Fentanyl, carfentanil and the opioid crisis: where do we stand now?

September 19, 2017, 9:50 am


Controversies and carfentanil: We have much to learn about the present state of opioid poisoning. Cole JB, Nelson LS. Am J Emerg Med 2017 Aug 24 [Epub ahead of print]


I have argued — both in a post on this blog and a column in Emergency Medicine News — that the idea of a “heroin overdose” is a completely outdated concept that is never coming back. The fact is, when a patient comes in today with a history of shooting or snorting heroin, none of us have any firm idea about exactly what drug or drugs were involved. The chance that such a patient was exposed only to heroin is virtually zero. More likely, the exposure was to some undefinable hodgepodge of fentanyl, carfentanil, U-47700, and Lord knows what else, perhaps with a pinch of heroin thrown it.

The authors of this brief, cogent article trace the history of the last 3 decades during which the epidemic of opiate over-prescription, overdose and addiction developed, and supplies of pure unadulterated heroin virtually disappeared. It is only over the last year or so that the fentanyl analogue carfentanil — a drug 100 times more potent than fentanyl itself — has been regularly detected in samples of “heroin.”

The authors aptly argue — as I did in the sources cited above — that this should change our mindset when managing opioid overdose patients:

“For the foreseeable future emergency physicians, EMS personnel, epidemiologists, law enforcement, and even drug dealers, should not think about heroin, fentanyl, carfentanil, or any specific opioid overdose; rather they will need to consider all issues under the generic rubric of opioid overdose,” forcing us to each learn new tactics even as we approach a decades old problem.”

These “new tactics” may involve reevaluation of the initial “wake up” dose of naloxone administered, and the amount of time these patients should be observed in the emergency department after reversal of opioid-induced respiratory depression.

This is a good short read, and recommended. The authors raise great questions. We will need additional data and studies before we can answer them definitively.

Related Posts:

Tox Podcast Episode #15: Naloxone in the Age of Carfentanil

Elephant-tranquilizer (carfentanil)-tainted heroin showing up in Ohio

Treating “heroin” overdose: the past is no guide


  1. RP Says:

    How is it economically sound for the drug dealer to be mixing in what I assume are more expensive components into the heroin? Unless they are achieving some sort of economies of scale shipping more potent drugs (smaller packages possibly easier to smuggle), how are they making money? Plus they are killing off their custome base. What incentive is there to do to this?

  2. Leon Gussow Says:


    Thank you for the questions. I think you’re looking at the economics of illegal drug distribution from the wrong angle. Heroin itself is actually rather expensive to produce and distribute: the poppy plan has to be cultivated, harvested, processed and smuggled. Much more potent synthetic drugs such as carfentanil have a much smaller volume per dose.

    As for killing off the customer base, that seems to be a byproduct of dealing with extremely powerful opioids and having poor quality control. In many cases, I suspect that the typical drug dealer on the street does not know what’s really in this product, or in what dose.

  3. milkshaken Says:

    As a pharma synthetic chemist to the economy of illicit drugs: In fact, synthetic super-potent fentanyl opioids are rather simple to make – it is comparable with MDMA. Any professionally equipped lab can make them easily.
    For fentanyl, it takes as little as 2 or 3 synthetic steps, the whole production can be completed within a week and the precursors are too common use so it is difficult to limit the access to the fentanyl precursors (like it happened in case of MDMA). The starting materials are very common, cheap, and not watched, so you run no risk ordering them.
    The enormous potency is what makes fentanyls so profitable. The recreational dose of fentanyl starts well below 0.5mg. For carfentanyl, the recreational dose is in single digit micrograms – amount that is too small to see with a naked eye. So if you are a shady custom synthesis lab in Shanghai, you could make ridiculous profits: 1 ounce of fentanyl is equipotent to one pound of top-quality heroin. 1 ounce of carfentanyl (which takes 2 more steps, so it is about twice more labor intensive than fentanyl to make) is equipotent to 50 kilos of heroin.

  4. ÐR Says:

    I’ve been told having customers OD on your stuff is actually great for business since it means you’re selling the good *&%^#. You’ll have addicts lining up your corner in no time.

  5. Leon Gussow Says:


    I agree. Synthetic opioids are much easier to produce than actual heroin, which is why it is a dead certainty that heroin itself is never coming back in any big way.


    i’ve heard rumors that for drug dealers having word out on the street that their stuff actually kills people is good for marketing, but have not seen any supporting evidence. I tend to doubt that this is true, by and large.

  6. Rich Hamilton Says:

    Was there ever such a thing as “just” heroin? In the 80’s it was often much less than 5% heroin and maybe mostly quinine. Agree that the fentanyl derivatives add a new wrinkle but in my experience, careful application of the approach by Christneson still differentiates those that are safe to discharge.
    J. Christenson, J. Etherington, E. Grafstein, et al. Early discharge of patients with presumed opioid overdose: development of a clinical prediction rule
    Acad. Emerg. Med., 7 (10) (2000;Oct), pp. 1110-1118

  7. Leon Gussow Says:


    Thanks for the comment. I agree that “pure” heroin never really existed — it was virtually always cut with something. But the use of synthetic opioids that are orders of magnitude more potent than heroin itself is relatively new over only the last few years.

