Skip to content

‘Treat and Release’ after Naloxone – What is the Risk of Death?


NaloxoneOften in the prehospital setting, naloxone is administered by EMS (or possibly a bystander) to reverse respiratory and CNS depression from presumed opioid overdose. The patient then wakes up, and not uncommonly, refuses transport to the hospital. The question is: Is it safe to ‘treat and release?’ Or, rather, what is the risk of death associated with this practice. A hot-off-the-press article, just published in Prehospital Emergency Care, addresses this question.

What is known

The folks over at The EMJClub Emergency Medicine Podcast (Dr. Brian Cohn, @emjclub) published an analysis of this topic back in 2014 with toxicology expert Dr. Evan Schwarz (@TheSchwarziee). They identified 4 studies that looked at this exact question.1–4 The studies varied in methodology, but all found similar results: the risk of death from recurrent opioid toxicity after naloxone administration was low, ranging from 0 to 0.13%.

Disclaimer: This data should not change practice and is only intended as a review of the available literature with analysis of their conclusions and limitations. We are discussing just one aspect of this multi-faceted, complex issue.

U.S. Prehospital Studies

The general strategy employed was retrospective chart reviews of medical examiner and prehospital records looking for occurrences of death after refusal of transport. The U.S. studies were conducted in the cities of San Antonio, TX and San Diego, CA.1,3 The upside here is that the EMS systems are similar to those in other parts of the country. The downside being that we don’t have centralized databases in most states and therefore patients could be missed if they presented to other counties within the state. No deaths were documented in either study.

Non-U.S. Prehospital Studies

Rudolph et al. published on their experience in Denmark.4 EMS there is a bit different in that physicians are present in the field to assess the patient and make transport decisions. Furthermore, there is a central database, meaning that patients were probably not missed if they presented elsewhere. In this study, they also included poison center records. Three patients out of 2,241 (0.13%) were identified as having rebound opioid toxicity that likely led to death. A similar study in Helsinki found no life-threatening events during a 12-hour follow-up period in 71 patients who refused transport after naloxone.2

Emergency Department Study

Watson et al. took a different approach.5 Utilizing a chart review strategy, they aimed to determine the frequency of opioid toxicity recurrence after an initial response to naloxone in sequential adult ED patients. The authors found that up to 45% had recurrent toxicity. Despite being an ED-based study, the results are difficult to interpret. Only 2 of the patients with recurrence had respiratory depression documented and neither received more naloxone. Most of the patients were oral opioid overdoses, rather than heroin. One take home point that is probably applicable: recurrence was more frequent with long-acting opioids, though it also occurred with short-acting opioids including heroin and codeine.


Prehospital Emergency Care

A New 2016 Prehospital Study

A new prehospital study, published in Prehospital Emergency Care, also assessed the risk of administration of naloxone with subsequent refusal of care.6 The authors conducted a retrospective review of all patient encounters by the Los Angeles Fire Department during July 1, 2011-December 31, 2013. The Coroner’s records were reviewed to determine if a patient with the same or similar name had died within 24 hours, 30 days, or 6 months of the initial EMS encounter. Of the 205 subjects identified, one (0.49%) died within 24 hours of the initial EMS encounter. The cause of death was coronary artery disease and heroin use. Two additional subjects died within 30 days, but the cause of death was either unknown or unrelated in both cases.

Key Point: All of the studies were retrospective and may have missed patients, particularly recurrent toxicity that didn’t lead to death. It is always safest for patients to be transported to the ED for evaluation.

Application to ED Clinical Practice

  1. If a patient presents to the ED after receiving prehospital naloxone for opioid toxicity, it is worth observing them for at least an hour (longer dependent on the situation). Be sure that after the naloxone has worn off, s/he doesn’t have recurrent opioid toxicity. Only one of the studies evaluated ED patients and found a higher rate of recurrent toxicity compared to the prehospital studies. The primary outcome in the prehospital studies was death. We can monitor more closely in the ED and can provide resources including substance abuse referrals and take-home naloxone.
  2. The most common opioid in the earlier studies was heroin. A one-time naloxone dose is generally sufficient to reverse heroin with a limited threat of recurrent toxicity. However, the opioid epidemic has changed, such that heroin is only part of the current problem. Prescription medications, fentanyl, and other opioids can be longer acting than naloxone’s 45-60 minute duration of effect. Adulterants also play a role, as highlighted by the recent CDC report on increased deaths related to fentanyl. The Levine study aimed to reevaluate the earlier data in light of the current times, but only captured patients up through the end of 2013.6 Although they found a low rate of death in 205 patients, recurrent toxicity may have been missed by their inclusion criteria.
  3. The EMJClub Emergency Medicine Podcast summarizes the prehospital data nicely:

