The EMS Lighthouse Project

[00:00:00] I have a confession to make. Up until this month, I was pretty damn sure there was no role for naloxone in cardiac arrest. Now, here's how my logic went. Bear with me here, just for a second. Opioid induced cardiac arrest typically goes something like this. Opioids bind with the mu receptors, leading to respiratory depression and loss of airway reflexes.
That leads to profound hypoxia and hypercapnia. Now, the hypoxia leads to bradycardia, which turns into bradyacystolic cardiac arrest. The hypercapnia drops the pH and the resulting acidosis inhibits, well, basically the entire cellular mechanisms of life. Things we consider bad. Now, the treatment for this is appropriate oxygenation and ventilation.
At this point, treating cardiac arrest with naloxone is a bit too late and is kind of sort of pointless. Because, well, the [00:01:00] horse is pretty much well and truly out of the barn by that point. If the only benefit from naloxone is reversing respiratory depression, And, we're ventilating the patient, then there's no potential benefit.
That was my feeling last month. And then two papers, one in resuscitation out of Portland, Oregon, and the other in JAMA Network Open. Those two papers came out in, well, two weeks. To cut to the chase, both are showing improved survival with Naloxone. Now stick around and we'll get into the details. But in the meantime, if y'all don't mind, just think about this.
Drop me some comments. What do y'all recommend is the best way to make crow more palatable.
Howdy y'all, I'm Dr. Jeff Jarvis. Welcome back to the EMS Lighthouse Project Podcast, [00:02:00] where we shine the bright light of science in the darkness of clinical practice. And from time to time, we allow me to admit that I might have been more certain about medicine than perhaps I should have been. And just maybe; I needed to shine some of that bright light on my practice.
Now, before we get into the papers for this episode, I want to point out that I wrote this episode in Mackinac City, Michigan. I got to plug Michigan and a great conference. I had the honor of presenting some of our NMSQA measurement work and spreading the good work on reducing lights and sirens use at the Michigan EMS QI and data workshop
in Boyne Mountain Resort. I want to thank the great folks at the Michigan Center for Rural Health for bringing me up. This conference was part of a series of grants they issued to EMS agencies in rural Michigan to work on an improvement project, and these agencies presented their work at this conference, and they [00:03:00] all used NMSQA measures.
I was very impressed with the work they did and the excitement and the passion they brought to improvement work. It really was a joy after years of working with these measures with my NEMSQA colleagues to see how they are being used in the real world to make care better for the citizens in Michigan.
Thank you to everybody who presented.. It was great getting to hear what y'all are doing, getting to learn from y'all, and making some new friends.
This was also my first time to visit Michigan. I have no clue how I've managed to miss it, but I did. And I wanted to make up for my prior omission. Bad, bad Jeff, not getting to Michigan. So, to make up for it, Kristi, my wife, and I stayed the rest of the week and got to explore the Upper Peninsula. We spent the night in Ontario, and we learned all about the Great Lakes and shipping up there.
I know this sounds [00:04:00] nerdy, but we really did have fun. We took a boat tour of the Soo Locks, and we saw how really cool it was to float these massive ass ships. up, like, 21 feet using only the power of water going downhill. Wicked cool. We got to spend the day on Mackinac Island, something that's always been on my bucket list and the country up there, it's absolutely beautiful.
And the weather was perfect. Now they called it summer, but it kind of felt like fall to us. It was 72 degrees. Absolutely amazing. And , I do have to admit, I grew up in the Texas Gulf Coast. Our water, it's kind of notorious for being, what's the best way to say this? Brown. Not clear, brown. I was incredibly surprised at how clean and clear the waters were in those lakes.
Wicked cool. I really enjoyed it. Anyway, bottom line, thank you all for having me. I had a blast. I am [00:05:00] now back in Fort Worth. It was actually a relatively cool day today. It was only 99 degrees. It was 107 when I left. I love this city, but I am not exactly wild about our weather. With that out of the way, let's dive into these two papers that are kind of making me wonder, is Tabasco sufficient to make crow edible? And actually for that matter, I've never eaten crow. Have you eaten crow? Is it like everything else and tastes like chicken? Maybe I don't need the Tabasco after all. Anyway, this first paper, it's called the Association of Early Naloxone Use with Outcomes in Non Shockable Out of Hospital Cardiac Arrest. This was published in resuscitation by Dr. Nathan Strong. and several Portland area colleagues, including Dr. Mo Daya, John Jew, and my buddy Ritu Sahni. It became available online in June.
