Skip to main content

Clinical Trials

A Clinical Disaster in France

Details are lacking, but there has apparently been a very bad accident at a clinical trial center in France. Reports are that a clinic in Rennes was testing (under contract) what the media are referring to as a <s>”cannabis-based pain killer”</s> drug (update: the French government is saying that’s not the case), but something has clearly gone terribly wrong. Several people are hospitalized in critical condition, with one said to be in a coma. This appears to have developed earlier this week.

And that, so far, is all anyone seems to know. All this raises a lot more questions than it answers, of course – one assumes that this was a first-in-man Phase I, but doses in such trials start small and go up, for the most part. We don’t know how many of the participants were affected (just the higher-dose groups?), or how long this took to happen. And obviously, we don’t know what sort of compound was being tested, but it does seem very strange for it not to have shown any potential for such trouble in earlier safety pharmacology and toxicity testing in animals. But then, we don’t really know what symptoms the affected patients have, either.

The last thing I can remember like this is the TeGenero experience back in 2006. That one at least was targeting the immune system, so there’s always room for out-of-control effects there (although, to widespread amazement, it’s coming back). But a pain killer? We need more information about what’s happened this time, and I hope it becomes available soon.

Update 2: latest reports are that this was a compound from Portugal’s Bial Pharmaceuticals. Their website lists a compound called BIA 10-2474 in Phase I for “neurological and psychological properties”, so that may be what we’re looking at. I’m unable to find out anything more on what this compound is or what it targets – one story has it that it’s a FAAH inhibitor, but that’s unverified.

Update 3: Le Figaro is now reporting a 128-person trial (90 dosed with compound, the others placebo). It is indeed the Bial compound mentioned above, and the drug had been in chimpanzees beforehand. The volunteers who had taken multiple doses are the ones affected, the article says (and I’m willing to bet that this means the higher-dose group, since there were probably several who had more than one dose?) The first symptoms appeared on Sunday, January 10. The most severely affected volunteer is said to be brain-dead (!), with the others severely affected, and a hospital spokesman says that at least three of the remaining five patients might have irreversible damage. 

Update 4: As per the comments, here’s an article from France that says that the affected patients show evidence (by MRI) of deep cerebral hemorrhage and necrosis. Needless to say, that is very, very bad indeed. No one’s sure yet if this is a compound-related toxicity or a mechanism-related one, but it’s worth noting that other FAAH inhibitors have been looked at with nothing like this happening.

Update 5: a good overview from Judy Stone at Forbes. She makes a good point: why was the placebo group so large in a Phase I?

Update 6: the worst-affected patient has now died.

Update 7: a roundup from David Kroll on what we know so far.

Update 8: some better-than-expected news on the hospitalized patients themselves.

Update 9: the clinical trial protocol, including the compound’s structure.

Update 10: the Royal Statistical Society is calling for still more disclosure, and for independent statisticians to be a part of the investigation (both of which sound like good ideas).

140 comments on “A Clinical Disaster in France”

  1. AndrewD says:

    According to the Guardian report, there is to be a News Conference this afternoon

  2. wim says:

    if it is a cannabinoid, and the active ingredient caused the troubles, that is some extraordinarily bad luck or malpractice, considering the “clinical trials” of millions of people using hundreds of different completely untested synthetic cannabinoids as “legal highs” without major acute health issues under a normal dosing regimen.

      1. wim says:

        I didn’t say “safe”. Of course taking untested drugs isn’t safe. Acute major adverse events, when typical, infrequent doses of cannabinoids are taken are nevertheless relatively rare. Millions of people have taken these drugs with no short-term issues. What you think of drug dealers and the hysterical articles you link are irrelevant to my original statement.

        1. Wile says:

          From the Newsweek item I linked. “…synthetic cannabinoid-related emergency department visits have risen in recent years, from 11,407 in 2010 to 28,531 in 2011 (the most recent year for which there are data), according to a 2014 report by the federal Substance Abuse and Mental Health Services Administration (SAMHSA).”

          I would call 28000 emergency room visits related to synthetic cannabinoid as being something that indicates that these are a very big deal. I believe your phrasing was: “…considering the “clinical trials” of millions of people using hundreds of different completely untested synthetic cannabinoids as “legal highs” without major acute health issues under a normal dosing regimen.”

          Therefore, you down played the risk of these in your post. I would call an emergency room visit a major acute health issue (can anyone say adverse event?). I think it is entirely relevant to your post. Again, there is no labeling as to what is a “normal dosing regimen”. I guess we’ll agree to disagree.

          1. Aaron Csicseri says:

            The reason that these synthetic coumbounds (like the JWHs) are associated with so many ED visits is because people who take these compunds are marijuana users who are either trying to save money, curt law, or beat a drug test. They expect the synth to produce a similar experience, when in fact most of these compounds are much more holucinagenic (think in terms of LSD, not THC). Subsequently, the patient has a anxiety attack and goes to the ER, when they were really in no physical danger at all (unless they have inteded to overdose). So, I agree more with Wim on this issue. I would like to note that synthetic canabinoids are not the same a “bath salts”..which are very dangerous.

    1. Bagnar says:

      It is not about cannabinoid, according to the ANSM (FDA european equivalent)

      BioTrial just publish something on Twitter:^tfw

      Tested compound is probably the BIA 10-2474. Confirmation will come soon.

      1. wim says:

        On press release it was confirmed that it was something “affecting the endocannabinoid system”. So with the information on Bial pipeline, confirmed FAAH inhibitor.

  3. petros says:

    Reports saying it is a Bial drug, one tweet has claimed it is BIA-10-2474

    1. salvarsan says:

      Not a cannabinoid, certainly not structurally, unless you consider Hoffman’s naphthoyl indoles to be cannabinoids solely by virtue of affinity to CB1,2 + NMDA receptors.

      Bial’s BIA 10-2474 was modeled on advil/ibuprofen intermediate metabolites and was intended for pain management.

      Search on “Chiral 1,3,4-Oxadiazol-2-ones as Highly Selective FAAH Inhibitors” in J.Med.Chem.

      Bial patent application at:

  4. Fabrice Pierre says:

    it is indeed a cannabinoid. According to local news (Ouest-France), the trial was run by the CRO Biotrial and the drug developer was a Potuguese company named Bial.