“The bulk of this data supports the ‘treat and release’ strategy adopted by many EMS systems, with the caveat that such a strategy be employed in select patients who have returned to baseline with stable vital signs and are capable of understanding the risks associated with discharge in the field. If patients want to go to the ED, this should still be encouraged as patients could be evaluated for drug related infectious diseases, as well as receive information about addiction treatment and other social services. Transporting the patient against their will, and holding them in the ED, is probably unnecessary and does not seem to be supported by available evidence. However if the patient took a longer-acting opioid such as methadone, it may be prudent to specifically warn them of possible risks associated with these agents as studies did not specifically look at the safety of a ‘treat and release’ strategy in patients exposed to long-acting opioids.”

  1. (Added August 25, 2016) Dr. Jeff Lapoint provided an important point in his comment below. Namely, naloxone in-and-of-itself is not benign. Risks associated with administration include delayed pulmonary effects and precipitated opioid withdrawal. Both of these conditions potentially warrant close observation and argue against the treat and release strategy.
  2. (Added November 17, 2016) A 2016 review drew the same conclusions as our ALiEM post. Specifically for heroin overdose, “In the absence of co-intoxicants and further opioid use there is very low risk of death from rebound opioid toxicity.”7 The authors go on to say, “For those patients treated in the ED for opioid overdose, an observation period of one hour is sufficient if they ambulate as usual, have normal vital signs and a Glasgow Coma Scale of 15.”

Bottom Line

  • We should not overturn the practice of ED observation for 4-6 hours. The data simply suggests that if a patient refuses transport at the scene or wants to sign out against medical advice after receiving naloxone, s/he has a low risk of death.
  • Keep in mind that the available data predate 2014, when fentanyl, carfentanil, etc. were not yet a big part of the scene. Therefore, all of the studies most likely included predominately heroin and oral opioids and do not account for the new, more dangerous adulterants.
Updated Aug 25, 2016 – The original title of this post was ‘Treat and Release after Naloxone – Is it Safe?’ While use of the word ‘safe’ is consistent with the terminology used in the studies, the primary outcome in most of them was actually death. As Dr. Jeff Lapoint pointed out on Twitter, lack of death does not equal safety. Therefore, the title was modified to more clearly match what was studied.
Updated Nov 17, 2016 with new review article.7
Vilke G, Sloane C, Smith A, Chan T. Assessment for deaths in out-of-hospital heroin overdose patients treated with naloxone who refuse transport. Acad Emerg Med. 2003;10(8):893-896. [PubMed]
Boyd J, Kuisma M, Alaspää A, Vuori E, Repo J, Randell T. Recurrent opioid toxicity after pre-hospital care of presumed heroin overdose patients. Acta Anaesthesiol Scand. 2006;50(10):1266-1270. [PubMed]
Wampler D, Molina D, McManus J, Laws P, Manifold C. No deaths associated with patient refusal of transport after naloxone-reversed opioid overdose. Prehosp Emerg Care. 2011;15(3):320-324. [PubMed]
Rudolph S, Jehu G, Nielsen S, Nielsen K, Siersma V, Rasmussen L. Prehospital treatment of opioid overdose in Copenhagen–is it safe to discharge on-scene? Resuscitation. 2011;82(11):1414-1418. [PubMed]
Watson W, Steele M, Muelleman R, Rush M. Opioid toxicity recurrence after an initial response to naloxone. J Toxicol Clin Toxicol. 1998;36(1-2):11-17. [PubMed]
Levine M, Sanko S, Eckstein M. Assessing the Risk of Prehospital Administration of Naloxone with Subsequent Refusal of Care. Prehosp Emerg Care. March 2016:1-4. [PubMed]
Willman M, Liss D, Schwarz E, Mullins M. Do heroin overdose patients require observation after receiving naloxone? Clin Toxicol (Phila). November 2016:1-7. [PubMed]
Bryan D. Hayes, PharmD, FAACT, FASHP

