The Donovan Memorial Lecture: ‘Fentanyl-Plus’: A New Era of Fentanyl Polydrug Combinations in The North American Overdose Crisis

Abstract

The 2025 ACMT Ward and Ryan Donovan Memorial Fund lecture was presented by Daniel Ciccarone, MD, MPH from Department of Family and Community Medicine at the University of California San Francisco (UCSF). This article is an edited version of his keynote address during ACMT’s 2025 Annual Scientific Meeting. During the course of his talk, Dr. Ciccarone discussed polysubstance use in the “Fentanyl-plus” era. The epidemiology and economics of the four waves of opioid overdose deaths were examined, and promising harm reduction strategies to assist in reducing overdose mortality were highlighted.

I was honored to deliver the 2025 Donovan lecture and am privileged to provide an edited transcript of it as a special contribution to the Journal of Medical Toxicology. In this, I explore polysubstance use in what I am calling the “Fentanyl-plus” era, a phenomenon that is occurring late in the four waves of mortality in the U.S.’s 20-plus-year opioid overdose crisis. In describing this era, I am talking about a broad range of phenomena: 1) intentional combination use, e.g., fentanyl and stimulants, including cocaine and methamphetamine; 2) unintentional polydrug use, i.e., the supply adulteration of fentanyl with new substances, e.g., xylazine and benzodiazepines; and 3) unpredictably mixed substances sold as counterfeit pharmaceutical pills.

For my research, I use quantitative techniques, including epidemiology and economics, to give me breadth, what I like to call the “view from 10,000 feet”. These answer the “how many?”, “where?” and “who?” questions. But if I want to answer “why?” and “how?” questions, epidemiology doesn’t help me that much. For that, I use street-based ethnography (i.e., the “view from a few feet”). That work is direct to the people, involving persons, I call them experts, with lived experience using drugs, to find out their perspectives. It gives context but also helps with causality, because epidemiology falls flat if it doesn’t meet criteria for plausibility, i.e., a logical social, behavioral, or biological mechanism that leads to the outcome of interest. So, I like combining the two disciplines [1–3]. In this essay, I will attempt to tie together the research data with insights from my clinical specialty, addiction medicine. And then, because the Donovan Lecture was part of a special session on harm reduction, we’ll talk about harm reduction strategies to assist in reducing overdose deaths.

The Four Waves of Opioid Overdose Deaths

Following the so-called Spanish influenza pandemic in 1919, we have benefited from a century of declining mortality in the US. Well, almost a century. Before the rise in deaths due to the COVID-19 pandemic, we saw a reversal of fortunes in which overall mortality rose; this was due to a specific cause: drug poisoning.¹

What has been driving overall US mortality up since 2000 are deaths due to specific subclasses of opioids. In looking at a graph of opioid-related mortality (Figure 1), we see these intersecting epidemic curves [4]. I have described these phenomena as the “triple wave” epidemic [5], followed by the “fourth wave” [6].

Figure 1. The four waves of the opioid overdose epidemic.

Credit: Reprinted from Friedman and Shover (2023) [4] under Creative Commons License.

In the triple wave, the first wave of mortality was due to prescription opioid pills. Wave 2 is the heroin epidemic beginning subtly around 2008 and picking up steam around 2012. And then wave 3 comes along, where heroin is adulterated with the synthetic opioid fentanyl, particularly on the East Coast, and we see a tremendous attack wave of deaths. Note: I’m using the terms synthetic opioids and fentanyl interchangeably. I’m talking about fentanyl, fentanyl analogs, and other non-fentanyl synthetic opioids, e.g., nitazines. Before I get to the fourth wave, at the center of this talk today, we need to understand how interconnected – indeed intertwined – these first three waves are.

What drove the first wave, the opioid pill wave, was excess prescription pills, which were, to a large degree, diverted to the street. The excesses in manufacturing, distributing, and marketing of prescription opioids have been well documented [7–9]. And we know diversion was a big part of this because street prices of prescription opioids plummeted [10, 11]. They became quite cheap. In Figure 2, we see an initial positive policy response where, as prescription numbers peak and start to come down, 2011–12, we see hospitalizations for opioid overdose going down as well. Our newly promulgated prescription guidelines, prescription drug monitoring programs, DEA arrests and seizures, and other pressures to reduce excess prescriptions and diversion to the street appeared to be working.