    1. Bagnar says:

      Sorry Fabrice, but you’re wrong. Or the ANSM is wrong, but this time, I will give them my confidence.

      Sadly, it will not rescue the guy almost dead….

      1. Fabrice Pierre says:

        you are right Bagnar. not a cannabinoid, according to the Health secretary.

    2. smotpoker says:

      Do Potugese companies only run cannabinoid trials?

  5. petros says:

    Rather than a cannabinoid it might be a FAAH2 inhibitor

    see the process patent WO2015112036 and references therein

  6. RS says:

    there are several clinical trials, sponsored by Bial, listed in EU Clinical Trials Register, including those for pain, at this link:

  7. Graham Brack says:

    10-2474 was previously reported to be a STAT3 signalling inhibitor in testing as a modulator for chronic pain.

    1. CWC says:

      If you are looking at niclosamide 10-2474, it appears to have been made by Focus Biomedicals (

      From what I can tell from SciFinder searches of BIAL-published studies, they have been making a lot of FAAH inhibitors lately.

      1. Stan says:

        Not niclosamide, that´s a tape worm drug, A WHO essential drug

  8. ps says:

    If it is a FAAH2 inhibitor, I will be curious as to how they did the preclinical trials. Unlike humans Mice don’t have Faah2 gene. The Faah gene that is shared by humand and mice is quite different from Faah2 in sequence.

  9. Matthias Grossmann, MD, PhD says:

    What, if they had used rats and dogs as usual? Same situation?

    1. Diver dude says:

      Hi Matt it’s Joe from the uk. I hope this wasnt one of yours?

  10. ps says:

    @Matthias Grossmann, MD, PhD
    Don’t have time to do a proper tree, but based on a cursory check Faah2 seems to be a primate thing. Dogs, cats, rats and mice don have it.

    According to the latest Reuters release it does seem possible that the drug is a Faad (Faad2?) inhibitor:

    “The minister said the drug was meant to act on the body’s endocannabinoid system, which deals with pain.”

  11. petros says:

    BBC saying 90 took part in the trial.
    Were the unfortunate 6 the highest dose group?

    1. wim says:

      according to press conference, yes, all the same group, trying highest dose so far

  12. david baker says:

    If it is a FAAH inhibitor maybe there is a liver failure as faah is highly expressed in the liver and presumably involved in lipid metabolism. This may not show up in toxicology tests as it will take time for the fatty liver to mature, but may be present in the humans and accentuated by the inhibition. Acute liver failure can lead to brain oedema.

    1. david baker says:

      Pictures of FAAH knockout livers stained with oil red O if interested

      I am not sure this would be a good target.

  13. Doug says:

    One has to be very careful in calling this a FAAH inhibitor. Selective FAAH inhibitors have been tested in the clinic and have been shown to be very safe.
    The Bial Pharma chemotype looks like it is quite reactive and undoubtedly inhibits more than FAAH especially at higher doses.

  14. ScientistSailor says:

    Here have been several FAAH inhibitors tested in reasonable-size PII clinical trials, with little effects of any sort. If this is an FAAH inhibitor, these results are not from on-target activity!

  15. CarinaB says:

    Anyone who can help me find the trial protocol? Couldn’t find an EudraCT number or anything…

    1. JonasAMW says:

      I doubt that you could find the actual *Protocol* online…

      1. spike says:

        Are Phase I studies listed in the EU Clinical Trials Database ( I thought that only information on Phase II-IV studies were publicly available.

  16. Holly says:

    Will you please report if it becomes known if any of these 5 were using marijuana during the FAAH inhibitor trial?

    1. Anon says:

      They weren’t unless somebody messed up badly. One of the exclusion criterium for these studies is the use of any drug and before dosing the subject is screened for those.

  17. diver dude says:

    Derek – to your comment about the TeGenero compound. Many of us are not surprised at all that it is back in development. When the data came in it became clear that the had compound worked as advertised. The problems were caused by the starting dose in the First in Man study being about 400x too high because the differences between the biology of the mouse model and man were not properly understood. I imagine the regulators will be all over it with a fine-tooth comb but that is as it should be.

  18. According to the French press, citing Pierre-Gilles Edan, the chief neurologist at CHU de Rennes, MR images show deep, hemorrhagic and necrotic brain lesions: Per patent information (, BIA 10-2474 is a potent FAAH inhibitor, with relatively high peripheral selectivity. How repeated and higher doses of BIA 10-2474 might cause catastrophic hemorrhagic strokes remains to be explained (obviously).

    Another curious tidbit, which needs to be confirmed: the 6 hospitalized enrollees knew each other.

    1. ps says:

      ” How repeated and higher doses of BIA 10-2474 might cause catastrophic hemorrhagic strokes remains to be explained (obviously).”

      How about this way:
      Br J Pharmacol. 2007 Mar; 150(5): 641–651. doi: 10.1038/sj.bjp.0707141
      PMCID: PMC1942073
      Endothelium-dependent metabolism by endocannabinoid hydrolases and cyclooxygenases limits vasorelaxation to anandamide and 2-arachidonoylglycerol
      W-Sv Ho1,* and M D Randall1

      1. H says:

        So what, you’re saying that a vasodilator would cause catastrophic hemorrhagic stroke? If anything I would imagine the opposite.

    2. M says:

      Yes it is somewhat curious how they would all six know each other. While it wouldn’t be unthinkable a group of friends might sign up together one would think the many layers of selection and randomization would make it unlikely they would end up in the same dosage group of the same trial!

      It could be a poor translation. Maybe they just knew each other from the trial (sitting in waiting rooms while undergoing tests etc.) – or from other trials at Biotral, since many participate in several trials as has come to light. Also, the participants sometimes stay in dorms up to the trial due to the controlled food requirements etc. so that would also make it possible to get to know each other.

  19. anonymouse says:

    For what it’s worth, what is the structure, chemically or biologically, of BIA 10-2474 ? Has it been published?

    1. JP says:

      I assume it is the same structure as in the process patent linked to in comments above. Barbara’s comment directly above yours has the link to the patent pdf

  20. metals wrangler says:

    DiverDude. that could be the classic case of “the dose makes the poison”. Hopefully FiH trials for any drug will start at NOAEL/100…

  21. Jonathan says:

    A cursory search didn’t turn up any preclinical PK. Is it possible that the repeated doses caused significant accumulation?