Bryan D. Hayes, PharmD, FAACT, FASHP

Chief Science Officer, ALiEM
Creator and Lead Editor, Capsules series, ALiEMU
Attending Pharmacist, EM and Toxicology, MGH
Assistant Professor of EM, Harvard Medical School
Bryan D. Hayes, PharmD, FAACT, FASHP

Latest posts by Bryan D. Hayes, PharmD, FAACT, FASHP (see all)

  • Jeff Lapoint


    I’m glad to see you added the above disclaimers and removed the bold highlights from the initial post. Your post is certainly informative and as always, well written. Indeed, there is a very low reported incidence of death after prehospital naloxone administration in the studies reviewed. I count myself among the many who look forward to your thoughtful posts.

    There is room for some clarification though. You posted a blog entry that has received over 7,000 views and tweeted it to almost 10,000 followers with the title “Treat and Release after Naloxone – Is it Safe?”. Safety, in this case, being defined as the lack of dying or, more accurately, not having the death be captured in a variety of retrospective studies. The title is misleading and the only text in bold of the initial post was the extremely low rates of mortality reported in one of the studies. Please consider changing the title or including a discussion about the risks of naloxone administration including delayed pulmonary effects and precipitated opioid withdrawal lest the casual reader or beginning learner mistake the post for an endorsement of clinical practice.

    In addition, there are problems with the data presented that were not acknowledged. How many of the patients really needed naloxone? Are we looking at group that wouldn’t have died regardless of whether or not they received naloxone? Are these studies just looking at people who were kinda sleepy or had consequential respiratory depression? In some of the studies we know in others not so much.

    Today’s opioids and their adulterants are quite different from anything we’ve seen before. Let’s focus our educational efforts on the safe and appropriate administration of naloxone to reverse respiratory depression and safe disposition after thoughtful observation.

    • Bryan D. Hayes

      Jeff, thank you for the thoughtful comments. Your point is well taken and we will consider a title change. Your concluding paragraph hits the nail on the head. As mentioned in the post, this is a complex issue and our review is focused on a small piece of the big picture. The practice of treating and releasing patients after naloxone is seemingly risky and dangerous. That being said, it is already standard practice in some EMS jurisdictions.

      Whether we want to pretend the data doesn’t exist or not, the fact of the matter is that once the newest study (reference 6) is published in print (currently Epub ahead of print), there was going to be a lot of discussion in both the prehospital and ED communities. As we know is the case in our busy lives, it is easy to skim an abstract and convince ourselves that the conclusions are accurate and applicable, sometimes without personally conducting a thorough investigation. Levine et al’s conclusion states: “The practice of receiving prehospital naloxone by paramedics and subsequently refusing care is associated with an extremely low short- and intermediate-term mortality. Despite an evolving pattern of opioid abuse, the results of this study are consistent with previously reported studies.” Boyd, et al take it one step further and conclude: “Allowing presumed heroin overdose patients to sign out after pre-hospital care with naloxone is safe.” So, instead, I thought it would be prudent to discuss the other related studies in light of the new article, especially considering the overreaching conclusions in several of them.

      The 0-0.13% number cited in our post does not reflect only one study’s findings, as you stated, but presents the range from all 4 of the prehospital studies prior to the 2016 article. Your question regarding how many patients actually needed naloxone is a valid one. Most of the studies described the details of the patient population. Boyd et al included patients with a Glasgow Coma Score of 8 or less. Signs of respiratory depression were considered to be a respiratory rate of less than 12 breaths/min, a peripheral oxygen saturation of less than 90% without supplemental oxygen (or less than 95% with supplemental oxygen) or cyanosis on arrival of the first responding unit. Vilke et al included patients who received naloxone, had an improvement in mental status, and signed out AMA. Wampler et al included patients who received naloxone, had a documented respiratory rate less than 12, and subsequently refused transport. Rudolph et al studied EMS units (MECU) in Denmark where a physician is present. At the scene, the MECU doctor establishes the diagnosis of acute opioid overdose based on the case history and typical clinical presentation. With over 4,000 patients (half coming from the study with an on-scene physician directing care), it’s likely that many of these patients did, in fact, need naloxone.

      As was mentioned in the post, the evolution of the opioid epidemic (with the adulterant issue you mentioned) adds even more complexity and significantly limits generalizability, even of the newest study’s results. We are neither defending the ‘treat and release’ strategy nor advocating for its use. What we are doing is providing a focused review of the available data on this topic and the historical context for the Levine article’s conclusions. We thank you for providing an additional perspective and we will add some of your points into the post, along with a title change.