Figure 2.

a. Opioid prescriptions by year

b. Opioid overdose hospitalizations by year

Credit, 2a: Reprinted from Manchikanti, Kaye, et. al 2017 [12] under Creative Commons License.

Credit, 2b: Dr. Jay Unick, University of Maryland, unpublished data

Unfortunately, inadvertently, these policy initiatives fomented disaster. Wave two, and subsequent waves, were fueled by a huge influx of persons dependent on opioids seeking alternatives to prescription opioids as the latter became scarce. We had an early signal of this, published in a paper titled Intertwined Epidemics, from work led by my colleague Jay Unick [13]. He found a national bidirectional statistical relationship between prescription opioid- and heroin-related mortality; a relationship implying population co- or alternating use of prescription opioids and heroin during a crucial period, 2012–14. We also documented the transitions in use of these two subclasses of opioids, on a personal level, through qualitative research. This lent plausibility to the correlations seen in the epidemiological research. This work was published in Sarah Mars’ groundbreaking paper, Every Never I Ever Said Came True [14]. Our research showed that as street-diverted pills became scarce, folks moved on to heroin, which was more available and less expensive. Here’s a quote, published in the paper, from “Mark”, a 51-year-old who has been using heroin for 5 to 6 years; so, interestingly, he’s an older person using heroin for only a short period of time. He was originally prescribed oxycodone for a knee injury, and he says:

…like a lot of people, you start on the pills, and then the doctor gives you some and some more, and then you get cut off by the doctor … [so] every morning we would go to the [dealer’s house] and they had both things, but … they were never out of heroin … [but] three times a week … they didn’t have the pills. So, I’d have to scramble around, and then I finally had enough and said, fuck. To hell with this, give me a bag, and I was off to the races.

[Credit: Reprinted from Mars et al. (2014) [14], with permission from Elsevier.]

Figure 3 is a photo I took in an alleyway in Philadelphia during our fieldwork there in 2012. The transition from pills to heroin often also involved a change in route of administration. The prescription bottle is not for an opioid, but the photo tells the symbolic story, nevertheless.

Figure 3.

Credit: Original photography, Dan Ciccarone

In summary, the increase in heroin use and overdose in wave 2 was driven by an increase in the at-risk population, i.e., heavy opioid consumers. This population was created in wave one largely through street diversion of prescription opioids, but as these became scarce due to robust supply restrictions, heroin demand rose. In 2016, the size of this population was estimated at 2.3 million people [15], a 40% increase from 2006 (estimated 1.6 million). However, the numbers kept going up. Early in the fentanyl wave, the at-risk population was estimated at 3.7 M in 2018 [16]. In a recent commentary in the New England Journal of Medicine, I argue that in our failure to recognize the true size of the at-risk population, our policies and expenditures failed to match the size of the problem [17].

And then came fentanyl. Fentanyl is a synthetic, lipophilic phenylpiperidine and an agonist at the opioid μ receptor. And it’s very potent because it’s lipophilic and has a high affinity for the μ receptor. A key point to make about wave 3 fentanyl use and consequences is that the fentanyl phenomenon was not initiated because of demand [18]. This was a supply-side phenomenon, where heroin was adulterated with fentanyl, and the fentanyl was not diverted from a pharmaceutical supply; it was clandestinely produced [19]. And this foresees what we will see in the fentanyl plus era: the addition of powerful additional substances into the primary marketed drug. First, fentanyl adulterating heroin and later, e.g., xylazine adulterating fentanyl.

At first, people using heroin didn’t like the “new” heroin; fentanyl-adulterated or - substituted heroin (FASH) was being sold as “heroin” [20]. Because we were doing field-based ethnographic research on heroin at the time, we were able to document the transition to FASH as it was happening. I even predicted it to a degree, as the grant I was operating under at this time, Heroin in Transition (HIT), funded by NIH-NIDA, discussed several heroin variants, including FASH, at the proposal stage. In the beginning, people were freaked out by the “new” heroin. Some people liked it, some people didn’t like it. So, for example, Tim, who’s been using heroin for 25 years, was interviewed in Baltimore, Maryland. He says:

The high is wonderful. It’s splendidly wonderful. It’s magnified heroin feeling by a great number.

[Credit: Reprinted from Mars, Ondocsin, and Ciccarone (2018) [20], with permission of Informa UK Limited, trading Taylor & Francis Group.]