This was a retrospective cohort study using the Portland Area [00:06:00] Cardiac Arrest Epidemiologic Registry. It's a regional database they use. It automatically imports data from their EPCRs. And I'm going to see if I can say this because I always massacre this name, Multnomah County, Clackamas, Washington, and Clark County.
So this registry is, they get their EPCR data in there and they link it to receiving hospitals so they have outcome data. They analyzed the data in that data set from January 1, 2018 to December 31, 2021.
They included all non traumatic cardiac arrest in adults and they defined adults as 18 and over, who presented with initial non shockable rhythm. They excluded patients with DNRs, those who had EMS witnessed arrest, patients with ROSC prior to receiving naloxone, and finally, and this is an important one, bystander or PD administered [00:07:00] Narcan with CPR and ROSC prior to EMS arrival.
They were basically those patients that we don't know if they were actually in arrest or not. Now, I bring this last part up because I can already hear you people out there talking about arrests that were magically resuscitated by PD with nasal Narcan. I can feel your thoughts about sleeping patients rudely awakened with a nose full of Narcan by a caring and compassionate PD officer.
Y'all really are some judging judgers. I hope you know that. Well, they got this group and they broke it down into patients who were exposed and unexposed. So the exposed group got Narcan and specifically it got Narcan, those patients got Narcan before IV or IO access. And they called this early naloxone.
The control group didn't get early naloxone. Now they may have gotten naloxone after the IV and IO, but if they had that, it was in the [00:08:00] control group. So early naloxone versus either no naloxone or late naloxone
Now, their primary outcome was ROSC at ED arrival and secondary outcomes were survival to hospital discharge and functional neurological intact survival defined the way we normally do CPC score of one. Before we get to the results, we need to take a quick dip into methodology ocean, or as the case may be methodology, Great Lake.
The fundamental problem both of these papers are going to have to try to overcome to see if naloxone is helpful in cardiac arrest. Naloxone isn't epinephrine. We don't routinely give it in cardiac arrest like we do with epi. We just give epi to everybody. We only give it to those patients that we think can benefit from it.
So the treating clinician [00:09:00] has to think there's something different about this patient. from those patients and about the only way I know they might think about this is if they suspect an opioid overdose patients who go into cardiac arrest because of opioids Those are highly likely to be fundamentally different than those who arrest from non opioid reasons.
Now, all studies hoping to compare the impact of a treatment to the impact of no treatment, they're going to have to have some way of dealing with indication bias, meaning the patients either had or didn't have the indication. Now, as we know, the cleanest way of doing this is with a double blind randomized controlled trial.
The downside, though, are these are expensive. They're time consuming. They require community or patient consent. Patient consent, notoriously difficult to get from people in cardiac arrest. And then, you know, the bottom line is they're just all around really hard to [00:10:00] do. But man, do they give you a, or they can give you anyway, it is possible to do crappy RCTs, but if they are well done, they give you a nice clean answer and can show causation.
Anything else, any retrospective study has to rely on some statistical approaches for controlling for these baseline differences. Now the two most common are logistic regression and propensity matching. And God knows we've talked about those in this pod before and I'm not going to bore you with it now.
But both of those basically help you determine the association, not causation, association between a treatment and the outcome after adjusting for differences and confounding variables that are likely to be important. Now, even though these are common and they're well accepted approaches, the more math you have to throw at a problem, the more suspicious you should be [00:11:00] of the results.
Because both of these two papers threw a lot of math at the problem. And because of that, I'm going to skip over both the math and, for the most part, the numerical answers they found. Now, why is that? Well, fundamentally, it's because I don't believe the accuracy of the numbers, given all the math and all the methodology limitations.
Now, that doesn't mean I don't believe there's an effect, just that the exact strength of that effect definitely is in doubt. Now, for that matter, the presence of a difference at all here really still should be suspect. But having multiple different papers with different populations and different methods can help improve our faith that that association is real.