  22. Past never past says:

    Reports seem to be coming in that BIA-10-2474 is an FAAH inhibitor of some type. And also that the AE is associated with severe inflammation, which they are trying to control with corticosteroids.

    A quick look at the Bial patent portfolio as well as recent publications suggests that BIA-10-2474 is probably an irreversible covalent inhibitor of FAAH. A big concern with this approach has been the possibility of generating an immune response to covalently modified epitopes that arise from either desired or off-target covalent modifications. Given that any gross toxicity should have been detected in the preclinical workup, there is a good chance this is what happened – an immunologic response that took a few days to get rolling, then had disastrous consequences (and couldn’t be aborted by discontinuing the drug, due to the long-lived nature of the covalent modifications).

    As I recall, a few years ago Pfizer advanced an irreversible covalent FAAH1 inhibitor into a Ph II for osteoarthritis, and they saw no effect, positive or negative. But it was a different compound and may not have been brain-penetrant.

    1. Joan says:

      Wondering why this AE wouldn’t show up in animal trials, if immune or inflammatory reaction is behind it? Reminiscent (at least to me) of PML in MS drug trials. Mice with EAE didn’t carry the JC virus…and didn’t develop PML, humans did and still do. Maybe a similar scenario? Heart-breaking, for all involved.

      1. cellbio says:

        Reactive compounds that lead to creation of a neo-epitope that drives an immune response can generate problems unseen in preclinical species, and not always related to dose and not universal in the population and without predictable timing. Might be an small AE, might be a big problem. Several large companies, those that trained the chemists I worked with, simply will not develop compounds which are reactive or which generate concerning levels of reactive metabolites for fear of some bad thing happening. The bad things that influenced this conservative approach were in the distant past, unknown to us, but now I see how events like this would yield a firm rule, regardless of the fact that one can point to covalent modifiers that are on the market. This trial and Zafgen too?

    2. spike says:

      There’s a conference (SMi’s 15th annual Advances and Progress in Drug Design conference) on Feb 15/16 in London at which Doug Johnson of Pfizer will give a talk (Feb 15th)
      1.30 – KEYNOTE ADDRESS: Reviving covalent drug design: Discovery of PF-04457845, an irreversible FAAH inhibitor with exquisite selectivity
      (Doug Johnson, Research Fellow, Pfizer).

      That should be a popular talk. Not sure whether the organizers moved fast or whether timing is everything.

  23. anon says:

    What did their preclinical study in chimps look like? Was there any sign of this?

  24. Jeff says:

    Just use the flowering plant and throw those pills in the trash.

  25. LizW says:

    Sponsor’s Bial Foundation supports parapsychology “research”

  26. PAINS_GUY says:

    #Doug – you are right about the reactive chemotype. If it is an acylimidazole as suggested on Wiki, sure it is much deactivated compared with hot hot CDI by virtue of the N-methylurea, but still, it is an imidazole carboxamide. I would be concerned about progressing anything like that (if it such a compound) and would always be on high alert.

  27. ahr says:

    bial’s current clinical trials:
    note that BIA 10-2474 – Neurological and Psychiatric Pathologies

    research on FAAH inhibitors from march 2014

  28. Istvan says:

    If, and I repeat IF BIA 10-2474 is indeed an imidazole urea, or to comply with IUPAC, is a 1H-1-imidazolecarboxamide, as calimed by wiki (as of 16 Jan) then cytochrome P450 interactions might
    be expected. For example, the agricultural fungicide prochloraz is a similar imidazole urea compound:

    In case you missed it, there is an informative NYT article:
    another one by Forbes:

  29. Picky Pixie says:

    Should “worth nothing” read “worth noting”?

  30. newnickname says:

    Diver dude mentioned the study that started (on purpose) at a dose 400x too high. My first thought is that was a non-deliberate ERROR by the clinicians / pharmacists preparing and administering the doses. Examples come to mind. Dana-Farber incorrectly dosed patients with cisplatin in the 1990s (one death; one injury). In “Beer Barrel Polecats,” the Three Stooges administered 9 cubes of yeast instead of just 3 cubes and had a tragic outcome: prison. Or it could be a problem in the production / manufacture and that batch of drug was tainted or mislabeled (mg/mL instead of g/mL). It might not be the drug at all (if given at the proper dose); it might be a human error someplace in the treatment plan.

    1. So am I right to summarise as follow (any uncertainties marked with asterisk)?
      BIA 10-2474 is an irreversible* inhibitor of a degradation enzyme pivotal* to CNS function. A SAD study using 90 healthies (inc 38 placebos!) has been completed; The first cohort of 8 healthies (6 test) have all started simultaneously, daily dosing at the maximum* SAD oral dose; on day 4 AEs are seen and dosing stopped.

      1. spike says:

        I may beg to differ that the Forbes article is “a good overview”.

        For example, the author states that “Serious adverse events on Phase 1 trials are very rare—on the order of “once every 26.3 years of individual subject exposure.”” That’s an interesting statement that takes on a very different meaning when you follow the link to the article that they cite (Clin Pharmacol Ther. 1978 Aug;24(2):127-32. – a review of studies performed in normal prison volunteers). Surely they could have found something more timely and relevant.

        With respect to the size of the placebo group, I’m not sure that I would agree that it was “so large”. The most common ratio for dosed:placebo subjects is 3:1 (this was the ratio used in the cohort in which the subjects were dosed so is fair to expect that the same ratio was used in earlier cohorts). Since 90 subjects were stated as being dosed with drug, one would expect there to have been 30 placebo subjects. Where the number of 38 placebo subjects comes from, I have no idea (I can only find it mentioned in the Forbes article but if anybody can verify that number I’d love to see the source of information). The author links to the BBC news story that mentions 90 subjects had been dosed with drug (other news stories state 108 subjects had been dosed to date) but the BBC news story doesn’t mention placebo subjects. I think that the 108 subjects comes from a BIAL press release and included test article and placebo dosed subjects combined. With time, the true numbers will come out.
        One learning from this tragedy is that we are all very good at speculating in the absence of data (after all, who can refute us) and that we shouldn’t believe all that we read.