      • Evan Schwarz

        First off, I’d like to thank Bryan for starting a discussion on a very important topic and Jeff and others for their very thoughtful and well stated commentary. As Jeff very rightfully points out, there are several limitations to this data including some studies that were not clear in their naloxone indication or their dosing strategy, although as Bryan points out the ones that did give their indications seem reasonable and what I’d expect from my EMS colleagues. They also seem like what I’d expect in real world situations. It’s also true that some of the articles conclusions may have overreached and may not have fully captured the nuances of this very complex topic but that doesn’t invalidate their results. In addition, we really have no idea what presumed opioid the patients took in any of the studies; the assumption is that their conclusions are based on short acting opioids such as heroin. As all of us are aware, the opioids used and any adulterants in them change over time. This will be true no matter how recent the study is. Today’s concerns are generally with the very potent fentanyl and fentanyl derivatives but I also can’t remember a time when we weren’t concerned about what else could be in the heroin supply.

        With that said, this is the best available evidence out there right now and the results are consistent amongst the studies which are published over a period of time. In addition there is no such thing as a perfect study. Even in the best circumstances, my patient in front of me will not be exactly like the patients in the study. In any situation, all I can do is apply the best available evidence to the patient in front of me and use that in conjunction with what I know and the patient’s preference to make the best decision for them while always making sure to apply a healthy dose of skepticism whenever I interpret any data.

        Does that mean anyone with a presumed opioid overdose should be given naloxone and then allowed to leave? Absolutely not. I also don’t think it means that just because a patient wants to leave, that we necessarily have the right to restrain them against their will if they have capacity to understand their decision. While dramatic, I know this is what I saw at times during residency. This is also what prompted Brian Cohn (@emjclub) and myself to first look into these studies as our local EMS was changing to a ‘treat and release’ system and the hospital was advocating placing all of these patients on elopement precautions and taking away all their possessions and holding them against their will. To us, this just seemed incongruent and we were trying to get a better idea about the risk to these patients to have an informed discussion with them. I also realize anything that challenges long held, dogmatic beliefs can be quite shocking and will naturally provoke passionate responses. Of course this happens on both sides. I’ve also received responses to this topic asking why this is even newsworthy as a ‘treat and release strategy’ to them is just standard care (please note I’m saying their standard care and not standard of care).

        In the end, I think everyone will have to determine how to apply this, if at all, to their own practice. As stated, there are many benefits of hospital transport. In addition to making sure they don’t have recrudescence or develop other complications such as pulmonary edema, they can also receive testing for infectious disease and receive addiction resources. Of course even if released, I’d still advocate for encouraging both medical and addiction follow up. For me, I’ve used the best available evidence to try to have the most informed discussion I can with my patients that I deem to have capacity to make an informed decision and with my EMS colleagues when they ask me about the pros and cons (of which there are both) about a ‘treat and release’ strategy.

  • Levi from Aus

    There was a presentation on this topic at Australia New Zealand College of Paramedicine conference last week. The presentation concluded similarly to this article – that is, that there is an extremely low (discernible) incidence of death in instances where naloxone has been administed and transport has been refused.

    I haven’t been able to find any of his publications to link in this comment, however the presenter was Nathan Stam from Monash University who retrospectively analysed Victorian data of heroin-related deaths over one year (Jan 2012 – 13), finding that in 198 cases of herion-related deaths, only one had occurred in the instance of naloxone having been administered and transport refused. I can’t recall what else he said, but I expect he would have cautioned that even this one case could be attributed to any number of factors which might have occurred between the time transport was refused, and death (including subsequent heroin use).

    It’s certainly an interesting topic given it’s so heavily cautioned in the university sector (or at least it was for me) that the pt is likely to relapse into unconsciousness/apnoea due to the comparitive half-lives of the drugs

  • It all boils down to your local narcotic use. If you haven’t seen a lot of fentanyl in the local syringes, and all they’re using is black tar, then this is pretty safe, as shown by the studies (ignoring the biases). With any other coingestant, you’re really pushing the boundaries of safety by a long shot. Thankfully in my area of the country, we haven’t seen any of the bad stuff, and these people leave all the time. Often without registering, so I can’t even generate a MSE form for them.
    I’m not the police, and I can’t make someone stay, even if they might be under the influence of something. And while waiting for the police, if the patient elopes, I’m not responsible (thank you NY case law).
    This is a pretty good reason to use smaller doses of 0.04-02mg, so that you don’t turn them into a narcanimal, but instead simply breathing with a pulse. It’s tough to change EMS culture of complete reversal, especially for the blue patient with a GCS of 3, but this would make this a lot easier to keep them around.