Note, he’s got a high tolerance to heroin, so the fentanyl gave him a new feeling, a new opioid sensation. But Rebecca, who’s new to the scene, in her 20s and has only been injecting for 2 years, living in Lowell, Massachusetts, she said:

I don’t like the feeling of it. I don’t like the high of it. I don’t like nothing about fentanyl.

[Credit: Reprinted from Ciccarone, Ondocsin, and Mars (2017) [21], with permission from Elsevier.]

And there were many people in her camp. Why were they freaked out? Well, part of it was the physiological feeling they get from it, but part of it is also the change in appearance (Figure 4a). The photo on the left shows what heroin in solution traditionally looks like. Heroin HCL solution is similar in color to iced tea or coffee; this appearance is so iconic that there were songs written about it, for example, the Stranglers’ Golden Brown song. Fentanyl salts, e.g., HCL or citrate, in solution look very different: sometimes brown but also clear, other times neon yellow (see Figure 4b), or milky-white. Folks could discern the fentanyl component through taste and appearance [21] and, in time, have adapted to it in interesting ways, e.g., “tooting” it first (i.e., nasal insufflation) to test the drug’s strength before committing to a full dose or injecting it [22]. And people remained respectful of it, fearful of it, having witnessed their peers overdosing and dying from it [21].

Figure 4.

a. Traditional “golden brown” color of heroin HCL solution

b. One of the new colors of fentanyl salt in solution

Credit, 4a: Original photography, Dan Ciccarone

Credit 4b: Original photography, Dan Ciccarone. Reprinted from Ciccarone, Ondocsin, and Mars (2017) [21], with permission from Elsevier.

That brings us to the fourth wave of opioid-related overdoses. In addition to polysubstance use, the focus of the next section, the fourth wave means something else: expansion to the western states, exposing fentanyl to new populations. This expanded the size of the population at risk and fueled the continuing rise in overdose deaths nationwide. California, for example, was spared the fentanyl problem in large part until 2018–2019, but it is catching up (Figure 5).

Figure 5. Any Opioid-Related Overdose Deaths compared to Synthetic Opioid-Related Overdose Deaths, California, Total Population - 12-Month Rolling (Crude Rates).

Credit: Figure creating using data from California Overdose Surveillance Dashboard. Prepared by California Department of Public Health (CDPH - Substance and Addiction Prevention Branch (SAPB). Accessed on 09/10/25. https://skylab.cdph.ca.gov/ODdash/.

The fentanyl scourge has expanded from a regional phenomenon in 2018 (Figure 6)—the Midwest and Appalachia, the Mid-Atlantic, and New England (with some involvement in CA and AZ). By 2021, we see it expanding further in California, as well as to Texas, and the Southwest.

Figure 6. Opioid overdose deaths by type of opioid, 2018 (top) and 2021 (bottom).

Credit: Kaiser Family Foundation analysis of Centers for Disease Control and Prevention, National Center for Health Statistics. Underlying Cause of Death 1999–2023 on CDC WONDER Online Database, released 2024. Data are from the Multiple Cause of Death Files, 1999–2023, as compiled from data provided by the 57 vital statistics jurisdictions through the Vital Statistics Cooperative Program, accessed through the KFF Dashboard, August 14, 2025.

What the fourth wave also means is the rise in the intentional co-use of illicit stimulants with fentanyl. Over the past 20 years, deaths attributed to cocaine have tripled, and those attributable to psychostimulants, mostly methamphetamine, have increased by a factor of 19 [23]. And what is driving these deaths in large part is co-use with fentanyl [24]. This finding was the impetus for our current study—Synthetics in Combination, aka SYNC, study (NIH-NIDA)—to look at what’s driving these increases in polysubstance deaths.

Fentanyl plus era:

We recognize that polysubstance use is common. On a commonplace level, coffee and chocolate are consumed together due to complementary taste profiles. Some people who say they don’t smoke will have a cigarette when they’re having a cocktail on a night out. On the illicit side of this, polysubstance use has always been common, but we must recognize that in the fentanyl plus era, we have historic levels of potency of the components, fentanyl, along with the additional drug, thus raising the risk environment. Methamphetamine, for example, currently has exceedingly high levels of purity and potency [25].

In thinking about this era, I want to make a distinction between intentional co-use, e.g., fentanyl plus-methamphetamine, and unintentional co-use, e.g., fentanyl plus xylazine. In the latter category, we also have counterfeit pills, which are often fentanyl plus the kitchen sink.

