I'm a critical care and cardiac anesthesiologist at the University of Chicago, and I'm here to finish out this session by talking a little bit about walking the tightrope, and by tightrope, I think I mean, you know, how to get out of the hospital, you know, without a long-term opioid use disorder. Two disclosures before I start. I am the critical care section editor for ANA, the journal, and so, if you will, I am paid to be skeptical from, well, what I read, and you may see that skepticism come across. And two, I am a cardiothoracic and critical care anesthesiologist, so if I do have an experience base, it is primarily in acute postoperative pain. I've got to figure out a way to make this go forward. It's not. Okay. Maybe I should turn off the laser pointer. I can't forward a slide without it. All right. Three parts of this talk. I'm going to talk a little bit about we use, you know, I'm going to observe that we use opioids frequently in ICU and talk about why, you know, that we only recently have recognized the potential for post-discharge opioid use in patients who are opioid naive, previously something thought to be very unlikely, and then finish with a look at things we can do potentially to reduce post-discharge opioid use and how good are they. All right. The reason we use opioids in the ICU is because they work, and so I sat down, you know, sort of to create this show, and I made a list of all the reasons we use opioids, and that list started getting longer and longer and longer, and when you look at this list, you can see, you know, if you were talking to an alien from outer space, you'd tell them it's the perfect drug. Look, they suppress cough and gag reflexes, which is useful for inhibited patients. They don't uptend consciousness. They have multiple routes of administration, so if you lose the enteral route, you still have IV or even transdermal. They have synergistic sedative effects with everything else that you might use. They can ameliorate nearly all procedural pain, and if you do end up in a side effect, many of them can be reversed. In fact, you know, to drive that home, I'm just going to give you this clinical scenario. You have a 52-year-old man with COVID pneumonia. He gets intubated, you know, and despite Seroquel, Propofol, Dex, and Fusions, he has persistent coughing spells, and every time he does that, his O2 saturations dive. I've given you a choice of drugs that you can add. You know, I'm going to guess that most of us would add fentanyl here. A small minority might add Cistatricurum and just paralyze him so he doesn't cough, you know, but everybody agrees that you can add Olanzapine until you hit the dose limit, and you're not going to stop him from coughing, you know, with this situation, which is why, you know, in 2013, it was one of the easiest recommendations to make. These are the 2013 guidelines. This is me here, you know, and I can tell you we did argue about a number of things, but one thing we did not argue about was whether to recommend an algal sedation. So there's the question, there's the answer, you know, and the only reason it's a plus 2B and not a 1B is because there really wasn't a lot of data to see whether it worked, but it was definitely better than the Ativan drip that was currently in vogue at the time. Now, when you're young, you say, ooh, that's cool, I get to be on a guideline, but when you get older, you realize that, you know, that guideline carries some responsibility and that people sometimes actually read what you write, and thus it was that by 2016, there were a number of papers observing that people actually did use an algal sedation. That is to say, opioid-first sedation strategies, this one here in the middle, this one from the Canadians, and the one on the right, a survey of 50 United Kingdom ICUs, both arguing that opioid-based sedation was a primary element, you know, of sedation strategies in these ICUs. And that was 2016, and then came something that I have to say we never thought would be true, and in fact, when I was a resident, they told me this could not happen, and this is the observation in the perioperative sphere, that if you are admitted for an acute indication and you have acute pain and that pain goes away, then your opioid use is going to go away. Also, you will not end up addicted, but when we studied it closely, adjusted for factors such as preoperative opioid use and other drug use issues, you can see that there is a non-zero incidence of patients six months after their knee and after their hip arthroplasty who actually have, you know, are still requiring opioids. In fact, if you look at the ones who are on opioids preoperatively, the rate is shockingly high. Now, it's not just prospective studies, you know, in 500 patients. It was also in large databases, retrospectively reviewed. These two published in high-impact journals, both of them observing that there is a non-zero incidence of people previously opioid-naive who are taking opioids six or 12 months after their surgery. And you can see, you know, depending on the surgery, between 1 and 1.4 percent over here on the right, the incidence at 6 percent at 90 days, and a series of patient factors that, you know, it's not quite clear whether those still shake out of the literature, you know, but they suggest that there are predictors that can tell us, perhaps, you know, what might work. Now, you know, when you look in the ICU, that has already been reviewed, but I just want to observe that, you know, we are culpable in the ICU. This is the Donahue paper published in 2019 in the Annals of Internal Medicine, one of the most cited papers in the whole post-opioid discharge space, you know, looking at 12 hospitals, 150,000 patients, five years, and finding, you know, that 48 percent of patients got an opioid when they were hospitalized, you know, and that if they were had an ICU admission involved, and you can see those in the graphic here, that 60 percent of patients who had an ICU admission got their first opioid in the ICU. So if you were admitted to an ICU in this hospital system between those two years, you're very, very likely to get an