So with that massive caveat out of the way, I just want to say this first paper out of Portland used multivariable regression, adjusting [00:12:00] for age and gender and arrest location like public or private, witnessed status, bystander CPR, response times and what the initial rhythm was. Now, remember they only looked at non shockable rhythms.
So, the two choices for initial rhythm was PEA or asystole. And they also included, because there were, four counties with multiple agencies, they adjusted for agency. Now, they also had some spree, pre, spree? They had spree. They had pre specified. Man, I love those spree candies.
I haven't had those in a long time. Spree. I may have to go get some . So pre specified, pre specified, they had several pre specified, appropriate subgroups and they also had some post hoc sensitivity analyses. Basically, they sliced and diced this data in all of the ways they could to see, Does it make a difference in the outcome?
Now, these were all very good things to do because they [00:13:00] want to understand the effect in their data a little bit. I don't want to get into the details, but just say they did a really good job of trying. Overall, they found that the group who got naloxone, and this is important, it gets to this baseline difference thing, they found that the group that got naloxone was substantially younger, more public y, in other words, they arrested more in public, and had fewer arrests.
Now when I say substantially younger, I mean the group that got Narcan, 37 median age. The group who didn't, 61. And we know that all things being equal, The odds of a 37 year old surviving arrest are greater than a 61 year old. Now, I mentioned they had fewer witnessed arrests. That would typically tend to make us think that there would be lower survival in the Naloxone group that had fewer witnessed arrests.
So that would bias it away from the Naloxone. [00:14:00] But that age thing though, it really jumps out at me. With those baseline differences, we would likely expect the group getting Naloxone to have better outcomes, and that's what they found. Even after adjusting for these differences, they found a survival benefit with Naloxone.
Now, most results across the board were in the rough range of an adjusted odds ratio of anywhere between two and six. Now, if that's true, that's pretty impressive, but the if in that sentence is pretty important. I'm not ready to hang my hat on this. \ among those patients with suspected overdose, the adjusted odds ratio was in the thirties.
Odds ratios in the thirties, if real, are unheard of. But if there's anybody Narcan is going to help with, it's probably those that we're pretty sure have opioids on board. Still makes me doubt the whole mechanism thing, but it does make [00:15:00] sense that you would see more of an impact if there were actually an opioid on board.
Now the confidence interval around that 30 was huge, but it was significant. they found benefit across all of their subgroups and sensitivity analysis. P. E. A. Yes. Benefit. Asystole. Yes. Benefit. How about if, sorry to our law enforcement officer friends out there, but let's just say we don't trust them to give Narcan to the right people.
We're glad they're doing it, but maybe just maybe they're giving it when they don't need to. So if that's the case, if they were never in cardiac arrest to begin with, I'm guessing their survival is probably going to be higher than if they were. So to clear it up, they just got rid of any law enforcement- administered Narcan. Still a benefit. Even after doing that. They even looked at patients who are in rest in arrest for more than 10 minutes. Still a benefit. So this is a positive trial, [00:16:00] but man, does it have some limitations, primarily indication bias. So my takeaway is that this is a nice signal, but I'm not changing my practice on this paper.
In fact, after reading it, I wasn't even going to do a podcast on it, despite the fact that several of my friends were authors. It's just one regional data set with a lot of math thrown at it. You know, I kind of figured that if there's really something here It'll show up in additional papers. I didn't need to go crazy about second guessing my deeply held assumptions with only one paper.
Which brings me to paper number two that came out around the same time. I was saying, but it's only one paper little premature on my part. So this second paper, it's titled naloxone and patient outcomes in out of hospital cardiac arrest in California. Lead author is Dr. David Dillon and two of the [00:17:00] co authors are Drs.. Sarah Niederberger. and Jim Menegazzi. Now I bring that up because those last two were the ones who published a paper on early use of bicarb in cardiac arrest and they showed a benefit in unwitnessed non shockable rhythms when given early. That one definitely had me questioning myself and it even led me to do a deep dive on bicarb and ultimately it added, it led me to add bicarb back into our arrest protocol right after that initial dose of epi, only in unwitnessed arrest and those with non shockable rhythms. So at least two of the authors here kind of have a habit of making me reevaluate my practice. Now, the paper here is published in JAMA Network Open in August of 2024.