Intentional Co-use:

The primary example of international co-use is fentanyl plus a stimulant. The “speedball”, for example, is the historically popular combination of heroin and cocaine. It’s a combination that makes sense, as the cocaine acts as a booster for heroin in people who have developed a high degree of tolerance to heroin. The so-called “goofball”, the combination of heroin and methamphetamine, was unusual because it is physiologically challenging. People felt like the methamphetamine would literally overpower the effects of any co-used drug. So, the fact that it was happening in the age of FASH required exploration.

A recent paper, The Motives and Methods of Methamphetamine and ‘Heroin’ Co-Use in West Virginia, led by my colleagues Jeff Ondocsin and Nicole Holm, explores the rationales for this co-use [26]. (Note: ‘heroin’ refers to FASH.) Let’s start with a quote from “Mary”, who’s in her 30s, interviewed in Charleston, WV, and has been injecting for one and a half years, who prefers methamphetamine to ‘heroin’, but also likes to inject both together. She says:

M: But I mean, I prefer meth. […] I like heroin and stuff like that, but it’s not my choice preference. But I do it every now and then. And actually mixing it with meth is the better buzz, believe it or not.

Interviewer: How do you decide on a given day?

M: Just what we feel like. I mean it’s like, well how sick are we for heroin? How bad is the dope sickness and if the dope sickness is not bad we’ll choose meth because then you can fend off the dope sickness even more by being high on meth and you won’t feel it.

In this setting, the combination is not called a “goofball;” they call it a “meth speedball” to differentiate it from a “cocaine speedball.” It’s resurgently popular since about 2015. What makes this different from the heroin era is that the fentanyl is potent enough to meet up mano-a-mano with the methamphetamine. Some people say they like it, while others complain of the physiological conundrum of going up and down at the same time.

One of the main points of the paper is that methamphetamine co-use can be seen as a behavioral adaptation to fentanyl. People have a range of rationales and strategies as to how they use it. Some people don’t care for it but will use it if it’s free. Some people will use it casually, especially in certain social situations. Others use both regularly but split up, e.g., meth in the morning before work and fentanyl in the afternoon to relax. Some are doing both drugs concurrently. And one of the key takeaways from our research is that people are using methamphetamine to reduce fentanyl use, e.g., to skip a dose, or use smaller amounts. Many who are using fentanyl are respectful, even scared of the drug, and they are trying to reduce the risk from it. And we have epidemiological data to support these qualitative findings on adaptation. Using supply data from Ohio, teammate Dan Rosenblum found that as methamphetamine supply expanded at the county-level, overdose from synthetic opioids declined modestly [27].

Unintentional Co-use:

The premier example of unintentional co-use goes back to the original distribution (circa 2014) of illicitly manufactured fentanyl. FASH, fentanyl-adulterated and -substituted heroin, was sold as ‘heroin’. This was such a compelling notion that we published a paper on it, Sold as heroin: Perceptions and Use of an Evolving Drug in Baltimore, MD [20]. Often, FASH had some heroin in it, and the components varied on a daily or weekly basis.

Unintentional co-use in the fentanyl plus era includes both accidental contamination as well as deliberate adulteration. It is challenging to tease these apart as the availability of supply-side data, including e.g., purities of each component, is limited. One supply database that our team has access to (through multiple FOIA requests) is from the Ohio’s Bureau of Criminal Investigation (BCI) crime lab [27, 28]; this work was led by Dan Rosenblum. This is drug seizure data, with testing for all major drug categories. The sample characteristics and our methods of analysis have been published [28]. In unpublished data from 2020, of all cocaine samples, 11% had co-occurring fentanyl in them (Figure 7). For methamphetamine samples, 3.5% had fentanyl in them (Figure 8). This database doesn’t have purity, so we cannot tell if the co-occurrences are accidental or purposeful. We can state that in both cases, contamination/adulteration is growing over time.

Figure 7. Fraction of cocaine samples containing an opioid, Ohio BCI Crime Lab Data, 2020.

Credit: Dan Rosenblum, Dalhousie University, unpublished data

Figure 8. Fraction of methamphetamine samples containing an opioid, Ohio BCI Crime Lab Data, 2020.