  • Robert Simpson

    Great post Bryan. I’ve commented on the Melbourne experience of treat and release from an EMS viewpoint a number of times, but 140 characters is not conducive to explaining the situation…

    During the late 90s, Australia had a catastrophic glut of heroin on the streets. Stories abound of dealers selling “Starter Kits” outside high schools, and the death toll from heroin overdose far outstripped the road toll. One of our major newspapers published the daily heroin death toll on their front page every day.

    Most of the heroin trade (and therefore use) was conducted in a relatively small geographic area, centered on the CBD and inner North and Western suburbs (such as Footscray, where I work)

    As a result, ambulance crews in those areas were exposed to a ridiculous number of heroin overdoses every day, with each crew sometimes hitting double figures in a 10 hour shift. Due to this workload, guidelines were introduced that allowed the treatment with IM narcan in order to wake the patient up completely (as opposed to titrating to respiratory effort). I’m sure you can also imagine the impact it would have on the 3 or 4 hospitals in the areas if they received another dozen or so patients each, all requiring airway and ventilation management every day, on top of normal caseload (plus whatever walk-in/drop-off ODs presented as well)

    It should be noted that heroin in Australia is typically very pure, with most adulterants being mere bulking agents like lactose to increase yield for the seller. This is obviously not the case everywhere, with fentanyl and fentanyl analogues becoming an increasing issue in some parts of the world.

    So for the last 20+ years we have successfully treated and released uncomplicated heroin overdoses in the community, with excellent safety, and we continue to do so for the most part.

    So what actually do after identifying the heroin overdose is: ventilate well (+/- airway adjuncts), for as long as it takes for SpO2 to normalise, CO2 to normalise, and heart rate to normalise. Whilst doing so, we assess for any other issues (other drugs, injuries, prolonged downtime etc), and then administer 1.6-2mg of IM naloxone. The patient then wakes up, thanks us, we offer transport to hospital, and further narcan if this is refused. If they do refuse (and most do) we do our best to make sure they are safe and have someone to keep an eye on them, and offer them more narcan.

    Where I find (and I can only speak for myself here) a complicated overdose (polypharmacy or a prolonged downtime) or post heroin arrest, I will typically not administer naloxone, but instead provide supportive care (intubation, ventilation and so on) and transport for further assessment and management. I do this rather than have an agitated, vomiting, hypoxic brain injured patient thrashing around in the back of an ambulance after a dose of naloxone that could never be expected to work anyway. No-one dies of naloxopenia…

    We have a great deal of experience, and some upcoming data to demonstrate that in our context, this is safe and effective. As Levi mentions above, one of our University researchers has been doing some research matching coroner data with ambulance data (not just heroin specific either) and it appears that by and large we have been doing the right thing in Melbourne.

    And that is the final note: all of the above is specific only to our context in Melbourne where we have very pure heroin (pharmacologically speaking) and a history of catastrophic amounts of it being available. In the case of deciding to intubate, we have paramedics with a minimum of 6 years training and education (post-graduate), and a proven track record at intubation (97-100% success, with >90% first pass, and excellent procedural/patient safety) making these decisions and providing that level of care.

    Even so, with falling rates of heroin OD, there is discussion around altering our approach to include more transport with titration of naloxone to effect, as is common in other services. As Evan mentions, there are benefits to transporting to hospital, and if our system is better able to handle it (which it previously wasn’t) it is likely a good idea.

    The same may not hold true in all areas: as Justin points out, it depends on your local situation. I have no issue with different approaches being used, as EMS is not homogenous around the world. However I have typically heard nothing but disbelief and disgust from other medics at our approach, so I was very pleased to see your post!

    For anyone with an interest in the drug and alcohol scene in Australia, Turning Point ( is the foremost agency conducting research in this area. When the data is published on our safety with treat and release, I will post on, and let you know on Twitter.

    Thanks again!
    Rob (@AmboFOAM)