Good morning, everybody. It's a pleasure to be here. I'm going to be talking about the epidemiology of opioid use disorder. I have no conflicts of interest or disclosures. First and foremost, we're going to talk about the epidemiology of opioid use disorder. To me, that means the incidents and try to characterize the patients that we see that have opioid use disorder or opioid-related disorders. We're going to use those terms somewhat interchangeably today. And specifically in the ICU setting, we're going to talk about the impact of COVID-19 on the epidemic. And finally, identify some opportunities to reduce harm in these patients. So first, let's make sure that we're talking about the same thing. DSM-5 criteria, opioid use disorder is defined as this by the American Psychiatry Association. You need two of the 11 criteria listed on the right-hand side, largely behavioral-related. Based on that scoring system, patient will be characterized as mild, moderate, or severe. And it's important to note that this is not validated in the ICU. For coding purposes, we often use ICD-10 codes, specifically the F11 and T40 codes. F11 largely focuses on opioid-related disorders, whereas T40 is acute poisoning and ingestion. And for those of you that are caring for these patients in the ICU, you're also going to be dealing with these other terms, both physiologic and behaviorally-related. So tolerance, which is certainly going to pose a challenge when you are trying to manage acute pain in the ICU setting. And also withdrawal can pose also complications related to acute care of these patients. So many of you have probably seen this CDC graph before, and it just illustrates the increasing incidence in fatality related to the opioid use disorder. Now interestingly, you see a plateau around 2017, 2018, 2019, and that was very encouraging for us to see. And then right around 2020 again, it started to pick up. The slope of the line increases substantially. You'll see mirroring the any opioid group is the synthetic opioids other than methadone. This is largely represented by fentanyl, specifically. So when you look at fatal overdoses related to medications, fentanyl accounts for nearly 70% of those cases. On the right-hand side, we have a heat map so you can get an idea of where you may lie geographically. West Virginia is at the top of that list. Now this is kind of an interesting slide to look at. So when you break down the fatality rate by specific drug, you see that, again, fentanyl is leading the charge. But a couple years delayed behind fentanyl, and albeit a smaller magnitude, you see a similar slope associated with methamphetamine and cocaine as well. So let's take a closer look at those. When you look at the fatalities associated with methamphetamine, you can break it down by the presence or absence of fentanyl with that methamphetamine. And you can see that methamphetamine laced with fentanyl seems to be strongly associated with and driving the fatality of those overdoses. The same case can be made for cocaine. You see very similar slope when fentanyl is tied with the cocaine. Moving on to COVID-19, we saw kind of a plateau, like I said, 2017, 2018, and even into 2019. In 2020, we saw about a 30-plus percent increase in fatalities associated with opioids. There's several variables associated with this. These particular authors identified reduced access to interventions, increased stress and isolation, lack of resources, anxiety, and there's also concurrently a changing epidemiology of drug use landscape. So you can see on the left-hand side, 5,000 to 6,000 cases. The last several years have quite a bit of overlap. There's been a slight modest increase, but there hasn't been a substantial increase. And then you get to 2021, and this graph depicts the first six months of the opioid overdose in the red line. You see a substantial increase. When you break it down at the state level, you see that West Virginia, Kentucky, and Tennessee had the largest gains from 2019 to 2020, largely related to the COVID epidemic. So what does this mean for outcomes? When you look at patients who have opioid use disorder versus those that do not, you find that those with opioid use disorder have increased likelihood of hospitalization, ICU admission, and mortality. And we think this is related to the general immunosuppressive effects of opioids, but also they have less pulmonary and cardiac reserve as a result of their opioid use. So when they do get an infection superimposed with that, they're less likely to bounce back. This achieves statistical significance for all three of those variables. And finally, looking at ICU admissions for overdoses, this is from the Vizient Clinical Database. I'm a big fan of the Vizient Clinical Database. You can pull a lot of really interesting diagnosis codes and correlate it with medication. And over 160 hospitals contributed to this data set. Now I should mention that the data may be a bit skewed in that the hospitals participating in this data set are largely urban and academic in nature, so it's underrepresentative of the overall population as a whole, but they identify a rate of 5.2 overdoses per 10,000 ICU admissions. That's a rate of 0.052%. So ballpark, that's probably about reasonable, about what we might expect. Now when we look at those patients more closely, most often they're female, they're white, variable age, we don't see a strong signal there, and most of them have some form of commercial insurance. When we look at the primary diagnosis that these patients present with, aspiration pneumonia, they're unable to protect their airway related to respiratory depression, rhabdomyolysis, patients are found down in awkward positions, cut off circulation, anoxic brain injury, again related to breathing, septic shock related to IV drug use, and a high percentage, 10% of these patients actually require mechanical ventilation. So again, looking at incidence, what we've seen over the years, and I'll be at this data a bit dated, only running through about 2015, what you find is that the overall incidence of ICU admissions has gone up. Paralleling that, the frequency at which we see opioid related ICU admissions has gone up. Now I would suggest to you that the scale on this graph is just a little bit misleading, in that the rates or the slope of the line related to opioid use disorder incidence actually goes up steeper than the overall ICU admissions. Now as far as outcomes go, the graph on the right hand side, what you're looking at is survival. And while the general ICU patient population has a relatively stable flatline survival, the incidence of mortality has gone up substantially in the opioid related group in the last several years. When you take a step back in your regression analysis and try to find signals in this big data set, again I think you identify that there's some variables that might be related to the bias of the data set in general, large urban academic medical centers. It's hard