It also was a retrospective cohort review looking at adults. With non-traumatic cardiac arrest and broke 'em into groups based on whether they got Naloxone [00:18:00] or they didn't. this one uses a bit more data from 2015 to 2023, and they used the CARES Registry and we know the CARES Registry.
That's the Cardiac Arrest Registry to enhance survival. That's the group out of Emory that I think many of us use. They used submissions to cares from three California counties. San Francisco County, Sacramento County, and Yolo Counties. unlike the Portland study, they included all initial rhythms, not just PEA or asystole.
So they could have had shockable rhythms in here too. Their primary outcome was survival to hospital discharge and secondary outcomes included sustained ROSC. They also used both regression and propensity matching, controlling for age and sex and initial rhythm, comorbid conditions, witnessed status. And here's a cool one. Whether or not the arrest was felt to be drug [00:19:00] related. I already said I wasn't going to go deep on the methods, but let's just say I'm pretty sure this group used, roughly speaking here, all of the maths. Every last one. If there is a way to control for differences, I'm pretty sure they crammed it into this paper, willing or unwilling.
It was a complex methodology paper. I'm going to have to sit down and bribe Remle with some tacos to just discuss these methods, but let's just stick with all the maths. Remember, when you use all the maths, be a little suspicious. Now, just like with the first paper, there were some important baseline differences here.
Those getting Naloxone were again younger and healthier. Now, not shockingly here, Naloxone was I know y'all, hopefully you're sitting down here. Naloxone was more likely to be given if the medic felt like the arrest was opioid related. No shit. [00:20:00] Now, what did they find? again, there is a survival benefit to patients who receive naloxone.
The effect size here was smaller than the study in Portland. But again, I don't think those exact numbers are really important, just the direction both studies showed benefit. They both did the slicing and the dicing and still showed benefit regardless of the slice or the dice that they looked at. So what the hell do we make of this paper?
We now have two pretty well done studies, albeit with serious limitations, that suggest there is a benefit to giving naloxone to patients in cardiac arrest. Again, I was ready to ignore that single paper, but having two come out, it makes me at least sit up and take notice. it doesn't mean that I'm going to change my practice, but they definitely have my attention.
Now, before getting into how I'm ultimately thinking [00:21:00] about these papers and how I think we should or should not apply the results to our practice. Let's talk about two last things. Study limitations, and the potential mechanisms that Naloxone might have that are helpful, assuming the benefit these patients found is real.
So first, let's talk limitations. The biggest limitation and one I am just not convinced can be overcome with math, is indication bias. There are just a lot of things that can be improved with Naloxone. Two fundamentally different groups of patients here. Patients who were given naloxone were given it for a reason, even if the medic didn't document it what it was, or did document it, but did it in a way that didn't fit easily into a statistical model.
The age difference alone are huge. And I honestly don't think more math is going to solve this question convincingly. I think more papers with better methods are going to be required. And I'm not [00:22:00] saying this as a criticism of these papers. These are great papers. Really, the first time we've been looking at it, we need this.
To say, hey, maybe there's something here, dig more into it, with better methods, dare I say, perhaps an RCT. Obviously, the more papers, even the more non RCT papers we see saying the same thing, the more certain we can be. Now, I want y'all to note, I'm trying not to stoke up the great, "when must we have an RCT" debate here.
Because, Lord knows, those are painful and I often end up hurting afterwards. Now, another limitation here is that both of these populations were from similar parts of the country. They're on the West Coast. Both of those areas have significant opioid use problems that are leading to large increases in opioid deaths.
So results in those areas may not be generalizable [00:23:00] to parts of the world that are not suffering such large opioid problems. Now, let's go back to potential mechanisms. That's really the thing that's making me think so much about these papers. If the only way that opioids impact or lead you to cardiac arrest is through that ventilation thing and you're ventilating, I just don't see how there can be a benefit.