    In their study, Christenson et al derived a rule that they hoped would allow safe early discharge of patients who had received naloxone for presumed opioid overdose. The review for readers, their derived rule had 6 components determined 1 hour after naloxone was administered. The rule held that the patient could be safely discharged if at 1 hour after naloxone administration they:

    1) can mobilize as usual
    2) have oxygen saturation on room air >92%
    3) have a respiratory rate > 10 and < 20 per minute 4) have a temperature > 35.0 C and < 37.5 C 5) have a heart rate > 50 and < 100 per minute, AND 6) have a Glasgow Coma Scale of 15 They found that by applying this derived rule to the derivation set of 573 patients, it had a sensitivity of 99% and a specificity of 40% identifying patients who would not suffer an adverse event after the 1-hour mark.This rule has to my knowledge never been validated. It was derived from a set of patients of whom 86% admitted to heroin use during the 1990s when the location of the study — Vancouver BC — had a ready supply of "inexpensive high-grade heroin." Patients in the study received a naloxone dose of 0.4 mg IV or 0.8 mg IM. And by the way, there was no real follow0up on almost of third of the patients. For all these reasons, I do not think if justified to generalize the rule to a different time when different drugs are prevalent. The authors found that in their derivation set, 193 of 573 patients met criteria for discharge at 1 hours after naloxone. I am sure that even if the rule was strictly applied today, much fewer patients would meet criteria.

  8. Rich Hamilton Says:

    Hi Leon
    The patient that EM physicians needs help with is not the still sleepy “heroin” OD, it’s the wide awake, 2 hrs post naloxone, eating a sandwich, walking around the ED patient, demanding to be discharged patient. Making that patient stay for 2+ more hours because of fentanyl adulteration is an overstated concern in my opinion.

  9. Leon Gussow Says:


    In general I agree but would argue that we can’t look at studies done decades ago to support our decision. Also, what if that patient had ingested counterfeit opioid pills bought on the street, and had needed 10 mg of naloxone to become wide awake.

    At this point it seems to me that the decision to discharge after naloxone is clinical and based on experience. It is not evidence-based, and never will be since the agents involved change constantly.

  10. Istvan Ujvary Says:

    Milkshaken writes:
    “For carfentanyl, the recreational dose is in single digit micrograms.”
    I am just being curious. Are there any human studies or other data to support this? What actually is the ‘recreational dose’? Does it differ from noticable bioactivity that one would treshold dose?

  11. Charles Caffrey Says:

    Spoke to a guy addicted to heroin while waiting for my bus the other day in downtown Chicago. A single anecdote, but apparently dealers will drive around giving free samples of new mixtures in order to, in his view, “experiment” on addicts. He knows many who have overdosed from these free samples, and he believes dealers will then use this information to tweak the recipes appropriately. Also, he didn’t buy into the whole “if you OD on it, it must be good shit” thing. I wonder if any opioid addict actually buys into that, or if it’s just a myth perpetuated by the lay public and medical providers. In fact, he only buys from one dealer he knows and trusts in order to minimize his risk of getting a superpotent batch that may kill him.

  12. Leon Gussow Says:


    Good question. In my opinion, the concept of a “recreational dose” doesn’t apply to carfentanil, since it is not a “recreational” drug by any definition. But by way of comparison, some commenters say that the lethal dose of heroin is about 75 mg. Since carfentanil is thought to be 10,000 times more potent than heroin, that would make the lethal dose 7.5 micrograms.


    i’ve never seen any evidence to support the (in my opinion) urban myth that if a drug dealer’s product kills a customer or two it’s good for business. As for the person you talked to at the bus stop, I think it’s dangerous to count on the quality control of even a trusted dealer. I would think that in most cases the dealers themselves are not sure what’s in their product, or the precise amounts involved.

  13. Istvan Ujvary Says:

    To Leon on ‘lethal’ dose:
    I doubt that one can make such an extrapolation from analgetic doses, or rather antinociceptive ED50 values obtained in one particular assay using one rodent species, to animal let alone human toxicity. (BTW: in human PET studies, C-11 labeled carfentanil has been used in sub-pharmacological doses up to ~9 microgram.) In the single rodent tox study reported by Janssen-group carfentanil and fentanyl appeared to be equitoxic. Analgesia and respiratory depression could be regulated by different biochemical mechanisms/signal transduction pathways thus do not necessarily correlate. I think that – even after 200 years – there is much to learn about opioids in general. It is sad and somewhat ironic that recent human tragedies have reinvigorated research in this area. This is a too high price for progress!
    Of course I am not saying that the use of carfentanil of carfentanil or any other of the fentanils for whatever purpose is ‘safe’ without medical supervision.

  14. Leon Gussow Says:


    I’ll say again that any concepts of “recreational” or even “therapeutic” doses of carfentanil are misguided. The drug is potent and unpredictable, and anyone who takes street opioids or “heroin” is playing a very dangerous game. I’m not familiar with the Janssen study you refer to, but I believe that any attempt to apply rodent studies to humans is unjustified.

  15. Adam Says:

    I think it is ironic that you wrote “any attempt to apply rodent studies to humans is unjustified.”. Rodent studies are the basis for the carfentanil potency estimate that you cited — I’m not sure why it is OK to extrapolate from rodent data to humans for one purpose but not for other purposes. The other person who posted is correct that analgesic potency estimates for fentanyl derivatives are poor predictors of LD50 values. Carfentanil does have a wider safety index then fentanyl, but ultimately carfentanil is still extremely dangerous given its ridiculously high potency.