Credit: Dan Rosenblum, Dalhousie University, unpublished data

“Tranq”, the combination of fentanyl plus xylazine, is a good example of the deliberate adulteration of fentanyl with another drug. Since fentanyl consumers don’t add xylazine and are only exposed to it through their supply, this would be considered unintentional co-use. Xylazine is not a drug approved for human use. As an alpha-2 agonist (A2A), it’s analogous to the human drug clonidine, and it’s a CNS depressant. A2A can potentiate opioids and aid in opioid withdrawal, and that might give us some clue as to why it’s being added to the fentanyl supply. It is curious that by 2019, when Tranq first appeared, 5 years into the fentanyl phenomenon in the eastern US, this opioid potentiator was being added. And it is being added to fentanyl in the regions first impacted by fentanyl [29]. This is an important aspect of the “Fentanyl Plus” era. Perhaps we can apply a clinical metaphor. It’s population level, not individual, but xylazine is hitting the places where people are most likely to be highly tolerant to fentanyl if they have been regular consumers over the past few years. More recently, xylazine is spreading to the western US and, in addition, is being replaced by medetomidine in its original places [30].

Xylazine-related deaths are going up; however, most of them, 97–99% of them, are co-occurring with fentanyl [29]. It is unclear what the contribution of the xylazine component is to these deaths. The injection of Tranq appears to cause severe skin and soft tissue wounds; these don’t appear to be infections but rather injuries. These wounds can be skin deep, going down fully through the soft tissue, and sometimes all the way down to bone, requiring extensive care, e.g., street-based nursing, emergency room, or hospitalization response [31–33]. Xylazine and other A2A are kappa agonists, a subclass of opioids known to cause wounds [34]. In a recent paper entitled Tranq Burn, we explored the perceptions of folks using tranq [35]. They described injections as painful, leading to rapid vein loss, and the wounds as “burns.” In this mixed methods research, we also looked at the pH of street samples of tranq and found a range of low pH: 3.6–5.9, with a suggestion that the fentanyl component was of higher acidity than the xylazine component. Since A2A are known to cause vasoconstriction, this paper ends with a hypothesis that the combination of injection of acidic drugs, leading to venous sclerosis and alternative injections (e.g., peri-venous or central vein), and the vasoconstriction effect leads to necrotic injuries of the skin and soft tissue.

Moving on to pills containing fentanyl. The is no legitimate pharmaceutical fentanyl pill, so when we find a fentanyl pill, any pill containing fentanyl, we know that it’s counterfeit. Some of these counterfeits appear to be high quality; visually, they’re strong mimics. In collaboration with Joseph Palamar and Linda Cotler, we explored the supply of these pills using the High-Intensity Drug Trafficking Areas (HIDTA) Performance Management Process (PMP) database. In the first paper, we found that the number of pills seized increased from 42,202 in 2018 Q1 to 2,089,186 in 2021 Q4, and the proportion of all fentanyl seizures, pills, and powder form, which were pills, doubled in the same timeframe [36]. In a recent paper, we found that by 2023, an outstanding 120 million fentanyl pills were seized, and the proportion of fentanyl seizures in pill form approached 50% [37]. The thing we need to understand about drug seizures is that they just reflect the supply. Some experts suggest a 1:10 seizure-to-supply ratio. That means we are talking about a billion fentanyl pills in the U.S. opioid street supply in 2023. And these pills contain multiple substances from diverse chemical classes, including opioids, benzodiazepines, and other sedatives [38, 39]. Who’s using them? This needs further investigation, but there are strong hints that those using them are adolescents and young adults [40, 41].

Let’s move on to another example of unintentional co-use: fentanyl plus benzodiazepines. Using the Ohio Crime Lab database again to ask the question: of all fentanyl samples now, what proportion has benzodiazepine in it? The answer is around 8% and growing quickly (Figure 9) (unpublished data). In the Canadian setting, Sam Tobias and his group, looking at a very large drug checking database with 10,000 samples, found 55% of opioids have benzodiazepines in the mix [42]. In survey results from participants bringing in a sample, they found a 105% increased prevalence of self-reported overdose among those bringing in fentanyl plus benzo samples versus the fentanyl-only samples. This is a kind of double trouble, as you have a double-downer combination. Increased overdose risk shouldn’t surprise us, because we know that heroin plus benzo, heroin plus alcohol, are riskier combinations than heroin alone. So, currently, we have two common double-downer unintentional use combinations in North America: fentanyl plus xylazine and fentanyl plus benzodiazepine. On the flip side, there could be a substitution effect, just like we see with methamphetamine. If, for example, the use of xylazine leads to less fentanyl use, perhaps because it is longer acting, overdose risk might go down at the population level. But if that double-downer effect is stronger, then we should see overdoses going up. More research is needed on this.