opioid, and if, as observed before, if you do get an opioid, your chances, you know, of being opioid-dependent at six months was higher. Now, you think about the potential strategies for making that go away, you know, and so I've listed them, you know, just describing the space. Patient education is one. Tell the patient, you know, that it's not good to be on opioids, you know, in principle is a good idea, but what exactly do you tell them, and I'll get into that. So, you know, multi-modal, and, you know, that involves, you know, non-opioid adjuncts like ketamine, lidocaine, gabapentin. Use less opioid, and that's the goal of multi-modal strategies, because if you use less opioid in the hospital, maybe you'll use less opioid later. And then protocolized post-discharge opioid prescribing, so provider-focused interventions. And here, I'm going to put on my editor hat and say, be careful, you know, because if you have a plausible claim with retrospective evidence, and you really, really want it to be true, then the world gets sort of gray. You know, we editors say to ourselves, extraordinary claims require extraordinary evidence. If you want to say magnets cure cancer, you're welcome to, but your evidence better be very, very good, because the pretest probability is very low. But what about plausible claims that seem like they make sense, you know, and claims that we desperately want to be true? What about those? Should the standard of care be different for those? Here's an example. Here's a paper looking at opioid prescribing and hospital discharge and chronic opioid use one year later. You know, it's a descriptive study. They aim to characterize and describe. It was a retrospective cohort study, so, you know, a database that already existed. They found a very, very strong odds ratio four level association between opioid prescribing at discharge and opioid use one year later, you know. And the question is whether you're allowed to use the word contributes here, because contributes suggest the causal relationship, you know, and technically not true. But if it's something that, you know, is plausible, if it makes sense, if you really, really want it to be true, can you say that? Or, you know, as down here, they say opioid receipt increased for future chronic opioid use. That's also a causal word. It suggests a causal relationship. You know, can you really say that here? That's a good question. Moreover, when you're working with pain, it's sort of squishy, you know, and here are two nuggets from the anesthesia literature. You know, if you wouldn't, not an anesthesiologist, you probably wouldn't know that both of them asking the question, what do you say when you're about to stick somebody with a needle? Do you say, this is going to hurt like heck, or do you say, don't worry about it, this is nothing? You know, and in both of these, the one on the left published in 2007 about IV insertion, and the one on the right about putting in the epidural and, you know, giving the lidocaine real at the beginning, if you minimize it, they feel less pain. Sort of an interesting observation. So if you're going to counsel patients to questions, what do you say? You say, it's going to hurt, or do you say, it's not going to hurt? Now, patient counseling has a sort of enthusiastic fan base. I've extracted text from two reviews of managing post-discharge opioid use, and you can argue whether these are causal language or not, but they both refer to a orthopedic trauma study where, you know, they looked at patients who got counseling and patients who didn't get counseling and found that the ones who got counseling used less opioids at six weeks. And that's the top one here, you know, we're more likely to stop opioid use than the bottom one, quicker time to opioid cessation, you know, in patients who had undergone orthopedic trauma. Pull that paper, though, and you find it's retrospective. You find that it's actually one surgeon who gave the counseling and a second surgeon who didn't give the counseling. So there's an automatic confounder that there are two surgeons and there's no crossover between them. Finally, when you look at actually what they found, what you see is that at six weeks, yes, there was less opioid use, slightly, in the patients who got counseling, but that difference disappeared at 12 weeks and then afterwards. So, you know, is it a thing or is it not a thing? I leave that for you to discuss. Here's another one. This one is famously sort of described in all sort of suggestions that patient counseling potentially helps reduce post-opioid discharge use. You know, this is one in orthopedic surgery award, actually the effect of preoperative education on opioid consumption, including telling them that opioids were bad for them. You know, they said randomized to the study group was significantly associated with decreased consumption, you know, which is good, but then when you say read it further, they say it has some limitations. You know, we looked at self-reports, so those who were incentivized to report less opioid use probably did, you know, so there's a confound there that the odds ratio was actually very small and that at the end of the day, you know, when they looked at how much there was, it was very, very little. So now I can say, oh, come on, Avery, you're asking for too much. You know, this standard of truth is too high because when you think about it, it'd be very hard to do a prospective randomized trial of patient counseling because you've got to have one group that you don't tell them anything, and I don't even know if that's ethical in 2024. Add in crossover effects by, you know, caregivers, you know, who don't maybe know that, you know, the patient's in one side or the other, provider bias, dropout, loss of equipoise, and you find that you've got a study that even if you got it done, it wouldn't be convincing and if there was no effect, nobody would believe you because we all want it to be true. Patient counseling is plausible. Retrospective evidence does exist, you know, and little downsides apparent. So why don't we just do it anyway? Well, that's really where we