    2. spike says:

      new nickname,
      I think that you misread Diver dude’s post. They were describing the TGN1412 incident and noting that the post-event calculations indicated that the dose administered was far higher than a dose that would give a minimal pharmacological effect. They weren’t referring to the current incident.
      As for mentioning that this was deliberate – well, my ghast is well and truly flabbered! I highly doubt that you have any evidence to support such a statement.

      1. newnickname says:

        Trying to explain what I meant: From what I understand from DD’s post, the TeGenero group calculated a dose but it turned out to be toxic. It was not their intention to do harm. It was their *intention* to administer that calculated dose and that’s the “deliberate” that I meant. In the cisplatin case at Dana-Farber, it was human error in overdosing the patients. In the current case, it might have been a manufacturing or packaging error; it might have been a mistake in the hospital pharmacy; it might have been a labeling error (mg/mL v g/mL). I’m assuming that they are already checking the patients for the possibility of an overdose, the batch of drug for an unintended impurity (recall: trace impurity in Showa-Denko tryptophan that proved to be toxic) and other possibilities.

  31. Formerly UCB says:

    10-2474 = an acyl imidazole = chemically reactive. Surely pushing the envelope here. Idiosyncratic rather than target-based toxicity is a possibility. Note lilly are investigating a similar acylated tetrazole LY-2183240. Uh-oh!

    1. DJ says:

      Lilly is not developing LY2183240. This compound was first reported in 2005 and Ben Cravatt showed in 2006 that it is very promiscuous (Alexander JP, Cravatt BF. The putative endocannabinoid transport blocker LY2183240 is a potent inhibitor of FAAH and several other brain serine hydrolases. Journal of the American Chemical Society. 2006, 128, 9699).
      I wish everyone working on covalent inhibitors would follow the advice in the conclusion of this paper, “Our competitive ABPP studies of LY2183240 underscore the importance of defining proteome-wide selectivity patterns for bioactive small molecules to illuminate potentially unanticipated targets of these agents. Indeed, none of the additional enzymes targeted by LY2183240 display any discernible sequence homology with FAAH, suggesting that their common inhibitor sensitivity profiles derive from a higher-order relatedness in active-site structure and/or reactivity.”
      Also see “Strategies for discovering and derisking covalent, irreversible enzyme inhibitors”, Future Med. Chem. 2010, 2, 949.
      These chemical biology methods were instrumental in developing the exquisitely selective FAAH inhibitor PF-04457845.

  32. Istvan says:

    Be warned: There is not evidence whatsoever for the chemical structure of BIA 10-2474 being the one described in the patent* and also shown in wiki.
    A recent review on FAAH inhibitors describes other structures patented by Bial:

    “It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” by Sherlock Holmes

  33. David says:

    Even worse, there’s now been a fatality:

  34. John says:

    Dear Derek Lowe, as someone reporting from a science journal i find it startling that you would re-report the yet to be validated report that the drug is cannabis based. I am not a scientist myself but I do understand the scientific process of proving theories and all that good stuff. I note that you do go on to talk about the Bial compound which is an FAAH inhibitor. It may be prudent to remove your cannabis based statement from your article unless it’s absolutely confirmed. Firstly, just because the BBC reports it, does tint make it worth sharing elsewhere. There standards are slipping. Secondly, believe it or not, some people will read this and immediately assume that someone has died due to a cannabis based drug trial.

    Yours sincerely,
    Someone concerned with the truth.

    1. YesMen says:

      Expecting PR problems at the coffee shop, John ?

      things are if you want to massage the message, you’re in the wrong place. I wouldn’t trust anyone with your wording to be in a quest for accuracy or scientific truth.

  35. Mark Harrison says:

    Very sorry to hear about this awful tragedy. According to BBC News ( ) the man who was left brain-dead has now died.

  36. John, did you really read what Derek wrote? He made clear that the *early media reports* were calling it “cannabis-based” – which they did; that the French government says it is not; and that it appears to be an FAAH inhibitor, and therefore is expected to have effects on the endocannabinoid system. Notwithstanding a couple of clueless comments here (I liked the “flowering plant” one), most of the audience of this Web log is well aware that “endocannabinoid” doesn’t mean the drug is literally made from cannabis. I think that what Derek wrote responsibly counters, and does not perpetuate, the media’s inaccurate early description of the drug.

  37. John says:

    Hi Matthew, I did read it and my comment acknowledges that. I appreciate that the majority of people here are interested in this from a scientific perspective and will not take this statement as a fact, but some people may only be interested in the headline. I am not here for science reasons per se, I came across this page in my search for the truth as I felt that the ‘early media reports’ where heresay. Any media outlet could say what they wish about such things, so I personally think that we have a duty of care to be careful with which pieces of early media reports we decide to share with others. I do not doubt Derek’s intentions and hope I didn’t come across that way. I just prefer to get the facts 🙂

    1. istvan says:

      We all prefer the facts. And the fact is that media reports did refer to the test substance as ‘cannabis-based’ which, as we now know, was not correct. In a brief and fact-based summary such an error should be pointed out otherwise some people reading the blog would think that the writer is uninformed or is withholding relevant (mis)information.

  38. See the following in connection with the uneven expression of FAA2 across species:

    and a blog post advising some caution

  39. M says:

    Hi, really tragic case. Now the brain-dead man has died. I have updated

    I think authorities should release more info, especially if they have little idea what is going on. It might allow other experts to look at what happened from their perspective. I remember with TGN1412 people quickly found hints in articles and patent applications that suggested it could indeed cause a cytotoxic storm. Also they found a wrong animal model was used. I hope the doctors are at least spreading more detailed info among leading doctors from many fields in France. This might well require a multi-disciplinary effort (neurological, metabolic, immunological etc.).

    As for what happened – the fact that they are given corticosteroids might suggest an immunological mechanism. But maybe it could also be just to control the inflammation resulting from necrosis since corticosteroids have also been frequently used in stroke. It is an interesting idea it could be caused by neo-epitope from the drug’s covalent binding to FAAH. But it is somewhat surprising then that 90 volunteers get the drug at a low-dose with no ill effects, but that the increased dose so quickly cause this immune reaction in 5 out of 6, i.e. so relatively consistently. I would have expected more variability especially w.r.t. to immunity but of course that is not always the case and it could be that the 5 share most common HLA types whereas the 6th is different. And also I would have expected signs at the lower-dose but of course it depends on how fast the escalation was. Could also be it requires repeat exposures so it is not so dependent on size of single dose. It could also be the 6 were over-dosed and therefore there can be no comparison in effect to the previous lower-dose groups.