Figure 9. Fraction of benzodiazepine samples containing additional substances by year, Ohio BCI Crime Lab Data.

Credit: Dan Rosenblum, Dalhousie University, unpublished data

Insights from Addiction Medicine:

The question is “why now”? Why the rise in co-use (intentional or not) with fentanyl as the main component? The timing of these events is interesting and can be interpreted in light of “late stage” events in an opioid cycle. Illicit drug use can occur in epidemic cycles, e.g., consider heroin in the US in the 70s [43] or ‘crack’ cocaine in the 90s [44]. There is no agreed-upon theory for why these cycles begin (e.g., perhaps a robust supply of an interesting new drug) or end (e.g., perhaps initiates are turned off by the negative aspects of the not-so-new drug). The four-wave opioid phenomenon is impressive in its durability – 25 years thus far. Addiction medicine can provide some insights into the late-stage polysubstance use we are witnessing: regular exposure to fentanyl will lead to rising dependency and increasing tolerance in individuals. On the demand side, this might lead to greater consumption and searches for increased purity of fentanyl, as well as drugs to boost the opioid effect. Some combinations include drugs with synergistic effects, boosting the opioid component. Cocaine is a potentiator of opioids and has been used in “speedball” combinations in earlier heroin use waves, e.g., 1970s, 1990s, especially. in the late stages of those cycles. Methamphetamine, historically called a “goofball,” was rare during heroin waves, as it perhaps overpowered the heroin effect, but it reportedly is a great booster of higher-potency fentanyl. Persons we interviewed reported a more intense rush from the combination [26]. And they also reported methamphetamine and fentanyl co-use as an adaptation: to reduce the risk of fentanyl’s deadliness.

This response to increasing tolerance is also happening collectively, at the population level. Suppliers have adjusted the strength of their fentanyl mixtures over time. According to recent US Drug Enforcement Administration data, the purity of fentanyl in drug seizures has risen from ~5% in late 2019 to ~25% in mid-2023 [25]. And methamphetamine supply has expanded outside of its historical zone in the West [45, 46]. More perplexing are the “double-downer” combinations, clearly supply-side impositions, leading to unintentional co-use. Why add e.g., A2A? Perhaps there’s some potentiation going on. Xylazine and other A2A are probably opioid potentiators. Xylazine has powerful sedative effect, so there is also a synergistic sedative effect in combination with fentanyl. Benzodiazepines are also sedating. Most are illicit benzos, e.g., bromazolam, whose potency and effects are largely unknown. There is as yet no clear answer as to why the double-downer combinations with fentanyl. Supply-side motivation, including adulteration or substitution, to make more profit, doesn’t make a lot of sense because fentanyl is quite inexpensive. More likely, the answer is in the consumer experience at a time of historically unheard-of levels of opioid tolerance.

Harm Reduction in the Fentanyl Plus Era

In summary, the “Fentanyl Plus” era is a late-stage event in a complex, intertwined, multi-wave overdose epidemic. It continues to evolve with new drugs and combinations. Polysubstance use is the new norm, and because of the levels of potency of the drugs involved, the consequences of use are rising. And the intentionality of combination use is important to keep in mind in our continuing research.

How do we address this? First thing, don’t panic. Every time there is a new drug, we create a sort of moral panic [47]. But stigma remains our biggest enemy. It interferes with everything we want to do, whether it’s on the prevention side, the treatment side, the recovery side. Even the public nuisance side. Some people have been promoting a renewed stigmatization of drugs [48], but I strongly disagree with them. While there may be some benefit at the primary prevention level (don’t start using), stigma, and its internalization, shame, are great barriers to secondary and tertiary prevention (preventing harms and death); folks simply hide and avoid services, blaming themselves for their predicament. Harm reduction takes the opposite approach, one of radical love and acceptance. “Come as you are.” No shame.

What does harm reduction mean? The basic notion is: we treat people where they’re at. Not where we want them to be. Where they’re at, today, this week. And it’s not much different from the norm of modern medicine, which we call patient-centered care. We used to tell people what to do based on the “doctor is always right” idea. “You did this to yourself; you need to stop smoking, lose weight”, etc. Blaming and shaming don’t work to maximize behavior change. So, for treating people with heart disease or diabetes, we don’t blame and shame; we treat people where they’re at. Behavior modification to the degree possible. And a team-based approach with support to the degree that is needed for that person. And medication to the degree that it lowers risk, e.g., using a statin to reduce the risk of heart disease. Medication is harm reduction! An absolutist approach doesn’t work. We don’t tell the person who’s had a heart attack and eats a lot of meat that they need to go vegan. We treat the symptoms, treat the person, and help them prevent the next one with medication and behavior change. And it’s incremental, so some people can make large changes in their lifestyle while others can only make small ones. So, particularly in something as challenging as addiction, we look for small changes. And we aim for engagement and retention over time.