to pull out a whole, whole lot from that group. So this kind of prompted us to do our own analysis, and I suspect many of you have access to the Trinetics database, and that's all we did was try to query that database and get a feel for what the incidence is. It includes over 80 health organizations, 100 million different unique patients, and we queried 10,000 random patients from each of those organizations. What we found when we compared the OUD to non-OUD group is that they're mostly the same with the exception of age. The OUD group on average is about a decade younger. And obviously the incidence of diagnosis related to OUD. Otherwise, sex, race, ethnicity were very similar. So if you look at whether or not these patients have an ORD diagnosis and an acute poisoning diagnosis, so your F11 and T40 codes, we've increased over the last decade, and we're just about 1%. And that's probably a reasonable number. The Stevens paper that we showed earlier identified a rate of .052%. So we're at 1%, we're in the right ballpark. If you tease each of those variables out individually, you see a slightly higher incidence. Again, you see an increase over the last decade, but you're more in the neighborhood of 2% to 4%. And if you use the Lucis criteria, an either or criteria, it's as high as 6%. Now there are some limitations to this data set. These diagnoses need to be made within one month of an ICU admission, so it wasn't necessarily tied to the admission. But there are challenges in using that diagnostic criteria because the diagnoses aren't really validated in the ICU, so sometimes it'll happen after the fact. I think this is an important ballpark number to keep in mind because what you're really interested in is understanding the frequency at which you have an opioid-tolerant patient or somebody with opioid use disorder because their physiology is going to be a little bit different, and that's going to change the course of the care in the ICU. That's really what we're concerned about. So the characteristics of these patients. Obviously, we have hypertension, diabetes, asthma, heart disease that are pretty common between the groups. But when you look at the most common comorbidities tied to these patients that have opioid use disorder, you identify things like depression and nicotine dependence and cannabis use and other behavioral issues like that. There's quite a big difference between the top diagnoses when you stratify by OUD compared to when you stratify by a non-OUD patient where you see a lot of the more chronic comorbidities. Right around 2017 is when you as the ICU caregiver started to see more patients coming to the hospital on buprenorphine and methadone. For years and years and years, methadone was the workhorse. Right around 2017, we saw more buprenorphine use. This predates changes to the X waiver. We had mounting data to demonstrate that the drug was relatively efficacious, relatively safe, and had less prescribing criteria compared to methadone. Some of the reasons driving that prescribing. I mention this because in the acute care of patients on chronic buprenorphine, especially when using for opioid use disorder where the average dose might be 16 to 24, even 32 milligrams, right, those occupy quite a few new opioid receptors. And it makes the acute management of pain a little bit more complicated. That's in contrast to patients that are using buprenorphine for chronic pain where they're half a milligram, one, maybe two milligrams, and you don't see that same degree of occupancy. So looking at outcomes, readmission rates, about 470 patients looked at retrospectively to identify some of the outcomes. Their 30-day readmission rate was almost 20%. Their 90-day enrichment was over 30%. And those that came to the ED at 90 days but weren't actually admitted was 15%. I can tell you compared to the average patient, the non-OED patient, these numbers are a little bit higher. There is a little bit of a silver lining in that in regression, we identified that patients that were on or exposed to buprenorphine, that reduced their likelihood of being readmitted. So there's an opportunity there. Limitations to opioid-related disorder coding. There's a dyssynchrony between the DSM-4 and 5 criteria and the ICD-9 and 10 codes that are used. They don't map out well. Also when we do manual chart reviews and we try to look at that coding, there's also a dyssynchrony. So there's limitations to using the administrative and financial coding data and not using diagnosis. There's also opportunities for harm reduction. Regardless of the environment that you're looking at, there are opportunities to reduce the impact of opioid use disorder and what you see in the ICU. If you can get naloxone in the hands of patients regardless of the environment that they're in, it's going to reduce the likelihood of inadvertent overdose. Now reframing the question just a little bit, is ORD an iatrogenic condition following ICU? A little bit of a provocative question. And we've got five papers here that suggest that a pretty high proportion of non-surgical patients who come to the ICU who are opiate-naive and require mechanical ventilation actually get discharged on opioids. And when you take it a step further and you look at using the 24-hour prior to discharge prognosis of opioid utilizations to predict how much post-discharge opiate they're going to need, we see a lot of over-prescribing. Our organization is actually pretty similar to this. We're about 10% of these patients go home on opioids. I would say the answer to my question is that no, it's not an iatrogenic condition. Prescribed opioids certainly play a role in the opioid epidemic, but in this instance, I don't think it's driving it. What I would suggest to you is that this is an opportunity to make an impact on transitions of care, really scrutinizing whether or not those opioids are necessary or not. Regardless of environment, prescribed opioids have a harm associated with them. This is a large national registry done in the Netherlands and identified that upon multivariate regression, opioid prescribing is associated with unplanned ICU admissions and death as well. So we know that opioids are not benign. There's a few other studies that have found similar signals. Again, limited in that it's retrospective, but it is something for us to consider. So in conclusion, the epidemic continues to worsen, unfortunately, in part fueled by COVID. Accurate opioid use disorder rates in the ICU are really, really, really difficult to elucidate. There's a host of limitations largely related to coding. And finally, there are opportunities. So really getting patients exposed to opioid use disorder and naloxone, optimizing those transitions of care, and limiting unnecessary opioid exposure. I think some of our colleagues later in the talk are going to give you some of these details for us. With that, thank you.

opioid use disorder

ICU setting

fentanyl

harm reduction

naloxone distribution