    As for other explanations, like contamination etc. I don’t know what to think. There is no common contamination I know of that would give this result (brain necrosis etc.) so it would have to be some other by/contamination-product in the medicine that has this effect, so we are basically back to finding the mechanism behind this 2nd compound. So the only element this explanation adds is that it could explain why it didn’t occur with the first 90 volunteers if they didn’t get the contaminated product.

    1. M, congrats on the BIA-10-2474 Wikipedia post. As called out in my own post it looks like you were editing in that patent structure at the same I was for However, as far as I could discern neither SciFinder nor SureChEMBL connected (daisy-chained) WO2014017936 to WO2010074588 via CID 72734378. I’d be interested in comparing notes should you care to de-anonymise directly (I edited the CID number into your entry)

      1. M says:


        Thanks although I didn’t put in the patent or molecule info, so must have been someone else.

  40. Chiara says:

    I believe sometimes a large-ish placebo cohort is added to early clinical phased to gather additional data which could be used to support MoA theories and/or efficacy analysis. I may be wrong.

    1. NJBiologist says:

      In a number of fields, there has been a push to generate evidence of target engagement as early as possible. This pressure has been especially acute in analgesic development ever since the failure of a number of programs focused on new targets in the 2000s. It’s entirely possible that the trial protocol called for some quantitative sensory testing to assess acute analgesic effects. (That’s speculation; I don’t know anything about the trial besides what I’ve read here.)

  41. Toxchick says:

    I’m most surprised that they did a study in chimpanzees. That suggests that is a target not present in humans, and also that it is a rather old drug(chimp experiments aren’t really done any more, and how odd for a small molecule). I wonder if that is why they had difficulty in extrapolating a safe dose-it could be due to exaggerated pharmacology like TeGenero. I don’t think it could be a new-epitope. I think they were only 4-5 doses and 5 days into the trial. An immune response would take longer. I think the steroids were to manage brain swelling or something.
    Interesting and tragic that the whole cohort (presumably first in the multiple dose phase) was dosed together. That’s isn’t done with biologics, in my experience.
    Such terrible news. I hope we will be able to learn from the mistakes made like we did for TeGenero.

    1. M says:

      @Toxchick: Hi I agree that such a fast immune reaction would be unusual but with the immune system anything is possible.

      As mentioned I agree that the corticosteroids could have been given for other reasons, i.e. swelling etc.

      So if we were to speculate on a non-immune related route, what would it be? One option would be the drug inducing a massive stoke several places in the brain.

      There are two types of stroke, ischemic stroke from blood-clots forming and blocking blood flow in the brain, and haemorrgaic stroke from broken blood vessels etc. leaking blood into the brain.

      As for the first, there are many drugs affecting the nervous system (including recreational drugs) that has been associated with increased blood clot formation and indeed ischemic stroke. Such a stroke could lead to haemorrage and nercrosis as second-order effect, and thus could be consistent with wha is reported.

      The latter option (the drug directly causing haemorragic stroke) could potentially be induced by the drug intererring with platelets or other factors involved in blood cloting. There are known drugs that do this bu even they don’t cause an immediate haemorragic stroke even if they raise the risk. So if the AE were to be haemorragic stroke it would need to have a more direct affect in damaging the vasculature lining of the blood vessels etc. I have never heard of a drug doing this, and it also seems such a mechanism would take some time to manifest and that there would be many other effects not just related to the brain.

      So based on this, ischemic stroke seems to be the most likely option of the non-immune related ones.

      Finally, anyone remember this case:

      Now, this was a synthetic cannabinoid whereas in this case, the FAAH-inhibition should only increase the amount of the endo-cannabinoid. Still, it is the cannabinoid system that is affected in both casesand it could be the present drug was more potent than the other FAAH-inhibitors tried and could lead to unnatural levels of endo-cannabinoids. I would be curious if anyone knows the specifics of the above-mentioned case (i.e. from the Daily Mail article) and how one drag of synthetic cannabis could have such a disasterous outcome.

      1. M says:

        I forgot to mention that stroke could also be induced by decreased circulation such as through excessive vaso-constriction (rather than clot formation) and it seems to be the speculated means of action of synthetic cannabis.

    2. Chiara says:

      It is true that studies in primates are not so common, but if metabolism is substantially different dogs may not be a suitable model after the rodent studies.
      As long as there is a sound justification, primates are still used.

      1. alchemist says:

        sure, and I have done it in the past. But normally it is cyno, not chimpanzee… I doubt this was just a PK reason to choose chimpanzee

  42. Cueball says:

    I doubt that stuff was extracted from naturally grown Cannabis. When my son, for instance, had his cancer, back in 2013, he was given a drug based on synthetically produced Cannabis (to increase his appetite). This drug however caused side-effects which had be to “treated” with other pharmaceutical shit. I never heard from anyone, taking naturally grown Cannabis, anything about side-effects. No wonder people die from such shit.

    1. M says:

      Cueball, It is well-understood that the drug is in no way cannabis-based. The relation to cannabis is this: Cannabis contains compounds (such as THC) that bind to the cannabinoid receptors of the brain, thereby producing much of the effect associated with cannabis. However, the human body also contains natural compounds that bind to these receptors. One of these compounds is broken down by an enzyme called FAAH. The drug in this trial was supposed to bind to this enzyme so as to prevent the breakdown of the natural cannabinoid (= endo-cannobinoid) thereby enhancing its effect.

      So there could be some overlap in effect of cannabis with the effect of this drug. But as you can see it is a quite indirect relatinship and the drug itself is chemically not related to cannabis (or THC) at all.

      1. Cueball says:

        And i also read some of the other posts and news/press statements about it.
        Here in Germany there was no mention at all, that emend the news from Friday last week. And i doubt there will be. As usually. They make up some news, only to scare the shit out of people and as soon as that is proven to be wrong, no mention at all. In german there is a word to express how this makes me feel: Fremdschämen (to be ashamed for sth others have or have not done)

        1. M says:

          Yes true media reporting is often very bad — in all subjects.