So, what is harm reduction in the Fentanyl Plus era? Medication is harm reduction. This was controversial in the harm reduction community years ago, as there was some distrust between this movement and the medical establishment. This is, thankfully, no longer true. Medication for opioid use disorder, e.g., buprenorphine, methadone, and naltrexone-xl, leads to improvements, e.g., stabilization, retention in treatment, improved outcomes, and reduced risky behavior [49]. Illicit opioid use goes down due to reduced craving. It reduces all-cause mortality, not just from overdose, reduces bloodborne virus transmission, and reduces criminal activity. Medication is harm reduction.

Naloxone peer distribution is harm reduction. Two good studies show this. In Massachusetts, they randomized towns to get a take-home naloxone distribution program. Those towns had an almost 30% reduction in overdoses compared to the control towns [50]. In Scotland, a national naloxone distribution program was studied, pre-post intervention, and they also found a 30-something percent reduction in overdose deaths [51]. Other studies have demonstrated its cost-effectiveness [52, 53].

Is smoking fentanyl harm reduction? Fentanyl smoking is an innovation in places where fentanyl is being sold as is, i.e., the West. Smoking fentanyl is a big deal, not just in terms of a growing phenomenon, but also in terms of what it means for harm reduction, e.g., reducing HIV/HCV transmission by avoiding injecting, and possibly decreasing overdose risk [54]. Smoking opioids is growing in prevalence nationwide, so it is a phenomenon worth following [55].

Fentanyl salts, e.g., HCL and citrate, have high pyrolysis points, i.e., they are very heat-tolerant, above their vaporization points, which allows for good bioavailability with smoking/vaporization. (Note: Heroin is typically only smoked in its base form [56]; heroin HCL salt burns too easily.) Smoking is typically done using aluminum foil (Figure 10a) or a glass ‘bubble’. A pipe or straw (in the figure in red) is used to concentrate the smoke and bring it into the lungs. This technique allows smokers to consume fentanyl in multiple small doses. Conversely, another way to smoke fentanyl is to use a silicon bong with a titanium ‘nail’ (bowl) and a very high temperature torch to flash vaporize a chunk—that chunk in Figure 10b is about half a gram of street fentanyl—and then take a big lungful of it. This is trying to replicate the intravenous bolus effect of a large dose rapidly hitting the bloodstream [54].

Figure 10.

a. Process of smoking fentanyl on foil

b. Fentanyl smoking using a bong apparatus

Credit: Original photography, Dan Ciccarone. Reprinted from Ciccarone et al. (2024) [54], under Creative Commons License.

So, what are the implications? These innovations are hard to suppress. We tried for a long time in the heroin era to get people to smoke instead of inject, so as to reduce HIV risk. It didn’t really take off in the US. But this fentanyl smoking movement is happening organically and spreading by word of mouth. We expect that HIV, Hepatitis C, and injection-related bacterial infections will all go down as this innovation catches on.

The impact of smoking on overdose risk is more complicated. There are two recent studies, the earlier one showing approximately a 30% reduction in self-reported overdose risk [57]. And in a national analysis that just came out, my team found an approximate 30% reduction in fatal overdoses based on smoking versus injecting route of opioid administration [58]. The key factor, I believe, is the reduced bolus, i.e., the speed and concentration at which the opioid hits the bloodstream, from smoking. It should reduce overdose, because you’re taking smaller bits of the drug into your body, and at lower concentrations, so you’re getting less across the blood-brain barrier per unit of time.

