Treating “heroin” overdose: the past is no guide

November 22, 2016, 12:06 am


Do heroin overdose patients require observation after receiving naloxone? Tillman MW et al. Clin Toxicol 2016 Nov 16:1-7 [Epub ahead of print]


The stated goal of this study was to search the medical literature in an attempt to answer 3 main questions:

  1. “What are the medical risks to a heroin user who refuses ambulance transfer after naloxone?
  2. “If the heroin user is treated in the emergency department after naloxone, how long must the be observed prior to discharge?
  3. “How effective in heroin users is naloxone administered by first responders and bystanders?”

The authors searched Pubmed and Google Scholar for articles relevant to these questions, identifying literature going back to the 1990s. The paper does not report any stringent methods as to how relevant literature was selected, or how chosen papers were analyzed. Nevertheless, the authors concluded . . .

Screw it. By the time I was halfway through the paper, I really didn’t care what the conclusions were. It was clear to me that there was no way that this type of literature review — even if it had much better methods — could provide clinically useful answers to the questions asked. The fact is, first responders and emergency practitioners never deal with “heroin overdose patients.” They treat patients who may have taken an opioid that could be heroin, to some substance they thought was heroin, or something that was given or sold to them purported to be heroin. And “heroin” now may not be anything like the “heroin” on the street when much of the research reviewed by the authors was carried out. Today “heroin” may be some god-awful combination of U-47700, fentanyl, carfentanil, and who knows what else — maybe with some heroin thrown in.

We really don’t know the pharmacokinetics of many of these drugs, and there is no reason to believe that they will be the same as those of heroin. I would suggest that data from a decade or two ago have no certain relevance to what we’re seeing clinically today. Over the last several years we’ve entered a whole new world of street drugs, and unfortunately we can not rely on past experience to guide us.

Related posts:

Keys to the safe use of naloxone

Position statement on bystander naloxone



  1. Steven Aks Says:

    Hi Leon,

    Kudos to our colleagues, for tackling such important questions about naloxone, and attempting to work through these areas of controversy. The three topics selected could have easily been in three separate papers.

    Question one is very important “What are the medical risks to a heroin user who refuses ambulance transport after naloxone?” Indeed there is mounting literature that those who sign out AMA in the prehospital setting don’t tend to show up in morgues in the next 24 hours, but we must remember that there are a lot of problems with this patient group. We do not know that these patients have significantly overdosed, and if they did, what they have taken? I agree with the good Dr. Gussow that the current state of street opioids is a moving target with everything from fentanyl to carfentanyl out there currently. We need careful prospective study if we want to answer this.

    I’m not quite there to accept the conclusion for the answer to question number two: “If the heroin user is treated in the emergency department, how long must they be observed prior to discharge?” They conclude: “an observation period of one hour with documentation of normal respiratory rate and consciousness is sufficient to exclude risk of life-threatening sedation or respiratory depression.” We don’t really have enough literature to support this claim. The one paper by Christenson is the only one that studies the one-hour rule. It is important to note that, despite the best efforts of the authors, only 20% of these cases were directly followed up. It is very possible that they missed significant problems after discharge in these patients. Studying 2 to 4 hours post overdose observation periods prospectively would be very important to address the duration of action of naloxone and to be certain of relapse of sedation after a long-acting or ultra-potent opioid.

    It is also worth noting that the three scenarios raise three different liability standards. We have prehospital, emergency department, and lay-bystander groups. All of these groups have different standards of training and different standards of care. It would be useful to understand the impact of the different levels of training, indemnification, and liability of the groups and how that affects decision making.

    Again, I applaud the efforts by the authors with this collective review and hope that this inspires controlled prospective study with good outcome measures.


  2. Leon Gussow Says:


    Thanks so much for the comments.

    I would argue that “careful prospective study” will never happen in any useful way. Because “heroin” is now an undefinable concept instead of a specific substance, any such study would be outdated long before it could be published, and not applicable with any generality. The best way to deal with this mess will be to stay current on recent reports (thank you, Twitter and #FOAMed!), as well as alerts from local poison centers and public health departments. And then, based on all incoming anecdotes, make a reasonable decision as to the best clinical management. Again, the past will be no guide.

  3. Ian Mitchell Says:

    Hi Leon:
    I was a young clinician working in the ED during the Christenson group’s study on early discharge after naloxone administration. I enrolled a number of patients for the study myself. The followup for this study was quite thorough in terms of looking through provincial databases.
    “After several months, we compiled a list of all patients who had not been directly contacted and searched the medical records databases of 6 local hospitals where these patients could have subsequently presented for care. For those who had a repeat hospital visit within 24 hours of the safe discharge assessment, we determined the reason for the visit and documented any treatment provided. Finally, we searched the records of the Chief Coroner’s Office and the Department of Vital Statistics to determine whether any patients had died after ED discharge.”
    In addition to providing universal health care, our government’s single payer system makes it relatively easy to see if any bodies turned up from adverse outcomes after discharge. Possible exceptions would have been people who were able to make it to the US or to another province before dying or made it to rural BC before needing medical care again.

    I am strongly in agreement that this study cannot be used to support a similar current policy as there is very little heroin in our local fentanyl these days. I have heard scattered reports of clinical decline up to 6 hours post naloxone. Unfortunately, for many IVDUs, the practical aspect of keeping them in the ED for 6 hours makes such a policy doomed from the beginning. Given the massive doses of fentanyl that people are using, I suspect their pharmacokinetics are different from those getting surgical doses.

    In my department, they leave with a take home naloxone kit when they are awake and as functional as we expect them to get.

    Lastly, do you know of any tox confirmed cases of fentanyl in street cannabis? This has been hinted at by our politicians, but denied by health and police authorities.

  4. Leon Gussow Says:


    Thank you for your comments. I agree that when determining a reasonable observation period in these patients, it is important to consider feasibility and the most current information about what drugs are on the street and recent experience of other emergency departments.

    I’m not absolutely sure, but I think the situation should be relatively simple with IV drug abusers. The effects should be fairly rapid, and if the patient seems ready to go 2 -4 hours after administration of naloxone — when the antidote effect should be absent — at least conceptually those people should do well. In my opinion, anyone with resection who requires a repeat naloxone dose — or drip — in the ED should probably be admitted.

    The situation is more complicated with oral overdose. Recent evidence indicates that counterfeit opioid pills may contain massive amounts of fentanyl, and that some patients have experienced renarcotization many hours after naloxone has been stopped. (See our latest podcast on “Naloxone in the Age of Carfentanil.” I think at this time it is not clear what a reasonable observation period would be.

    I know there was concern in Canada about cannabis being laced with fentanyl. Apparently that concern was unfounded, and I am not aware of documented cases of this occurring anywhere else.