          It was somewhat interesting that the French Minister so strongly emphasized and clarified it wasn’t cannabis, it sounded like it was some really important point to get aross. Almost sounded like like they were afraid it would give cannabis a bad name!

          In actuality, it was probably because it would have been seen as a sort of embaressement to the establishment if they had allowed a cannabis trial and it had went wrong (as if this trial is any better, just because it doesn’t involve cannabis!).

          How did your son do with the cancer and all?

          1. Cueball says:


            all went well. He had to go thru chemo but could avoid radiation. He lost his hair and had some side-effects in his stomach and gullet. He is fine now. You can, of course, never tell if cancer comes back again. But i can assure you if it comes back again, i will not trust our state-wisdom-stamped docs again. i have read too much since that time.
            thx for asking 🙂

          2. M says:

            Happy to hear about your son. Sounds like Hodgkin’s, but great that he didn’t need rads. I surely hope he doesn’t relapse and that is fortunately very rare. But even then I would put my money on modern medicine. Hodgkin’s used to be invariably fatal, and is not curable in most cases, even late-stage.

          3. M says:

            As was pointed out below I meant to say “and is NOW curable in most cases, even late-stage”.

  43. alchemist says:

    I don’t think we can already start pointing fingers at the reactive nature of the molecule. Many people here point out the problem of neoantigen formation. But that (to my knowledge) is loinked to a risk of idiosyncratic tox. Having most (all) of one cohort developing such side effects in a matter of days does not look like idiosyncratic tox at all, like others have already pointed out.
    More likely to be an off target effec tthat was not adequately visible in preclinical species (with or without intervention of the reactive moiety). And this kind of thing can happen even without anybody screwing up: imagine an off-target for which the molecule has no cross-reactivity vs the preclinical species, and which is not a part of the standard selectivity panels. In a number of human targets I worked on it was hard to get compounds cross-reacting vs the human target (e.g. chemokine receptors). If that happens with a toxicophoric target you are screwed…

  44. Istvan says:

    A ‘manufactoring error’ theory is being investigated:
    The English-language article is based on the original in French:
    Nevertheless, it would be suprising to find an undetected toxic contaminant.

    1. M says:

      Thanks for the info. It is of course prudent they are investigating this, just as it has been mentioned here as a possibility. So frem that perspective there’s zero news in the article. But maybe for that reason also, one can’t help speculating if there’s more to it – i.e. if they already have indications this is what happened. Also given that they go into some details about the manufacturing chain (Italy, Hungary) etc. If one were to look for a contamination the first thing I would do is not to look at these steps, but instead to look at the batch the medicine came itself, identify the possible contaminant and then trace backwards to see where that could have come from. If you don’t know if there’s a contamination at all, let alone what it is, it’s hard to see what they could actually check. I guess they could check some paperwork etc. and of course run tests on the ingredients but again then it really would make more sense to do the same for the final product first! And if you can’t find any abnormal thing there, would it really make sense to look at the manufacturing chain? Only if you were completely desparate and were grasping for straws.

      This also mention transport a few times. Transport could be relevant from at least two perspectives. It could cause a mixing up in drug along the way (maybe switching the active ingredient for something else) or improper transport could cause degradation into a toxic product. Of course the latter would be very bad news for a drug if it could easily decompose into a lethal drug (unless it was really gross improper transport at play, and not just some slight overheating etc.). So the mixup seems more likely.

      I’m wondering how much news gets published in French that doesn’t get translated to English. With TGN1412 there was a lot more information out in the days after the disaster.

      1. AF says:

        Does not seem to be much real info in french. Less than on this
        blog … I can’t trace where the corticoid information comes from.

        1. M says:

          Hi, I can’t find those articles either. But I did find it myself Saturday after someone above mentioned it. As I recall it was from a quote from Pierre-Gilles Edan, but I’m not sure if it was from the original press conference Friday. He specifically mentioned it was to control “inflammation”. Could be it is easier to google for it on the french media if it is the translations that have been taken down. I will post a link if I find it again.

        2. M says:

          I found it now, in French:

          “On essaye de contrôler la réaction inflammatoire par des corticoïdes, mais l’antidote de ce médicament n’est aujourd’hui pas connu”

          “We are trying to control the inflammatory reaction with corticosteroids but the antidote for this drug is not known at this time. ”

          I’ve seen the quote about a lack of antidote in English media as well, but they left out the other bit. Of course it is unclear from the quote if the inflammation is directly caused by the drug or if it is an indirect effect from e.g. stroke.

  45. Istvan says:

    Regarding the Forbes-overview by David Kroll:
    Again, most of what is written there is mere speculation. And the statment that “PubMed lists no publications of any type from Bial on any FAAH inhibitor” is misleading: Not all medicinal chemistry journals are referenced by PubMed (Chemical Abstact still remains the most comprehensive and reliable source of chemical information). In fact, there is a peer reviewed paper entitled “Design, synthesis, and structure–activity relationships of 1,3,4-oxadiazol-2(3H)-ones as novel FAAH inhibitors” published in MedChemCommun in 2011:!divAbstract

    As few days ago I also located this recent review “Fatty acid amide hydrolase inhibitors: a patent review (2009 – 2014)” referring to Bial patents in the area:

    Unfortunatelly, we need to wait until Bial discloses the structure of BIA 102474 and only then can one make educated guesses on the biochemistry and pysiology involved in or responsible for the unexpected toxicity of large doses of the substance, pure or contaminated.

  46. M says:

    This article also calls for more transparency – it seems outside researchers are completely in the dark.

  47. Christian L says:

    If I understand the covalent reaction mechanism correctly, then we have a 3-(4-imidazolyl)phenylurea molecule as a “leaving group”. This by itself looks big enough to have some effect.

    This is of course highly speculative, but the resulting molecule is for example an almost exact substructure of, which inhibits a Ribosomal protein S6 kinase (p90RSK). RSKs have been reported to be important in preventing apoptosis.

    1. M says:

      So as I read what you say, this by-product might inhibit an enzyme that inhibits apoptosis. So the net effect of this would then be an increase in apoptosis. I don’t know if this could lead to the symptoms observed, maybe if the effect was especially specific to the vasulature in the brain etc (due to the drug being bio-available there), but that’s of course highly speculative. But at any rate, it is a side-effect what would definitely need to study.