However, there are counterpoints to be made. The use of a bong with flash vaporization tries to replicate the IV bolus effect of a rapid and strong rush. In doing this, the risk of overdose remains high [54]. Fortunately, the cost of the bong setup and native understanding of its risk limit its popularity. There is another overdose risk; this is one created by the smoking technique. Following the smoking process, a residue is left on the heated aluminum or glass surfaces (Figure 11a). The residue looks burnt, which comes from the caramelization of the sugars used as fillers. But the residue is reusable, retaining some active fentanyl according to folks we interviewed. It is reusable simply by reheating the pipe or foil. Some people will scrape it out, save it, and use it later, even reporting that the accumulated product is stronger, better [54]. There is a lot we don’t know about this residue: its components, whether untouched fentanyl salt or a pyrolysis product, or their potencies. An obvious concern is accidental exposure to folks who don’t have fentanyl tolerance, e.g., consumers of methamphetamine will use the same glass bulbs to smoke and might accidentally be exposed to fentanyl and potentially overdose. Alexander Bazazi in San Francisco is following a case series where methamphetamine users were accidentally exposed to fentanyl and overdosed [59]. Following this, we need strategies to reduce pipe sharing. One strategy is to label the pipes (Figure 11b); the rubber bands could symbolize, “This pipe’s my pipe, it’s not your pipe,” or “This is a fentanyl pipe, not a meth pipe.” This strategy hasn’t taken off yet. Bottom line, I want to think that smoking is harm reduction, but we need more data and more debate about this. If you want to read more about smoking fentanyl, check out our paper, Innovation and adaptation: The rise of a fentanyl smoking culture in San Francisco [54].

Figure 11.

a. Glass pipe with fentanyl residue

b. Rubber bands used to label pipes, distributed by a harm reduction program

Credit: Original photography, Dan Ciccarone

Drug checking is harm reduction. Canadian programs have been inspirational leaders in this [60–62]. Leadership in terms of promoting it, understanding the complexities of it, and promoting technology transfer, e.g., teaching others. US harm reduction programs are also making advances in drug checking [63–67]. It includes intimate point-of-use drug checking, e.g., immunoassay test strips, which have high sensitivity and modest specificity, so there is potential for false positives [68]. Drug checking also includes venue-based testing, which utilizes Fourier-transformed infrared spectrometry (FTIR) machines, which have better specificity but low sensitivity. So many of the venue programs do both in combination.

The biggest theoretical—and I think there is some evidence base for this now—impact is going to be for people who want to avoid fentanyl [69]. For example, cocaine users, who have an 11% chance of being exposed to fentanyl in Ohio, may want to know if their preferred drug is contaminated. And counterfeit pill consumers, our young people, who are smoking pills, need to know whether there is fentanyl in that pill.

There are also behavioral strategies that count as harm reduction. These are organic strategies created on the street and promoted by word-of-mouth as adaptations to the risks inherent in using fentanyl. These strategies, which we documented in a paper called Toots, Tastes, and Tester Shots, include: using smaller, more dilute, but more frequent intravenous shots; trying a little bit inhaled through the nose first (tooting) to see how it feels, and then deciding whether to consume more; doing a tester shot to see how it feels in your body first before doing more [22]. Bmore POWER, a wonderful harm reduction group in Baltimore, created a campaign: “Fentanyl acts fast. Go slow.” [70].

Conclusion:

“Fentanyl-plus” is an era of polysubstance use marked by high levels of potency of the constituent drugs. Its raison d’être can be best understood with insights from addiction medicine, especially rising collective tolerance to potent fentanyls and supply-led adjustments. Supplies for each plus drug are rising, some as part of adulteration strategies and some independently.

It is imperative that we do not treat each drug wave as a ‘moral panic’ [47]. The multi-wave opioid overdose phenomenon should be considered as a whole: intertwined and overlapping phenomena leading to high levels of mortality and morbidity.

There has been a policy failure to understand the size of the at-risk population [17]. This is a prime principle of public health, especially in estimating rates, but also in determining the size of the response needed. This historic cohort represents a challenge to a treatment and prevention system that was overwhelmed at baseline, and which will require historic levels of funding to meet demand. These funds can come from governmental sources as well as opioid litigation windfalls to state and local governments.

Substance use treatment is more complex with polysubstance use; stimulants, benzodiazepines, etc., have treatment options [71], but implementation gets challenging as many will require specialist addiction medicine services. We need an enlarged addiction medicine workforce, as well as telehealth and “warm line” support of primary care providers treating patients with substance use disorders.

Harm reduction is an engagement philosophy. It is a part of a continuum of care that spans prevention, treatment, and recovery. To be more effective, harm reduction needs crisis levels of funding, implementation, and technical assistance. And finally, on a personal note, harm reduction is love. I’ve been privileged to work with some of the most amazing people working in harsh conditions with little funding, providing compassionate care in this disaster that we call fentanyl. They do it with grace and dignity. They are my heroes.