were with inalgal sedation in 2013, which is why we're here in the first place. And let's say one day we find that, you know, withholding opioids leads to reluctance to take opioids leads to a limited functional recovery. Then what are we going to say to those patients, you know, who now can't move their arm because their shoulder hurts? You say, is that just the cost of doing business? So left unanswered is the standard of evidence for saying this works or it doesn't work. Now multimodal strategies, I think, have already been gone over in this space before. I'm just going to direct your attention to these three papers, the one on the left, shoulder blocks don't reduce the long-term use of opioids after shoulder surgery, one of the most painful surgeries we do. Over there on the right, ketamine for cytoreductive surgery doesn't help either. And the one in the middle, you might say there's a word reduces here and that's something, but on the other hand, that's the only paper in the entire lidocaine space that says anything about it. Now what about provider-level education? And with this, I'm going to finish. You really, really want it to be true. You know, just tell people to write for fewer doses. I'm a cardiac anesthesiologist, you know, this is a look at, you know, opioid use after cardiac surgery. And you can see there's a pretty much linear line between how many OMEs a patient gets prescribed, you know, on discharge and the likelihood six months later of being on opioids chronically. I actually took care of a patient for a stone or wound debris just last week and he said, I don't know, doc. He gave me a hundred oxys to take home with me. I'm like, really? He said, yeah, I took one. I felt dizzy. I got constipated. I didn't take any other ones. I said, good, good for you. And this study looks like you can ratchet down the immediate prescribing of postoperative opioids. You know, this is new persistent opioid use after postoperative intensive care. You can see that over a time period, there's a decrease in the instance of it, you know, from 2000 to 2016, you know, and this is a look at the effect of the Opioid Safety Initiative, which is a veteran's hospital initiative on opioid prescribing, and, you know, these are for very, very high dose, 100 OMEs per day level, you know, prescriptions, and you can see that that level went down over time. So there is some suggestive evidence that it might work, but when you study it, and this is about as good a study as they come, 11,000 patients in each group, you know, this is at the Mass General Hospital. They went, you know, system-wide, so every section of the Department of Surgery there, they said, you guys are prescribing too much opioids, stop, you know, and so, you know, with signs like this, they start to describe and change the culture. What they ended up was reducing opioids prescribed at discharge in nearly every single opioid section, in every single surgical section except for vascular surgery, and you know those vascular surgeons, you know, and you can see the percentage of patients that received any opioid referral prescriptions, unfortunately, did not change pre versus post. I'm just going to leave you with one more thing before I finish, you know, and that is, you know, the idea of OxyContin, you know, which, as Purdue Pharma famously said in 1999, is the one to start with, the one to stay with, you know, and, you know, I actually, because there's construction on the Kennedy Highway, you know, I end up listening to podcasts, and so I ended up listening to Peter Berg, who is the creator and director of the Netflix series Painkiller, you know, on his Joe Rogan podcast, which I do listen to, and, you know, he said, you know, I was curious, I took an Oxy to see what it felt like, he said it was fantastic, he said, I can totally see how people get hooked, it's like being dropped into a vat of warm honey. Jim Zachney here, the lead author of this paper on the left, you know, he worked in the University of Chicago for many years and spent some time describing the likability space of different opioids, and what he found was that OxyContin was more likable than other opioids you might use. This one in the middle observed, you know, in patients who are pre-existing users of opioids that they liked Oxy better than hydrocodone, and the one on the right also saw a choice of significantly more users than hydrocodone because the quality of the high was viewed to be much better, you know, by 54 percent of the sample compared to just 20 percent of hydrocodone users. So maybe we could use a drug that's not Oxy. To summarize, because I think we're at the end of the time, opioids are a mainstay of mental status management in the ICU because of all the different reasons I've listed, flexible, titratable, no ceiling to the analgesic effect, well understood, commonly used, everybody knows how to use them. Unfortunately, a consequence which we did not realize until it popped up in the mid-2010s is a post-discharge opioid dependence instance which, depending, you know, on which paper you read, ranges from 1 to 10 percent, you know, six months later, and about 40 percent of patients discharged from the ICU at discharge from the hospital end up on an opioid prescription. That rate belies the actual harm of post-opioid use disorder because diversion can extend the number of patients who are potentially affected. Therapeutic options are limited. The effect of patient education is unclear. Hospital strategies, likewise, things that decrease immediate opioid use don't seem to decrease long-term opioid use, so maybe the game is best played after the patient's already left the hospital. And that provider-focused strategies can reduce opioid use on discharge, that's been proven. The question is whether they have any effect on opioid use six months later because that's sort of an unexplored issue, as is whether likability ratings for OxyContin, which are off the charts compared to other opioids, may potentially play a role also. Thank you very much.

opioid dependence

pain management

ICU opioid use

multimodal strategies

OxyContin alternatives