      1. Christian L says:

        Yes, that’s a quick hypothesis I came up with. But I don’t know enough about this to say whether it makes sense.

        My main point here is that we seem to have one equivalent of this product molecule for every protein that we covalently inhibit – maybe even more if the drug gets hydrolyzed by other means. It’s big enough to qualify as “drug-like”, and thus I think we should also look at this compound in any target search, whether in vitro or in silico.

  48. Bob says:

    Hi all,

    I don’t understand half of what you are all saying (I’m learning while reading), but this is so far the most interesting technical discussion I’ve found about this incident.

    Keep commenting ! Don’t let this comment section die ! 🙂


  49. tangent says:

    I get the general idea of how covalent modification could trigger an immune response, but I’m not having web-searching success at looking examples of compounds that have caused this in the past. Can anyone help?

    1. M says:

      As a starting point, see:

      Otherwise, google things like ‘covalent drug risks’ and many papers come up.

      Note that the risks discussed with covalent drugs is not limited to the immunological concerns.

    2. NJBiologist says:

      Pointing your favorite search engine at “hapten” will give you a good overview. For specific haptens, try urushiol or dinitrophenol (OK, the first is a natural product, but the second was a well-known diet aid in its day).

      1. toxchick says:

        The classic example of immunogenic haptens is desflurane:

        1. NJBiologist says:

          The ether anesthetics are interesting examples.

          Possibly because of timing (DNP was on the market in the 1930s, DES was introduced in 1992), there is a vast literature on haptenation built on DNP. If you ever have a day to kill, Baruj Benacerraf (and others, but his papers are easy to find) published some good stuff on this in the 1960s–it’s fascinating to watch them using pretty basic tools to work out effects of carrier vs. hapten, specificity for hapten congeners, etc.

          1. tangent says:

            Thanks for the examples of haptens, these are fascinating to read the details on. I see nickel is another fun one.

  50. Justa Retiree says:

    To M in reply to M
    January 18, 2016 at 10:49 am

    You told Cueball that Hodgkin’s is “is not curable in most cases, even late-stage.”
    I’m sure you meant NOW curable.
    My mother passed from Hodgkin’s in 1956 when she was 36 and I was 8. In those days it was a sure death sentence, thus my lifelong interest and joy in hearing about the great progress and success in treatment of this disease.
    @Cueball, his post is correct except for that word. I hope your son continues to be well. Trust science based medicine for the treatment of Hodgkin’s.

  51. M says:

    @Justa: Yes, thanks for point that out – I meant NOW, and I hope that was clear from the context. Sorry to hear about your mother, many good people lost their lives to Hodgkin’s in those days. It was only in the 60’s the first chemotherapy regimens capable of producing a cure were developed, and they were ver harsh. Today’s treatments (ABVD, BEACOPP etc.) are even more effective, delivering high cure rates with fewer side effects.

    1. Istvan says:

      Re: Pingback
      Oh, no! “Synthetic marijuana drug trial goes wrong”
      Very very confusing post. The readers of Natural Society deserve an intelligent and, most of all, correct article. How could this happened?

      1. Istvan says:

        sorry above: “…could have happened.”

        I have added a speculative comment to David Kroll’s blog at Forbes.

  52. Silence Dogood says:

    The “covalent drug” hypothesis many of you mention above seems very plausible in this case. I know all of us working in drug discovery look forward to the testing of this idea it as further facts come in (compound structure, chemical and biological reactivity, etc.) The disastrous outcomes of this Phase I trial seem at first very perplexing–why were some individuals receiving active compound left utterly unharmed (by neurological testing and MRI) while others in same or similar dosing cohorts left with possibly permanent brain damage or even died? Is it possible that the trial drug was so reactive a chemical species that off-target toxicity shows a high degree of biological variability in humans associated with metabolic state, GI transit, blood flow, expressed non-target proteins varying by individual based on genetic background, diet, etc.? Such patient-to-patient variability of random, toxic off-target effects of this drug might be expected for a highly reactive species, whose capacity to modify proteins is largely unselective for the FAAH enzyme.
    As discovered by a commenter above, if the structure of the BIAL 10-2474 looks anything like this (!divAbstract) then subjects in the study would have been exposed to a highly electrophilic oxadiazolone ring whose elimination from systemic circulation would be almost certainly and exclusively a function of chemical degradation by biological nucleophiles. We really need to know the structure of BIAL 10-2474. Now.

  53. eCB guy says:

    Its important to also note that the lesions and scarring seen in the brain may not have happened from a stroke, per se, but could have been due to the drugs off target effect at molecules that regulate brain fluid balance. The cursory descriptions of the damage seem to be reminiscent of vasogenic edema where there is an accumulation of water in the cerebral space that results in the development of vacuoles and lesions throughout the brain. It is possible that a non-specific effect of this drug could have somehow interacted with aquaporin 4 channels in astrocytic feet, which regulate water flow into the cerebral space, causing the development of cerebral edema. The presentation seems similar to what has been seen with epsilon toxin, which can also cause vasogenic edema. While very little is known about FAAH-2 there is almost no way this could involve a direct effect at FAAH. Hundreds of people have gone through phase 1 testing of very selective FAAH inhibitors in multiple pharma companies (Janssen, Pfizer, Merck, Sanofi) and no AE like this has ever been reported. Not once. Even with the repeated dosing regimens that have been examined (which have been up to 28 days). For this to happen in 5 of 6 people is too specific and would definitely have emerged if this was due to FAAH inhibition in any manner. This has to be a non-specific effect of the compound itself, or as others have suggested, some form of contamination or bad dosing. It is disappointing that there is not more news coming out, including some better description of the symptoms and specifics on the damage that developed so that scientists can begin thinking about this more broadly. Given that FAAH inhibitors do represent a very good target for some disorders (apparently not osteoarthritic pain though) it will be a shame if lack of information taints the public against this target.

  54. Pharmacologist says:

    While we are all left at speculation, If the drug was an irreversible FAAH inhibitor, it would not have been the first. What worries me usually is metabolism (non-linear PK and/or transport). Or if polyreactive metabolites are generated (more that a reactive, but selective parent drug).

