mg dose cohorts, in effect, skipped a dose based on extrapolation from the single dose part of the study, and that the progression to 50 mg was too large a jump. On reviewing the subject data from the trial itself, the committee noted that BIA 10-2474 showed non-linear pharmacokinetics at doses between 40 and 100 mg (that is, the molecule appeared to be accumulating at higher doses) and that most likely the elimination mechanism had become saturated. Thus dosing at 50 mg daily was - each day - 40 times more than required to achieve complete inhibition, and in practice this dose level resulted in accumulation.^[15]

Trial status at the time of serious adverse events

According to Bial and the Rennes University Hospital, by the time the serious adverse reactions emerged, 116 subjects had been recruited and 84 other volunteers had received the drug during the trial, with no serious adverse events being reported.^[1][4][5] The single dose part (up to 100 mg BIA 10–2474), the fed vs. fasting part, and the first four dose groups of the multi-dose part of the study had each been completed in 2015. Notably, the administration of 20 mg BIA 10-2474 once daily for 10 consecutive days elicited no serious adverse events in the six volunteers who received it.^[18][31]

Dosing of eight volunteers in the highest multi-dose group in the BIA 10-2474 trial commenced on 6 January 2016. Six of the participants received 50 mg per day of the drug while two received placebo.^[1][4] The first subject became ill on the evening of the 5th day of dosing (10 January). On the following day, the other subjects received a 6th dose at 8:00am before the trial was suspended later that day (11 January).^[31][32]

French authorities

The Agence Nationale de Sécurité du Médicament (ANSM) announced an investigation, and that an inspection of the trial site was already underway.^[3][50] and formation of a specialist committee of pharmacologists, toxicologists and neurologists to review all existing data on FAAH inhibitor drugs.^[18] The French health minister Marisol Touraine, who visited the trial site in Rennes and spoke with the victims' families, called the events "an accident of exceptional gravity" and promised to investigate the matter via the Inspector General for Social Affairs (IGAS), with a final report due end of March.^[4][27][50] A judicial investigation was also initiated after the Office of the Chief Prosecutor for Paris announced it would look into potential charges of involuntary injury through its public health enforcement section, seeking assistance from the Gendarmerie in Rennes and the Ministry of Justice's Office for the Environment and Public Health.^[50]

The ONIAM (Office National d'Indemnisation des Accidents Médicaux), responsible for medical injury compensation, stated only 10 accidents had occurred during clinical trials over the past 15 years per its records, and that those cases had "consequences infinitely less serious" than the incident in Rennes.^[57]

Initial reports from the authorities suggested that the causes of the brain injuries was likely to be due to the mechanism of BIA 10-2474 and the doses employed in the trial.^[15][21] ANSM Director General Dominique Martin said "It is clearly the molecule that is the cause", following the publication of the ANSM expert committee's initial findings on 7 March 2016.^[58]

In May 2016, the French Health Minister announced several new measures for clinical trials in France, including the establishment of an expert group within the ANSM for review of first in human and early phase studies.^[59]

Agencies outside France

A European Medicines Agency (EMA) spokesperson said in January 2016 that "EU authorities will look carefully at the findings to determine if further measures are needed to protect health of clinical trial participants. Until EU authorities have the full picture, it is not possible to say whether any revisions to EU guidelines are required".^[62] Later, in July 2016, the agency proposed to revise its key first-in-human clinical trial guideline, the last significant revision having been published in response to the 2006 TGN1412 clinical trial that had had similarly dire effects in its trial volunteers.^[63] The EMA stated the proposal aimed to address risks posed by complex trials, such as the one undertaken in Rennes, having 'several steps of clinical development within a single clinical trial protocol'.^[64]

The European Investment Bank, which provided 110 million euros in funding for Bial's FAAH inhibitor programme stated it had been in contact with the company about the incident, but that "it would be premature to consider recall of the EIB loan at this stage".^[65]

The US Food and Drug Administration issued a statement that it was in contact with its counterparts the ANSM as well as the EMA and announced investigations into FAAH inhibitors as a drug class."FDA is in the process of collecting and reviewing safety information pertinent to FAAH inhibitors under investigation in the US. FDA will work with sponsors to ensure the safety of participants in clinical studies and take regulatory action as appropriate." Later, in August 2016, the Agency issued a further statement: "The Agency has found, based on the available information, that BIA 10-2474 exhibits a unique toxicity that does not extend to other drugs in the class, called fatty acid amide hydrolase (FAAH) inhibitors."^[66]

The German drug regulator, the Bundesinstitut für Arzneimittel und Medizinprodukte (BfArM) issued a statement 19 January that no clinical trials with FAAH inhibitors were underway in Germany, but that it had authorised seven such trials previously, which were completed without serious incidents.^[67]