So there's got to be something going on. Now, the beauty of good research papers is that they, when they're good, have good discussion sections. that help us put the findings into context of previous literature, and they give us some framework around which to think of the results. Now, both of these papers recognize that, assuming medics were appropriately oxygenating and ventilating the patients, and I'm sure they were, that naloxone reversing respiratory depression isn't likely to be the mechanism responsible for these [00:24:00] benefits. They do suggest several alternative reasons that naloxone may work. Now, apparently there's even some work out there, saying some of these may be accurate. For example, there's some work that indicates opioids might lead to vasodilation and act as direct myocardial suppressants. Now we certainly know some opioids can cause hypotension, particularly morphine, so this is plausible. Even after ventilations have been assured, if we reversed this vasodilation and this direct myocardial suppressant. That might help achieve ROSC.
And in patients without suspected opioid related arrest, Naloxone might be blocking endogenous endorphins, and that could increase circulating catecholamines. If this were the case, we would expect to see a survival benefit with Narcan, even in patients without suspected arrest. And remember, that was one of the sensitivity [00:25:00] analyses that the California paper reported.
And they found it helped non- opioid related arrest also. Now, it is possible that that was based on suspected opioid, arrest and maybe we just aren't very good at figuring out what is or isn't. But still a signal there. So these results suggest benefit and they gave some possible alternative Explanations for how they might work.
So what am I gonna do with this information?, I'm certainly not gonna change my protocols to say just give naloxone to everybody hell It's more impactful than epinephrine not jumping on epi here I could say that, but I'm not gonna. I'm just not convinced yet that there's something real here. I do think that there is one thing I'm going to do differently though, and this is where that crow [00:26:00] eating comes in.
I've been a pretty vocal member of "Team Don't Give Narcan in Cardiac Arrest" for years now, I might have been pretty cocky about it if I'm being honest. Sometimes I think there's still too much of 20 year old paramedic Jeff in me, and I'll admit I can be cocky and maybe even pushy from time to time.
Just from time to time, though. Lord knows I have taught that there is no good reason, once you've started ventilating the patient, to give Narcan. And now these two papers have me changing my outlook. So I'm not ready to say it definitely helps. But I'm willing to allow that it is reasonable to think that it might and I'll even go so far as to say if you suspect an opioid ideology feel free to give it instead of like hunting you down and say, "you know that can't possibly help right?"
That would be pushy Jeff I'm gonna suppress pushy [00:27:00] Jeff on that for now and I'm most definitely going to acknowledge that We really have to admit we just don't know here at least until we get some better studies I often tell my residents and I tell my EMTs and paramedics that I get to work with that I think the most important characteristic a clinician can have is humility.
The humility that they, that I, might not have the right diagnosis, or I might not have the right treatment, or in this case, I might not understand the true mechanism by which naloxone might be helping cardiac arrest patients. So, with that in mind, I'm hoping perhaps a bit of maybe Tabasco Spiced Up Crow can give me a bit of a humility booster.
All right, that's what I got for you guys. Thank you for listening and as always, thank you for what you do for your communities day in and day out. Take care, [00:28:00] y'all.

[00:29:00] So when I said that's all I got for you now, take care y'all, like end of the podcast, I might have misspoken just a little bit. Certain things have come to light and I'm going to blame this on Dr. David Wampler. Dave, Damn you. We were almost on to the next stupid podcast we want to listen to on our play feed But no you you my friend had to write an invited commentary to the JAMA Network open article we just discussed I Swear guys everybody.
I'm sorry blame it on Dave But the commentary was so good. I really want to point this out What Dave talks about in this it's it's really short you have no excuse not to read it Open access. Citations are in the show notes. Please go take a look at it. It's less than 500 words. What Dave gets to is what are the potential mechanisms by which we're seeing this benefit [00:30:00] of naloxone in out of hospital cardiac arrest.
Perhaps, just perhaps, there is something other than mu receptor blocking. So the two papers we talked about hinted on this. Dave goes a little bit more into detail like the biochemist he is. So. Let me give you the citation. It is Naloxone and out of hospital cardiac arrest more than just opioid reversals.
And again, this is by David Wampler. David is the director of research at UT Health Science Center, San Antonio EMS program. He's a paramedic, PhD biochemist, great guy. Take a look at this. It was August 20th, 2024 in JAMA Network Open. Link in the show notes below. Now I mean it all done. Thanks a lot and take care.