  55. Sarah E says:

    The French newspaper Le Figaro has published the clinical trial’s protocol. Scroll down through the French and download the English PDF.
    On page 26 of the protocol BIA 10-2474 is listed as:
    3-(1-(cyclohexyl(methyl)carbamoyl)-1H-imidazol-4-yl)pyridine 1-oxide

    1. Istvan says:

      Merci beaucoup!

    2. Silence Dogood says:

      It’s basically a reactive carbamoylation reagent; close cousin of CDI.

  56. Bagnar says:

    Here is the correct structure.

    “According to the document, BIA 10-2474 is 3-(1-(cyclohexyl(methyl)carbamoyl)-1H-imidazol-4-yl)pyridine 1-oxide.” (The document is the detailled protocole)

  57. Michael Scherz says:

    I’ve read the protocol – was pleased to see the correct chemical structure finally – and note that the protocol (p. 27) describes NOAELs (in ng*h/mL) for mouse, rat, dog and monkey (without distinction whether chimp or cyno), with treatment durations up to 6 months in rats, up to 3 months in mice, dogs and monkeys. I’m amazed that such a seemingly reactive imidazolyl urea can achieve such NOAELs over a 3-month period. The protocol does not describe the adverse effects observed at doses *above* the NOAEL, and provides no guidance to the investigator on which, if any, drug-related signs or symptoms of toxicity are to be expected from the toxicological findings.
    The protocol also points to a weak inhibition of cytochromes 2D6 and 3A4 at concentrations below 30 microgram/mL. The NOAEL’s range from 20’000 to 180’000 ng*h/mL, i.e. at or well above this threshold. BIA 10 -2474 is also described in the protocol as “extensively metabolized”, but metabolizing enzymes are not detailed.
    It seems conceivable that the drug has (irreversibly) inhibited its own metabolism, and may have far exceed the planned exposure in the 5 affected invidividuals of the 6-member active treatment cohort. It will be very important to obtain the pharmacokinetic data of the compound from these affected invididuals, in order to sort out what has happened here, and how to avoid it in the future.

    1. spike says:

      I highly doubt that Tox studies were performed in chimps. The ethical issues with that (or even any studies in chimps) are huge. What struck me about the Tox program as described in the protocol was that it was very extensive and wasn’t designed to support a FIH study. I’m not sure that I have seen a six month study used to support a FIH study in HVs before. Having two non-rodent species (4 week and 3 month) is also unusual. The exposure in the dog is strange. Highly unusual to see a gender-related difference in exposure (and in this case it was different in the two studies). Also, the exposure in the two studies at the NOAEL was similar (50 mg/kg in the 4 week and 20 mg/kg in the 3 month).
      Similarly, having a full repro package before FIH is unusual. I doubt that it was a case of front-loading the program planning on success. It seems that there were reasons that so many studies were done. The IB is the document that we need to see but I expect that there was something interesting in the Tox program.

      1. Spike says:

        Forgot to note that having two 3 month Tox studies in rodents (rats and mice) is also unusual. That was one costly Nonclinical program. Looks like they were “species shopping”.

  58. Pharmacologist says:

    Interesting is that the radiolabel ADME in rat and dog did not suggest that large amount of parent or metabolites are bound irreversibly or for very long time in tissues.

    Excretion occurred mainly in the first 24-48 h and was essentially complete by 72 h. Though would be interesting to see pictures and actual levels per tissue over time.

    1. Michael Scherz says:

      The radiolabel studies are single (and usually low) dose studies. The consequences of sustained high-dose exposure need not mimic these single dose experiments; and the details provided on p. 28 of the protocol are not enough to judge how carefully the radiolabel studies had been designed.

    2. DJ says:

      It will also be important to know where on the molecule the radiolabel was incorporated. For all we know, it could be on the imidazole leaving group.

  59. AF says:

    It seems that of the 6 people who were hospitalized, the oldest died (49 years old), and the youngest was asymptomatic. Information found in
    a single local newspaper source.

  60. Anonymous Researcher snaw says:

    Interesting recent article has information that I had not seen elsewhere:

    1. eub says:

      Thanks, some interesting information there.

      It really does not sound from this like an aggressive ramp-up of dosage to the disaster. Even a cumulative or delayed effect requires a rather narrow window — no reported effects at 4 x 50 mg cumulative, one catastrophic response at 5 x 50, majority by 6 x 50. Are there other examples?

      I wonder about possibility of an additional factor in common across the affected group, something that had not been present earlier in the 50 mg series but now took effect. An environmental factor or something.

      Article says dosed up to 100 mg as early as October, but by implication those subjects were not in the group harmed later.

      The first hospitalization was Jan 10 following a 50 mg dose, which I believe they’re saying was the fifth 50 mg dosing, to an eight-patient group.

      Jan 11 the other seven volunteers received the sixth 50 mg dose (and I’m sure that decision will be revisited in hindsight). Later that day the first patient was in a coma. Dosing was then stopped.

      Five of the seven others (? not sure I follow the numbers) were hospitalized in the couple of days following. The French drug safety agency was notified Jan 14, after at least the second hospitalization, and I’m sure they will be reviewing precisely when.

      I would love to know what they were doing with their unusual animal tox species suite. FAAH2-related, or something else?

      —– Quoted:
      According to the timeline released by ANSM, the first patient was dosed in early July with 0.25 mg. Those dosages were then increased, with some participants receiving as much as a 100 mg does in October 2015. No issues seem to have come to light (or were reported to ANSM) until January 10, 2016. On that day, the fifth of administration of a 50 mg to eight participants caused one volunteer to show symptoms that required hospitalization later that evening.

      The following day, seven other volunteers received the sixth treatment dose in the morning. That day, the participant hospitalized the previous evening received an MRI and was in a coma. At that point the trial was discontinued per the agreement between BIOTRIAL and the sponsor. Between January 13th and January 15th the five other volunteers were also hospitalized, with ANSM being notified of the adverse events on January 14th.

  61. Mandrake says:

    Turns out the compound in question is a pyridine N-oxide with a few molecular relations to a toxin found in mushrooms and also to paraquat

    Why did this company decide of all the FAAH inhbitors they had to test the most unstable one in man?????

  62. Arianne says:

    From tomorrow i’m gonna lose weight. Summer is comming girls!
    Who starts with me? I found a good method on the web and want to give it a try, simply type in g00gle; fitMarikka advices

  63. Bill says:

    Cravatts whole idea failed….

Comments are closed.